POPULATION BIOLOGY

OF TROPICAL INSECTS
POPULATION BIOLOGY
OF TROPICAL INSECTS

ALLEN M. YOUNG
Milwaukee Public Museum
Milwaukee, Wisconsin

PLENUM PRESS • NEW YORK AND LONDON

 Library of Congress Cataloging in Publication Data
Young, Allen M.
Population biology of tropical insects.
Bibliography: p.
Includes index.
I. Insect populations - Tropics. 1. Title.
QL491.6. Y68 595.7'05248'0913 82-7562
ISBN-13: 978-1-4684-1115-7 e-ISBN-13: 978-1-4684-1113-3 AACR2
001: 10.1007/978-1-4684-1113-3

©1982 Plenum Press, New York

Softcover reprint of the hardcover 1st edition 1982
A Division of Plenum Publishing Corporation
233 Spring Street, New York, N.Y. 10013
All rights reserved
No part of this book may be reproduced, stored in a retrieval system, or transmitted
in any form or by any means, electronic, mechanical, photocopying, microfilming,
recording, or otherwise, without written permission from the Publisher
"Internarse en un bosque es reconciliarse con la vida. El asombro
resbala de hoja en hoja. La fecundidad telurica palpita y destella, como
una libelula dorada, en los oblicuos rayos solares que se filtran por entre
los densos ramajes."
.... From "Bosques Y Vagabundos," by Carlos Luis Altamirano
La Nacion (seccion Ancora), San Jose, Costa Rica
November 12, 1978
 PREFACE

In this book I have tried to bring together the major developments in the study of
insect populations in tropical environments. In some ways, this task has been a
difficult one because conceptually it is virtually impossible to limit a discussion
of insect ecology to the tropics, since the same concepts, theories, and hypoth-
eses concerning the mechanisms by which habitats support insect populations
often apply both to temperate and to tropical regions. Thus one might argue
effectively that a book such as Peter Price's Insect Ecology represents a more
comprehensive treatment of insect ecology, including the tropical aspects. Yet
because there has been a tremendous amount of new study on insects in the
tropics in recent years, and because there has also been a strong historical interest
in tropical insects, judging from early museum expeditions and medically and
agriculturally oriented studies of insects in the New and Old World tropics, I
believe there is a place for a book dealing almost exclusively with tropical
insects. But logically so, such a book by necessity incorporates data and informa-
tion from Temperate Zone studies, if for no other reason than because insights
into the properties of tropical environments often emerge from compariso'ns of
species, communities, or faunas between temperate and tropical regions. An
understanding of insect populations in the tropics cannot be divorced from a
consideration of Temperate Zone populations.
The book is unabashedly biased to the New World tropical studies because
of my interests and more proximal awareness of research on insects being con-
ducted in this region. But a valid attempt is made to incorporate research studies
from the Old World tropics.
The basic mission of this book is to bring together various concepts and
studies related to the issue of explaining the spatial and temporal patterns of
insect diversity in the tropics, considering the breeding population of a species to
be the major ecological unit beyond the individual within the population. The
population is interacting with both biotic and physical components of the envi-
ronment, within and between habitats, and regionally. Much of tropical insect
population biology is therefore concerned in different ways with describing pat-
terns of distribution of individual species in relation to resources and other
environmental factors.
There is also a need for detailed descriptive data on the natural history of

vii
viii PREFACE

individual species in particular habitats in the tropics. What are some of the
major attributes underlying observed patterns of distribution? Can we recognize
the roles of biotic control agents in influencing the populations of insect species
as well as the role of abiotic factors? We recognize that the tropical regions of the
world are extremely heterogeneous in terms of climatic factors and the patterns of
fluctuation of these factors. The kinds of selection pressures affecting an insect
species on a mountain top in the tropics may greatly differ from those affecting
the same species or a closely related species in an adjacent lowland area. Within
major climatic areas, varying subclimatic conditions will be arising from
localized weather patterns and effects of topography and kind of vegetation
cover. Selection pressures on many insect species may be very different in a
strongly seasonal region than in a relatively nonseasonal tropical region.
The fine tuning of insects and their resources, as exemplified by many
plant-insect interactions in the tropics, may reflect coevolved interactions be-
tween insects and their resources, and in recent years there has been a great
interest in defining the traits or properties of such associations. Emerging from
such work is the concept of the individual green plant as an extremely
heterogeneous resource for insects, in the sense that defense systems against such
herbivores vary among different parts of the plant and as a result of highly
localized environmental conditions at anyone point in time and changes in the
environment through ecological and evolutionary time. Herbivore pressure may
be a function of the ability of insect species to adapt to the defensive chemistry of
plants selected as hosts. Herbivore pressure is also a function of (1) diurnal or
seasonal cycles in defensive chemistry metabolism and (2) the degree to which
herbivores synchronize feeding with phases of low concentration of defensive
compounds in specific plant structures targeted as food. Much of our understand-
ing of insect-plant interactions comes from studies in the tropics.
Tropical insect studies also point to the need to consider the success of an
insect species in a particular habitat or geographical region to be a function of
behavioral and physiological flexibility instead of evolutionary adaptations, i.e.,
genetic variation in populations. Neotropical species of Drosophila may be
capable of feeding on a variety of qualitatively different kinds of food not so
much because populations of these flies are heterogeneous genetically, but be-
cause behavioral flexibility enables individual flies to effectively exploit the
range of available food types or breeding sites.
Understanding the occurrence of regions in the tropics of high species
richness for insect taxa is a function of the analysis of historical (biogeographi-
cal) events and ecological factors. In recent years there has been increased study
of forest refugia in Central and South America in relation to periods of recent
glaciation, and such studies have increased our understanding of why pockets of
high species richness are surrounded by areas of relatively low richness. Ecologi-
cal factors, including the occurrence of natural and human-induced changes in
habitats in the tropics-the latter through agriculture, lumbering, and population
PREFACE ix

expansions-also shape the distribution of insect species and the kinds of re-
sources they exploit in a particular area. An attempt is made to examine the
properties of insect populations in such man-made environments as well as the
effects of natural succession on species.
Basic research studies on the mechanisms of insect-plant interactions in the
tropics provide mankind with new information about the unique and regulatory
properties of natural products, data that have the potential to be translated into
applied use through biotechnology as related to the development of naturally
based repellents, insecticides, and fungicides. Tropical floras represent a millieu
of novel and diverse mechanisms of chemical defense against herbivores, and the
insects exploiting such resources have developed detoxification enzyme systems
and other means of countering the defenses of plants. Such interactions provide a
basis for the development of new and effective methods of control for pests and
pathogens associated with Temperate Zone crops and forests. The study of tropi-
cal insects and their interactions with their resources has great potential to the
application of natural products to new research in agriculture, horticulture,
pharmacology, etc. But the first step in reaching such applications is to determine
what is out there, to examine the patterns of distributions of species or groups of
species and genera, to find out what causes the observed patterns, and in the case
of insect-plant interactions to elucidate the mechanisms holding such relation-
ships together and their genetic, physiological, or behavioral basis.
I have attempted to summarize a bulk of the pertinent literature on tropical
insect popUlation biology, with an emphasis on research done within the last 50
years. This book can serve as a helpful starting point for undergraduate and first-
or second-year graduate students with interests in ecology or environmental
biology.
I wish to thank all of the biologists who consented to copyright permissions
from their publications. Mary Joan Young offered encouragement during the
preparation of the manuscript. Dr. Theresa A. Noeske provided painstaking
technical assistance. Karen Heerhold did an excellent job of typing the final
draft of the manuscript. The guidance and cooperation of Ellis Rosenberg and
his staff at Plenum are also appreciated. Gerald R. Noonan reviewed Chapter 10.
I am grateful for the encouragement and strong vote of confidence given to me
by my good friend Oskar R. Zaborsky with the initial plan to write this book. My
own field work experience in the tropics covers a 12-year period, with support
from the Organization for Tropical Studies, Inc., the Associated Colleges of the
Midwest and Lawrence University, the National Science Foundation, Friends of
the Milwaukee Public Museum, and the American Cocoa Research Institute. I
owe a great deal to these people and sources of funding for making my field
work possible.

Allen M. Young
Milwaukee
 CONTENTS

1. The Faunistic Richness of Insects in the Tropics: A Brief Overview 1

2. Individual and Population Responses to Environments . 19

2.1. Ecological Response . . . . . 19

2.2. Natural Selection . . . . . . 22
2.3. Some Properties of Populations. 22
2.4. Ecological Diversity and the Tropics 26
2.5. Some Mechanisms Underlying Tropical Insect Density 32
2.6. Strategies of Response to the Environment 33
2.7. Natural Selection and Properties of Populations 40
2.8. Environmental Predictability and Response Patterns 50

3. Machinery of Environmental Response Mechanisms in Insects: Key

to Evolutionary and Ecological Diversification 61

3.1. Notes on Insect Diversity 61

3.2. Comments on Insect Physiology 67
3.3. Integration of the Insect Phenotype by Hormonal Systems: A Brief
Review 71
3.4. Diapause and Migration in Insect Populations: Escape in Time
and Space. 74
3.5. Temperature, Photoperiod, and Other Factors Related to Insect
Activity: Further Comments 82
3.6. Notes on Mating and Communication in Insects 88
3.7. Rhythmicity in Tropical Insects . 100
3.8. Sociality and Defense in Social Insects . 104
3.9. Density-Related Behavior in Insects 111
3.10. Notes on the Physiology of Feeding in Insects 114

xi
xii CONTENTS

4. Ecological Aspects of Plant Defenses against Insects .137

4.1. Plant Defenses and Coevolution: An Overview. . . . . . 137

4.2. Evolution of Unpalatability and Implications for Food Chains 148
4.3. Notes on Defoliation in the Tropics . . . . . 158
4.4. Notes on Seed Predation in Tropical Habitats . 160
4.5. Notes on Plant-Ant Interactions in the Tropics. 164

5. Distribution Patterns of Insects in Tropical Habitats 173

5.1. The Concept of Local Adaptation . . . . . . 173

5.2. Examples of Habitat Differentiation in the Tropics 178
5.3. Insect Communities in Tropical Habitats . . . . 195
5.4. Plant Resources and Habitat Selection in Tropical Insects 201
5.5. Disturbed and Undisturbed Habitats: Some General Effects 207

6. Population Responses to the Environment in Tropical Insects 217

6.1. Adaptation to Habitats. . . . . . . . . . . . . . . . 217

6.2. Demography and Fitness in PopUlations: Implications for Tropical
Insects . . . . . . . . . . . . . . . . . . . 219
6.3. Further Notes on Population Regulation . . . . . . . . . 230
6.4. Some Examples of Census Histories in Tropical Insects . . . 239
6.5. Population Responses to Environmental Heterogeneity in Tropical
Insects . . . . . . . . . . . . . . . . . . . . . . 245
6.6. Competition as a Selective Force in PopUlation Responses . . 259
6.7. Spatial Environmental Heterogeneity and Demes in the Tropics 266

7. Effects of Seasonality on Insect Populations in the Tropics 273

7.1. Temperate Zone and Tropical Seasonality . . . . . 273

7.2. Some Regional Patterns of Seasonality in the Tropics 280
7.3. Some Seasonal Distributions of Tropical Insects . . 299
7.4. Molding of Ecological Traits in Tropical Insects by the
Dry Season . . . . . . . . . . . . . . . . . . 322
CONTENTS xiii

8. Dynamics of Organization of Insect Communities in Tropical

Ecosystems . . . . . . . . . . . . . . . . . . . . . 335

8.1. Notes on Tropical Ecosystems . . . . . . . . . . . . 335

8.2. Patterns of Organization in Tropical Forests: Some Comments 337
8.3. Notes on the Dynamics of Tropical Insect Communities . . 341
8,4. Some Ecological Factors Generating Patchy Distributions of
Insects in Tropical Habitats . . . . . . . . . . . . . 375

9. Insect Species in Agricultural Habitats in the Tropics 387

9.1. Plant-Insect Interactions and Tropical Agricultural Habitats:

Some Comments . . . . . . . . . . . . . . . . . . 388
9.2. Insect Pollination in an Agricultural Habitat in the Tropics: The
Case of Cacao . . . . . . . . . . . . . . . . . . . 394
9.3. Grazing and Related Activities by Insects in Agricultural Habitats
in the Tropics . . . . . . . 405
9.4. Tropical Insect Pest Populations 420

10. Biogeographical and Regional Evolutionary-Ecological Effects on

the Maintenance of Tropical Insect Faunas: A Brief Perspective 429

10.1. Historical Events and Ecological Processes . . . . . . 429

10.2. Altitudinal and Other Macrogradients in Tropical Regions 435
10.3. Notes on Altitudinal Gradients in the Tropics . . . . . . 442
10.4 Biogeographical Patterns of the Origins of Tropical Insect Faunas 446
10.5. Some Mechanisms of Ecological and Evolutionary Differentiation 451
10.6. Some Examples of Differential Adaptation to Tropical Lowland
and Montane Areas by Insects . . . . . . . . . . . . . 455

References . . . . . . . . . . . . . . . . . . . . . . . . . 461

Index 495
 CHAPTER 1

THE FAUNISTIC RICHNESS

OF INSECTS IN THE TROPICS
A Brief Overview

The primary purpose of this book is to discuss the various ecological factors that
influence both temporal and spatial properties of insect populations in the tropics.
Factors such as diet, habitat, altitude, climate, seasonality, and geography are
considered to be the major determinants of both past and extant distributions of
tropical insect populations. Comparisons of distribution patterns to underlying
mechanisms are often made with insect populations of the temperate zones, since
overall insect diversity is relatively greater in the tropics than in temperate zones.
This book attempts neither to exhaustively cover all of the literature on insects in
tropical environments nor to provide a complete, comprehensive unison of
broad-scale historical or biogeographical explanations for insect distributions in
the tropics with all explanations based upon contemporary ecological interac-
tions. Rather, given the amount of different approaches applied to accounting for
the diversity of tropical organisms in general, and insects in particular over the
past three centuries, the book draws together some of the pertinent studies relat-
ing to the population structure and dynamics of individual species and how such
properties of insect populations relate to the more general issue of the great
diversity of insect species in tropical regions.
My approach recognizes the fusion of both conceptual and empirical studies
from population genetics and population ecology, resulting in the science of
population biology. As pointed out by Lewontin,(!) the adaptive role of genetic
variation in natural populations remains largely unknown or assumed. The prob-
lem at hand is somehow to relate genotypic variation, phenotypic plasticity, and
fluctuating parameters of the environment with the maintenance of enzyme
polymorphisms in populations. Based primarily upon studies of certain classes of
soluble enzymes there is evidence of considerable genetic variation in popula-
tions, as exemplified by the pioneering studies of Lewontin and Hubby. (2) The
problem, therefore, is to relate natural selection to the observed highly variable
loci studied in natural populations. It is also essential to determine levels of
phenotypic variation and to relate these patterns to ecological factors. Enzyme
2 CHAPTER 1

polymorphisms have been conceptualized to adapt species to heterogeneous envi-

ronments.(3) Different phenotypes may be better adapted to particular features of
the environment in different ways, reflecting some pattern of variation in fitness
among the genotypes present in the population. Genotypic differences in fitness
result from phenotypic differences manifested biochemically, physiologically, or
behaviorally. Such differences may be evident within the population by indi-
vidual variation in life table parameters such as fecundity or developmental rate.
PopUlation biology is viewed as one of five major component sciences of
evolutionary biology, the others being biogeography, community biology,
natural history, and evolutionary genetics (including analyses of developmental
and physiological mechanisms underlying modes of response of organisms to the
environment). I also recognize that obviously none of these components of
evolutionary biology is mutually exclusive. In fact, it is somewhat difficult to
suggest such a breakdown of a complex field of natural science. Yet it is my aim
to draw upon primarily the major avenues of contemporary thinking in popula-
tion biology to explain some mechanisms of environmental response in insect
species in the tropics and to do so in such a manner that a clearer view emerges of
why insect species are so diverse in the tropics. Very little of this information is
my own or is new. Most of it is the result of many tropical biologists' conducting
both short-term and long-term field studies, using experimental designs based
upon ecological theory developed in this century.
Another major component eventually rests upon the long history of collect-
ing and watching organisms in the tropics. The major museums around the world
have extensive collections of tropical insects, a result of intensive collecting in
the past. The early literature also contains some interesting natural history of
specific taxa of insects in the tropics. The current emphasis in tropical biology in
general is the elucidation of the temporal and spatial properties of popUlations
and communities based upon testable hypotheses generated from contemporary
ecological thinking. Yet as with most taxa of tropical organisms, there is a huge
gap in our knowledge of old-fashioned natural history as well as long-term
census histories of individual populations. Such data are critical in understanding
the uniqueness of the tropical regions of the world and essential as a prerequisite
for making sound recommendations on the conservation of tropical habitats,
terrestrial and aquatic. It seems that tropical biology has passed through different
stages of emphasis and approach; some sort of feedback loop may be operative.
The early collecting and watching (natural history) phase gave way to detailed
population studies of disease vectors in the case of insects, and today an em-
phasis is on experimental approaches to long-recognized issues. But as we at-
tempt to do more experimental studies, we find the gaps in our knowledge of the
natural history of the organisms being manipulated. Clearly all of these ap-
proaches are essential.
Although the science of population biology developed largely as a discipline
FAUNISTIC RICHNESS OF INSECTS IN THE TROPICS 3

of the Temperate Zones, the great diversity of insects and other organisms in the
tropical regions of the world has been recognized for some time. A survey of
Amazonian streams reveals about the same number of insect species as some
Temperate Zone streams, as shown in a study by Patrick.(4) Patrick(4,S) suggested
that both temperate and tropical streams contained about the same number of
niches for small organisms, including insects. She explained the greater number
of fish species in tropical streams as a result of a greater area being available for
these larger organisms, and therefore sufficient room for a greater number of
species. In the case of butterflies, the survey of Ehrlich and Raven<6) clearly
indicates a far greater number of genera and species of these insects in the tropics
(Table 1.1). In fact, the data also show that some families are limited solely to
tropical regions. As we will examine in later chapters, the great diversity of
butterflies in tropical regions is largely explained by Ehrlich and Raven and
others by the expression of specialized larval host plant utilization patterns.
Another example of diversification with a single genus in the tropics is
illustrated by Paulson's (7) analysis of oviposition behavior in the dragonfly genus
Micrathyria. Six Central American species were found to have very different
types of oviposition behavior in terms of how and where eggs were placed in the
environment. Paulson suggests that such behaviors are related to the ecological
pressure of predation on eggs and young nymphs. Emerson's classic treatises(S,9)
on the zoogeography of termite genera points out that the Indomalayan, Ethio-
pian, and Neotropical regions support the greatest number of termite genera and
that the Ethiopian region alone accounts for 47% of all genera known. The large
expanse of tropical rain forest characteristic of these regions is believed to, be the
major factor promoting the diversity of termite genera found there. Emerson
points out that tropical Africa was a major center for termite diversification but
also stresses the role of ecological processes such as interspecies competition and
termite dispersal by man as other factors affecting distribution. Emerson(S)
suggested that introduced species of termites are usually eliminated by native
species in the Belgian Congo but that the development of dwelling and agricul-
tural habitats by man provided a new niche for such species. Peters's monograph
on the taxonomic diversity of mayflies found in tropical regions OO ) indicates
considerable divergence within the order over large geographical areas.
LinsleY°J) reports that more than 28% of the North American cerambycid
beetles occur in the Neotropical region, the highest percentage for the worldwide
regions surveyed, although the greatest number of species of the group occur in
North America. The group exhibits considerable divergence in feeding habits,
strata of forest habitat preferred, breeding conditions, disease resistance, and
many other factors. There is also evidence that different species are active as
adults in different seasons where wet and dry seasons occur. WatsonOZ ) pointed
out that all 30 species of Odonata known from the northwest of western Australia
are of tropical affinity and consist of two groups, namely, those that breed in
4 CHAPTER 1

TABLE 1.1
A Summary of the Worldwide Geographical Distribution of the Major
Groups of Butterflies"

Approximate number of

Taxon Generab Species Distribution

Papilionidae 24(22) 575-700

Baroniinae 1(1) Central Mexico
Parnassiinae 8(8) 45-55 Holarctic and Oriental; greatest
diversity, Asia
Papilioninae 15(13) 480-640 Worldwide, mainly tropical;
greatest diversity, Old World
tropics
Pieridae 58(40) 950-1150
Coliadinae 11(8) 225-250 Cosmopolitan; greatest diversity,
tropics outside of Africa
Pierinae 43(29) 650-750 Cosmopolitan; greatest di versity,
tropics
Dismorphiinae 3(3) 80-120 Primarily Neotropical, one small
Palearctic genus
Pseudopontiinae 1(0) 1 West equatorial Africa
Nymphalidae 325-400 4,800-6,200
(ca. 202)
Ithomiinae 30-40(10) 300-400 Neotropical; Tellervo, Australian
Danainae 10-12(10) 140-200 Cosmopolitan; greatest diversity,
Old World tropics
Satyrinae 120-150 1,200-1,500 Cosmopolitan; greatest diversity,
(ca. 70) extratropical
Morphinae 23-26(12) 180-250 Indomalayan and Neotropical
Charaxinae 8-10(8) 300-400 Tropicopolitan, few temperate
Calinaginae 1(1) 1 Oriental
Nymphalinae 125-150 2,500-3,000 Cosmopolitan
(ca. 85)
Acraeinae 8(6) 225-275 Tropical; greatest diversity,
Africa
Libytheidae 1(1) 10 Cosmopolitan
Lycaenidae 325-425 5,800-7,200
(ca. 167)
Riodininae 75-125(17) 800-1,200 Tropical, few Nearctic and
Palearctic; Metropolis,
Neotropical
Styginae 1(0) I Peruvian Andes
Lycaeninae 250-300 5,000-6,000 Cosmopolitan; greatest diversity,
(ca. 150) Old World tropics
Total 730-930 12,000-15,000
(ca. 432)

aReproduced from Ehrlich and Raven. (6)

"Number in parentheses indicates number of genera for which food plant records are available.
FAUNISTIC RICHNESS OF INSECTS IN THE TROPICS 5

temporary waters and those that breed in permanent bodies of water. He suggests
that limits to the distributions of these species are determined largely by the
specificity in larval habitats. Widespread species in his study were those occupy-
ing temporary waters, i.e., species with strong tendencies for dispersal. Species
associated with permanent riverine habitats exhibit less dispersal tendency. He
concludes that the distribution of some species of the region is accountable in
terms of past distribution of permanent water habitats. Thus in this study the
explanation of present-day distributions of dragonflies in this region takes into
account relative dispersal tendencies of species utilizing two very different types
of larval habitats and the interpretation of past and present changes in the distri-
bution of these habitats. That study emphasizes the historical role of biogeo-
graphical events upon which is superimposed prevailing ecological circumstances.
Clearly, such studies indicate the role of biogeographical patterns in the
general distributions of insect groups in the world. These and other studies point
to the need to consider not only contemporary or ecological properties of genera
or species in explaining why a particular organism is found where it is, but also
the strong historical component, namely, those biogeographical events resulting
in
in an "early" settling of taxa an area and the degree of dynamic changeover at
the present.
The great diversity of tropical insects makes it imperative to keep in mind
both the historical and contemporary components. of environmental factors ac-
counting for the observed diversity patterns in nature. Sample sizes also confuse
the picture and analyses. Apparently, the more intense the collecting or survey-
ing of a group in the tropics, the more species or other taxa are found .. Thus
Park(13) reports a large increase in the number of pselaphid beetles from Jamaica
as the direct result of increased collecting there and suggests that cases of appar-
ently impoverished psephalid beetle faunas on other Caribbean islands may be
due to poor collecting. Many other published examples of both ecological and
biogeographic-type surveys account for the high generic and species abundances
of insects in the tropics. But clearly the maintenance of such high diversity
results from different kinds of mechanisms in different groups. If the so-called
ecological niche of a species is defined as the collection of resources used by
individuals in a population of a species through space and time, at least part of
the explanation of why more species of insects occur in the tropics relates to
divergence in the ways certain resources are exploited among closely related
taxa.
Part of the justification for a book on the popUlation biology of tropical
insects comes from the well-known observation that the tropics generally contain
a greater number of closely related species for many genera, and also more
genera in different families of insects. Somehow, therefore, there must be rela-
tionships between the popUlation structure and dynamics of individual species
and the ability for many more species to occupy the same environment or habitat.
6 CHAPTER 1

A great abundance of species found in a particular region can result from the
following factors: (1) higher rates of speciation, (2) lower rates of extinction, and
(3) the ability of tropical environments to absorb a greater number of species.
The last factor implies a strong historical component in the sense that closely
related species may accumulate in a particular region over a long period of time.
But what is a long period of time? MacArthurl I4 - 16 ) suggested that a combination
of spatial differentiation of tropical forests and patchy geographical distributions
of individual species promotes a greater diversity of species in the tropics, and
that these effects are most pronounced in tropical environments characterized by
high productivity and low seasonality. MacArthur's explanation of high species
abundance in the tropics(!6) indicates two major components to such diversity: a
pattern of local diversity, in which some portion of the overall diversity of a
group in a region is accounted for by numbers of species within and among an
array of habitats, and a global pattern, in which micro- and macrobiogeographi-
cal effects add to the effects of local diversity. Thus questions of tropical diver-
sity must be defined in terms of the area under consideration and the ability to
distinguish among different kinds of habitats within an area. As will be discussed
later, the issue of a relationship between the degree of environmental predictabil-
ity and of ecological specialization also determines the species abundance level
of a particular environment(!5, 16).
Historically, the lowland tropical rain forest has intrigued biologists with
explanations of high species abundance as indicated in the writings of Dob-
zhansky(I7) and Corner.(!8) The observed high species abundance in such tropical
environments has indeed promoted a wealth of elegant and detailed,studies of
tropical insects and other organisms. Thus, even though it goes beyond the scope
of the tropics per se, Gotwald's monograph on the morphological diversification
of ants,(I9) with an emphasis on mouth parts and feeding, takes on particular
significance since the tropics contain the greatest diversity of ants. The detailed
studies on the nesting habits of Trigona bees by Michend 20 ) and Wille(ZI) reveal
that diversification in nesting habits, rather than feeding habits, has been a major
pathway of environmental modification resulting in the maintenance of high
species abundance both locally and regionally. Both Fittkau(22-24) and Illies(25)
have discussed the ecological diversification of aquatic biotypes in Amazonian
streams, suggesting strong correlations between various biotype classes and the
kinds of aquatic insects and other macroinvertebrates found in each. Their studies
again provide an example of regional diversification of insects and other inverte-
brates in the tropics, indicating something everyone knows, namely, that geo-
graphically distinct environments sometimes have very different collections of
taxonomically related organisms. Trophic relationships generally not present in
Temperate Zone communities, such as the associations of certain dipterans with
swarm raids of army ants studied in Central America by Rettenmeyer,(26,27) have
added immensely to our understanding of just how "ecologically saturated" the
FAUNISTIC RICHNESS OF INSECTS IN THE TROPICS 7

tropical rain forest environment is in terms of unique life styles of insect species
found there.
The picture that emerges is that the tropics are indeed a heterogeneous
mixture of different types of environments and that within each of these envi-
ronments, the many different organisms together exhibit a range of generalized
and specialized life styles. I define life style as what the individual organism does
during its lifetime, the range of such "itineraries" represented by all the indi-
viduals within a population, and the underlying reproductive capacities and
breeding system in the population. A thread that weaves all of this together is
how the individuals in the population allocate energy to different components of
life style.
The great difference in tropical environments is exemplified by Hubbell's
analysis of trees in a tropical dry forest(28) and comparisons to tropical wet forests
and Temperate Zone forests. He demonstrated (see Fig. 1.1) that tree species are

100'

~
10 "

~~
"
: ~ "" Tropical wet forest
Q) ~ \ '",. ,;/' (Manaus, Brazil)
"
~
!:
~
50. '~" .'
E
- .01 \ Tempe;ate moist " " ,
:\ • forest '"
!' \. (Smoky Mtns .• Tenn.) ' .........
.001 ~ \ '"'' " Tropical dry forest
\, (Guanacaste, Costa Rica)
Temperate montane

.0001
(subalpine) forest
(Smoky Mtns., Tenn.)
\
o 20 40 60 80 100 120 140 160 180 200
Species Rank

FIGURE 1.1. Ecological dominance-diversity curves for two tropical and two temperate forests.
Importance values for the temperate forests are based on annual net production; importance values for
Costa Rican dry forest are based on cross-sectional areas for all stems of a given species, and for the
Amazonian forest from above-ground biomass. Although the comparisons are not exactly based on
the same data sets, the pattern is clear: dominance of individual tree species is greater in temperate
forests even though some species of tropical trees exhibit clumping. [From S. P. Hubbell, Science
203: 1299-1309 (1979). Copyright 1980 by the American Association for the Advancement of Sci-
ence.]
8 CHAPTER 1

often clumped in tropical forests as they are in temperate forests but that frequent
disturbances in the tropical dry forest environment create clumped distributions
of many species. Comparisons of such data with tree species distribution in
lowland tropical wet forests suggest that the distribution of plant resources for
insects are distributed differently in different kinds of tropical forests. Such
heterogeneity is expected to result in different life styles and population
dynamics for insect species. As the great majority of insects are herbivores or
plant predators, it is the interrelationships of plants and insects that probably
account for the spatial and temporal structuring of insect popUlations in the
tropics.
The recent studies on the coevolutionary interactions between plants and
insects suggest that insects contribute in a significant way to the distribution of
abundance of plant species, especially in the tropics,(6.29) but also that plants
influence insect distributions as well. It has been known for some time that
geographical and habitat diversification within plant taxa often result in similar
shifts in distributions of pollinators such as bees. Hurd and Linsley have traced
the spread of certain Cucurbitaceae and pollinating bees in the Americas.(30) The
elegant studies of Gilbert°1J have also indicated that certain species of pollen-
feeding Heliconius butterflies have acted as major selective agents in the popula-
tion biology of some Cucurbitaceae. (See Chapter 3.)
The environmental heterogeneity of the tropics is also reflected in the kinds
of climatic regions found there and the extent to which biologists have conducted
studies in them. Thus the comparison of the recorded insect faunas of the
Galapagos Islands and British Isles by Linsley and Usinger(32) indicates that a
much higher number of families, genera, and species of insects are found in the
British Isles. They ascribe the depauperate condition of the insect fauna of the
Galapagos to accidental colonization by species from the mainland of South
America. The fact that much of the Galapagos are scrub and desert may also
contribute to the relatively low abundance of insect taxa found there. One might
speculate that the bees studied by Linsley and Usinger were tracking the envi-
ronment in the sense of changing their geographical distributions in response to
the spread of both wild and cultivated cucurbits. Certain types of obligate associ-
ations between ants and plants in the tropics as studied by Janzen(33) reflect
similar joint distribution patterns . Yet for other groups of insects, tropical island
faunas are considerably diverse, sometimes even more so than the adjacent
mainland. Paulson(34) suggests that the higher number of Odonata species in the
Bahamas than in southern Florida and the Florida Keys is due to the absence of
some undetermined environmental factor that typically reduces both species
number and populations of these insects in the latter regions. From various
studies of faunas on tropical islands, it is apparent that neither absolute island
size nor area is the major determinant of faunistic size and complexity within
different trophic levels. Another major component of diversity is the collection of
FAUNISTIC RICHNESS OF INSECTS IN THE TROPICS 9

habitats present. The interaction of regional size, habitat diversity, and climatic
factors molds the communities associated with regions. In this book I hope to
provide examples of the effects of climate, region size, and habitat availability on
the diversity of insect faunas found in the tropics, in addition to examining the
effects of these environmental factors on population ecology.
Kohn and Orians(3S) have pointed out the need to examine the ecology of
closely related species. When dealing with tropical insect faunas, it is necessary
to recognize those characteristics related to taxonomic affinities and those ex-
pressed as a result of ecological conditions. Given the greater number of closely
related species in the tropics, such analyses are very critical in attempting to
explain both global distributions (sensu MacArthut 16 » and local distribution of
species. Again, there is the need to recognize the historical component of oc-
cupying a particular environment or range of environments, and the contempo-
rary selection pressures and nongenetic response opportunities for insects.
Because of their great diversity in the tropics, relatively short generation
times, and often readily distinguishable phenotypic features, insects in general
are ideal organisms for studies concerning distributions of species in relation to
resources and environmental changes. The great diversification of ants in the
arboreal habitats of tropical forests has provided the evolutionary opportunities
for other arthropods to converge on them in terms of mimetic resemblance. An
example of this phenomenon is given in the carefully documented study of
ant-mimicking salticid spiders (see Reiskind(36». In some groups of insects in
tropical regions, there is the opportunity to study various life styles and habits of
individual genera and species that indicate possible evolutionary pathways of
ecological specialization and associations among closely related forms (e.g.,
Erwin and Erwin(37». Some forms may exhibit behavior and habits intermediary
in character state to other kinds of life styles. Canopy orientation, for example,
has been shown to be the primary orientation mechanism underlying home
range behavior in an Old World tropical ponerine ant.(38) Yet at least one New
World tropical ponerine, Paraponera, is an active canopy forager unlike the
Old World Paltothyreus studied by Holldobler, in which foraging is a horizontal
or lateral activity on the forest floor. One might suggest that actual canopy for-
aging in a form such as Paraponera might be an intermediate step in more
specialized forms of foraging as exemplified by Paltothyreus. (38) Tropical insect
faunas overall provide the opportunity for comparative study of the behavior
and ecology of closely related forms or convergent forms, because there is gen-
erally a greater diversity of insect taxa in tropical regions. Paramedical ecologi-
cal studies of disease vector insects in the tropics have illustrated that local high
diversity of vector dipterans such as blackflies supports a larger spectrum of
disease microorganisms (e.g., Takaoka(39». Sometimes the pattern of distribu-
tion of the closely related insects is defined along habitats since different forms
may diverge to occupy different habitats (Figs. 1.2 and 1.3).
10 CHAPTER 1

FIGURE 1.2. Border strip of mixed primary and advanced secondary tropical rain forest in north-
eastern Costa Rica. The low vegetation to the left is regenerating young secondary forest about 3
years old. The species composition and distributions of populations of individual plant species differ
greatly between the two areas, providing a form of spatial environmental heterogeneity for insect
communities associated with these habitats.

Perhaps an even more relevant reason for their study in the tropics is that
insects are a major link in the food chains of tropical communities and ecosys-
tems. The botanical work on the structuring and floristic composition of tropical
forests cannot be divorced from the impact of insects on such systems. A bidirec-
tional flow of information about the environment results in responses by the
receiver. Thus making an attempt to account for the distribution of certain tree
species in a tropical forest may not be entirely possible without examining the
popUlation biology of pollinators and seed dispersal or seed mortality agents such
as certain insects. Likewise, the distribution of the insects may not be explained
without an examination of the tree species acting. as resources for the insects.
Perhaps the greatest contribution of tropical biology in the last 20 years has been
the elucidation of the interdependence between plants and animals in shaping the
static and dynamic properties of tropical forests. Thus although Marston Bates
studied the natural history of anopheline mosquitoes many years ago, with an
emphasis on the role of these insects as disease vectors, studying these insects
today may require a different approach. One might wish to examine whether the
males of these mosquitoes are in fact effective pollinators of certain understory or
even canopy flowers in the tropical rain forest environment.
FAUNISTIC RICHNESS OF INSECTS IN THE TROPICS 11

FIGURE 1.3. Mixed primary and secondary tropical rain forest in northeastern Costa Rica. The
forest patch shown here is dominated by the canopy tree Pentaclethra macroloba (Leguminosae)
from primary forest, mixed with another dominant, Goethalsia meiantha, characteristic of advanced
secondary forests in this region of Central America.

The role of insects as food items in tropical ecosystems cannot be disputed.

Apart from the role of herbivorous and pollinating insects as major selective
forces in the evolution of tropical plant assemblages ,<40-42) many insect species are
sources of nutrients for other organisms, including parasitic and predatory ar-
thropods and many vertebrates. Just as the structuring of fruiting seasons in
conjunction with the spatial distributions of individual tree species in tropical
forests is clearly a major selective force in the development and maintenance of
bird species diversity in tropical environments,(43) the availability of insects, in
diverse forms, with diverse feeding and life style habits, and activity patterns
compose a variegated ecological fabric exploited by other organisms. Thus it is
not surprising that studies of stomach contents of crocodiles from tropical Afri-
can lakes generally indicate that insects compose the bulk of food for young age
classes of these animals. In such systems, the predators are active at times of the
day when major foodstuffs such as belastomids and dragonflies were active or
completing final ecdysis, respectively.
Hubbell(28) and many others have called attention to the need to study the
flora of certain regions of tropical America. If we are going to study the impact of
12 CHAPTER 1

insects on plant species in the tropics, we should have reliable information on the
names of the plants and their ecological distributions across regions and habitats.
There is also a need for more studies of high-altitude plants and animals in the
tropics. A great deal of emphasis has traditionally been on lowland systems,
although the trend is changing. H. G. Baker and G. W. Frankie's analyses of
flowering and pollinator phenologies in Costa Rica's Monteverde cloud forest
will provide new information on the structuring of tropical vegetation from that
region. A. M. Shapiro's life history and physiological studies of high-altitude
pierid butterflies in the Andes will provide an explanation as to why certain
groups of insects thrive there while others do not. Such studies complement
nicely the earlier survey approach of detailed accounts of cloud forests such as
that of Myers(44) in Panama.
Out of all of the above, 1 hope to convey the importance of various ap-
proaches to the best study of insect populations in tropical environments. First,
there is clearly the need for more detailed natural-history studies, particularly on
host plant relations and herbivorous predators and parasites. There is also the
need to continue to develop research programs designed at examining plant-
animal interrelationships and effects of climate on them, in less popular regions
such as high altitudes and deserts. There is a need for more studies of aquatic
ecosystems and how they are affected by habitat disturbances. And there is of
course the need for the elegant conceptual and experimental studies that have
developed largely through the Organization for Tropical Studies programs in the
past 20 years.
A great deal of new information coming from the study of insect populations
in the tropics has contributed in a significant way to our general view of the
natural world and man's impact on it (Fig. 1.4). There has been a reshaping of
traditional ecological concepts and a refreshing reexamination of the strategies by
which insects and organisms respond to changes in the environment. A basic
theme of this book, and one that does not necessarily reflect the popular
viewpoint, is that we must examine what insects do in "adapting" within (I am
avoiding the phrase "adapting to' ') the collection of intrinsic and extrinsic var-
iables in the environment, not necessarily in terms of genetic or evolutionary
response, but in terms of response plasticity at the behavioral or physiological
level. Short-term responses to environmental changes might be more frequent
and more the rule than evoking long-term or evolutionary responses. Another
goal of this book is to emphasize the heterogeneity of the tropics. For this reason,
I include a discussion of seasonality, and as with the other chapters, I draw upon
specific examples to illustrate a particular phenomenon.
Tropical environments vary tremendously in terms of climatic conditions,
particularly in terms of monthly distributions of precipitation and total annual
precipitation. There can be considerable variation from year to year for total
precipitation, and the length of a dry season may also vary considerably in
FAUNISTIC RICHNESS OF INSECTS IN THE TROPICS 13

FIGURE 1.4. Clearing away of advanced secondary tropical rain forest in Costa Rica, for conver-
sion into cattle-grazing pastures. Such activities create a dynamic condition in the wet tropics in
which shifting distributions of herbivorous insect species follow major perturbations in the local or
regional environment.

different years. Thus climatic heterogeneity is present within and among dif-
ferent tropical regions .(45) A glimpse at the precipitation regimes for Central
America illustrates this point.(45) Even over smaller areas such as French Guy-
ana, precipitation distribution varies considerably.(45) Such patterns coupled
with other environmental factors impose an instability on tropical ecosystems.
Further change is imposed by habitat disturbances resulting from both natural and
man-made modifications in the local environment. There is also tremendous
altitudinal variation with profound effects on the distribution of organisms, as
discussed by Janzen.(46) Such considerations dispel the notion of a stable and
constant tropics. There is a resulting uneven distribution of species over the
tropics. In the next chapter I examine briefly how organisms respond to the
environment and environmental change. Thus Chapter 2 is essentially an intro-
duction to the concept of adaptation , and it forms a basis for the rest of the book.
Chapter 3 provides a brief examination of insect physiology and structure in
terms of survival and reproduction within tropical environments, with an em-
phasis on feeding and possible diapause strategies associated with the dry season.
This discussion blends into a discussion of plant-insect interactive mechanisms,
including plant defenses against insects and the development of coevolved as-
14 CHAPTER 1

sociations. As with most of the chapters, the information is topical in format,

since the literature on this subject is growing rapidly and other recent publica-
tions such as Janzen(42,47l provide more in-depth and conceptual discussions.
From this rather qualitative overview I examine modes of adaptation in insect
populations and give examples of insect population studies from the tropics, and
include a few from Temperate Zone regions that are relevant to the issues dis-
cussed.
Chapter 4 is devoted to plant-insect interactions in tropical ecosystems,
with an emphasis on herbivory and pollination. There has been a flurry of
research in both aspects of plant-insect interactions in the past 20 years. I also
attempt, in other chapters, to examine relationships between insect diversity and
habitats, and also devote one chapter to selectional and historical responses of
insects to resources over habitats and regions in the tropics.
In some of these chapters, for example, Chapter 9, I incorporate some of my
own research data. In this chapter, I suggest a model of cacao pollination and
production, examining the question of optimal yield from the viewpoint that
cacao cultivation may cause changes in pollination strategy in the wild. In Chap-
ter 7 I include a brief discussion of different tropical regions based on climate.
The book is clearly oriented toward a lot of recent work done in the New World
tropics, although I attempt to incorporate studies from other regions as well. My
overall theme is that complex abiotic and biotic factors mold the spatial and
temporal distributions of insect species in tropical environments. We must look
at mechanisms promoting abundance of individuals and high species number in
both habitats and over regions. The approach is basically ecological aJ.1d natural-
historical sensu stricto, but I attempt to incorporate some aspects of evolutionary
genetics whenever appropriate and comfortable for me to do so. Research on
tropical insects, and their relation to other portions of the environment increases
our understanding of why the tropical regions of the world contain the greater
taxonomic diversity and associated patterns of ecological specialization for many
groups of insects than do the other major faunistic regions of the world. Some
groups of insects in the tropics may be deliberately imported to the northern
latitudes as part of research on the effectiveness of introduced predators or
parasites on pest species associated with food crops (e.g., Marsh(48l).
This book presents an overview of some of the major discoveries of how
insect species are integrated into the complex and varied environments of tropical
regions. Janzen(42l has emphasized the dynamic nature of research in tropical
biology, and the future directions of it. There is a need to study in detail the
ecology of individual species, especially from the viewpoint of analyzing the
other organisms with which each species interacts, and the effects of seasonality
and other local weather conditions on populations. Such studies are coupled with
tedious natural history studies at the autecologicallevel. A great deal of tropical
biology has examined interactions of organisms over relatively short periods of
FAUNISTIC RICHNESS OF INSECTS IN THE TROPICS 15

time. There is a need for long-term studies on the census histories of animal and
plant populations to detect cycles in abundance and other phenomena. The con-
servation of distinct tropical habitats (Fig. 1.5) in different tropical regions will
ensure the resources essential for understanding the complex nature of food
chains. At the opposite end of the spectrum, but not exclusive, is the analysis of
whole ecosystems in tropical regions. It seems to me that both approaches are
necessary for understanding the temporal and spatial properties of populations of
component species in tropical ecosystems.
The description of population census histories, patterns of resource exploita-
tion, and strategies of fitting into the environment for insect species in the tropics
must be linked to the issue of explaining the patterns of diversity of insects within
and among habitats, and over larger regions of the environment. Gradients in
diversity of insects across such units of the environment are generally greater in
the tropics. We expect to find habitats or regions in which the numbers of insect
genera and species associated with various trophic levels in communities are
considerably higher than for comparable Temperate Zone environments, and we

FIGURE 1.5. Understory of premontane tropical rain forest in Costa Rica.

16 CHAPTER 1

also expect to find others in which such diversity is low. In the sense of Vander-
meer's(49) definition of a community as a set of "operational trophic units," the
collection of individuals of an insect species composing a population or subpopu-
lation associated with a certain habitat is the fundamental component of tropical
diversity patterns that warrants analysis.
The population is conceptualized as a collection of reproductively capable
individuals as well as juveniles (having an age distribution) changing to varying
degrees through space and time. Depending upon habitat, seasonality, trophic
position, resource availability, character states, and other factors, some popula-
tions of insects might be physiologically or genetically programmed for traits
favoring colonization, while others are in some sort of equilibrium with other
components of the community and therefore programmed for fitting into the
environment at hand. We do not expect all species within a given trophic level to
be in similar states of popUlation dynamics or degrees of integration into com-
munities. As will be developed in this book, the degree to which insect popula-
tions are influenced by biological and physical features of environments will vary
considerably among different habitats and regions in the tropics. The existence of
a wide range of habitats, from both natural and man-made perturbations to local
environments (Figs. 1.2-1.4), has provided the opportunity for insect species to
diversify in life styles and for overall diversity of insects to increase in the
tropics. Such environmental mosaics establish different sets of changing en-
vironmental conditions or challenges for insects and other organisms. We expect
patterns of adaptation within popUlations and for whole sets of popUlations to
vary considerably in different types of tropical habitats, and to be influenced by
seasonality or a lack thereof.
We can examine some of the patterns and mechanisms of adaptation of
insects to tropical environments, but we can only scratch the surface. Many
species have not been studied, and the trophic organization of different com-
munities has not been determined. We seem to have a lot more information about
trophic organization at the plant-insect interface than for other levels of commu-
nity organization. Extensive data on tropical litter communities are lacking, save
for a few such as the interesting study in which Stanton(50) compared temperate
and tropical litter communities. Stanton compared the species richness and diver-
sity of litter mites in three habitats in Costa Rica and Wyoming and found a
greater species richness and diversity of mites in forest litter at both localities. A
greater diversity of mites was found in the tropical forest habitat than in all other
habitats sampled, and a greater proportion of mite species in the Temperate Zone
samples was classified as colonizers. (50) Such data from decomposer com-
munities provide further support for the greater diversity and species richness for
organisms in tropical forests. There is also a lack of information on the resource
utilization patterns and popUlation dynamics of parasitic wasps and flies in the
tropics and on the degree to which such organisms regulate populations of her-
bivorous insects in tropical environments. The biogeographical studies of mam-
FAUNISTIC RICHNESS OF INSECTS IN THE TROPICS 17

malian ectoparasites by Wenzel and Tipton(5J) and others have provided some
data on the broad distribution of parasitic arthropods in the New World tropics,
but the omnipresent need is to examine the role of parasites in the structure and
dynamics of individual communities.
Long-term studies of insects are needed to determine the degree to which
insect populations are cyclic irl the tropics. A plethora of short-term studies has
outweighed long-term studies of 5 or more years. Data on the communities of
organisms in both standing water and running water aquatic environments in the
tropics are also wanting. Given the predatory role of both larval and adult
Odonata and the diversity of Odonata in some tropical regions,(34) such studies
might be particularly elucidating in terms of the impact of these insects on other
species in terrestrial and aquatic habitats. There are many other examples, but at
least one more deserves mentioning: the need to gather extensive field data on the
natural history and ecology of "showy" insects from the tropics, insects that are
likely candidates to end up in museum or private collections or specimen
dealerships throughout the world. The riot of gaudy-colored butterflies, beetles,
and other insects from the tropics has created an international market for these
insects. Yet before we can establish realistic guidelines for protecting individual
species of these insects, or setting aside large tracts of land for ecological pre-
serves, we need to determine the population structure, resource requirements,
and degree of "ecological flexibility" of each species in different but representa-
tive portions of its geographical range. Such a body of data is growing for some
butterflies of the New World tropics but is generally lacking for butterflies in
other tropical regions of the world.
This book is also not an attempt to summarize all that is known about the
ecology of insects in general. One could argue that a book on tropical insects
somehow excludes a great deal of knowledge on insects from other regions. I
take the view that the time is right to bring together some of the interesting and
informative work being done on the tropics, and a good deal of this work
concerns the interactions of insects, particularly herbivores and pollinators, with
other organisms. A cornerstone of this view is that the tropics represent a unique
plethora of strategies concerning the ways in which organisms fit into the envi-
ronment, and in some ways examining issues of popUlation regulation, competi-
tion, effects of abiotic factors on organisms, mutualisms, and other ecological
phenomena is easier than in Temperate Zone systems. This is not to say that one
excludes or minimizes the importance of similar studies in the Temperate Zones.
Just the opposite is true. The beauty of tropical studies is the opportunity for
meaningful comparisons with Temperate Zone studies. Thus this book, drawing
heavily upon some of the major directions in research on tropical insects, sup-
plements other books concerning insect ecology, general ecology, and popula-
tion biology. Realizing the magnitude of the task of bringing together the litera-
ture on tropical insect populations, it is my sincere hope that omissions of
seemingly pertinent material from this book are not construed as an oversight.
 CHAPTER 2

INDIVIDUAL AND
POPULATION RESPONSES
TO ENVIRONMENTS

2.1. ECOLOGICAL RESPONSE

The degree to which an insect or other organism becomes integrated into the
environment is dependent in part upon how long the species has occupied it. The
extent to which a species becomes a resident of a certain community is a function
of many factors, of which the following often prevail: (1) the number and
abundance of ecologically similar and/or closely related species present in the
community, (2) the ability of the species to obtain sufficient resources for sur-
vival and breeding, (3) the ability of the species to withstand biotic mortality
agents such as pathogenic microorganisms, predators, and parasites, and (4) the
ability to cope with shifts in (1) through (3) resulting from seasonality and to
cope with the abiotic changes in the environment brought into play by seasonal-
ity.
The studies of Lack(52.53) have provided ecologists with evidence that
closely related species sometimes exhibit ecological isolation, achieved by geo-
graphical separation among the species, by habitat, or by feeding habits where
ranges overlap. A basic assumption of such analyses is that the resources re-
quired by each species are limited but sufficient to result in competition for them.
The degree to which ecological isolation plays a role in determining the geo-
graphical and habitat distribution patterns of species is also governed by the
effects of seasonality on alternative food supplies or other resources in short
supply. Thus although many tropical rain forest understory butterflies in Central
and South America exhibit some degree of specificity for feeding on certain
kinds of rotting fruits and molds, a seasonal period of high fruit abundance for
particular tree species may result in a form of "ecological convergence" in
which the butterfly species jointly exploit the abundant fruit supply. A similar
argument can be made for some genera and species of butterflies visiting in-
florescences in tropical forests (Figs. 2.1 and 2.2). The most difficult issue in
terms of detecting mechanisms of ecological isolation in habitats is the degree of

19
20 CHAPTER 2

FIGURE 2.1. An unidentified hesperiid butterfly visiting the inflorescences of the bush Serjania cf.
atrolineata Saw. & Wright (Sapindaceae) during the dry season in the central highlands of Costa
Rica. In this region and during the lengthy dry season, adult butterflies aggregate along shaded and
moist strips of vegetation along streams, where they feed on nectar from flowers. Many species of
plants in such habitats bloom during the dry season . Breeding tends to. diminish as the dry season
advances and some butterflies, and presumably other herbivorous insects associated with foliage,
enter into a reproductive diapause. Under such conditions, diapause represents a physiological
response to a decrease in correct oviposition substrates and larval food (i.e., young leaves). Depend-
ing upon the species of insect and foodplants, such a pattern varies considerably in seasonal tropical
environments.

resolution of analysis. The literature is replete with examples of studies illustrat-

ing that one's definition of the niche of a species is a major factor in determining
whether or not two or more species have niche overlap and if so to what extent.
Through detailed natural-history studies in the case of tropical insects, it
will be possible to determine the temporal and spatial variations in (I) what
resources are utilized, (2) what percentage of time the average individual spends
on each resource, and (3) what other species use the resources and the time-
frequency distribution of utilization. The issues and problems of defining the
INDIVIDUAL AND POPULATION RESPONSES TO ENVIRONMENTS 21

FIGURE 2.2. The nymphalid butterfly Metamorpha stelenes visiting the inflorescences of the bush
Serjania cf. atrolineata in the central highlands of Costa Rica during the dry season. The preponder-
ance of blooms of this bush represents a concentrated food resource for adult butterflies at this time of
the year in this region, and because breeding tends to decline as the dry season advances, adult
butterfly popUlations become increasingly characterized by very worn individuals such as the one
shown here. i.e., the age distribution of adults is skewed toward the older classes. The larval host
plants of M. steienes, various genera and species of herbaceous Acanthaceae, tend to cease vegeta-
tive growth and tum brown as the dry season advances, thereby dropping out of the local assemblages
of potential oviposition sites and larval food sources for this species and a close ally, Anartiafatima.

ecological niche have been reviewed and expanded by Vandermeer. (49) How a
particular insect species responds to the environment is shaped by the factors
mentioned above, namely, interactions with potential or actual competitors,
availability of required resources, effects of biotic and abiotic mortality factors
affecting popUlation size and structure, and the effects of spatial (within, be-
tween, habitats) and temporal (seasonal) variations in these effects of environ-
ment. The size and the age structure of the population also shapes the response of
individuals to these environmental parameters.
22 CHAPTER 2

2.2. NATURAL SELECTION

Clearly, a major approach to the issue of how organisms respond to the

environment has been dominated by the theme that a great deal of such response
in fact involves evolutionary change. The theory of gradual change by evolution-
ary processes is based largely on the premise that natural selection acts on the
individual in the population. For the processes of evolutionary change, it is
generally believed that the population is the most important level of biological
organization. Since each population represents a gene pool with varying degrees
of confinement (lack of gene flow), the evolutionary process is held to be
changes in the genetic makeup of the population over time. Different populations
of the same species can change in different ways through time or follow different
adaptive pathways, resulting in distinct or similar adaptations or traits. Lewontin
and White(54) demonstrated such effects in populations of the Australian
grasshopper Maraba scura.
A major feature of populations is the distribution of individuals in space and
time, described by habitat or regional surveys and census histories, respectively.
The size or density of a population is determined by the interaction of birth rates,
death rates, and dispersal. Changes in the environment often result in changes in
these characteristics, either toward increasing or decreasing population size, if
not maintaining stable conditions. Some of these effects are shown in Fig. 2.3.
Changes in the genetic makeup of a popUlation also arise from genetic drift.
Wright(55,56) demonstrated rigorously the effects of drift on the makeup of popula-
tions, showing that drift effects are most likely in small populations (i.e., small
in size relative to other popUlations of that species). As each population consists
of a collection of individuals or genotypes, it is the array of genotypic variation in
the population that determines, under the view of evolutionary adaptation, the
capacity of the species in that particular environment to cope with the environ-
ment. Such reasoning has resulted in a fusion of concepts and models from
population genetics and population ecology (Table 2.1) that forms the basis of
scientific study in population biology.

2.3. SOME PROPERTIES OF POPULATIONS

Two major aspects of popUlation ecology are the census history generated
by a population placed in an unlimited environment through colonization or
range expansion, and the interactions of popUlations of two or more species
existing in the same environment and at the same time. In terms of the tropics,
one might expect that the observed accumulation of closely related and ecologi-
cally similar species of insects in some communities, such as indicated by Jan-
zen(41,47), and the existence of a broad range of newly created secondary and
INDIVIDUAL AND POPULATION RESPONSES TO ENVIRONMENTS 23

I
I
I
I
I
I
I
W I
N I
I

1ii I
I

z I
o I
I

ti
I
I
.... ..
I
..J I
:::l I

......
I
Il. I
~ ....
I
I
I
I ......
I
I
I
.. ..
I
I
........
I
I
..
.........
I
I
... ..
TIME
,
FIGURE 2.3. Relationship between time and insect population size. Depending upon the kinds of
selection pressures operative on an insect population. population census histories will vary considera-
bly. The dotted line shows a population in which increase in population size is gradual, perhaps as the
result of strong selection for life history characteristics favoring a low reproductive potential (r), or
perhaps as a result of intense predation of a density-related form that keeps the population from
realizing the carrying capacity of the environment. The dashed line shows a population on the course
for overexploiting its resources, through a high reproductive potential or as the result of relaxed biotic
control mechanisms. The solid line shows the three major phases of population census history:
growth, stabilization or realization of carrying capacity of the environment, and decline. If an insect
population realizes the carrying capacity of the environment and produces an evolutionary adjust-
ment, such a population may become vulnerable to extinction if selection pressures shift in such a
way so as to require increased reproductive potential by the species (a higher r). All three kinds of
census history are expected in the tropics, where the environment is considered heterogeneous in
terms of selection pressures operative. The dashed-line "strategy" may be characteristic of coloniz-
ing episodes, when an insect species enters a new adaptive zone (habitat) and requires an increased
rate of population growth.

agricultural habitats would indicate that interacting sets of species and coloniza-
tion events are common phenomena in the tropics.(S7) We return to the data on
these in subsequent chapters. But returning to Fig. 2.2, it is seen that the entry
of a population into a new environment with unlimited resources is characterized
by four stages: (1) rapid growth of the population in the new environment, (2) a
period of population stabilization, perhaps with minor fluctuations in abundance,
(3) a population decline, and (4) extinction.
The vast amount of animal extinctions evidenced in the paleontological
record is in itself proof of the adaptive role of extinctions, in that major changes
in the environment flushes out some taxa while others are able to survive and
24 CHAPTER 2

TABLE 2.1
The Classical Models of Adaptations in Populations, from Both
Population Genetics and Population Ecology

Population genetics Population ecology

Hardy-Weinberg equilibrium Population parameters

Theory of inbreeding Models of regulation
Models of selection Density dependence
Balanced polymorphism Density independence
Directional selection Predator-prey systems
Stabilizing selection Population-environment interactions
Selection in and of populations Theory of community and
Breeding structure regulation of diversity
(species packing)
Social systems

exist in altered environments. The flux of secondary growth habitats in tropical

regions suggests' that both local extinctions of some insect species and shifts in
adaptation in others are a major component of the dynamic condition of such
habitats. One might expect that changes in food availability and quantity, along
with abiotic changes such as increased exposure to sunlight and precipitation,
constitute some of the selective forces in newly created tropical terrestrial
habitats.
The census history of a population-the record of change in size of the
population through time-involves periods of population growth, stabilization,
and decline. Many studies have shown that both intrinsic and extrinsic factors
cause changes in the population dynamics of insect species and other organisms.
It is the interaction of changes induced by the membership of the population itself
and by changes from other populations or from abiotic factors that account for the
observed census history. The relationship between population density and popu-
lation size has been studied by many population ecologists, including Thomas
Park, who has used the Tribolium flour beetle to study insect population
dynamics (e.g., Park(58». As population size increases, so does density and
therefore intraspecific competition for a set of resources in the environment. The
mortality rate in laboratory cultures of nymphs of the American cockroach,
Periplaneta americana, increases with increasing population density, with
mortality generated by periods of high population growth in which nymphs are
injured and killed at ecdysis, retardation of the population growth rate due to
density-related effects adversely affecting survival, and eclosion of adult
roaches.(59) Dense populations are characterized by adult roaches with lower
average body weight than low-density populations. (59) Pimentel(60) suggested that
a genetic feedback mechanism regulates the size of populations for herbivores,
INDIVIDUAL AND POPULATION RESPONSES TO ENVIRONMENTS 25

predators, and parasites. Under Pimentel's model, in the plant-herbivore system,

for example, a herbivorous insect species specialist on a particular plant species
establishes a sphere of selection pressures influencing the defense strategy of the
plant against herbivores (see Chapter 4) and through successive generations such
selection pressure changes the genetic makeup of that plant population. (60) Such a
change subsequently affects the density of the herbivore population. Pimentel
envisions food chains as sets of interacting genetic feedback mechanisms.
Much of the diversity of tropical herbivorous insects is accounted for by
specialist feeding behavior in which individual species tend to have restricted
host plants (see Chapter 4). In one sense, the concept of coevolution between
plants and animals is a restatement of the genetic feedback mechanism proposed
some years ago: as herbivore density increases on a particular plant species,
particularly where the herbivore in question is a specialist, and if the pressure is
sustained over subsequent generations of the host plant species, selection will
favor the expression of increased resistance to the herbivore in subsequent gener-
ations, and such changes will cause one of three responses in the herbivore
population: (1) it will adapt to the new conditions of the host plant; (2) it will
experience an ecological shift to another host plant species with a less developed
defense or to qualitatively different species capable of being overcome without
large losses in fitness; or (3) it will experience extinction. The evolutionary
approach suggests that the outcome is largely determined by the genotypic varia-
tion in the population. Brereton(6]) has reviewed some examples of self-
regulation in animal populations, suggesting that some populations are capable of
such regulation. Such self-regulation is believed tQ be a response to changes in
the environment such as crowding and fewer resources. Deleterious crowding
may induce physiological responses in individuals within a population, such as
reduced average fecundity or increased tendency toward dispersal, with the result
of slowing down the growth rates of popUlations. The entomological literature
from the Temperate Zone is full of examples of the relationship between density
and emigration in aphid populations associated with various plant species in
seasonal regions.
The population dynamics and pattern of environmental exploitation by in-
sects in the tropics is molded by the historical events resulting in an accumulation
of ecologically similar species in a particular region, and by the more contempo-
rary ecological factors that affect the size and distribution of populations within
and among habitats. It is difficult to view both types of forces as separate entities.
The centers of evolution for many groups of insects are located in the tropics
(Haffd62 »), and superimposed upon this pattern is the existence of past forest
refugia in South America during the Pleistocene in the case of the New World
tropics. Such conditions have molded the general distribution of insect taxa in the
tropics. Pianka(63) had distinguished between the historical accumulation of
species in the tropics and the contemporary ecological factors that promote
26 CHAPTER 2

species abundance in the tropics. A brief discussion on the prevailing views on

tropical diversity follows as a prelude to further considerations of how insects
and other organisms respond to environmental conditions.

2.4. ECOLOGICAL DIVERSITY AND THE TROPICS

Pianka(63) reviews the major theories on the origin and maintenance of

tropical diversity, discussing the time theory, the theory of spatial heterogeneity,
the competition hypothesis, the predation hypothesis, and the climatic stability
hypothesis. None of these approaches is necessarily mutually exclusive, and it is
likely that the explanation of diversity must be considered for different sets of
taxa at a time and that each of the above explanations may hold to some degree in
each instance. The spatial heterogeneity approach professes that regions in the
tropics tend to be more diverse in terms of topographic relief and number of
habitats (macrospatial heterogeneity). The question of microheterogeneity is not
included (within habitat diversity). MacArthur(JS) has discussed the importance
of habitat heterogeneity in explaining tropical diversity.
The competition hypothesis advances the notion that more species compete
for limited resources in the tropics, thereby promoting ecological specialization
and ro<?m for more species to fit in. The idea was put forth by Dobzhansky,(J7)
who suggested that competition is the most important mechanism regulating the
high diversity of the tropics. The predation hypothesis, supported by data from
Paine,(64) Connell,(6S) and Janzen,(29) says that intense predation keeps popula-
tions of some species at certain trophic levels well below the carrying capacity of
the environment, thus allowing more space for other species to fit in. Whereas
the competition hypothesis requires some sort of equilibrium population sizes
among potential competitors to drive selection leading to specialization, the
predation hypothesis is essentially a nonequilibrium model. The climatic stability
explanation advances the idea that the generally stable and benign climates of
lowland tropical areas have fostered the evolution of specialized species, result-
ing in a greater number of species occupying various habitats. Under both the
competition and time stability hypotheses, ecological specialization is thought to
be the response to the environment for many species, thereby making it very
difficult to distinguish between the two hypotheses by controlled experiments.
The productivity hypothesis developed by Connell and Orias(66) predicts that
greater production results in greater diversity, all else being equal. Presumably
increased productivity provides the biomass necessary for eventual ecological
diversification into many species, each exploiting a different part of the environ-
ment. A corollary of this hypothesis is that the increased temporal heterogeneity
of the tropics provides different periods (seasons) to which different species in
the process of evolving can adapt, thereby allowing more species to fit into a
INDIVIDUAL AND POPULATION RESPONSES TO ENVIRONMENTS 27

particular region. The explanation may hold in those tropical regions where there
is considerable temporal heterogeneity.
This hypothesis is most interesting from the standpoint of primary produc-
tion and possible ramifications for the development of insect diversity. Pianka
suggests that plants use their own homeostatic mechanisms and storage capacities
to increase the stability of primary production, which in tum buffers plant species
against climatic variability. Under the productivity hypothesis it might be en-
visioned that herbivorous and pollinating insects form a major portion of the
sphere of selection pressures driving the diversification of plant species in the
tropics, resulting in an increased accumulation of both plant and insect species.
Increases in productivity by necessity involve decreases in reproductive ability,
as discussed by Cody(67) and others. If a greater amount of energy is allocated to
vegetative structures and maintenance, less is available for the elaboration of
gametes. Clutch size is a measure of reproductive ability. Examples exist of
organisms of reduced clutch sizes in tropical forms. Reduced clutch sizes imply
smaller populations and therefore more space for other species. Furthermore, the
reduction in clutch size reduces total energy requirements and increases the
chances to survive in less productive areas of the tropics (Pianka(63»).
MacArthur(5 ) points out that the majority of tropical environments are less
seasonal and more productive, and that the regions with strong seasonality and
lower productivity, such as tropical deserts and mountain tops, tend to have
fewer species. Clearly such patterns reflect differences in intrinsic ability of
individual genotypes or phenotypes in populations to cope with grossly different
sets of environmental factors. Thus a desert or mountain top in the tropics is
considered to be a marsh environment in the sense that primary production is
lower due to less plant biomass present in most such regions, and that the
constant or predictable severe climatic conditions establish an ecological bar-
rier for many taxa. One might predict that the reduced faunas of such regions are
in fact a select group of specialists in the broad sense, since the resource base is
restricted to relatively few classes of resource types, and that such species have
developed those physiological and morphological traits needed to thrive in such
climatically harsh regions. MacArthur(5 ) also points out that both large regions
and small areas in the tropics tend to have greater numbers of species than
comparable temperate ones. He explains these patterns by suggesting that there
has been a longer history of speciation in the tropics driven largely by interac-
tions among species. The increased rate of speciation in the tropics has resulted
in greater accumulation of species there; interspecies competition has precluded
the development of communities characterized by one or few dominant species.
Such effects are operative over small areas, resulting in great differences in
faunas among different areas. MacArthur believes that marginal habitats in small
areas are filled with species resulting as a spillover from the high productivity
and nonseasonal conditions of the lowland tropical rain forest environment in
28 CHAPTER 2

general. MacArthur's graphical representation of the accumulation of a greater

number of species in both large and small areas of the tropical environment is
shown in Figs. 2.4 and 2.5. That environments with higher productivity favor the
development of additional species capable of utilizing marginal resources is
shown in Fig. 2.6.
Depending upon the amount of phenotypic flexibility for exploiting a range
of resources by the members of an insect population, and spatial and temporal
changes in the abundance of the species, island populations of insects may be
more susceptible to both natural and artificially induced events that result in
successful extinction of the species. Susceptibility to extinction will vary through
the different stages of the taxon cycle. (68) Mainland populations of related or
identical species are more likely to possess a greater tenacity for maintaining
populations against environmental events promoting extinction. Presumably on
the mainland a species may occupy a range of microhabitats or exploit a wider
constellation of resources in some cases, although such patterns are certainly not
universally true for all insects in the tropics. If the species in question has a broad
geographical distribution and there are deliberate attempts to eradicate it (i.e., a

IX:

FIGURE 2.4. Relationship between area and abundance of insect species. The abundance of species
in a community is the result of both regional and habitat selection pressures determining the abun-
dance of species in an area. R is the rate of extinction plus emigration (curve E) and immigration plus
speciation (curve I). Lower extinction and higher speciation rates are predicted for the tropics,
resulting in a higher equilibrium number of species in an area than in a comparable Temperate Zone
area (dashed line). But historical factors may prevent communities in both areas from attaining such
equilibrium conditions. [From R. H. MacArthur, Bioi. J. Linn. Soc. 1:19-30 (1969). Reprinted with
permission. Copyright by the Linnean Society of London.]
 HABITAT

u p

FIGURE 2.5. Relationship between productivity and insect species richness in communities: A
model proposed to explain the local patterns of species accumulations in communities. In local
communities, a major factor determining species richness is the productivity of the community.
Recognizing unproductive (U) and productive (P) communities, the extinction rates will be different,
i.e., higher in unproductive communities. Immigration (I) falls off as the number of species increases
in either community as the result of competition among species. The net result is a greater equilibrium
number of species in productive communities. Relatively more productive communities may be
characteristic of some tropical terrestrial habitats, although a spectrum in productivity is expected for
the tropics. [From R. H. MacArthur, BioI. J. Linn. Soc. 1: 19-30 (1969). Reprinted with permission.
Copyright by the Linnean Society of London.]

FIGURE 2.6. Utilization efficiency of various species in a community (the small bell-shaped curves)
plotted against a spectrum of resources used by the species (the R axis) in a community. The large
curves depict densities of resources in productive and unproductive environments; T is the threshold
density needed for these species to maintain breeding populations in this community. Because the
resource productivity curve is wider in a productive environment such as the tropics, more resources
are available and more species can accumulate. [From R. H. MacArthur, BioI. J. Linn. Soc. 1:19-30
(1969). Reprinted with permission. Copyright by the Linnean Society of London.]
30 CHAPTER 2

pest species or disease vector) from a particular area of that range, such attempts
will undoubtedly be futile, as exemplified by the attempts to eradicate the
yellow-fever-carrying mosquito Aedes aegypti from Colombia. (69) Under such
conditions, reinvasion rates into an area where an eradication program was
erected will be frequent and often successful.
One of the major problems of interpreting the speculations of why a greater
number of species within a particular trophic level may occupy a particular
region or collection of habitats is the necessity to gather information on the extent
to which the species overlap in the exploitation of resources. The considerations
of MacArthur point to some overlap in resource utilization (Fig. 2.6), and classi-
cal ecological theory states that the populations of two or more species cannot
coexist if they overlap completely in their utilization of the environment.
Competition for such resources will probably either cause the potentially
competing species to diverge in habits or cause one species to become extinct.
Obviously, partial niche overlap is expected to be a common phenomenon in
natural communities, particularly in species-packed ones such as those associated
with certain kinds of habitats and collections of habitats in the tropics. Thus we
might predict some overlap in the guilds of host plant species utilized by different
species of Hemiptera or Homoptera through an array of contiguous tropical
habitats. Certainly the co-occurrence of high-density accumulations of the
nymphal skins of several species of cicadas at the base of Pentaclethra trees in
the Caribbean lowlands of Costa Rica, as reported by Young,(70,71) suggests some
level of niche overlap. The overlap in the case of these cicadas might be in terms
of emergence sites or feeding sites associated with the root crowns of certain tree
species acting as resource patches. But what is the distribution of individuals of
each of these cicada species over several different tree species in different
habitats? How much overlap is there? These crucial questions form the basis of
further study. If indeed certain tree species are limited resources for tropical
cicadas, to what extent has competition played a role in causing the observed
patterns of diversity, where diversity in this instance is defined simply as the
number of cicada species and their associated nymphal skin densities associated
with each tree species?
The familiar logistic equation of population growth is used as a means of
conceptualizing the degree to which a species has saturated the resources of the
environment. In this equation, dN/dt = rN [(N - K)/ K], dN/dt is the rate of
change in population size expressed as a function of the exponential growth
factor, rN, the instantaneous rate of increase (derived from life table parameters)
times the existing population size, and adjusted by (K - N)/ K, the amount of
resource space available to the popUlation now or at some future date. The
asymptotic condition derived from this expression (Fig. 2.7) shows the carrying
capacity, K, of the environment, the total number of individuals that the envi-
ronment can support. The value for K is highly sensitive to all environmental
INDIVIDUAL AND POPULATION RESPONSES TO ENVIRONMENTS 31

W
N
(ij
Z
o
~
..J

UNITS OF TIME (CENSUS HISTORY)

FIGURE 2.7. Alternate actions of colonization of a new habitat or adaptive zone. Many coloniza-
tions result in rapid extinction (curve I), while successful invasions may involve a period of over-
abundance (curve ill) with eventual attainment of equilibrium following a period of oscillations.
Presumably such populations are characterized by individuals programmed with very high levels of
reproductive potential and where biotic control mechanisms are generally lacking. Such a species
may be at an advantage if entering a community where competitive displacement can be initially
overcome by having a high r, or intrinsic rate of increase. Curve II is probably least typical for such
invasions, depicting a population census history pattern expected under conditions of high environ-
mental certainty and an absence of competition from resident species. *Date of founders arriving in
new environment.

factors and is expected to be in a dynamic state of flux in most habitats. Expla-

nations of high species abundance of insects and other organisms in the tropics
involving competition imply the existence of populations at or near their respec-
tive carrying capacities.
The individuals composing each popUlation are expected to have somewhat
different ways of utilizing the available resources in the environment, and this
range in variation, sometimes referred to as niche breadth, is dependent largely
upon the genetic and physiologicaVbehavioral differences among the individuals,
as developed below. Pianka has argued that there should be no gross differences
in micro spatial heterogeneity between comparable habitats in the Temperate
Zones and the tropics. Yet a major issue in understanding tropical species abun-
dance patterns is the determination of within- and between-habitat niche breadth
for herbivorous insects. Such data are needed to provide information on how
32 CHAPTER 2

diversity is maintained within habitats and through a collection of habitats (i.e., a

small area, sensu MacArthur(14).

2.5. SOME MECHANISMS UNDERLYING TROPICAL INSECT

DENSITY

Starting out with the assumption that the lowland tropical rain forest envi-
ronment is relatively benign to most organisms and that productivity is high, the
need arises to distinguish between the roles of competition and predation in
structuring the distribution of insect species, particularly herbivores and pol-
linators, over a collection of primary and secondary growth habitats. The
sweep-sample studies of Janzen(72) have provided considerable data on the
species abundance and trophic structure of insect communities associated with
different kinds of tropical vegetation and under different climatic conditions.
Such studies are an important step toward designing other studies manipulating
the potential effects of competition and predation. But the task at hand is im-
mense. It would be ideal to know something about the life tables of selected
herbivorous Coleoptera, Homoptera, and Hemiptera in such communities as well
as the degree to which each species exploits various plant species in the habitats
studied. Just as with the studies of host plant partitioning by Temperate Zone
Homoptera by Denno,(73) it would also be necessary to say something about the
ways each species utilized each host plant, particularly where considerable niche
overlap is evident.
It would be interesting if population densities turned out to be greater within
herbivorous Coleoptera for individual host plants where other species were ab-
sent. Such data would suggest competition. The ways in which a plant-feeding
coleopteran views each potential host plant species might be considerably dif-
ferent from the ways such resources are viewed by herbivorous hemipterans and
homopterans (sensu Janzen(72). Thus we might expect that individuals within
each taxon would respond differently to the distribution of the resources and that
there might be a general response pattern for each species during a specified time
period.
The following hypothetical but plausible example offers a basis for discuss-
ing how responses can be made by insects exploiting plants in tropical habitats. A
light gap within lowland primary rain forest encompasses an area of about 0.25
ha and has existed for 5 years. A chrysomelid beetle species A is a specialist on
one of several species of herbaceous pipers found in the light gap. The population
of the beetle is generally confined to the light gap, and individual host plants are
seldom defoliated, if at all. Apparently plant predation is kept in check in part by
the physiological needs of the beetles and by some level of predation by ants and
other organisms on the beetles. Other closely related and ecologically similar
INDIVIDUAL AND POPULATION RESPONSES TO ENVIRONMENTS 33

species of low understory chrysomelids occur in the surrounding forest. Suddenly

the forest is cleared away, and the light gap has become a large patch of secon-
dary growth. A second species of chrysomelid moves onto the host plant species
occupied by species A. Species B beetles possess some competitive edge over
species A beetles, and both exhibit a tendency to utilize the same parts of the host
plant. Neither is able to handle other parts of the host plant species. Eventually
the population of the host plant increases in size owing to more habitat being
available, and species B beetles are able to colonize a major portion of the plants,
while species A cannot. Species A beetles eventually shift to a second species of
piper, and the herbivore population takes hold and grows. Eventually each beetle
species enters into some sort of equilibrium with its host plant species. Ecologi-
cal isolation, in the form of divergence between the two beetle species for
different host plant species, has resulted presumably as a result of competitive
contact between the species. Niche overlap with respect to feeding is thereby
minimized as a response to an environmental perturbation. Given the dynamic
state of affairs of habitat disturbance in tropical regions, such ecological charac-
ter displacements may occur frequently in real systems. There is very little data
on such shifts. But the documentation of such ecological shifts in the distribution
of tropical herbivorous insects may provide information on response mechanisms
to changing environmental conditions.

2.6. STRATEGIES OF RESPONSE TO THE ENVIRONMENT

An insect or any other organism can respond to an environmental change

either genetically or nongenetically. It may ignore the perturbation and survive,
or go extinct locally. The response strategies are of interest here. Natural selec-
tion is defined as the differential survival of individuals through different genera-
tions within a popUlation of a species. Many kinds of nongenetic variation
influence popUlations.
1. Variation in individuals through time, including age variation, resulting
in different niches being utilized, as seen in the life cycles of many
holometabalous insects; seasonal variations in the individual and whole
generations of individuals such as different phenotypes in insects with
seasons, including color forms and winged and wingless forms.
2. Social variation such as caste structure in some social insects.
3. Ecological variation, including shifts in habitat, flux of temporary
habitats, some types of density-dependent variation, growth form varia-
tion, and host-induced variation.
The above sorts of variation observed in insects and other organisms need
not involve changes in the genetic makeup of populations, as they are short-term
34 CHAPTER 2

adjustments to changing environmental conditions. That is not to say, however,

that such responses may not eventually result in genetic changes over long
periods of time. Obviously, with few exceptions such as monozygotic twins (and
even here there might be somatic mutations), no two individuals in sexually
reproducing species are genetically identical. Some portion of this genetic varia-
tion among individuals in a population contributes to variation in the phenotypes
present in the population. Phenotype is defined as the morphological, physiolog-
ical, behavioral, and developmental traits of the individual. It is the result of both
genetic, somatic, and nongenetic factors influencing the structure, biochemistry,
and behavior of the organism. Sometimes the variation from individual to indi-
vidual within a population is rather discontinuous or abrupt with respect to the
character states studied, a condition sometimes referred to as phenotypic
polymorphism. Often the variation is continuous among individuals and far more
difficult to detect. Discontinuous genetic variation, often a basis for observed
discontinuities in phenotypes, is an intrapopulational phenomenon.
The genetic mechanisms underlying some phenotypic polymorph isms are
complex, sometimes, in the case of insects controlled by large series of multiple
alleles or by pseudo alleles at closely linked loci. A possible selective advantage
of a large number of alleles being involved is the increased likelihood of reducing
the frequency of homozygotes through successive generations. Under some en-
vironmental conditions an absence of homozygotes might be advantageous. In
fact, the most frequent phenotype in a polymorphic population is by no means the
result of a dominance effect of particular alleles. Genetic dominance is not
equivalent to numerical dominance of a particular allele or linked alleles. In some
cases the most common allele in a series of multiple alleles is a recessive,
possibly maintained by a selective advantage of numerical equivalence of various
phenotypes. Furthermore, what might appear "recessive" in terms of visible
components of the phenotype may in fact be dominant in terms of the physiologi-
cal phenotype, especially since most genes are selected for in terms of physiolog-
ical effects.
Theodosius Dobzhansky and his co-workers have investigated in both field
and laboratory studies the systems of chromosomal polymorphisms found in
various species of Drosophila flies. Drosophila and other dipterans such as
Chironomous and Anopheles possess giant salivary gland chromosomes-
polytene chromosomes-which permit the direct observation of gene sequences
as the result of differential staining properties of the chromosome material,
leading to distinct banding patterns associated with different loci. A significant
discovery from both temperate and tropical studies of natural populations of
Drosophila species by Dobzhansky and Pavad74 ) and Da Cunha and Dob-
zhansky(7S) and others was the demonstration that many different gene arrange-
ments can exist within a population, suggesting the maintenance of considerable
genetic polymorphisms in populations. For example, the third chromosome of
INDIVIDUAL AND POPULATION RESPONSES TO ENVIRONMENTS 35

Drosophila pseudoobscura has at least 16 known gene arrangements and

Drosophila simulans more than 50. Such variation is different from variation
arising from single loci, since in the former, whole gene arrangements are in-
volved. Some of these studies revealed that central populations tend to be
chromosomally polymorphic while marginal popUlations tend to be
monomorphic. Since chromosomal inversions tend to suppress crossing-over and
therefore recombination, such spatial changes in the genetic structure of these
insect populations, particularly in Central and South America, were considered
as highly adaptive; populations entering into or occupying marginal habitats
would have the genetic resiliency (through recombination) to cope with new
environmental conditions, whereas central popUlations existed in more or less
stable conditions and locked gene arrangements in the form of chromosomal
polymorphisms preserved the genetic component of the phenotypic architecture
best suited for those conditions. A difficulty with such studies is the challenge to
correlate specific portions of the phenotypes in such different populations with
specific factors in the environments. As with loci on structurally normal chromo-
somes, the matter is confounded further by pleiotropic effects.
The field experiments of Lewontin and Birch(76) and others have shown that
in both temperate and tropical environments evolutionary change does not
imply the gradual perfection of a particular genotype and resulting phenotype,
but rather the trends toward maintaining some level of genetic variation in the
populations. Thus if some genotypes of a tropical nymphalid butterfly endow
drought-resistant character states to the phenotypes, and others endow optimal
traits for humid conditions and the population exists in a humid or wet forest
environment, the species may have a chance to survive an unusually long dry
season, particularly if the immature stages are incapable of entering into
diapause. The type of environmental saturation predicted by MacArthur<15,16) for
the formation of species to occupy marginal habitats or resources under condi-
tions of high productivity implies a need for some sort of mechanism to allow
species to shift to new resources. A popular view is that genetic variation in
populations permits such events to occur. The colonization of marginal habitats
is made possible by the maintenance of genetic variation or some form of
phenotypic variability, perhaps non genetic . Genetic variation arises from various
sources, including the integrity of genetic factors such as alleles capable of
changing in frequency and kind as the result of mutation, gene flow, and recom-
bination, and the preservation of these sources of variation by behavioral and
ecological factors as well as cytophysiological factors.
Barring inevitable consequences of small popUlation size, maintaining some
degree of phenotypic flexibility or variability within breeding popUlations would
appear to be ecologically and biogeographically advantageous for species. It is
almost trivial to say that all environments change and therefore species in them
change correspondingly. Maintenance of some flexibility allows the individuals
36 CHAPTER 2

in a population to "predict" future environments in the sense of being

preadapted in some ways to environmental perturbations. Thus we know that the
colorful pierid butterfly Perrhybris lypera thrives as a larval specialist in the
understories of some Central American rain forests. Young(771 has described
some of the specialized phenotypic attributes of this species, including a highly
specialized form of oviposition and utilization of host trees in the lauraceous
genus Ocotea. Presumably such a species has experienced a long evolutionary
history in this environment. If this species possesses the ability to survive some
forms of environmental perturbation due to genetic characteristics, then one
might predict that some level of genetic variation is maintained. That level of
genetic variation might be greater than some minimal "noise" level that one
might expect most species to possess, as the ultimate device for coping with
environmental changes. Since Perrhybris like most other organisms orchestrates
its own environment, and orchestration of the environment for a single species
depends upon the activities of other species and the physical environment, it is
really the individuals composing the popUlation that possess the ability to fit into
the environment wherever possible, and in such a way as to maintain a breeding
popUlation. This is another way of saying that each species, especially in those
environments where species abundance is high, essentially constructs its own
environment, as opposed to adapting to it per se.
In such cases the tools of environmental construction include a combination
of genetic and non genetic characteristics, some of which vary from phenotype to
phenotype within the population. The classic models of selection summarize
three basic ways that the genetic component of popUlations responds to environ-
mental changes. An insect specialist can be predicted to experience strong
stabilizing selection in which most individuals in the popUlation possess the
optimal phenotype (Fig. 2.8). The response of different genotypes to selection
and the operation of ecological constraints on population size impose some loss
of fitness to the population, but otherwise the trend is toward decreasing
phenotypic variance. In terms of herbivorous tropical insects, one might antici-
pate stabilizing selection to be most operative in instances where a specialist
insect species possesses a relatively low reproductive rate, to have good dispersal
tendencies to locate patches of host plants, and to tend to utilize large patches of
host plants. The latter trait lowers the likelihood of overcrowding and density-
dependent selection imposing high losses of individuals as a food supply dwin-
dles.
The case of chrysomelid beetle species A and B described earlier in this
chapter provides an example of directional selection, the instance in which a
species makes an ecological shift to a new resource, favoring a different
phenotypic class. Environmental perturbations may result in episodes of direc-
tional selection (Fig. 2.9). There is a shift in the optimal phenotype. Thus the
shift of beetle species A from its original host plant to a second species of host
 Op
PHENOTYPIC SCALE

FIGURE 2.S. Relationship between stabilizing selection and phenotype variation. Stabilizing selec-
tion, operative through successive generations within a breeding population, results in a reduction of
the range of phenotypic variation in the population. Some optimal phenotype, Op, and a small range
of phenotypes around it, is optimal for the environment. Stabilizing selection is expected for habitats
or environments in which selection pressures have been predictable over long periods of time. Yet
most populations will experience more than one general kind of selection, and the phenotype in any
one generation represents a compromise of such conditions.

DIRECTIONAL SELECT ON

OPl OP2
PHENOTYPIC SCALE

FIGURE 2.9. Directional selection. When one resource base is limited and the species can poten-
tially utilize a second kind of resource in the same environment or habitat, assuming that the species
possesses the population growth potential to exploit both kinds of resources, and the population is
expanding at such a rate as to have the potential to overexploit the original resource, natural selection
may result in a switch to the second resource if it is more abundant than the first and if the first
resource is dwindling in supply. Thus through time, one kind of optimal phenotype (Op!) is replaced
by another (Opz), representing an adaptation to a new resource.
38 CHAPTER 2

plant depended upon residual phenotypic variation in the population that resulted
in another phenotype becoming optimal for the new host plant species. In such
cases we are talking about the sum total of those biochemical and physiological
characters that allow a herbivorous insect species to feed successfully on a new
resource, without appreciable losses in fitness. We examine some aspects of this
topic in the next chapter. Depending upon the source of the phenotypic var-
iability in the population, a specialist insect species mayor may not be able to
undergo such shifts. If Perrhybris is virtually totally genetically specialized to
feed on Oeotea leaves as a larva, as a result of strong stabilizing selection,
insufficient genetic variation in the popUlation may restrict exploitation of other
lauraceous and nonlauraceous host plants should the Oeotea population be elimi-
nated in favor of cacao or bananas. The first response might be nongenetic rather
than behavioral and physiological, but if it fails, genetic response becomes
critical if there is sufficient time to adapt. If the genetic variation is not there,
there would be little or no opportunity for directional selection to favor another
phenotype in that habitat or region, and the species would become extinct or
migrate.
An intermediate step, in the case where directional selection is possible, is
the breaking down of a population structure in which one phenotype is consid-
ered optimal. This process is known as disruptive selection (Fig. 2.10). Disrup-
tive selection can be viewed from the standpoint of plant-herbivore interactions
as the initial stages of a divergence in host plant exploitation patterns, away from
one host plant previously considered optimal under previous environmental con-
ditions. If sufficient variation is present, a species might attempt to exploit
several alternative resource types, a stage of disruptive selection. If suitable for
maintaining a breeding population in terms of interactions with other species and
the physical environment, the genetic structure of the insect population might be
stabilized with a series of several optimal phenotypes, each one associated with
one or more similar resource classes, together which provide the species with
sufficient resources to maintain a breeding popUlation. If not all phenotypes are
optimally suitable for maintaining the population, endowing grossly different
levels of survival or reproductive capacities, directional selection is expected to
take over if one of the phenotypes is better than the others in the ecological shift.
Thus there might be selection for the Neotropical pierid Dismorphia virgo
to exploit several species of Inga trees in a mountain forest ravine, but individu-
als associated with one species in particular do better than the others. Perhaps the
leaves are more digestible or contain fewer defensive systems, placing less of an
energetic drain on the species. The original impetus for an ecological shift might
have resulted from a disappearance of the original host plant, changes in population
age-structure of the original host plant, or the arrival of a new herbivorous
species on the original host plant, leading to competition. All three modes of
selection are operative in natural popUlations of insects and other organisms, and
INDIVIDUAL AND POPULATION RESPONSES TO ENVIRONMENTS 39

DISRUPTIVE SELECTION

PHENOTYPIC SCALE

FIGURE 2.10. Disruptive selection. If a population is expanding at a rate greater than what can be
thriving on a single resource base, and if there is some phenotypic variation present in the population,
selection may favor the exploitation of a second kind of resource in the same environment, if it is
present and accessible to the species. This disruptive selection results in a divergence into two
optimal phenotypes, Op\ and 0p2, within the species, so that the population's membership exploits
two different sets of resources. Disruptive selection of this sort may also result from intense competi-
tion associated with an original resource, selecting for the exploitation of a second, more cryptic
resource within the same region or habitat.

because of the high species abundance of insects in many regions of the tropics,
such forces undoubtedly contribute to molding the resource associations of insect
species there.
But it cannot be assumed that all forms of environmental orchestration in the
tropics are of this origin. A great deal of so-called adaptation is probably taking
place at the level of the phenotype, without genetic change. We can say that the
implied cornerstone of Darwin's theory of evolution is that some sort of variation
exists within populations of species, and this spectrum of variation permits
responses to environmental changes that may otherwise result in extinctions. The
models of selection discussed as well as other factors such as drift affecting
populations tell us that allelic frequencies, and therefore genotypic distributions,
change in different generations within a population. And such changes represent
significant departures from the constant proportions in allelic frequencies pre-
dicted by the familiar Hardy-Weinberg equilibrium. Where p is the frequency of
one allele in a two-allele system and q is the frequency of the other, the distribu-
tion of genotypes for each generation is predicted to be
p2 + 2pq + q2 = 1
40 CHAPTER 2

The model assumes infinite popUlation size, random mating, and constant en-
vironmental conditions. It is clear from the foregoing discussions that popula-
tions and environments simply do not have such properties, and therefore the
various factors affecting the genetic structure of populations discussed above
provide a basis for a more complex approach to how individuals respond
phenotypically to changes in the environment. Later, we will examine some
effects of resource patch structure on the genetic and nongenetic response pat-
terns of species. Our emphasis will be on the patterns of resource utilization by
insects feeding on plant species in tropical habitats.

2.7. NATURAL SELECTION AND PROPERTIES OF POPULATIONS

The basic model of environmental change response by organisms is the

effects of selection on individuals within a population. Such a model has these
components: (1) a population consists of individuals, (2) there is genetic varia-
tion among the individuals, (3) differential fecundity and/or survivorship among
the individuals biases allelic frequencies, and (4) adaptations evolve that enhance
individual fitness. A model for the adaptation of whole popUlations as the units of
selection substitutes "group" or "population" for "individual" and "commu-
nity" for "population" in the above statement. Lewontin (78) has reviewed the
properties of selection acting on individuals and whole popUlations. Others (e.g.,
Wade and McCauley(79); Wilson(80l) have discussed some mechanisms on the
evolution of phenotypic features at the popUlation level and have argued that whole
communities can be selected for.
Considering the above model for the response patterns of individuals within
popUlations to environmental changes, several studies have shown that various
forms of frequency-dependent selection exist in populations. Much of this work
comes from studies of insect popUlations in general. Ayala(SI.s2) showed that
when Drosophila populations become very crowded under laboratory condi-
tions, evolutionary changes occur within the populations that allow densities to
increase. In such systems the adult flies appear to have evolved an increased
"tolerance" to high densities: while there are no increases in the fecundity or
survival of larvae and pupae in such cultures, adults live longer and the number
of flies emerging each week remains constant. Such changes in adult flies are
believed to be due to a shift in gene frequencies, and therefore phenotypes
(assuming that individuals within the popUlations are responding genetically to
increased densities) as a function of population density. It is the interaction of
population growth in density and natural selection that brings into play a
frequency-dependent selection under these conditions. The selective value of
genotypes changes with density changes in the population.
Anderson(83) has examined how the genetic structure of a popUlation might
INDIVIDUAL AND POPULATION RESPONSES TO ENVIRONMENTS 41

change during the usual logistic growth pattern (Fig. 2.7) and concluded that the
population during the phase of maximal growth is best described as the weighted
arithmetic mean of genotype rates, and that the carrying capacity phase is the
weighted harmonic mean of the individual carrying capacities of genotypes.
Natural selection alters allelic frequencies so as to maximize major parameters of
population growth at critical times during the census history of a population.
Lewontin(84.85) has examined the effects of increasing density on viability in
Drosophila, and such studies generally show a relationship between density and
selection. Lernd86 ) discusses some genetic properties that tend to mullify the
effects of selection that change with changes in population density and other
factors. Mechanisms such as recurring mutations and nonadditive genetic var-
iance reduce or neutralize the effects of selection precluding the ability of the
species to respond to environmental changes.
In addition to population density invoking frequency-dependent selection,
the genetic structure of populations is also altered by a different kind of environ-
mental factor, namely, interspecies competition. We envision that insect popula-
tions, at different times, and under different conditions, will be subject to the
selective forces invoked by both intraspecific and interspecific competition. The
degree to which these forces operate on particular populations is obviously a
function of many factors, including resource availability relative to population
growth parameters for a single plant-feeding species, the impact of biotic and
abiotic mortality factors on the herbivore popUlation, resource accessibility
(patch structure) in time and space, phenotypic flexibility enabling the herbivore
to utilize other resource types of precluding it from doing so, and the presence of
competing species or potentially competing species if the environment changes.
A specialist herbivorous insect having a popUlation structure characterized
by either intense predation on immature stages or a very low reproductive rate,
and exploiting an abundant food plant species having an accessible patch struc-
ture, may be occupying an essentially empty universe. Such conditions approach
Anderson's(83) weighted arithmetic mean of genotype growth rates underlying
overall population growth. This period of rapid population growth is analogous
to the early phase of a colonizing episode.(87) If mortality is lacking and reproduc-
tive rates are high, intraspecific effects may increase, especially if dispersal ability
and success likelihood are low. Assuming a sufficiently short generation time,
such a condition favors selection for density-tolerant genotypes as the population
approaches its carrying capacity. Thus one might argue that the ithomiid butterfly
Hypothyris euclea in the premontane tropical wet forests of northeastern Costa Rica
sometimes experiences rather intense colonizing episodes, with increased shifts in
breeding populations from small, disturbed habitats in forest understories to expan-
sive secondary habitats largely dominated in some areas by its larval host plant Sol-
anum rugosul11. The medium-sized tiger-striped butterfly (Fig. 2.11) places very
large numbers of eggs on the host plants, and larval survival is high (Young(88,89»).
42 CHAPTER 2

FIGURE 2.11. The Neotropical butterfly Hypothyris (Ithomiidae) exploiting decaying animal tis-
sues as a food source in Central American tropical rain forests. (a) H. euclea feeding on a piece of
trochanter muscle from a grasshopper leg and (b) feeding on the soupy tissues of a Eurytides pupa
(Papilionidae: Papilioninae) previously gouged out by a bird.
INDIVIDUAL AND POPULATION RESPONSES TO ENVIRONMENTS 43

Although there is some egg mortality, presumably from a pathogenic fungus,

adult densities in secondary habitats are considerably higher than elsewhere,
where smaller breeding populations occur. Within a few weeks following a
concerted burst of egg-laying activity, a high percentage of the host plant indi-
viduals are severely defoliated by larvae (Figs. 2.12 and 2.13), and subsequently
the butterfly population declines rapidly. (89) The system appears to be an example
of high population growth ability of the herbivore, but with an absence of
density-related mortality factors, eventually resulting in a density-related deple-
tion of the resource base. No other ithomiids use this plant species as a larval host,
and few other herbivores are present. The tough, sandpapery leaves may allow
this plant species to drop out as a major resource for most insects, and presum-
ably the cluster oviposition of H. euclea has evolved as a means of ensuring
group feeding in the larvae, an activity which probably involves cooperative
processing of leaf tissue before it is digested. Most other ithomiids place eggs
singly on host plants and scatter them widely. Another, season-related case is
that of the dry-season defoliation of the ornamental monocot Canna edulis (Can-
naceae) by the larvae of the skipper butterfly, Calpodes ethlius, at Mazatlan, in
Sinaloa, Mexico. Although the large final instar larvae and pupae experience
severe parasitism by a chalcid wasp, earlier instars severely defoliate most
patches of this ornamental during the early dry season (Fig. 2.14), and some
preliminary evidence suggests a reduction in fruit set resulting from this preda-
tion on C. edulis (A. M. Young, unpublished data).
It was J. B. S. Haldane who suggested that when population size or density
is small, intraspecific competition is greatly reduced, thus permitting the evolu-
tion of those phenotypic traits responsible for increased success in interspecific
competition. The laboratory experiments of Park(S8), Lernet86 ), and others with
Tribolium and other insects have pointed to the role of initial popUlation size and
genetic composition of populations as major determinants in the outcome of
interspecific competition. Depending upon the amount of genetic variability
present in the initial populations of each species and initial popUlation sizes,
obtaining reverse outcomes in the competition process is possible. The
Drosophila competition studies of Ayala(90) demonstrate that the rarer species can
evolve a competition advantage and reverse its density or even outcompete a
formerly superior competitor. When population size or density becomes very
low, perhaps as a result of intense competition with another species or from other
factors, an increased competitive ability may evolve during this period when the
population was on the verge of extinction.
Although the role of interspecific competition in tropical communities is not
really understood, the above considerations elucidate the ways this ecological
process can mold the pattern of resource utilization by closely related or ecologi-
cally similar species. Competition among species can be searched for in terms of
herbivorous insects associated with plant species. As more and more tropical
44 CHAPTER 2

FIGURE 2.12. Gregarious behavior in the Neotropical ithomiid butterfly H. euclea. (a) Eggs
deposited in large cluster on a leaf of the larval foodplant Solanum rugosum (Solanaceae). (b) Third-
instar larvae and gregarious defoliation of food plant leaves.
INDIVIDUAL AND POPULATION RESPONSES TO ENVIRONMENTS 45

FIGURE 2.13. (a) A natural habitat of H. euclea in premontane tropical rain forest in northeastern
Costa Rica. Low-density adult populations are associated with clearings and light gaps where scat-
tered individuals of the food plant occur. (b) Defoliation by larval H. euclea typically seen in
secondary growth habitats where the food plant occurs in large patches. Such defoliation is in-
frequently seen in forest habitats.
 46 CHAPTER 2

FIGURE 2.14. The Neotropical hesperiid butterfly Calpodes ethlius defoliating Canna lilies at
Mazatlan. State of Sinaloa. Mexico during the dry season. (a) Adult. (b) dead pupa showing parasitic
wasp (probably braconid) exit hole. (c) defoliated leaves. and (d) patch of larval foodplant.
INDIVIDUAL AND POPULATION RESPONSES TO ENVIRONMENTS 47

FIGURE 2.14. (Continued)

forest habitats are altered and converted into secondary or agricultural habitats, it
is possible to study the possible role, or lack thereof, of the competition process
as insect species colonize resources in new habitats. In some ways this situation
is analogous to the taxon cycle phenomena associated with the biotas of oceanic
islands rather than habitat islands on tropical mainlands.
The taxon cycle is a succession of usually four distinct phases or stages in
the change in the composition of the biota on an island and/or islands. Wilson(91J
and others have developed the concept of the taxon cycle to explain the observed
distributions of ants and other organisms on various archipelagos. A discussion
of taxon cycles logically follows a discussion of how competition alters the
genetic structure of popUlations, since once again we are examining the question
of what forces govern the distributions of species. How insects experiencing
colonizing episodes respond to new environmental conditions is a function of
both the genetic makeup of the colonizing individuals and the resource
heterogeneity and diversity of the native (resident) fauna of the area being col-
onized.
Lewontid87l has shown that genotypes endowed with the means to promote
rapid popUlation growth make good colonizers to new environments. An array of
colonizing individuals programmed for high population growth will be able to
take hold of a new resource and even possibly dislodge species already using it.
Thus genotypes selected for colonizing ability make initially good competitors
when entering a new adaptive zone. Such genotypes are endowed with the ability
48 CHAPTER 2

to produce phenotypes capable of being generalists and taking hold of an array of

resource types in the new environment. Such ecological release is characterized
by population expansion into several habitats and gives the species a broad niche
breadth. But such conditions are usually ephemeral, since this form of "swamp-
ing" competitive ability soon wanes as the native biota, particularly ecologically
similar species on the one hand and predators and parasites on the other, respond
by ecological shifts to compete with the colonizer. Thus an initially broad habitat
distribution contracts, and eventually local extinctions of the colonizing species
prevail. The colonizing population is conceptualized as passing through these
stages of the taxon cycle: (1) expansion of the population into many habitats and
perhaps over several islands, (2) differentiation of the populations in each habitat
or island to adapt to local environmental conditions, (3) spotty extinctions of
some popUlations, resulting in a fragmented distribution of the colonizing
species, and (4) an expression of endemism in the surviving popUlations as
differentiation continues and the genotypes present adapt to prevailing condi-
tions. This final stage represents the evolution of a different kind of competitive
ability, one that selects for genotypes most adaptive to the species when the
population is at or near its carrying capacity, following the considerations de-
veloped by Anderson<83) and others such as Gadgil and Bossert. (92)
In terms of energy allocation, one might predict that selection favors those
genotypes in the populations that enhance adaptation to the properties of a con-
tracted niche for the species, including resource specialization and development
of defensive mechanisms to discourage predation. Again, the colonizer becomes
part of the environment, and the environment is no longer new to the arriving
species. This is another way of saying that species themselves are the architects
of their own environments. Thus fitting into a habitat does not necessarily imply
that a species is adapting to the environment at hand, but rather that it responds to
environmental factors in such a way as to siphon off the needed resources while
avoiding mortality-generating factors that would significantly lower the size of
the population to a point where extinction would occur.
It is not difficult to extend such considerations of taxon cycles to the ques-
tion of herbivorous insect species in the tropics undergoing ecological shifts to
either new host plant species in different habitats or to new populations of the
same host plant species in new habitats. The high species density condition of
some guilds of insect species associated with certain host plant species may select
for regular or episodic colonization attempts when interspecific competition be-
comes intense. Yet from the above considerations we would expect that those
genotypes present that are best adapted to competitive interactions when popula-
tions are at or near carrying capacity levels are not the genotypes best suited for
colonizing episodes. Perhaps for this reason we might expect popUlations under
such conditions to maintain some genotypes programmed for colonization, i.e.,
INDIVIDUAL AND POPULATION RESPONSES TO ENVIRONMENTS 49

those endowing phenotypic traits such as high reproductive ability and good
dispersal abilities. Such residual variation in the population, coupled with disrup-
tive and then directional selection, could bring into play the colonizing strategy.
Those genotypes that optimize the ability of individuals to overcome defensive
systems of the host plant, to maximize the retrieval of nutrients from the host
plant with maximal efficiency, and to have the ability to act behaviorally and
physiologically in a manner that ensures survival in the face of competitive
threats, represent one segment of the mixed strategy calculated to exist in most
insect populations.
The other segment of the population might best be adapted to dispersal,
increased energy allocation to reproduction, greater flexibility in feeding on more
than one resource type, and a lack of ability to compete with ecologically similar
species. When the colonizing genotypes arrive on the new resource, a taxon
cycle might be set into motion, depending upon the guild of insects already
associated with the resource and other environmental factors. We might also
expect that the colonizing genotypes are better able to withstand new features of
the physical environment.
The counteradaptation that fonus the last stage of the taxon cycle, in which
resident species restrict the niche breadth of the colonizer, may result in an
equilibrium being established with a new species added to the community. This
situation is analogous to MacArthur's idea of how new species are added to
communities in the tropics. The equilibrium is established not only with other
ecologically similar species but also with predators and parasites affecting popu-
lation size. Sometimes, however, the counteradaptation may result in the col-
onizing species becoming extremely rare, in which case the selection pressures
are released and the species can once again increase in abundance, and the cycle
begins all over again. Such secondary expansions of species are well documented
from actual archipelagos such as the West Indies, but more studies are needed to
examine such cyclic phenomena in the dynamics of insect communities as-
sociated with different tropical habitats.
Janzen(93J has analogized the properties of colonization as suggested for
islands with the issue of how insect species colonize host plants. A major aspect
of the island model is the explanation of the number of closely related species
found among islands in an archipelago. The discussion here has focused on the
differentiation of populations either by habitat or individual host plant species
within a habitat. Clearly such conditions, coupled with mechanisms of reproduc-
tive isolation, can result in speciation in the case of tropical herbivorous insects.
But the discussion fares just as well in tenus of such patterns representing ways
that insects respond to the environment, without considering speciation. The
speciation process undoubtedly involves many other factors as well, as discussed
in Mayt94J and elsewhere.
50 CHAPTER 2

2.8. ENVIRONMENTAL PREDICTABILITY AND RESPONSE

PATTERNS

From these considerations, it is seen that tropical terrestrial habitats, with

their varying diversities of plant species and patch structure of plant species,
represent complex environmental grains promoting a spectrum of adaptive path-
ways and response patterns in herbivorous insects, pollinating insects, and pred-
atory and parasitic insects exploiting the former two groups. Particularly in high
species density habitats such as some of those found in the lowland tropics, it is
to be expected that not all individuals within a population of a particular insect
species are adapting to the same features of the environment. Rather there exist
mixed strategies of environmental response requiring mixed sets of behavior.
Different phenotypes are interacting with the environment in different ways.
Concomitant with such diversified response patterns within a population is
the development and expression of an integrated breeding system that ensures
that some portion of the phenotypes reproduce each generation. Mixed behaviors
may indicate a mixed pattern of adaptation to environmental factors. The tropical
cicada Zammara smaragdina is a large, sexually dimorphic, mottled green and
black insect (Fig. 2.15) with a very distinct song. Most adults emerge within
primary or advanced secondary forest (more than 25 years old) habitats, and few
if any emerge in more disturbed habitats (Fig. 2.16). In forest habitats males
form conspicuous clusters in trees, where they chorus synchronously throughout
most of the day. However, many individuals sing individually in isolated small
trees in open pastures and near dwellings. The two very different singing be-
haviors seen in this cicada may represent a mixed response to the environment.
Perhaps males in large choral groups in or near breeding habitats represent one
response to the need to attract females and mate, and the disjunct and isolated
distribution of others in open fields is another response to ensure that females
"missed" by the first strategy are mated by the second form of singing behavior.
It is noteworthy although not necessarily surprising that males emerge before
females and wait for the latter to emerge. The two different forms of singing
behavior may represent two different mechanisms to increase the likelihood of
matings to achieve a frequency necessary for maintaining the popUlation. The
conspicuous strong sexual color dimorphism in Neotropical Parides butterflies
(Fig. 2.17), coupled with the marked behavioral dimorphism between the sexes
in which males patrol specific portions of the habitat to court females (A. M.
Young, unpublished data), is another example of sex-related mixed strategies in
tropical insect populations. There are other examples. In still other butterflies of
the tropics, the adult sexual differences in wing coloration are only slightly dif-
ferent, as shown by Morpho peleides (Fig. 2.17), although in other species of
Morpho such differences are more pronounced.
The sometimes tremendous size range in a collection of adult insects of one
 2
c
<:
6
c
»
r-
»
z
c
~
o
~
c
S
(5
z
,;J
rn
~
zrJl
m
rJl

d
m
~
:ii
o
z
3:
m
Z
iii

11111111111111111 1IIIII1 i 11'1

i 51 t;!
II \til ~1
METRIC 1 2 3 41 7 81 91
FIGURE 2.15. The Neotropical cicada Zammara smaragdina, from premontane tropical rain forest in northeastern Costa Rica. The
species is markedly sexually dimorphic in body coloration: the male (two specimens to the right) is bright green and brown, while the <II
female (left) is olive green and brown.
...
52 CHAPTER 2

FIGURE 2.16. The Neotropical cicada Z. smaragdina (inset). which is associated with mixed
primary and advanced secondary forest in northeastern Costa Rica as well as some areas of primary
forest dominated by the canopy tree P. macro/aha (Leguminosae). Shown here is a secondary habitat
of this insect, dominated by the tree G. meiantha (Tiliaceae). The subterraneous burrowing nymphs
of this cicada are believed to associate with the root crowns of G. meiantha and P. macrolaha while
secondarily exploiting intermingled root systems of other plants.

species from a habitat may reflect differences in what individuals do with the
environment. In northeastern Costa Rica within a 2-week period near the end of
the short dry season in one year, there was a 20% range in size differences in
forewing length for male individuals of the butterfly M. peleides for a total of 23
individuals measured and released, and with sizes clustering into large and small
individuals. One can argue that such variation reflects differences in larval nutri-
tion or growth efficiency, but laboratory-reared M. peleides using one host plant
species and under the same environmental conditions results in similar size
variation in male butterflies. Mixed sizes may reflect mixed adaptations to mat-
ing and other activities, and such variation can be assumed to be an advantage for
the species in a particular habitat.
Because insects tend to have short generation times, be small in size, and
exhibit rapid increases or changes in population census histories that correlate
INDIVIDUAL AND POPULATION RESPONSES TO ENVIRONMENTS 53

FIGURE 2.17. Two examples of NeotropicaI butterflies (Lepidoptera) exhibiting coevolved larval
food plant associations in tropical forests. (a) Morpho peleides (female above, male below) exploits a
variety of Leguminosae as larval food plants and belongs to the superspecies Achilles, which has a
broad geographical range in Central and South America. (b) Parides iphidamas, (female above and
male below) and another one to four species of the same genus, depending upon the locality, share a
variety of species of Aristolochia (Aristolochiaceae) vines as larval food plants, although habitat
differences often separate some populations in an area.

well with abiotic and biotic factors of the environment,(9S-97) the observed ex-
perimental results of rapid responses to selection in laboratory studies certainly
justify the assumption that insects exhibit evolutionary changes in popUlations in
response to changing environmental conditions. Such results do not mean, how-
ever, that all observed response mechanisms in insects and other organisms
represent deep-seated genetic changes in the popUlation through time. Given the
long history of migratory ability, one might expect that the changes in physiolog-
ical conditions in overwintering populations of the monarch butterfly, Danaus
54 CHAPTER 2

plexippus, as studied by Tuskes and Brower,<98) do in fact represent stable por-

tions of the genotype. Migratory behavior in this insect is a fairly predictable
event through successive annual second-generation populations. Thus a major
portion of the physiological traits accompanying overwintering behavior are
believed to be directly controlled by genetic factors, resulting in a good correla-
tion between genotype and phenotype with respect to these traits. Such reasoning
may eventually prove to be false, but there is no reason not to believe that some
portion of the phenotype consists of stable traits endowed by the genotype. In the
case of the monarch, some additional physiological "slop" is also undoubtedly
brought into playas a trigger of genetically programmed physiological changes
associated with overwintering. Such additional factors may represent nongenetic
responses to changing conditions in the environment that bring into play the more
extensive physiological remodeling associated with overwintering.
Many other forms of environmental response by insects may operate at the
phenotype level of organization rather than at the genotype level. Whole collec-
tions of species may exhibit such response systems. It is not unreasonable to
suggest that short-term and rapid shifts in tropical insects from specialist to
generalist feeding patterns and vice versa are commonplace and represent be-
havioral and physiological shifts rather than genetic changes in the population.
Slobodkin and Rapoport<99) have presented the various response patterns that
organisms can make following an environmental perturbation (Fig. 2.18). They
distinguish between individual and population responses to the perturbation.
Population responses are relatively slow responses involving the reshaping of life
table parameters and genotypes in the popUlation. Such adaptations are believed
to evolve over several generations if the selection pressure arises frequently in the
population and if behavioral and physiological responses do provide the quick
recovery mechanisms following each perturbation. Behavioral and physiological
responses enable individuals to shift habits rapidly without any appreciable loss
in fitness.
Clearly there are many examples of generally species-specific behaviors in
the insects, such as the one shown by the analysis of mating behavior in the
dipteran genus Drosophila by GrossfieldOOO ) and others. Within species, how-
ever, we might expect some degree of response plasticity, depending largely
upon major shifts in types of food and other resources. It is not surprising to note
that complete diet shifts in some tropical organisms from fruit to prey
(Smythe([O])) represent a facultative response to shifting availabilities of fruit.
Many changes in the environment represent ephemeral events rather than perma-
nent changes, and such conditions necessitate only a short-term response by
species found there. In the model developed by Slobodkin and Rapoport,(99l
permanent changes in the environment may require genetic changes in the popu-
lation.
Thus shifting distributions of insects associated with a certain range of
 INDIVIDUAL AND POPULATION RESPONSES TO ENVIRONMENTS 55

BEHAVIOR

w
U)
Z

PHYSIOLOGY ~
13a:
....I
~
C
;;:
PHYSIOLOGICAL a;g;
ACCLIMATIZATION

DEATH RATE
CHANGES

SELECTIVE MORTALITY w
U)
AND FECUNDITY z
~
13a:
DEEP GENETIC CHANGES ~
IN ANATOMY, INNATE 6:
BEHAVIOR, ETC. :l
~
Q.
~
TIME ----+

FIGURE 2.18. Mechanisms of response by organisms to the environment, ordered by the time
required for full response when the organism perceives an environmental perturbation. In the study of
patterns of insect species in tropical habitats, natural-history and ecological studies are often needed
to determine the relationship between environmental factors and responses other than behavioral.
Feeding preferences, for example, may have a strong physiological response component, and
switches to different food plants may not involve a genetic change or evolutionary response. The
adaptive strategy of an organism is a mixture of these different kinds of response mechanisms, a
portion of which are evolutionary over a specified period of time. [From L. B. Siobodkin and A.
Rapoport, Q. Rev. Bioi. 49:181-200 (1974). Reprinted with permission.]
56 CHAPTER 2

solanaceous plants (genera and species) occupying a certain tropical habitat

provide the environment heterogeneity that promotes rather short-term shifts in
host plant associations in some insects. One might argue that a particular as-
semblage of chrysomelid beetles associated with solanums in the habitat consists
of species generally adapted to feed on most solanums. But individual dif-
ferences among each potential host plant species require some physiological or
behavioral adjustments in the feeding behavior in some of the beetle species not
usually associated with a particular host species. Whereas the general predilec-
tion for solanums might be a genetic characteristic of the beetle species, with the
optimal phenotype within the population of each species being a "good solanum
feeder" as the result of stabilizing selection, response mechanisms at the
phenotype level allow species to switch host plants facultatively. Different re-
sponse patterns fade in and out depending upon the host plant species attacked.
The question of physiological and genetic regulation of such feeding has
been reviewed by Feenyyo2) Such response mechanisms also imply an ability to
switch from generalist to specialist and vice versa. Especially in the tropics, with
a plethora of distinct habitat types, it is likely that many insect species frequently
encounter new resources. The ability to utilize such resources becomes a matter
of possessing the behavioral and physiological means of incorporating them into
niche structures. I define behavioral mechanisms with respect to plant feeding as
the ability to recognize and respond positively toward particular species, and
physiological mechanisms as the ability to possess the appropriate complement
of digestive tract chemical agents and other factors responsible for proper pro-
cessing of food. This topic is explored further in the next chapter. Geograph-
ically, one might look for cases of ecological shifts within species as the result of
behavioral and physiological responses to new environmental conditions rather
than as reorganizations in gene pools. Under these conditions, one can envision
behavioral and physiological morphs of species rather than of genetic varieties.
These considerations have overtones for studies in biogeography in which the
analysis of ecological shifts within individual species or genera becomes part of
arguments accounting for evolutionary divergence of groups. Not all observed
divergence is evolutionary. Levins(Jo3) has discussed the role of developmental
flexibility in the manifestation of phenotypes, suggesting that such phenotypic
flexibility provides a form of homeostasis within populations in response to some
patterns of environmental factors.
Depending largely upon the climate and other abiotic features, some tropical
regions of the world are considered to be both predictable and benign environ-
ments for organisms (Slobodkin and Sanders<IQ4». "Predictable" is used in this
context to describe the general absence of recurring catastrophes or sudden,
irregular changes in the environment. Fluctuating environments can be either
predictable or unpredictable, and environments in general can be benign or
harsh. Benign environments are those in which climate tends to promote biotas,
INDIVIDUAL AND POPULATION RESPONSES TO ENVIRONMENTS 57

whereas harsh environments preclude some biotas and promoting few. Species
occupying harsh environments such as the tops of the Andes Mountains or
tropical deserts can be specialists to both the abiotic and biotic features of these
environments, if fluctuations are predictable through time. The rigors of harsh
environments of low predictability select for more frequent extinctions of species
and less speciation, so the species abundance levels in most trophic levels are
relatively low. There should also be selection for dispersal or migration of
species from areas of low predictability areas to those of high predictability, but
not the reverse. But such invasions are probably rare or unsuccessful due to
competition from resident species in predictable environments. Predictable envi-
ronments tend to be saturated with species, and species abundance is high.
Successful invasion of predictable environments is thought to be possible only
when perturbations in the environment take place rather than as a result of
ecological displacement or presence of unfilled resource patches.
Species in predictable environments tend to be specialists, while those in
unpredictable environments are generalists. Generalists in such environments are
believed to possess physiological and ecological buffers for the genotypes as a
means of coping with a broad range of environmental conditions. Predictable
environments tend to promote speciation and lower the likelihood of extinction,
since most predictable environments are benign.
Such conceptualization of environments suggests that lowland nonseasonal
regions of the tropics accumulate more species than seasonal regions, high-
altitude habitats, and deserts. The major variable is the severity and pattern of
climate, setting up selection in favor of life history traits and other phenotypic
features that maximize survival in response either to physical conditions of the
environment or to biotic features as in the case of species in predictable environ-
ments. Such considerations help to explain the biogeographical patterns of
species in conjunction with detailed analyses of past environments and centers of
origin of taxa. Successful response to harsh or unpredictable environments is
more difficult physiologically than it is to benign or predictable conditions. A
harsh environment might also be predictable, as we might envision habitats at the
poles, but again species abundance equilibrates at lower levels than for benign
and predictable environments. In an extreme case, the role of daily rising tem-
peratures and moisture in determining the faunistic composition of biotic com-
munities is associated with harsh environments such as that found in Antarctica
and discussed by Wise and Gressitt. ([05) The restricted availability of both
moisture and warm temperatures lowers the biotic richness of such regions.
Insect species living in the habitats associated with high altitudes and desert
regions of the tropics are believed to be generalists, since few resources are
available under such conditions. There is little opportunity for speciation, and the
generalist strategy, involving the expression of behavioral, physiological, and
ecological buffer systems, precludes the exposure, so to speak, of genotypes to
58 CHAPTER 2

selection pressures. In one sense, unpredictable and harsh environments in the

tropics relax selection pressures associated with high speciation and low extinc-
tion rates. Ecological interactions such as competition become less frequent in
such environments. Being poikilotherms, insects experience such environments
as being harsh in terms of temperature and moisture extremes. (This topic is
explored further in Chapter 7 in a discussion of insect evolution in the tropics.)
But such considerations do not reduce the importance of considering the role of
short-term behavioral and physiological responses by insects in tropical habitats
with high species densities. In such areas, the combined effects of adaptation at
the genetic level and also phenotypic flexibility in terms of development, be-
havior, and physiology mold the distributions of species on resources. In harsh
and unpredictable areas, there is probably less opportunity for evolutionary
change to take place in taxa already established there. More data are needed on
the distribution and abundance of species at high-altitude regions and deserts
within the tropics.
Let us illustrate these ideas. The average carabid beetle species high up on
the slopes of the Peruvian Andes might be a generalist on grass seeds, along with
some level of opportunistic foraging on other resources, but the larger pool of
species found in Amazonian forests exhibit somewhat species- or genus-specific
feeding habits. Some species might be strictly predaceous as either ground-
cover- or canopy-dwelling forms, while others forage on specific types of seeds
or other plant products. The presence of benign and predictable features of the
lowland nonseasonal wet forest, coupled with the existence of many habitats,
both natural and manmade, promotes diversification in these beetles and other
insects, so that accounting for the local diversity of the group under such condi-
tions becomes a matter of detecting habitat distributions of species and diversifi-
cation of life styles within habitats surveyed. Barring the rugged physical terrain
of higher altitudes, the task of accounting for the taxonomic diversity of carabid
beetles in terms of ecological conditions present would be easier since fewer taxa
are present and habitat diversity is low. A complete explanation, of course,
would also take into account biogeographical considerations.
The phenotype is a complex mixed system of genetic, developmental,
physiological, and behavioral traits and processes that endow each individual
within a popUlation with some ability to thrive in a given environment. The
environment represents a dynamic array of biotic and abiotic factors fluctuating
in time and space and setting up the sphere of influences affecting phenotypes.
Because the phenotype involves behavioral traits, the environment is manipu-
lated by the individual. The energy budget and food expenditure, life table
characteristics of the popUlation, and pattern of environmental variability mold
the reproductive strategy of the species in a particular habitat or collection of
adjacent habitats. In most insects the combination of larval-derived and adult-
derived nutrients influences reproductive effort. Thus the evolution of fecundity
INDIVIDUAL AND POPULATION RESPONSES TO ENVIRONMENTS 59

traits is molded by the efficiency with which both larvae and adults captivate
nutrients, of which some portion will go into construction of body tissues and
maintenance of body functions associated with the production of gametes. Such
parental investment patterns are expected to vary between males and females.
The evolution of sexually distinct phenotypic traits imposes another issue of
complexity in the manner in which individuals within a population respond to the
environment. The high nitrogen requirements associated with the production of
eggs in insects is a major physiological factor influencing the feeding habits of
adult females. A variable environment might select for life table characteristics
favoring colonization, that is, high egg production or decreased developmental
time. Lewontin<87l has demonstrated that slight changes in fecundity have pro-
nounced effects on the growth rate of populations, more so than phenotypic
switches in characteristics such as developmental rate and longevity. The envi-
ronment may establish selection for increased fecundity in an insect species, in
response to either unpredictable conditions associated with invasions of new
environments or increased predation rates within existing popUlations. Thus the
evolution of energy budget characteristics is closely linked to the evolution of life
history traits within populations. Selection is expected to be especially pro-
nounced on females, since increases in fecundity are the most sensitive means to
increase the growth rate of a popUlation. For these reasons we also examine life
history, physiological properties of insects in general, and other parameters of
popUlations in order to understand the phenotype and range of phenotypes in a
popUlation interacting with resources for insect popUlations in tropical environ-
ments.
 CHAPTER 3

MACHINERY OF
ENVIRONMENTAL
RESPONSE MECHANISMS
IN INSECTS
Key to Evolutionary and
Ecological Diversification

3.1. NOTES ON INSECT DIVERSITY

Any attempt to discuss the high abundance and ecological success patterns of
insect species in tropical environments must, in my view, take into account two
major sorts of driving forces that have contributed to the elaboration of insects as
a group and to the elaboration of the multiplicity of life styles represented within
the group. I define life style in this context as the population parameters and
habitat associations of particular genera and species. within families of insects.
Population parameters include age-specific fecundity and mortality schedules,
developmental rate, and patterns of dispersion. The habitat is that portion of the
overall environment of a region that supports breeding populations of particular
species. The habitat may also include a large spectrum of "transient" species
that do not breed there. I assume that insect species settle into those habitats that
permit them to maintain levels of population abundance that usually prevent
extinctions. Much of this book is devoted to discussions of various aspects of
such patterns. In the present chapter, however, we examine some of the major
properties of the generalized insect phenotype that have pre adapted insects in
general to the patterns and structure of environments. Thus an understanding of
insectan success in terms of adapting to various kinds of environments warrants
an appreciation for the vast evolutionary history of the group as well as contem-
poraneous ecological factors affecting phenotypic diversification at generic and
species levels.
Many introductory textbooks in zoology stress the tenacity and evolutionary

61
62 CHAPTER 3

success of the arthropods as a group, and within the group, the explosive diver-
sification of that major terrestrial mandibulate group, the insects or hexapods.
There is still considerable controversy (e.g., Schram(!06)) over the probable
pathways of arthropod evolution in general. The traditional view of an essentially
monophyletic origin of the group is giving way to a polyphyletic view, which is
growing in popularity among some invertebrate paleontologists. The new infor-
mation driving such a change in view appears to be detailed analyses of em-
bryological patterns and functional morphologies. Acknowledging the apparently
obscure evolution of the arthropods, zoologists feel comfortable in stating that
the insects arose from some arthropod an stock about 350 million years ago, and
probably in the lower Devonian or earlier. The popular view for the origin of the
insects is that they evolved from some proto-myriapod ancestor. Some evidence
from morphology and embryology favors the view that the insects evolved from a
myriapodlike ancestor by a form of neotony in which the rate of development of
successive ontogenetic stages was slowed down, resulting in a reproductively
mature adult with a larval body plan. In such a manner zoologists analogize the
first instar of some Symphyla with the adult jnsectan body plan. A good review
of these basic ideas is given by Wille (07 ) and in many zoology textbooks.
The paleontological record indicates that as many as 11 orders of insects, 7
of which are now extinct, were already present by early phase of the Upper
Carboniferous, followed by the appearance of 19 orders in the Permian. It is
noteworthy that in both periods there were some representatives of neopterous
Exopterygota: insects capable of complete wing flexing and with the
hemimetabolous form of postembryonic development. The two basic forms of
postembryonic development in the modern-day insects, many forms of which
appeared in the early Mesozoic and into the Jurassic, are (1) the more evolution-
arily primitive condition in which the individual passes through three life
stages-egg, larva (nymph or naiad), and adult, a pattern of ontogeny known as
hemimetabolous development-and (2) holometabolous development, in which
the individual passes through egg, larva, pupa, and adult stages. The addition of
the pupa stage is believed to be a qualitatively new feature in insectan develop-
ment, although its precise evolutionary origin is not known. (08 )
In most insects with holometabolous development, there is great dissimilar-
ity in the life styles of the larval and adult stages, a feature which permits
individual species to partition resources effectively between the two stages, and
which results in considerable selection in different directions within the life
cycley09) The evolution of the neopterous Endopterygota, modern-day insect
groups with holometabolous development and in which wings and other appen-
dages develop internally rather than externally, represents the first great explos-
ive radiation of the insects, presumably in response to selection pressures arising
from the interactive effects of habitat diversification, predation, and perhaps
MACHINERY OF ENVIRONMENTAL RESPONSE MECHANISMS 63

some forms of abiotic mortality. A common point of discussion in this context is

the paleontological correlation between the evolution of angiosperms and groups
both of pollinating and of plant-feeding insects. Those groups of social insects
such as ants and termites which usually do not enter into feeding and pollinating
relationships with plants appeared relatively early in the evolutionary history of
insects, about the middle of the Cretaceous, whereas other social insects such as
wasps and bees, most of which depend upon nectar, pollen, and other nutrient-
rich plant products, did not appear until much later, in the Oligocene. One can
speculate that the high incidence of ant associations with extrafloral nectaries of
many plant species in the tropics today represents a derived character state
associated with ecological factors promoting further development of new or
secondary feeding habits in these insects. Carpentet llO ) has recently reviewed the
evolutionary history of the insects.
With a modem-day group as evolutionarily and ecologically diverse as the
insects, it is interesting to note that considerable diversification was present in the
early history of the group, and that while whole orders have become extinct, such
events were followed by an even greater diversification of orders and groups
within orders. The interplay of (I) different life cycle patterns, (2) the diversifica-
tion of mouthparts, digestive tract physiology, and feeding habits, (3) the evolu-
tion and elaboration of wings, and an associated musculature permitting various
forms of thermoregulatory behavior, (4) defensive and courtship-related pos-
tures, and (5) sophisticated chemosensory mechanisms associated with feeding
and mating and other forms of communication, superimposed upon changing
arrays of habitats and climatic conditions, must have been the major features
underlying the evolution of insects. To this laundry list we add the evolution of
body size ranges and the ecological advantages and constraints imposed by body
size and the generally poikilothermic condition of most insect species.
When examining the insects, one is tempted to reinforce the view of MacAr-
thurJ5) and others that has accounted for the high number of species in the
tropics, since the high species and generic numbers of insects found in lowland
nonseasonal tropical regions reflect considerable diversification in the five gen-
eral types of characteristics mentioned above. The availability of many diverse
habitats in the lowland tropics and the associated diversity of resources for
insects in them have promoted the diversification of the group, especially at the
generic and species levels. Wellington(96) has reviewed the general effects of
climate on insect populations and has stressed the need to look for microclimate
effects on the distribution of insects in particular regions. He suggests that while
one might say a characteristic insect fauna is associated with particular regions,
one must look at shifting distributions within and between regions in terms of
prevailing climatic conditions, which vary considerably annually. Wellington's
considerations suggest that some sort of density-independent randomness shapes
 64 CHAPTER 3

TROPICAL RESOURCE DIVISION

II 11111111111111 I 11111111111 111111 111111

H
AMOUNT NEEDED FOR ONE
PARASITIC HYMENOPTERAN POPULATION

EXTRA TROPICAL RESOURCE DIVISION

III III
FIGURE 3.1. Patterns of resource exploitation (spaces between vertical bars) in comparable spectra
of tropical and extratropical habitats. In tropical habitats there is a greater division of resources
among herbivorous insect species than in comparable extratropical habitats, and the average parasitic
hymenopteran requires the exploitation of more than one species of host in order to maintain its
population. Since population sizes of individual herbivorous insect species are greater in extratropical
habitats, fewer such species are needed to support a single species of parasitic hymenopteran. [From
D. H. Janzen, Proceedings of the XV International Congress of Entomology, pp. 84-94 (1976).
Reprinted with permission of the Entomological Society of America.]

the insect faunas of whole regions. The greatly shifting climatic conditions of
Temperate Zone regions suggests that such effects might be less intense for
tropical faunas, although little data are available to examine these effects.
When examining tropical insect faunas we must be cognizant of the prevail-
ing climate of particular regions, especially with respect to seasonality and year-
to-year trends in seasonality (see Chapter 7), and must examine the habitat
diversity within a region, and the structure and diversity of vegetation within
each habitat. Janzen(47l has considered the problem of explaining why there are
so many species of insects, particularly in tropical regions. His suggestions

~
I-
4:
I-
ro
4:
FIGURE 3.2. Correlation between
I time and species richness of a habitat.
4: As resources become used the rate of
LL
o increase in number of species de-
Ul
Ul
clines, eventually leaving a commu-
W
Z
nity in equilibrium with available
I resources. [From D. H. Janzen, Pro-
Cd
n:: ceedings of the XV International Con-
Ul
J,I,l gress of Entomology, pp. 84-94
U
W (1976). Reprinted by permission of
Ul
D.. ~ _____________________________________
the Entomological Society of Amer-
TIME -> ica.]
MACHINERY OF ENVIRONMENTAL RESPONSE MECHANISMS 65

elucidate the roles of habitat diversity, distribution of resources, productivity,

and other factors on the species richness of tropical areas.
Starting with the fact that insects are generally poikilothermic and relatively
small, so that ratios of body surface area to volume in very small insects impose
severe dessication problems in dry environments, physiological constraints as-
sociated with low temperatures, seasonal shifts in temperatures, and low
moisture availability essentially disappear in the relatively nonseasonal lowland
tropical rain forest environment. The major issues of environmental response,
adaptive in both a long-term genetic sense and in the context of short-term
physiological and behavioral response systems, in the lowland tropical rain forest
region becomes one of adjusting primarily to biotic factors.
Janzen(47J developed a hypothesis for tropical resource division in which an
absence of dominant species in tropical habitats promotes generalist feeding
strategies in insects exploiting resources. Common species generally have large
populations, providing a large resource base for several species in the next step of
a food chain. However, because insect species in the tropics tend to have small or
highly fragmented populations, it is difficult for predators or parasites on these
species to be specialists. Thus parasitic Hymenoptera in the tropics are expected
to be generalists (Fig. 3.1), whereas in Temperate Zone habitats the increased
bi~mass associated with some species of plant-feeding insects in certain types of
habitats provides sufficient resources for several specialist parasitic hymenopter-
ans. But as Janzen points out, such patterns of resource utilization in the tropics
will also be constrained by the degree of similarity among host species for
parasite species in the habitat. The phenotypic traits of morphology, physiology,
and behavior impose constraints upon predatory and parasitic members of food
chains, constricting potentially broad niche breadth with respect to host species.
Janzen(47J also argues that the increased productivity of some tropical habitats
permits the evolution of more species of herbivorous insects and that, depending
upon the patch structures of the various plant species present, herbivore popula-
tions will be disjunct to varying degrees.
In such a manner it is possible that a large number of specialist herbivore
species accumulated over evolutionary time in some tropical habitats (Fig. 3.2)
and that development of higher levels in food chains, such as parasitic insects,
are molded largely by the structure of the herbivore community associated with
the vegetation in a habitat. Add to these effects the availability of varying types
of habitats in the lowland tropics in particular, resulting in a greater pool of species
and genera, especially with regard to "herbivore type" species such as stem and
foliage feeders, flower feeders, root feeders, and seed (fruit) feeders. Com-
munities characterized by a high number of species will accumulate as a result of
increased speciation rates promoted by a more heterogeneous environment where
constraints imposed by abiotic factors are usually minimal. Janzen's models also
show that environments with low primary productivity will accumulate a lower
66 CHAPTER 3

FIGURE 3.3. A hypothetical relationship be-

tween photosynthesis (upper curve) and noc-
turnal respiration (lower curve) as a function
l' I - - - -_ _ PHOTOSYNTHESIS of increasing elevation in the tropics. The dif-
>- ference between the curves is proportional to
~

'> the amount of harvestable productivity. The

i= model shows that plant productivity, as a
L)
« resource for insects, should be greatest at in-
<i.
L)
termediate elevations in the tropics, where
(5
o harvestable productivity is maximized. [From
..J
o D. H. Janzen, Proceedings of the XV Inter-
t7i RESPIRATION
>- national Congress of Entomology, pp. 84-94
I
D- (1976). Reprinted by permission of the Ento-
~-------------------------
ELEV ATlO N -> mological Society of America.]

number of insect species. Thus, higher elevations in the tropics generally have
fewer insect species than lower elevations, owing to a slowing down of photosyn-
thesis and harvestable productivity resulting from cooler days (Fig. 3.3). As
harvestable productivity increases, that is, that portion of plant biomass that can
be converted to nutrients by insects, the heterogeneity of resources (host plants
and host plant parts) increases (Fig. 3.4) and more insect species can be sup-
ported.
J anzen(47J argues that resource heterogeneity, within and between habitats,

HARVESTABLE PRODUCTIVITY

FIGURE 3.4. Predicted pattern of change in the number of resource patches large enough to support
an insect population, as harvestable productivity increased. [From D. H. Janzen, Proceedings of the
XV International Congress of Entomology, pp. 84-94 (1976). Reprinted by permission of the author
and the Entomological Society of America.]
MACHINERY OF ENVIRONMENTAL RESPONSE MECHANISMS 67

and the level of primary productivity are the major determinants for insect
species richness. Each kind of insect is expected to view the pattern of resource
heterogeneity and productivity differently, depending largely upon its lifestyle,
including its energy budget and allocation of nutrients to reproductive and non-
reproductive activities. Janzen argues that the small size of insects in general
imposes a greater habitat and resource heterogeneity on insect species and that
their ability to modify their habitats declines.
Borrowing briefly from optimization theory (e.g., Maynard Smith(\l1), a
safe assumption is that the individuals (phenotypes) within a popUlation of an
insect species will experience the prevailing environment (habitat) in such a way
as to maximize their survival and reproduction (fitness). The habitat will impose
a grain of resource availability and heterogeneity that will be viewed differently
by different insect species. The grain of resource availability refers to the relative
abundance of different types of resources available to a species, and how these
resource-types are distributed temporally and spatially. That' 'view differential"
will be largely governed by the necessity to optimize energy retrieval in a manner
consistent with a host of phenotypic characters determining fitness, namely,
age-specific fecundity and mortality schedules, developmental time, dispersal
ability, and the amount of flexibility associated with exploitation of re-
sources. OI2 ) Thus popUlation structure and dynamics is largely determined by
productivity and distribution of different types of resources within a trophic level
(resource and habitat heterogeneity). The higher number of species of insects in
tropical habitats adds to these considerations the effects of competition for re-
sources among species. Competition is expected to add, in effect, another layer
of heterogeneity in the sense of promoting even greater divergence in life styles
among competing species.

3.2. COMMENTS ON INSECT PHYSIOLOGY

The internal physiology of insects in general is governed primarily by en-

vironmental constraints influencing responses to temperature, moisture, food,
mates, and predators and parasites. The precise problems of environmental re-
sponse vary greatly within a species as a function of geographical region, habitat,
prevailing local weather conditions, and resource availability and distribution.
We can speak of the generalized physiology of a particular group as being very
different from those phenotypic changes resulting from selection within single
popUlations of a species. For those components of the niche of which competing
species partake, we also expect selection for the development of physiological
processes associated with coping with the competitors. Physiological adaptations
cannot be entirely divorced from behavior, and we can view a great deal of insect
behavior as resulting from physiological events or else leading to physiological
68 CHAPTER 3

changes as short-term responses to environmental perturbations (see Chapter

10).004)
Physiological processes constitute a major set of events that buffer the
genotype of an individual from the effects of the environment. The individual
responds to the environment by behavioral, physiological, and genetic
mechanisms. It is the joint effect of behavior and physiology that buffers the
genetic system from the potential forces of selection, including outright extinc-
tion. Insects generally exhibit these classes of responses to stimuli from the
environment: visual detection of resources and nonresources; sophisticated and
often directional locomotor movements in response to food, enemies, shelter,
changes in the physical conditions of the environment, escapes and colonizing
movements to new environments; periods of rest when active responses to the
environment are minimized; sophisticated chemosensory responses to odors and
tastes associated with resources and nonresources; and establishment of set post-
ures on substrates and mediated by proprioceptors.
It is beyond the scope of this book to discuss detailed aspects for all of these
classes of processing of information from the environment by insects in general.
It is, however, illustrative to discuss some features of some of these response
classes, especially as related to how insects respond to abiotic and biotic features
in habitats. Much of the discussion in some of the subsequent chapters elaborates
upon the relative role of abiotic and biotic factors in molding the population
structure and habitat associations of insects in the tropics.
Studies (e.g., Swihart°J3» have revealed the role of color in the visual
mechanisms of some tropical butterflies and other insects. Butterfly species have
been demonstrated to be visually sensitive to colors present on the wings of
individuals of each species, and natural mechanisms regulating such abilities are
probably rather specific for a c'ertain set of colors. The role of ultraviolet sensitiv-
ity in the visual systems of insects has been discussed by Eisner. o 14) All of these
studies suggest that insects in general have evolved some rather sophisticated
visual discriminatory abilities in both the visual and the ultraviolet spectra. These
studies also point to the role of these mechanisms in adapting each species to the
environment. The great diversity of butterfly wing color patterns in the tropics,
coupled with the prevailing low population densities of many species and cooc-
currence of several similar species or complexes of similar-appearing models and
mimics in the same habitats, places premium selection pressure on the elabora-
tion of highly discriminatory visual systems. Under these kinds of environmental
conditions, such discrimination permits species-specific mating to take place
with minimal interference among species. The joint role of visual cues in the
phenotype- and species-specific odors associated with one sex or the other,
functions to ensure mating frequency at some level suitable for maintaining
breeding populations. Such physiological properties place constraints on the kind
MACHINERY OF ENVIRONMENTAL RESPONSE MECHANISMS 69

of spatial structure there is to the adult population of each species, compromised

with constraints arising from the distribution of resources such as adult food
supplies and larval host plants, in the case of many species of herbivorous
holometabolic insects.
The elegant experiments of D. M. Wilson, such as Wilson and Hoy's(llS)
analysis of locomotory control mechanisms in the milkweed bug, Oncopeltus,
and the beetle, Zophobas, have shown that the degree of relationship between
visual systems and locomotor systems varies tremendously in different insects.
Optomotor responses are well developed in some insects and poorly so in others.
Such studies suggest that other forms of input modality exist in insects where
optomotor response mechanisms are lacking. The behavioral patterns of insects
involve a variety of input modality systems, including those based upon vision,
odors, tastes, and tactile pressures. As will be developed later in this chapter
(Sections 3.4,3.5) and Chapter 7, the migratory ability of insects such as On-
copeltus depends upon the ability to discriminate changes in photoperiod. In such
insects,(IlS) optomotor response systems are well developed and account for
much of the overall behavior of the insect. Optomotor response and other forms
of sensory input into the neuromuscular systems of insects constitute a significant
portion of the generalized insect phenotype concerned with a multitude of short-
term and long-term (migratory) responses involving walking or flying. They
form the basis for the expression of many behaviors often very specialized in
tropical insect populations, including mate location, territorial defense, spacing
patterns associated with distribution of food and other resources, and general
proneness to dispersal and movement among resource patches when patches are
small and widely scattered relative to the population size of the insect species and
the average-sized individual within the population.
Many studies have pointed to the effects of water stress on the ability of
insects to survive under prolonged periods of dryness. Insect populations in very
seasonal tropical regions are annually subjected to extended periods of water
stress during the dry season. However, Janzen,OI6) Frankie,OJ7) and Frankie et
al. (J 18) have shown that the tropical dry season is a period of maximal activity for
some insects, particularly those visiting and pollinating the flowers of certain tree
species that flower in the dry season. Insects active in the tropical dry season tend
to be large, since small insects are more vulnerable to dessication. Thus the
tropical dry season acts as a sort of ecological filter in which a substantial portion
of the insect (and overall arthropod) community drops out from particular
habitats where the effects of dryness are paramount on plant populations and
general vegetation cover. See Chapter 7 for a more detailed discussion of season-
ality in the tropics.
Physiological constraints of size, cuticular structure, and ability to absorb
metabolic water demonstrated by EdneyOI9), Fraenkel and BIewett020 ), Berridge
70 CHAPTER 3

and Gupta(!2J), and many other workers, drive selection for mechanisms that
permit species to escape in space or time from very dry conditions or permit some
species to become permanent residents of perpetually dry areas such as tropical
deserts and some mountain tops. The harsh nature of dry environments has been
known for some time owing to laboratory studies on the effects of different
humidity regimes on body weight and developmental time. Lower humidities
necessitate more food to produce a certain body weight, and the developmental
period is extended in the larval stage. The increased food intake at lower
humidities permits a greater intake of water in response to the environmental
stress imposed upon the insects. The study of Fraenkel and Biewett020J provides
a good example of such effects.
It is also evident from various studies on the fine structure of tissues as-
sociated with water uptake in insects (e.g., Berridge and Gupta(12!l) that selection
must have been strong in the early evolutionary history of the insects to develop
efficient means of water uptake, necessitated by small body size and the invasion
of various terrestrial habitats. The tropical desert today represents an example of
a harsh environment in which communities are expected to have fewer insect
species than comparable areas of the wet lowland tropics. Clearly a major adapta-
tion permitting insects to colonize desert environments has been the evolution of
spiracular control mechanisms that permit regulation of body water content in
response to changes in humidity (Fig. 3.5). Insect species in tropical deserts are
expected to be generalists toward foodstuffs that optimize water intake,
metabolic or otherwise, but specialists toward being active at those times of the
day or times of the year in which water stress is somewhat alleviated.
From the studies of EdneyOl9) and others it is evident that insects have
adapted to desert environments in various ways (See also Fig. 3.5.). Desert
beetles tend to have impermeable cuticles and because they tolerate high temper-
atures can be active on the ground surface during the day. Desert roaches, on the
other hand, have thin, permeable cuticles and live beneath the sand, where they
can absorb water vapor. Such traits represent ecological specialization in re-
sponse to stressful abiotic features of the environment. The impact of such stress
on insects in deserts promotes generalized feeding habits. The specialized adapta-
tions, notably those that are morphological, may preclude such species from
invading wetter regions of the tropics. Thus not all organisms can shift from
unpredictable to predictable environments as concluded by Slobodkin and San-
ders. (04 ) Specialized adaptations may preclude the ability of a desert-dwelling
insect to respond to selection favoring ecological range expansions into less
stressful wet tropical regions. Some specialization is also expected in terms of
muscles and muscle physiology related to the expression of postures associated
with seeking shade and rapid movements in response to predators under condi-
tions of maximal exposure against the ground cover.
Subsequent sections of this chapter examine other physiological response
 MACHINERY OF ENVIRONMENTAL RESPONSE MECHANISMS 71

1.2

1.0

0.8
FIGURE 3.5. Relationship between
water loss from tsetse flies (Glossina
morsitans) and relative humidity ~ 0.6
when the spiracles are permanently a
open or blocked. The rate falls off in
drier air if the spiracles are not dis-
Air
turbed and allowed to regulate. In- ~ _____A
sects in dry tropical environments,
within certain limits, can regulate
water loss by such a mechanism. 0.2
Spiracles blocked
[From E. B. Edney, Science _e
156: 1059-1066 (1967). Copyright
1967 by the author and the American o~-=~===;~e~-~ __~__~
Association for the Advancement of 100 80 60 40 20 o
Science.] Relative humidity (%)

systems in insects, including integration of the phenotype through hormonal

control systems, feeding behavior, diapause and migration, and defense
mechanisms.

3.3. INTEGRATION OF THE INSECT PHENOTYPE BY HORMONAL

SYSTEMS: A BRIEF REVIEW

Many entomology and zoology textbooks discuss the processes of feedback

regulation of postembryonic development in insects. A pertinent review of con-
cepts and evidence is given by Schneiderman and Gilbert. (J22) The early
physiological studies of Wigglesworth and others have elucidated the roles of
several hormones in the regulation of metamorphosis in insects. The basic model
consists of the release of activating hormone from neurosecretory cells in the pars
intercerebralis portion of the brain, and its subsequent storage in a gland called
the corpora cardiaca in the head. The production of activating hormone and its
flow to the corpora cardiaca is regulated by genetic mechanisms coupled with
cues from the environment. The activating hormone undergoes timed releases
from the corpora cardiaca with target effects on the prothoracic gland, an endo-
crine gland, which synthesizes and releases a growth hormone that induces the
molt cycle at specified intervals of growth. Another endocrine gland, the corpora
72 CHAPTER 3

allata, synthesizes and releases the juvenile hormone, which is responsible for
differentiation of larval tissues at each molt and the suppression of pupal or adult
differentiation. The basic mechanism is the same in both hemi- and holometabol-
ous insects, as indicated by the early studies of Wigglesworth<I22a) with the tropical
hemipteran Rhodnius and Caroll Williams's studies with saturniid moths, <122 b)
most notably Hylaphora cecropia. The three endocrine glands involved in the
control of development in the insects are derived from the embryonic ectoderm,
and two, the prothoracic and corpora allata, are derived from adjacent body
segments. The two structures are virtually identical in cellular structure, and the
hormone products elaborated from both also related chemically. The fact that
growth hormone (ecdysone), activating (brain) hormone, juvenile hormone, and
the synthetics farnesol and cholesterol all activate the prothoracic glands indi-
cates their chemical relatedness in experimental studies with different insects.
In some insects it has been shown that by itself, activating hormone or
juvenile hormone can act as an effective functional substitute for ecdysone and
induce molting under controlled conditions. Various other studies, such as those
in which the activating hormone is substituted for ecdysone in cockroach nymphs
induced to molt by brain implantations following surgical removal of their por-
thoracic glands, provide further evidence for the common origin and similar effec-
tiveness of the three major hormones. Simoes and Uemura(23 ) have reviewed the
effects of these hormones on chromosomes in the dipterous insects. The giant
polytene chromosomes of these insects exhibit puffing patterns associated with
gene activity and protein synthesis. Puff patterns tend to be specific for different
tissues and vary with different developmental stages. The experiments showing
that ecdysone injections during intermolt periods change the patterns of chromo-
some puffing provide evidence of a feedback control mechanism on gene regula-
tion in these insects, since puffing reflects pulses of RNA synthesis.
The studies of Krishnakumaran et at., (124) have demonstrated a definite
pattern of DNA synthesis in each tissue during each molt cycle of saturniid
moths. Krishnakumaran et al. found that DNA synthesis occurred in some tis-
sues such as Malpighian tubules and nerve tissue soon after the release of ec-
dysone during larval molt periods. Yet in other tissues such as imaginal wing
disks and muscles, DNA synthesis during the larval period did not correlate with
ecdysone secretion. Such observations point to the highly tuned system of pro-
grammed tissue differentiation during the life cycles of most insects. The factors
responsible for bringing these processes into play come from the environment.
Many studies have shown that the timing of molt cycles and differentiation of
adults are highly synchronized with the environment, and some cue or set of cues
sets up a pathway of control with ultimate effects at the molecular level. Thus the
demonstrated inhibitory role of juvenile hormone or famesoic acid derivatives
that have juvenile hormonelike activity (Spielman and Skaff(25), in the dif-
ferentiation of adult insects is a function of the substances themselves but also of
MACHINERY OF ENVIRONMENTAL RESPONSE MECHANISMS 73

the integration of systems of insect development with the environment. An active

area of research is the attempt to use the juvenile hormone or synthetic mimics to
inhibit maturation of insect larvae as a means of biological control in natural
populations.
Insect species in less seasonal and warm environments such as tropical
lowlands are exposed to different seasonal-related environmental cues from
species in the Temperate Zone. Studies have related the effects of change in
photoperiod and temperature on postembryonic insect development. Such cues are
less operative near the equator, where the length of day is fairly constant
throughout the year and temperatures show little fluctuation outside of diurnal
cycles. Thus one might argue that insect development in such regions is strongly
genetically programmed for life cycle length and that there is little or no correla-
tion of the timing of developmental stages with environmental changes. The
sharp changes in photoperiod and temperature associated with temperate regions
impose an additional and highly adaptive layer of control on insect population
dynamics and natural history. The long growing season of the lowland tropical
rain forest regions promotes the development of overlapping generations within
each species without synchronization cues from temperature and photoperiod. In
the more seasonal tropics, we might expect that the synchronizing environmental
cue is precipitation or the lack of it. Thus insect communities associated with
field habitats in tropical regions where the dry season is long and intense, result-
ing in many plant species becoming deciduous and therefore less accessible as
food resources for herbivorous species, are expected to comprise species with
popUlation structures adjusted to the dry season. The tropical dry season may
represent a period of reduced popUlation growth for insect species that cannot
migrate to less stressful habitats such as some riparian forests and forest habitats
at higher altitudes. Such species may enter into a state of reproductive diapause
as adults or as diapausing immatures.
In the lowland seasonal tropics we must look to the absence of precipitation
and its secondary effects as the probable mechanisms for causing significant
changes in the popUlation structure of single species and for entire communities
associated with habitats where the stressful effects of seasonality are most in-
tense. In some ways insects experiencing such conditions are analogous to in-
sects in the Temperate Zone, even though the environmental changes causing
seasonal effects are different and the problem of surviving freezing temperatures
is absent. How insect enzyme systems are affected by the tropical dry season war-
rants considerable study, since temperature-induced changes in enzyme activity,
so characteristic of Temperate-Zone species, are absent.
Much of the successful social evolution in the Hymenoptera has been medi-
ated by both endocrine and exocrine control systems (Wilson(SOl). For exam-
ple, Roseler et ai. (126) have shown that the social hierarchy in Poiistes paper
wasps is regulated in part by differing degrees of activity of the corpora allata of
74 CHAPTER 3

females. In this system dominant females have more active corpora allata than
subordinate females. Many tropical social paper wasp nests have multiple found-
resses, suggesting that such a control mechanism may be particularly adaptive in
wasps with multiple foundress colonies.

3.4. DIAPAUSE AND MIGRATION IN INSECT POPULATIONS:

ESCAPE IN TIME AND SPACE

A cornucopia of field studies have demonstrated that insect population sizes

in some tropical regions and habitats fluctuate throughout the year. Tropical
regions consist in great part of very seasonal climatic conditions imposed by the
interaction of prevailing winds, topography, and oceanic currents. MacArthur(l6)
summarized the major climatic regions of the Earth and the general properties of
tropical regions. (For a detailed analysis of climate and weather conditions in
Central and South America, consult Schwerdtfeger. (45» It is clear that geograph-
ically and climatically the tropical regions show considerable variation in physi-
cal features of the environment, which in tum profoundly affect the biotas found
in them. It is also clear that some tropical environments are less benign and
predictable than others in terms of annual fluctuations in climate. Tropical re-
gions thus contain quasi-temperate areas as well, which can be classified as
tropical deserts, tropical mountain tops, and lowland tropical regions where for
most of the five or more months of the dry season virtually no rain falls. Other
tropical regions undoubtedly can be classified as harsh and unpredictable in terms
of organismic response to the environment over contemporary and evolutionary
time. To some degree, all three regions mentioned above provide conditions that
may contain a prolonged interruption of the growing season for insect species,
but such effects are most likely found in the highly seasonal lowlands.
One might expect that insects and other organisms found on tropical
mountain tops and in tropical deserts experience harsh conditions throughout the
year and over many successive years. In some ways, such conditions are more
benign than an environment in which there is a rather gross change in the
environment for about one-half of the year. Organisms on tropical mountain tops
or in tropical deserts can thus evolve adaptations that permit them to maintain rather
stable popUlations in which the average phenotype within a popUlation is a
well-buffered generalist. The challenge in the lowland seasonal environment
appears to be more of an ecological challenge for species. As pointed out earlier
(Section 3.4), the long interruption in the growing season in seasonal tropical
lowlands is somewhat analogous to the interruption in the growing season in
Temperate Zones, even though the environmental cues regulating physiological
response systems are expected to be very different. To cope with tropical season-
MACHINERY OF ENVIRONMENTAL RESPONSE MECHANISMS 75

ality, insects and other organisms may develop mechanisms that allow for escape
in time or space within the context of the annual alteration of seasons.
The winter season of Temperate Zone countries and the severe tropical dry
season represent periods of population decline for some insect groups (Fig. 3.6).
The discussion is limited in the case of the seasonal tropics to those insects that
cannot physiologically remain active in lowland habitats throughout the severe

non-seasonal tropical field

LU
N
1i5

z
o
~
-l
::J
c..
o
c..

SUCCESSIVE MONTHS

strongly seasonal tropical field

WET DRY WET

SEASON SEASON SEASON

FIGURE 3.6. Predicted patterns of seasonal changes in the size of an insect population in the seasonal
and nonseasonal tropics and the temperate zones. Both temperate and tropical seasonality promote
rapid increases in population size at the beginning of the growing season, and enforced decline during
the period of highest stress, the northern winter or tropical dry season.
76 CHAPTER 3

dry season characteristic of tropical dry forest zones. Insect populations in the
rather nonseasonal lowland wet tropics are expected to have even levels of
abundance throughout the year, with minor variations in abundance superim-
posed by biotic mortality factors, which also tend to have constant effects
throughout the year under such conditions (Fig. 3.6). Temperate Zone insect
populations, especially those species that are plant feeders, may exhibit a spring
pulse of moderate-to-high peak adult abundance and a second peak of moderate-
to-low adult abundance later in the summer or early fall (Fig. 3.6). The interac-
tion of a favorable spring period promoting luxuriant growth in host plant
species, rather synchronous emergence and breeding in insects, and favorable
weather, and a lag in the growth of predator and parasite popUlations, promote
high abundance in species with two or more generations per year (growing
season). A reduction in biomass of fresh plant tissues, buildup of predator and
parasite popUlations, and perhaps increased dryness interact to yield a lower peak
of abundance later in the summer or fall. The distance between peaks within a
growing season is a function of generation time, which in turn is determined by
environmental effects on fecundity and developmental time. As the summer
progresses, biotic mortality factors play an increasing role in affecting herbivore
popUlations, whereas the spring colonists are largely free of such effects and
popUlations are regulated primarily by density-independent factors such as the
pattern of successive dry or moist days, warm or cold days. In the seasonal
lowland tropics, wet-season peaks of herbivores are expected to be about the
same between successive wet seasons (Fig. 3.6), since the interplay of both
abiotic (relatively weak) and biotic factors (strong effects of predators and para-
sites) are expected to be about the same each year. The size and shape of the
popUlation growth curves for insect species active in the wet season are largely a
function of generation time. Perturbations in overall habitat conditions, as a
result of selective clearing of vegetation, complete clearing, residual effects of
insecticides, etc., are expected to disrupt such evenness in abundance in the
direction toward smaller populations.
The following discussion is limited to those insect species which cannot
remain active in their usual habitats during the dry season. Diapause, a shutdown
in many physiological processes that promote the intake, storage, and use of
nutrients for both vegetative and reproductive functions, is the mechanism that
allows an insect to escape in time from a period of environmental stress so great
that outright mortality of individuals occurs or fitness is reduced so much that
popUlations experience extinctions. The environmental control of the onset and
termination of diapause has been investigated extensively for insect species in
Temperature Zone regions. A typical control pattern for many Temperate Zone
insects is the triggering of diapause by successive days of reduced photoperiod in
the fall, and a chilling below a certain critical temperature range, followed by a
succession of increasingly warmer days to break the diapause. The mechanisms
MACHINERY OF ENVIRONMENTAL RESPONSE MECHANISMS 77

of photoperiod and temperature-mediated control are not the same in all insects at
a given latitude. One speaks of "long-day" and "short-day" phenotypes when
discussing diapause in Temperate Zone insects. Long-day conditions, such as
early-to-mid-summer conditions of daily photoperiod length at the higher
latitudes, usually promote rapid development to the adult stage, whereas short-
day conditions, typified by late summer and fall, promote the expression of
diapause in the second generation and adapt the organism for overwintering; the
studies on developmental rate and timing of diapause in the Viceroy Butterfly,
Limenitis archippus, by Clark and Platt(27 ) are illuminating. A host of secondary
factors, including diet and temperature, also induce diapause in some insectsY2S)
Different strains of an insect species from different portions of its range may
undergo diapause at different stages in the life cycle (e.g., Kappus and Ven-
ard(129». Such studies also point to the subtle interaction of temperature and
photoperiod in the induction of diapause. Wigglesworth(3O ) first demonstrated
that the corpora allata were necessary for the maturation of eggs in Rhodnius.
Siew<l3)) showed that the adult life-span of the chrysomelid beetle,
Galeruca tanaceti, can be divided into the physiological phases of prediapause,
diapause, ovary maturation, and oviposition and that each phase is controlled by
a specific level of hormone activity. At the termination of diapause, there is a
marked increase in the neurosecretory system, thus promoting oocyte maturation
and bringing into play the physiological and behavioral syndrome of oviposition.
A fascinating study of the Argentine stem weevil, Hyperodes bonariensis, by
Goldson and Emberson(l32) demonstrates the selection pressures favoring the
evolution of reproductive hibernatory diapause in some environments and not in
others. It was shown that this insect has some popUlations in Argentina ex-
periencing the diapause and others that are not. When an insect species has a
wide geographical and topographical range, some populations develop diapause
phenotypes when adaptive and others do not, the latter a result of mild winter
conditions. Yet when the stem weevil invaded those parts of New Zealand with
mild winter conditions, the diapause condition remained, presumably as a "re-
lict" physiological trait left over from Argentinian gene pools. One might expect
that such a condition would change toward a loss of this trait, depending upon the
manifold effects of genes that regulate it on other aspects of the weevil's
phenotype. Sims(33 ) has reported on an interesting relationship between larval
host plant suitability and diapause conditions in the swallowtail butterfly, Papilio
zelicaon, in California. The species is generally univoltine in habitats where its
native umbelliferous host plant species occur, but multivoltine where it uses
introduced plant species. The latter condition favors the expression of diapause in
the overwintering population, and Sims demonstrated that induction of diapause
is regulated by photoperiod and development, and termination by temperature.
Apparently strong selection favors the exploitation of the introduced plant
species (a condition of environmental change related to a biotic component of the
78 CHAPTER 3

environment and one necessitating the development of multivoltinism and as-

sociated diapause).
Field studies are needed to determine which insect species in tropical com-
munities enter into diapause, the life stages involved, and the environmental cues
triggering and breaking the diapause. Some butterfly species of the lowland
Guanacaste region of Costa Rica, where the dry season lasts almost 6 months,
undergo reproductive diapause as adults. Changes in moisture availability, adult
nutrient availability, and larval host plant condition, which affect adult butterflies
and other insects of tropical regions whose dry season is both long and severe,
are worth studying. Such studies may also have implications for understanding
how insects in tropical desert regions remain active under conditions of extreme
and prolonged dryness. The average generation time of many insects in tropical
habitats is less than 30 days, imposing a pattern of population dynamics incom-
patible with the longer environment oscillations reSUlting in seasonality. The
adaptive challenge is to survive a time period in which several generations could
have occurred if the dry season were absent.
In the neotropical pierid butterfly, Eurema daira, wet- and dry-season morphs
appear to be in lowland Guanacaste in Costa Rica, as studied by Paul A. Opler (per-
sonal communication). Wet-season morphs are present in the first half of the dry
season, and these individuals seek out moist forest refugia along streams and also
migrate vertically into surrounding higher elevations where conditions are wetter.
The dry -season morph increases in abundance in the second half of the dry season
and enters a reproductive diapause, exhibiting few if any movements associated
with seeking wetter conditions. Such a mixed strategy ensures against complete
popUlation decimation during the dry season, since migrants apparently retum to
the lowlands with the new wet season.
Short-term vertical migratory movements for many butterflies in this region
of Costa Rica and in other tropical regions might be more prevalent than once
believed. The energetics associated with diapause in tropical insects warrants
study, since a species occupying a seasonal tropical environment must allocate
energy to physiological mechanisms regulating diapause and thereby act as a sink
for resources normally allocated to other aspects of the phenotype, such as
reproductive physiology and defensive behavior. If overall insect species rich-
ness and trophic level complexity are less in seasonal lowland tropical habitats
than in nonseasonal wet habitats,<272.137l such tradeoffs in energy allocation
within the popUlations of individual species might be compensated, so to speak,
by reduced predation and parasitism rates during periods (wet season) of optimal
popUlation growth. There is a need to determine whether season-specific changes
in the xylem and phloem chemistry of plants or shifts in moisture content profiles
of soils in tropical areas with distinct dry seasons provide the environmental cue
for triggering of the rather synchronous emergence of some insects capable of
being active during the dry season. The cicadas studied by Young(70.71.135-139l
MACHINERY OF ENVIRONMENTAL RESPONSE MECHANISMS 79

offer a system for such studies, since distinct patterns of seasonal emergence of
adults are evident in many species. Similar systems might be responsible for
shifts in population structure' for plant-chewing insects as well.
Dingle(J40-142) has shown that species of milkweed bugs, Oncopeltus, from
nonseasonal tropical regions exhibit no diapause induced by photoperiod. The
Temperate Zone migratory species, Oncopeltus jasciatus, undergoes photo-
periodically induced diapause, as seen in some, but not all, populations. Dingle
suggests that cases of photoperiodically induced diapause in some tropical
species are those in which the photoperiod regime acts as a cue for other seasonal
cycles, such as rainfall. Various species of tropical locusts are known or sus-
pected to respond to small changes in photoperiod as a cue to shifting conditions
of rainfall. Yet, as Dingle points out, other tropical insects enter diapause in
response to cool temperatures. Periods of short photoperiods in Central America
correspond to periods of reduced rainfall-dry seasons-but milkweed plants
tend to be available at such localities throughout the year as a resource for
Oncopeltus. The rugged topography of Central America and portions of northern
South America create a spatial pattern of widely scattered dry areas at certain
times of the year, causing some migration of the bugs and therefore a patchy
distribution of the populations. Since resources are always available and the bugs
are good at dispersing, they do not require a photoperiodically induced diapause
to survive. Even as conditions become unfavorable in an area, there is evidence
that the bugs are capable of long periods of starvation, so long as water is
available. Thus one expects that the early dry season is not a period of intense
migration in tropical milkweed bug populations. Such resistance gives the bugs
ample time to disperse to new sites or to withstand short-term environmental
perturbations such as destruction of the habitat (e.g., through slash-and-burn
clearing).
Metabolically, such periods of starvation are equivalent to diapause, and the
production of juvenile hormone is blocked in both. Dingle(J42) suggests that the
evolution of photoperiodically induced diapause in the genus, an obvious pheno-
typic trait adaptable to invasions of Temperate Zone environments, requires only
the linking of short-day input from the environment to the neurosecretory meta-
bolic systems. He suggests that the northern limits of the tropical species of
Oncopeltus are set by the inability to evolve a response to photoperiodic cues.
Since milkweed seeds and plants tend to be available at the end of the sum-
mer and at the time the bugs escape to more favorable latitudes, lack of food
cannot be the cue. Under such conditions, an attempt to account for the additive
genetic variance contributing to the age of first reproduction as a function of
photoperiod would be difficult, and therefore producing diapause strains in these
species from selection would be difficult. The factors contributing to this genetic
variance unknown at present. Most tropical populations of milkweed bug species
exhibit no photoperiodic diapause.
80 CHAPTER 3

Migration refers to the alternate strategy of escape from unfavorable condi-

tions of the environment. It is a gross change in those behavioral and physiologi-
cal traits of the phenotype associated with the exploitation of resources in the
immediate environment. By definition, migration is an escape in space, to a
different habitat or geographical region where environmental conditions favor
either a new burst of population growth on the part of the migrants or colonists or
a period of vegetative' 'waiting out" and eventual return to the original environ-
ment, where new flushes of population growth ensue as environmental condi-
tions return to a favorable state.
Johnson (43 ) maintains that migration is the most effective means of insect
dispersal, the well-directed, massive movement of insects from one area to
another, during which other kinds of activities cease or are held to a minimum.
Migratory flight is a well-developed phenotypic trait molded to the dispersal
needs of each species and is best developed in winged insects. At the age of
maximum reproductive value,<I44) young adults, especially females, are the best
candidates for migrations. Johnson(43 ) refers to the' 'cogenesis-flight syndrome"
as the temporary reduced ovarian development during the time of maximum
flight activity, even though young adults will eventually be the best colonizers
once they arrive in the new habitats. The interplay of individual behavior patterns
and prevailing weather conditions determine the migratory routes of most in-
sects.
Migration can be viewed as a major change in the genetic structure of a
breeding popUlation, if some portion of the popUlation splits off and migrates to a
new environment. We need to examine the correlation between genetic control of
dispersal and traits that make good migrants or colonizers, such as early age of
first reproduction, high fecundity, and short developmental time. Few data are
available on the effect of migration on the breeding structure of insect popula-
tions and on the extent of migration of individuals within successive generations.
We need to examine the factors promoting migration in different insect taxa in
order to understand the biogeographic patterns of these groups. Knowing the
factors that may initiate migration in a particular group may help biogeographers
interpret past and present taxa distributions.
Southwood(45 ) has stressed the long-range displacements associated with
migration in some tropical insects. Migration in tropical insects may be more
frequent than we know and seems best documented in pest species such as some
African locusts. The heterogenous topography of regions such as Central
America establishes a system of climatic heterogeneity, some components of
which might favor both vertical and horizontal migrations in flying insects such
as butterflies, beetles, and hemipterans, those groups most associated with vege-
tation subject to considerable change by drought. Dingle(42 ) found that three
species of tropical Oncopeltus exhibited far less long-duration tethered flight
than the well-known Temperate Zone migrant o. jasciatus. He suggests that
MACHINERY OF ENVIRONMENTAL RESPONSE MECHANISMS 81

tropical habitats are less seasonally variable and that such environmental predic-
tability should minimize the adaptive importance or tendency toward migratory
flight. Strongly fluctuating habitats should result in greater frequency of mi-
gratory behavior in insects. But I believe that Dingle was referring to the lowland
nonseasonal wet tropics. Lowland dry areas may select for some level of mi-
gratory ability, even if relatively short term and of short distance, in some
tropical insect species. Root and Chaplin(46 ) found considerable short-distance
movement in a mixed population of two Oncopeltus species in Colombia.
We need to study a few partially known cases of long-distance migration in
tropical insects such as the day-flying moth Urania julgens(47 ) and allied
species. Short-distance migratory movements might be prevalent among insects
occupying habitats in the lowland seasonal tropics where a dry season occurs for
5 or more months each year. Few data exist on the movement patterns of insect
species in such regions. A good general working hypothesis for such studies
would seem to be the degree of environmental unpredictability incurred by the
dry season as a driving force in the phenotypic response of movement to new,
less stressful habitats. To what extent do such movements occur in these tropical
communities? Are the majority of species involved, herbivorous insects? What is
the impact of such movements, if they exist, on the population structure of
species left behind and upon the community structure as a whole? These seem to
be among the great unanswered questions about the impact of tropical seasonality
on the expression of temporal and spatial escape mechanisms in insect species.
To what extent do genotype and phenotype reorganize when insect species
develop migratory movement patterns? What are the population consequences?
Dingle(42 ) has elegantly demonstrated the interplay of density, photoperiod, and
temperature regimes on the life history and population growth of Oncopeltus. He
showed that high temperature, long photoperiod, and low population density
result in a lowered age of first reproduction, an age-specific fecundity schedule
highly skewed toward young age classes of adult bugs, and therefore a higher
rate of population increase. Photoperiod and the effects of population density
influence population growth only on the adult bugs; temperature affects the rate
of development. Dingle concludes that such effects result in a phenotype highly
responsive to the need for migratory movements and argues that such characteris-
tics make this insect an ideal colonizer of milkweed patches during spring in
northern latitudes. Migratory movements are induced by short days (fall), cool
temperatures, and high densities, all of which delay reproduction and thereby
provide the time the species needs to migrate to more suitable regions as winter
approaches. Such conditions in Oncopeltus during the fall generations represent
incipient diapause. Such data point to rather short-term changes in the genetic
structure of milkweed bug populations at northern latitudes, in which spring and
fall adult populations exhibit very different response patterns to the environment.
Similar shifts in populations may occur in tropical milkweed bug populations
82 CHAPTER 3

where selection has favored the evolution of a dry-season form highly adaptive to
short-distance migration to wetter sites. Such morphs, however, may not possess
great changes in the phenotype in terms of those parameters underlying popula-
tion growth, although some response is predicted if migratory ability, even in
short-distance movement patterns, requires a shift in energy allocation. Alterna-
tively, it might well be that the response mechanism resides at the phenotypic
level and represents no significant shift in genetic structure of the population over
a few successive generations. We need to dissect the response patterns of On-
copeltus, in both temperate and tropical populations, into those components
resulting from gene frequency shifts within two or more generations and those
arising from physiological-stage changes. The planned studies of Dingle and
others should provide the necessary genetic analysis of milkweed bug popula-
tions to elucidate these effects.
The best-documented examples of migration in tropical insects come from
the classic studies of African locust populations. In these cases, large shifts in the
phenotypes present in migrating or nonmigrating populations clearly involve the
interplay of shifting availability of food and increasing population densities,
representing intraspecific competition within populations. Migration is induced
by these changes in food supply and density and synchronized with seasons
favorable for the growth of host plants elsewhere near the equator. Short-distance
migratory movements in tropical insects might prove to be more frequent than
previously documented, notably for those regions where lengthy dry seasons
promote escape from prevailing environmental conditions.

3.5. TEMPERATURE, PHOTOPERIOD, AND OTHER FACTORS

RELATED TO INSECT ACTIVITY: FURTHER COMMENTS

The effects of rather small changes in daily photoperiod on tropical or-

ganisms have not been well investigated. Comparisons of the effects of photo-
period at different latitudes are available. Stubblebine et al. (148) found that the
tropical leguminous tree Hymenaea courbaril exhibits more growth in the seed-
ling stage when grown under long-day or tropical regimes than under short-day
regimes, and found good correlation between latitude of seed origin and vegeta-
tive growth. Incidence of diapause induced by photoperiod was found to vary in
different geographical strains of the moth Pectinophora gossypiella for material
from populations ranging from EI Paso, Texas, to Venezuela, with the frequency
of diapause being greatest in the EI Paso poulation.(149) Changes in diet, such as
lipid content, and in temperature did not alter the response to diapause or lack
thereof.
Temperature is another niche component that influences the distribution of
insect species in different environments. Tropical regions exhibit considerable
MACHINERY OF ENVIRONMENTAL RESPONSE MECHANISMS 83

diurnal fluctuation in ambient temperatures promoting displacement in activity

periods for different groups of insects in tropical habitats. Some insects are active
in the day, others at night. The development and expression of physiological
mechanisms that allow activity at cooler temperatures in the tropics permit the
co-occurrence of a great number of taxa in the same habitat by partitioning the
environment temporally into daytime and nighttime components. Very high tem-
peratures in the tropics such as those associated with the midday period promote
lethargy and general inactivity in many insects and other organisms. The ob-
served peak activity periods at dawn and dusk in some insects, such as certain
species of cicadas and butterflies, suggest a synchronization of activity with
times of least stress from excessively cool or high temperatures. Youngmo )
suggested that some tropical cicadas may have optimal chorusing at such times,
borrowing from a model of optimal sound transmission developed for vertebrates.
The degree to which insects are subjected to various forms of environmental
stress will shape the diurnal patterns of activity. Whitford et al. (151) found
greatest foraging at the soil surface in the morning and evening in the Chihua-
huan desert ants Novomessor cockerelli and Novomessor albisetosus when air
temperatures ranged from 20 to 40° C. The limiting factor was demonstrated to
be water, since the foraging period was extended when ants were given supple-
ments of seeds. Such conditions of physical stress from the habitat may affect the
activity patterns of various insect species thriving in high-altitude regions of the
tropics as well as lowland desert areas within the tropics. Populations are ex-
pected to be more physically controlled (sensu Slobodkin and Sanders <I04») than
populations in wetter regions. Lowland hot and wet habitats are expected to be
the most benign in terms of physical stress conditions, even if a dry season is
present.
Except when small insects are dispersed passively at high altitudes, where
temperatures are cooler and the ability to be active is probably reduced,(2) flight
initiation, speed, and direction are generally influenced by temperature and
wind. SnowOS3 ) demonstrated that Anopheles and Culex mosquitoes in Africa
lose the ability of directed flight in winds with speeds exceeding flight speeds.
Environments which place less thermal stress on insects are usually those
where the greatest diversification of species occur. The lack of such stress, and
other forms of abiotic stress, promotes speciation in response to biotic features of
the environment. Maximum allocation of resources goes to specialized adapta-
tions that allow each species to utilize smaller portions of the habitat and that
usually involve a reduction in population parameters promoting the growth of
large popUlations. Downes(Js4) has argued that the habit structure and resource
availability in Arctic environments make possible a greater diversity of insects
than actually found there.
But the cold cannot be tolerated by the physiological mechanisms of most
insects, even those species derived from Temperate Zone faunas. Downes
84 CHAPTER 3

suggests that the diversity of insects will never increase appreciably in the Arctic
due to speciation in response to thennal stress. The greatest diversity of insects
occurs in the tropics, where species are still well within their physiological
tolerance of temperature. The well-documented, highly synchronized
emergences and swarming of Arctic chironomids (see Olivet I55 ») suggests
another barrier to the accumulation of more species in the Arctic: the compres-
sion of activity into relatively short seasons precludes the development of en-
vironmental partitioning by different tolerable seasons.
Temperate Zone environments also require the evolution of low-temperature
tolerance mechanisms to adapt to chilly fall and spring conditions. Young(J56)
demonstrated the rather unpredictable nature of the first killing frost date in
Appleton, Wisconsin, which has a range of as much as 30 days, and the pattern
of its occurrence clearly shows the instability of temperature associated with
northern latitudes. Insects must be able to thrive under progressively cooler
conditions as the first killing frost approaches. Such adaptations provide time for
insects to prepare for the winter. The analogous issue for tropical insects would
be the evolution of low-temperature tolerance at high altitudes, such as in the
Andes. Mutchmor<157) has shown that insect species capable of low-temperature
tolerance exhibit greater ATPase activity. Thus cold-tolerant species can main-
tain some level of muscle activity at temperatures that nonnally would cause
inactivity in less tolerant species. Cold tolerance is manifested in different life
stages .0 58 ) Mutchmor and Richards (59 ) have shown that several kinds of insects
exhibit cold tolerance by maintaining certain levels of activity of muscle apyrase
but that others are incapable of such physiological regulation.
The initiation of flight in insects, analyzed by Nachtigall and Wilson(J60) for
dipterans, involves the stimulation of a succession of different muscles at the
appropriate millisecond and the subsequent integration of the whole system. In
some insects with highly waterproofed cuticle, the heat generated from activated
flight muscles is not dissipated into the environment and such insects appear to
have no special mechanisms for protection from overheating.(16L162) The coats of
fine hairs of bees, hawkmoths, and other large insects provide effective insula-
tion against convective heat loss during flight,06J) and one might expect such
large-bodied insects in the lowland tropical forests to be active during cool
periods of the diurnal cycle, since heat generated from flight can be retained.
Larger insects should also be found on tropical mountain tops, where it is cool.
The patterns of wing muscle activity during the characteristic warmup periods of
many Lepidoptera are generally the same patterns of muscle activity exhibited
during flight, although in some the burst lengths of muscle firing are shorter
during the wingbeat period.(163) Such an adaptation permits a greater generation
of body heat immediately prior to flight and would be advantageous in cool
environments. Some butterflies exhibit regulation of the thoracic temperatures by
behaviorally changing abdomen and wing positions, orientation to the sun, and
MACHINERY OF ENVIRONMENTAL RESPONSE MECHANISMS 85

perching. (64) Quiet basking in sunny spots raises body temperature in Papilio
polyxenes far more readily than shiveringy64) Higher temperatures promote
greater circulatory exchange between thorax and abdomen, dissipating heat from
the thorax. Heinrich(165) has demonstrated the large amount of heat production
associated with preflight warmups in the sphinx moth, Manduca sexta (Fig. 3.7),
resulting from rapid wing vibrations.
Whereas P. polyxenes is active at sunny times of the day and can therefore
capitalize by behavioral shifts to thermoregulate in sun patches (and cool off
away from them), most sphinx moths are active in uniformly cool conditions of
dusk and must rely on wing vibration movements for preflight warmups. Such
behavior is a requirement for successful flight in large insects. Heinrich points
out that such thermoregulation is a key factor in shaping the energetics of forag-
ing in many flower-visiting insects. The great differences in temperature between
day and night in the lowland tropics require considerable preflight warmup for
larger insects that forage early in the morning or at dusk, while smaller insects
are limited to intermediate periods but away from the hottest period.

I
40

0:

~ 3 30
..,"
o I
.
Q.

~ 2 20 ~

r ...
.!!

.D
;;:
10
•
O~-- __-L______ ~ ____ ~~ ____ -L~ ____ ~ ____ ~ ______ ~

10 20 30 40
ThQracic temperatur" (·C)

FIGURE 3.7. Rates of wing vibration and calculated rates of heat production in preflight warm-up in
the sphinx moth. Manduca sexta. in relation to thoracic temperatures. [From B. Heinrich, Science
185:747-756 (1974). Copyright 1974 by the American Association for the Advancement of Science.]
86 CHAPTER 3

Such differences have profound effects on the flowering times of many

tropical plants. (66 ) The flowers of some species must be synchronized to open at
those times of the day when pollinating insects can fly with minimal physiologi-
cal constraints. Thus hawkmoth-pollinated species have flowers which open at
dusk or late afternoon, and some species pollinated by the large euglossine bees
open before dawn, when the bees are foraging. The higher the muscle tempera-
ture, the greater the number of flowers that can be visited. Under conditions of
large food supplies, bees may use large amounts of energy for thermoregulation.
Heat produced while a bumblebee is on flowers, under conditions of moderate
ambient temperatures, helps to maintain a thoracic temperature near the
minimum required for flight between flowers. Such regulation breaks down at
high ambient temperatures as thoracic temperature rises. (65 ) During the period of
visiting flowers, the bee must obtain a certain number of calories per minute to
remain flying. Heinrich(65 ) has reviewed the energetics of foraging in
bumblebees. At cool temperatures, the energy investment required for foraging is
usually greater, since a certain body temperature must be maintained (Fig. 3.8).
The abdominal temperature, only a few degrees higher than the air temperature,
is maintained by passive heat flow only. Until the air temperature reaches 25°C,
however, maintaining thoracic temperature at an average of 2SOC requires a
greater energy investment than for intermittent flight from flower to flower.
The coevolution between pollinating insects and plants in the tropics takes into
account the energy requirements of pollinators.
The decreased population densities of many plant species in the lowland wet
tropics is due to competition, predation, and other factors (see Chapter 8), and
usually a succession of plant species is flowering throughout the year at a given
habitat. The lowland wet tropics support fewer bee species than other regions
such as arid environments, and most bee species found here are generalists for
visiting many different kinds of flowers. Trigona bees, for example, are ex-
tremely opportunistic foragers on many plant species. The apparent absence of
oligolectic behavior in lowland wet forest bees results from their being long-lived
and multivoltine, making them switch flower species throughout the year. Speci-
ation in bees(67 ) is presumably blocked in such regions by these effects, since
specialized behaviors and resource utilization do not occur. Janzen(40) has dis-
cussed two different patterns of flowering in tropical plants that promote out-
crossing even when population densities are low. Some tropical plants exhibit a
large amount of synchronous blooming at a specified time of year, and such com-
pressed flowering is often synchronized with the dry season. Such flowering
creates a signal for many insect visitors, and some of these have high energy
requirements. Janzen shows that such mass flowering tends to show up in forest
canopy and open areas. In the second pattern of reproduction in tropical trees, a
plant species produces a few large flowers continuously from an inflorescence
that may last several months. Large-bodied pollinators, including euglossine
 MACHINERY OF ENVIRONMENTAL RESPONSE MECHANISMS 87

50.---------------------------------------------------------~
, 50
,, "
,,
,
"
iii
""
;;;
::> 40 40 iii
...w
iii ::>

~
.,<.>
~
a:
"e
w
....•.•../ .... .,
<.>

~
a:
"ew
w
w
,
It;

-----------.-
,,
~ 30 30 ~
"ffi , S
,, l:'
,,
Q.

......
~
,
:IE

,,
~
"iz ,,
<.>
c:;
oo ,; "a:
o
~ 20 ;' 20 ;:

,,','"
'O ~ __________ ~
"
,"
,,__________ L __ _ _ __ _ _ _ _ _L __ _ _ __ _ _ _ _ _L __ _ _ _ _ __ _ __J 'O
o 10 20 30 40 50
AIR TEMPERATURE (DEGREES CelSIUS)

FIGURE 3.8. Patterns of energy investment necessary for a bumblebee to maintain a level of body
temperature required for flight. Thick solid line: abdominal temperature, averaged in both sunlight
and shade. Dashed line: Average thorax temperature, (the bee's flight muscles are in the thorax).
Thin solid line: limits of thorax temperature . [From B. Heinrich. Flight Energetics oflhe Bumblebee
(1973). Copyright 1973 by Scientific American , Inc. All rights reserved.]

bees, other large bees, sphinx moths, and other animals , exploit this system.<l66)
Each flower provides a large energy reward, which in tum provides sufficient
energy for visitors to fly among them, even though great distances might be in-
volved. The flowers are usually morphologically specialized to exclude visitors
other than their pollinators, and their pollinators tend to be generalists and forage
at flowers not a part of such a system. With increasing altitude in the tropics, the
number of pollinating insects declines as a result of unfavorable ambient tempera-
tures adversely affecting small insects and insects in general. Some tropical moun-
tain tops tend to have bumblebees as pollinators.I!66) Large pollinators require more
energy, thereby necessitating that nectar be accessible for outcrossing to occur in
a particular plant species on a tropical mountain top. Outcrossing is often impos-
sible and apparently unnecessary, since a large percentage of the montane flora
may be autogamous, a situation analogous to Temperate Zone plant species that
bloom in early spring.
88 CHAPTER 3

Many studies (e.g., Weit 168 ») have demonstrated the effects of temperature
on insect population dynamics. Wide fluctuations in the normal temperature
daily cycles in the lowland and mountain tropical habitats are expected to be
infrequent and therefore probably have no major effect on the population
dynamics of most tropical insects. Owing to an absence of unpredictable wide
fluctuations in ambient temperatures throughout much of the tropics, shifts in the
degree of stability between predator-prey and host-parasite interactions are
largely a function of changes in communities resulting from selection pressures
created by habitat modification and succession. Many studies [e.g., Miller(J69,170)
(dragonflies); Lamb 07 l,172) (aphids); Wall(73 ) (cockroaches)] have analyzed the
effects of water stress on insects, indicating that insects are sensitive to changes
in the availability of water and that some forms of species shifts in tropical
communities may in fact be the response to such a stress. The physiology litera-
ture is replete with examples of water stress in arthropods. The great changes in
species richness and composition of insect communities associated with pasture
vegetation in a very seasonal region of Costa Rica (Janzen and Schoenet 134 »)
have shown that small insects drop out of the community in the dry season.
Seasonal lowland tropical habitats are expected to impose considerable environ-
mental stress from lack of water, even though temperature regimes may be the
same between seasons.

3.6. NOTES ON MATING AND COMMUNICATION IN INSECTS

In those insect populations where individuals belonging to a particular

species have a rather patchy spatial distribution owing to patchy distributions of
preferred host plants or other kinds of resources, specific sets of behavioral
patterns that structure courtship, copUlation, and overall conspecific communica-
tion are expected to be integrated into phenotypes. Such adaptations are also
expected to be present in species with clumped population distributions, and
depending upon the degree to which such traits are genetically determined, such
behaviors might be specific for a species as a whole rather than varying consider-
ably among different populations of that species. Given the high species richness
of insects in the lowland wet tropics, one might expect that specific forms of
behavior are different in each species, particularly when examining closely re-
lated species or genera. Specialized behaviors associated with mating and other
reproductive activities under general conditions of low population densities are
expected for habitats where species richness is high. The mating strategy of the
aposematically colored metallic green and orange (the band is orange) coreid
bug, Paryphes blandus, entails low densities of adult bugs breeding on the small
patches of Anguria warscewiczii (Curcubitaceae), the host plant used by de-
MACHINERY OF ENVIRONMENTAL RESPONSE MECHANISMS 89

FIGURE 3.9. Copulating pair of the Neotropical hemipteran Paryphes blandus (Creidae) on the
food plant of the nymphs, Anguria warscewiczii (Curcubitaceae).

veloping nymphs (Fig. 3.9). The host plant is patchy spatially, and breeding
must be a function of bugs' successfully locating widely dispersed patches . There
will be strong stabilizing selection for such traits if the species population is well
integrated into the community. If the species population is not well integrated,
one might predict that there will be a mixed strategy, in which there exists a set of
behavioral patterns that assist integration into the existing community, and
another set that is most adaptive toward shifts in habitats associated with coloni-
zations. Strong intraspecific competition, associated with inability to disperse
into other habitats, would also favor the expression of mixed behaviors as an
adaptation for different segments of the population to utilize different portions of
the environment.
Both interspecific and intraspecific competitive pressures arising from lim-
ited aerial space, ground space, and space for general spacing patterns are ex-
pected to mold the behavioral patterns of insects, especially in tropical habitats
where species densities are high. The vertical structuring of butterflies in tropical
forests (Papageorgis(174» and the general observation that some species of
90 CHAPTER 3

FIGURE 3.10. The Neotropical butterfly Hymenitis andromica lyra (Ithomiidae) visiting the flow-
ers of Eupatorium (Compositae) at the border of a rain forest in the mountains of Costa Rica.

Morpho butterflies fly near the ground and others in the canopy ,075, 176) may
reflect differentiation among closely related species to utilize different portions
of the environment for courtship, oviposition, and other activities.
Many species of ithomiid butterflies that dwell in the tropical wet forest
understory tend to have rather transparent wings, as do some of the satyrid
butterflies in the same habitats. The studies of Pliske(I77- 179) have revealed the
role of specific chemical substances in shaping courtship-related communication
MACHINERY OF ENVIRONMENTAL RESPONSE MECHANISMS 91

between males and females in ithomiid species. The males of various ithomiid
butterflies routinely visit dried or fresh flowers from which they apparently
obtain precursor substances for pheromones associated with courtship (Figs. 3.10
and 3.11).
When a cluster of species associated with a habitat exhibit similar
morphologies, reproductive isolation and mate recognition are determined
largely by olfactory rather than visual signals between sexes. Ithomiid and
satyrid butterfly species in deep forest understories might have converged in
cryptic coloration to escape predators that hunt visually, and the evolution of
chemical communication systems in such groups compensates for a loss of
species specificity that might otherwise be associated with color and morphology
under other conditions. In other insects visual cues may operate at a distance to
bring potential mates into close juxtaposition, but actual courtship is then
mediated primarily by olfactory and tactile cues at close range. Many naturalists
know that the males of several species of Morpho butterflies will be attracted to
the same decoy, suggesting that cues other than vision result in the species
recognition process. Patchy distributions of nonfloral food sources of Morpho
butterflies, such as fermenting sap flow wounds in trees in tropical forests,
function to bring males and females together. Although conditions are generally
subdued on the forest floor, where these butterflies feed in aggregate, sudden
flashes of wings may function to attract others (and perhaps mates) to these
feeding sites (Fig. 3.12). The cryptic undersides of the wings render the feeding
and virtually motionless butterflies very inconspicuous against the forest floor.
These butterflies, and several other allied groups, exploit rotting fruits and soupy
fungal growths on sap flows on or near the forest floor, establishing selection
pressure for cryptic coloration and behavior while feeding. High percentages of
feeding morphos reveal symmetrical fragments of missing wings, suggesting
attempted attacks by small insectivorous vertebrates such as lizards, even while
the butterflies feed. The role of olfactory and tactile cues in butterfly species that
partake of mimicry complexes is also clear.
Many investigations, such as those of Brower and Jones,(J80l have described
the olfactory secretions of male danaid butterflies. In this system a courtship-
inducing substance is produced in glands of the abdominal hair pencils and
deposited, through a specific behavioral sequence, into wing pockets. From there
the secretion disseminates into the habitat to attract females. Vane-Wright(J81l has
discussed the phylogenetic importance of wing scent pockets and other structures
in the evolution of the satyrid and morphid butterflies. Aldrich et aZ. (J82) and
Aldrich and Blum (83 ) have studied the production of courtship scents from ab-
dominal glands in male coreid bugs. They suggest that such adaptations facilitate
the colonization of new habitats if the male bugs precede females and then attract
them into the new zone.
In other systems, such as the mating swarms of male lovebugs (Diptera:
FIGURE 3.11. The ithomiid butterfly Mechanitis isthmia visiting an unidentified bloom at the border
of a coffee plantation in the central highlands of Costa Rica. Because its larvae feed on various species
of Solanaceae, the conspicuous orange, yellow, and black tiger-stripe coloration of the wings is be-
lieved to be aposematic . This ithomiid is a familiar species of secondary habitats throughout much of
southern Central America and also South America.
MACHINERY OF ENVIRONMENTAL RESPONSE MECHANISMS 93

Bibionidae) studied by Thornhill (84) in Mexico, swarming by males establishes a

syndrome of competition for females, which are attracted into such swarms. The
swarms are established before females emerge, and the swarms are stratified verti-
cally in terms of male lovebug sizes, with larger individuals near the bottom. Larger
males near the bottom of swarms have greater access to emerging females and
copulate more frequently. The conspicuous swarming behavior of many tropical
dipterans, most notably of the families Culicidae, Cecidomyiidae, and Ceratopo-
gonidae, in which swarms are dense and often columnar, warrants further study
considering the findings of Thornhill. Such swarms may be highly structured in
terms of the mating strategy. Yet relatively few data exist on the mating strategies
of nonsocial insects, with the obvious exception of many studies of Odonata. In
many species, males tend to aggregate near emergence sites or critical resources to
attract females, which generally have less dense populations. It is difficult to
generalize across the board for all insects, but data from butterflies, hemipterans,
flies, cicadas, and others in the tropics suggest mating behaviors in which female
individuals exhibit cryptic behavior and in which males exhibit single or mixed
strategies of attracting females for mating.
Such patterns of adaptation are likely to be present in some groups outside
of the tropics, but these might be due to the tropical origin of the groups or
specialized habitat conditions favoring male competition for females. (85 ) In
studying the distribution of males and females of tropical forest butterfly species
that feed on rotten fruit on the forest floor, Young075.186-189) and Young and
Thomason(190) found that males predominate at such baits, for all species studied.
Furthermore, females tend to show up at times different from periods of peak
male abundance. Mating does not occur at bait sites. Females, particularly mated
ones, presumably have different daily activity patterns from males, probably
associated with oviposition and avoidance of predators such as birds. The ob-
served high densities of males of some butterflies, such as M. peleides,075,190> at
fruit baits suggest that males aggregate in the habitat to attract females, even
though this behavior might not occur at baits; males are prone to aggregate. The
tropical rain forest contains a wealth of rotting fruits at different times of the
year, a situation which undoubtedly promotes swarming behavior in many flies
and other insects attracted to the fruit for food and oviposition. The presence of
vultures in the tropics probably reduces the likelihood of courtship systems
evolving in many insects that could feed on carrion.o 9J)
Swarming in the social insects has been reviewed by Wilson(80) and others.
Naumann (92 ) has examined the role of trail odor marks in determining the
swarming direction in Polybia wasps in Ecuador. Polybia nests are frequently
raided by Eciton army ants, as are nests of Polistes, (93) resulting in the forma-
tion of "absconding swarms" to differentiate from those swarms associated with
the founding of new colonies. (See Wilson(80) and West-Eberhard(94 ) for detailed
discussions of the evolution of social behavior in insects and the analysis of
94 CHAPTER 3

FIGURE 3.12. Cryptic feeding behavior in Morpho butterflies . (a) A group of several morphos feed-
ing at a small sap wound in the exposed root of a canopy-size legume tree near Puntarenas, Puntarenas
Province, Costa Rica. (b) A disturbed butterfly that has opened its wings, revealing a flash of iridescent
blue in a patch of sunlight.
MACHINERY OF ENVIRONMENTAL RESPONSE MECHANISMS 95

FIGURE 3.12. (Continued)

chemical communication in social insects.) In general, the high species richness

of social wasps and ants in the tropics reflects considerable phylogenetic and
ecological diversification in behavioral mechanisms regulating a variety of col-
ony integrative activities such as alarm responses, recruitment, swarming, and
96 CHAPTER 3

reproduction. Such features of colonies are molded largely by prevailing ecologi-

cal conditions superimposed on the historical development of character states
within taxa of social insects. For example, Crewe and Blum09S ) have discussed
the phylogenetic significance of alarm pheromones in the Attine ants. They
discovered a "common chemical heritage" based on 3-ketones and 3-alcohols
for the genera studied, and these properties were found to agree with the accepted
phylogeny based on other characters. Other studies (e.g., Blum et alY96» have
analyzed the roles of certain chemicals in activities such as trail laying and
robbing in social insects.
The behavioral repertoires of social insects must be interpreted in terms of
both historical and contemporary ecological factors affecting species and genera.
The great diversification of foraging behavior exhibited by ants, especially in the
tropics (Carroll and Janzen(97 » reflects considerable specialization in terms of
food gathering. Eisner l14 ) and others have stressed the convergent occurrence of
similar chemical substances in different insects and other arthropods (largely the
result of the substances being the most effective in carrying out a particular
function, and being able to be metabolically produced directly or indirectly by the
internal physiology of the organisms in question.) Thus citral is used by stingless
bees in tropical forests to mark trails and by robber bees in tropical forests to
disorient their victims. (96)
It is impossible to review here all of the interesting features of insects in
tropical habitats in terms of behaviors associated with mating and communica-
tion. To understand the diversification of such behavior one must examine as-
semblages of closely related taxa of insects in a particular habitat or range of
habitats and relate the observed behaviors to differences in how individuals in
popUlations of each species utilize available resources in space and time. The
types of mating and courtship behaviors that evolve will be a function of the past
evolutionary history of the group and the present-day selection pressures shaping
resource utilization, degree of intraspecific and interspecific competition, preda-
tion, energy budget, and population structure as related to seasonality or lack
thereof.
A recent emphasis on sexual selection mechanisms in populations(85 ) has
emphasized the importance of intermale competition for females in insects and
the role of such mechanisms in the evolution of mixed mating strategies within
single-species populations. We need to reexamine and describe the mating
strategies exhibited by different groups of tropical insects to determine the impor-
tance of intermale fighting or competition as a driving force in providing or-
ganisms with "better fits" into their environments. In the context of this book it
is important to realize that such data are ordinarily lacking in tropical insects and
some kinds of tropical habitats are the places where such specialized forms of
mating behavior are most likely to be found, particularly if such behaviors are
related to the high densities of species usually associated with such habitats.
MACHINERY OF ENVIRONMENTAL RESPONSE MECHANISMS 97

Outside of the social insects, butterflies, flies, and cicadas figure strongly as
suitable insects through which to examine some of these ideas.
It is also necessary to consider the role of selection in determining the
internal reproductive physiology of insects as a major component of reproductive
behavior in general. The integration of the generalized insect phenotype in terms
of the fine balance of hormonal processes regulating embryonic development(!98)
is at the very foundation of reproductive physiology in insects. Nutrition, light,
and mating figure prominently as regulative factors among the wide variety of
internal and external environmental factors that determine egg maturation in
insects (Engelmann(J99»). Reisen and Aslamkhan(20o) have found that the gono-
trophic cycles of some tropical mosquitoes are regulated by blood meals, whether
or not females have been inseminated.
The considerations of Cody(67) stress the high premium placed by selection
on the evolution of clutch sizes, notably in tropical organisms. If there is a
greater amount of biotic regulation of insect populations in the tropics, particu-
larly for lowland wet environments, selection will favor those clutch sizes within
single popUlations of each species that permit these populations to remain in
some fashion of equilibrium with the popUlations of other organisms composing
the niche. Cody argues that smaller clutch sizes evolve in the tropics to ensure
that popUlations stay at their carrying capacities. Unstable conditions tend to
select for larger clutch sizes associated with changes in other population growth
parameters that promote a higher rate of popUlation growth. Yet a growing body
of evidence suggests that the very instability of some tropical regions promotes a
greater diversity and therefore should select for population growth parameters
favoring colonization and invasions to new habitats.
The great range of habitats in the lowland wet tropics, including some that
are relatively new and unstable, provides sufficient habitat heterogeneity to pro-
mote colonizing episodes by insect species, particularly and initially at the herbi-
vore and decomposer levels of community organization; this means the elabora-
tion of phenotypes programmed for increased clutch sizes or at least a mixed
strategy favoring phenotypes of low and high clutch sizes. There appears to be
too much environmental heterogeneity in the lowland tropics to suggest a uni-
form hypothesis of reduced clutch sizes. SimpsoU<20I) has discussed the role of
historical regional events in shaping areas of high diversity in South America,
suggesting that large-scale environmental heterogeneity accounts for increased
diversity.
A. H. Ehrlich and P. R. Ehrlich(202) suggest that multiple matings are
probably frequent in most butterflies, including tropical species, but that multiple
spermatophores are used for food rather than for mating. The existence of multi-
ple matings in these insects seems to be indirect evidence for the existence of
mating systems in which males compete for females. In some tropical butterflies,
mated females exhibit specific displays designed to discourage multiple matings.
98 CHAPTER 3

High predation rates from birds and other vertebrates on tropical butterflies are
debatable. Much of the theory of mimicry and development of wing coloration in
the Lepidoptera is based upon the relative roles of such color patterns and
associated flight behaviors in mating and predator avoidance or evasiveness.
Studies by Brown, and his associates(203-205) and others with heliconiine and
ithomiid butterflies with tiger-stripe wing color patterns have shown that birds
actively eat butterflies, as has the study by Brower et al. (206) of overwintering
monarchs in Mexico. MOlpho and other butterflies feeding on rotten fruits in
tropical forests tend to exhibit symmetrical wing damage patterns at the posterior
edges of the hind Wings<I75,lSS,lS9,207) (Fig. 3.13). Young OSS ) has interpreted such
patterns to imply predatory attacks primarily on butterflies that are resting or
feeding rather than flying. The evolution of false eyespot patterns in these fruit-
sucking butterflies as an adaptation for increased crypsis for late-afternoon feed-
ing has been discussed by Young. OSS ,lS9) Allied species active higher up in the
forest tend to have reduced or no eyespots on the ventral sides of the wings,
suggesting a lack of selection favoring these markings away from the forest floor.
Mating in some butterflies involves males emerging before females within a
cohort and waiting for females to emerge. This behavior is well known for some
species of Heliconius butterflies (e.g., Alexander(20S). Young(77) suggested that
aggregates of freshly eclosed males of the highly sexually dimorphic tropical
pierid P. lypera wait together in the forest understory near the emergence sites of
the females and then compete for mates. This species exhibits cluster oviposition
and highly gregarious larvae, conditions promoting aggregated pupation and
synchronous adult emergences. Males of Morpho amathonte exhibit extremely
predictable flight routes through forest clearings on a daily basis, perhaps as an
adaptation for attracting females waiting in vegetation along such routes (A. M.
Young, unpublished data). Such a system may disperse males so that inter-male
competition for the same females is lowered. In the above example with Perr-
hybris, unmated males appear to exhibit a form of lek behavior by aggregating
to attract females. The behavior of pupal mating is well known for some species
of the Neotropical butterfly genus Heliconius (Fig. 3.14). Males will be attracted
to the yet uneclosed pupae of females, presumably as a result of an odor pro-
duced by the latter, and mating takes place shortly after eclosion and expansion
of the wings. Such behavior may reduce the amount of time required for mating
in the wild and thereby increase the time available for egg laying. The selection
pressure driving such behavior would be predation on the adult butterflies and
perhaps predation on the immature stages as well. The adaptive significance of
such behavior, coupled with the simultaneous development of the ovaries, is that
females can begin to oviposit soon after eclosion. Ringd 209 ) discussed the lek
behavior of males in some species of Hawaiian Drosophila. In Drosophila
grimshawi, male displays are density dependent on frequency. As population
density increases, so does the frequency of male displays. The presence of
females increases the frequency of communal displays and aggression in males.
FIGURE 3.13. Eyespot markings on the undersides of the wings of Morpho butterflies (a) compose
a form of crypsis when these insects are feeding in shaded areas of the forest floor. When light rays
hit the wings (b) the vertebrate-like eyespots are enhanced and may function otherwise to discourage
attack by small insectivorous vertebrates. Attac~s from the rear of a feeding butterfly may be largely
unsuccessful, leaving symmetrical gaps in the wings.
100 CHAPTER 3

FIGURE 3.14. Pupal mating in the Neotropical butterfly Heliconius sara. (Photographs from Finca
La Selva, Puerto Viejo, Heredia Province, Costa Rica, courtesy of Robert Borth.)

3.7. RHYTHMICITY IN TROPICAL INSECTS

The large change in temperature associated with day-night cycles in the

lowland tropics provides a background of abiotic effects governing the expres-
sion of clear-cut diurnal rhythms of activity in many insects. Often different but
closely related species with similar ecological habits exhibit different temporal
activity rhythms. Insects of Temperate-Zone areas also exhibit different diurnal
patterns of activity, but in the tropics, where insect diversity is great, such
behavior is probably very adaptive in the sense of reducing competition among
ecologically similar forms that exploit similar or identical resources. The noctur-
nal habits of some Temperate-Zone insects, such as many Orthoptera, may form
MACHINERY OF ENVIRONMENTAL RESPONSE MECHANISMS 101

a relict trait that was previously adaptive in tropical environments. Basically four
major groups of insects, classified according to activity patterns, are seen
superimposed upon the diurnal cycle in the tropics: (l) insects active in the day,
(2) insects active at dawn and predawn, (3) nocturnal insects, and (4) insects
primarily active at dusk. Not all of the classes are mutually exclusive, since some
groups of insects, such as cicadas, can include species with individuals active at
dusk as well as during the day. Young086 ,189) found that fruit-sucking tropical
butterflies in forest habitats exhibit clear daily patterns of feeding. An example is
M. peleides, a familiar, widespread species belonging to the superspecies achil-
les of South America. Males tum up in greatest numbers between noon and 2:00
P.M. on sunny days to feed at deliberately placed baits of rotting bananas on the
forest floor. Females are less predictable, turning up, always in very small
numbers, at all hours of the day. On some days a burst of feeding activity in both
sexes occurs near dusk. Ca/igo butterflies, on the other hand, appear in peak
numbers at the same baits early in the morning and steadily decline throughout
the morning.
Although Hermann(21O) has shown that workers of the giant tropical ant Para-
ponera clavata are nocturnal foragers in Ecuador, Hermann,(21O) Young,<211l and
Young and Hermann(212) discovered considerable diurnal foraging in cloudy or
overcast weather conditions in northeastern Costa Rica. In Young's studies, there
was usually a prominent burst of nest exits in late afternoon and through the dusk
period. Young attributed the observed patterns offoraging rhythmicity to prevail-
ing weather conditions. There is some evidence, however, that this species and
other ants possess endogenous activity patterns. (213) Tropical ants in general
might be genetically programmed for specific endogenous activity periods, but
these are modified by the influence of prevailing weather conditions experienced
by nest members each day. Many tropical ants exhibit definite diurnal cycles of
foraging activity, while some species are "generalist" foragers, being active
through most of the day and night.
The highly synchronized nocturnal chorusing patterns of many Orthoptera is
well known. In tropical forests, the greater vertical and horizontal structural
complexity of vegetation in forest habitats provides a pattern of environmental
heterogeneity for diversification of habits among the Orthoptera. Diversification
in feeding habitats and habits is accompanied by species-specific songs and
chorusing patterns. The evolutionary significance of singing in the Orthoptera
has been reviewed by Walker.(214) There are more genera and species in most
families of Orthoptera in lowland tropical wet forests, including groups such as
mantids, phasmids, and roaches where sound production is less developed than
in other lineages. Convergent or phylogenetically generated nocturnal activity in
the Orthoptera is made possible in part by the great diversification in singing
habits, coupled with ecological differences associated with feeding. There is a
greater incidence of seed-eating orthopterans in the tropics. Many tettigonids are
opportunistic predators on insects at lights in the tropics.
102 CHAPTER 3

The "sundown cicada," Fidicina mannifera (Fig. 3.15), is a large, robust

insect that exhibits short bursts of singing at dusk and occasionally at dawn. In
the lowlands of northeastern Costa Rica, the males begin to sing at about 5:50
each evening and sing to about 6:05 or thereabouts. On the Pacific side of the
country, the cicada sometimes sings intermittently throughout the day, as seen
for one population in lowland Puntarenas Province, and possibly further north in
Guanacaste. These differences in the pattern of daily singing between popula-
tions on either side of the central Cordillera suggest that abiotic factors such as
daily regimes of temperature change are responsible for the timing of chorusing.
A host of other cicada species, such as Fidicina sericans, Fidicina pronoe, and
Quesada gigas (Fig. 3.16), also exhibit bursts of dusk singing on the eastern side
of the Cordillera, but these species also chorus during the day. The males of these

FIGURE 3.15. The Neotropical cicada Fidicina mannifera. Lowland tropical dry forest populations
in Costa Rica are characterized by smaller average adult body size and lighter markings on the body
than adults from lowland tropical rain forest.
 ~
o
:z:
Z
!;B
-<
o."
m
:z
<
:5
o
:z
3:
m
:z
~
r-
;a
m
C/)

~
:z
C/)
m
3:
m
o
:z:
>
:z
(jj
3:
C/)

FIGURE 3.16. Some characteristic cicadas (Homoptera: Cicadidae) of lowland tropical rain forest in Costa Rica. From left to right: Quesada gigas, ...
Zammara smaragdina (male and female, respectively), Fidicina pronoe, Fidicina sericans, and Fidicina amoena. S
104 CHAPTER 3

and other species position themselves in aggregates in trees(215) perhaps in such a

manner so as to optimize sound transmission to attract mates. Young(!50) has
hypothesized that dawn or dusk aggregated chorusing in tropical cicadas repre-
sents a convergent adaptation of different genera and species to chorus at times
when sound transmission is the highest. Selection for optimal mate attraction in a
multispecies system has resulted in convergent chorusing, and participant
species have distinct songs. Costa Rican lowland and premontane tropical rain
forests support a rich diversity of genera and species of cicadas, and different
species in the same area often exhibit differences in peak annual emergence
period and habitat association. When several species occur in an area, there is
considerable range in popUlation densities as well.
The sweep sample studies of Janzen(72) have shown that whole groups of
insects in different trophic levels of communities replace one another between
day and night in various tropical habitats. Such data emphasize the role of
diurnal cycles in the structure and composition of communities of insects as-
sociated with vegetation. Detailed reviews have shown the role of abiotic factors
in shaping the diel rhythms of activities in many insects. The interesting consid-
eration for the tropics is the possible selection pressure created by limited re-
sources as a force in detennining differences in peak activity periods of ecologi-
cally similar species.
Another major set of factors regulating activity rhythms in tropical insects is
the response of plant species to diurnal cycles. A clear case is that of syn-
chronized flowering in many tropical plants. The recent studies of H. G. Baker
and I. Bakd216,217) and Baker et al. (2IS) have revealed diel rhythms of nectar
production in several tropical plant taxa. Pollinating insects and other animals
seeking nutrients from nectar will have foraging periods synchronized with peak
nectar production.(21S) As discussed in Section 3.5, such insects must have the
physiological means to be active at these times of the day in order to obtain suf-
ficient nutrients. In short, the very sharp temperature gradient over a 24-hr period
in tropical lowlands provides a cue for those physiological mechanisms respon-
sible for synchronized flowering in many species. These conditions in tum pro-
vide selection shaping the internal physiology and foraging behavior of pollinat-
ing insects. The existence of different flowering strategies over 24-hr periods in
many tropical plants provides a source of environmental heterogeneity promoting
diversification in peak foraging periods of flower-visiting and pollinating insects.

3.S. SOCIALITY AND DEFENSE IN SOCIAL INSECTS

Since it is not possible or appropriate to discuss here the biological features

of social insects in depth, the reader is referred to Wilson,(SO) who provides
considerable infonnation on the definition of true social behavior in insects and
discusses the evolution of sociality.
MACHINERY OF ENVIRONMENTAL RESPONSE MECHANISMS 105

Because the social insects, including both the Hymenoptera and Isoptera,
reach their greatest taxonomic diversity in the tropics, some brief comments are
made here concerning generalized features of these insects, possible mechanisms
of diversification, and expression of defense and other forms of behavior gener-
ated at the level of the colony. In subsequent chapters we return several times to
examples of the impact of social insects on tropical communities.
The tropics provide a greater structural and chemical diversity for highly
adaptable social insects such as ant species to respond to in the context of
ecological specialization. One need only think about the great diversity of nest
site patterns exhibited in ant species in the tropics. Many ant species are
specialists at arboreal nesting, making nests in hollow twigs, epiphytes, and
other structures off the forest floor. Brown and Wilson,(219) for example, have
traced the evolutionary changes in the foraging behavior of the Dacetine ants,
suggesting shifts in both foraging habitats and degree of generalized versus
specialized predatory behavior. The dacetines exhibit well-delineated patterns of
morphological-ethological coadaptive systems associated with foraging in dif-
ferent kinds of habitats.
The general hypothesis is that the great opportunities for evolutionary diver-
sification at subfamilial, generic, and specific levels of taxonomic organization
have been concomitant with the need for niche specialization in terms of nest
sites and to some extent feeding habits. Perhaps more than most other groups of
social Hymenoptera, ants have exhibited the most morphological and behavioral
diversification in response to diverse ecological resources. Diverse feeding
habits, some generalized and others specialized, are seen in the ant groups of the
tropics: scavengers, predators, exudate gatherers, nectar gatherers, pollen feed-
ers, and parasites. Termites, another large group of social insects, apparently
have less diversification, being limited to functioning as decomposers of dead
wood and other cellulose detritus in tropical forests. The tremendous ecological
diversification in hymenopteran social insects was initiated with a huge adaptive
radiation more than 50 million years ago, endowing the groups with these three
features: systems of castes and functional roles within the colony that change
with adult age; elaborate systems of chemical communication that include signals
of alarm, recruitment, and recognition of nest mates and non-nest mates; and
elaboration of nest structure and nest positioning in the environment as means of
regulating temperature and humidity inside nests.
Wilson(SO) discusses the various stages in the evolution of eusocial behavior
in the social insects, as well as promising directions in research that bring
together the analysis of social systems with population dynamics and evolution-
ary genetics. According to Wilson,(sO) three major traits are characteristic of
eusocial insects: individuals of the same species cooperating in caring for young;
reproductive division of labor, with sterile or semi sterile individuals working on
behalf of the female reproductive caste (queen); an overlap of at least two
106 CHAPTER 3

generations in life stages capable of contributing to colony labor, so that off-

spring assist parents for some portion of their life-span. The greatest diversity of
eusocial insects occurs in the tropics, along with many examples of presocial
behavior and colony organization in some insects. The study of tropical ants and
social wasps provides the opportunity to determine the relationship between
modifications in a basic design of eusociality adapted to particular ecological
factors, including next habits, food resources, predation and other forms of biotic
mortality on whole colonies, and influences of abiotic conditions.
It is noteworthy to appreciate the extreme ecological diversification of social
insects in the tropics. Wilson(SO) points out that the center for speciation in the
vespine paper wasps is tropical Asia, where as many as 50 species occur. From
here, the group has penetrated the Temperate Zones, although fewer species are
present. The long growing season of the tropics has promoted the development of
multiple-foundress systems in some social paper wasps, a uniquely tropical phe-
nomenon which allows the production of large workers castes to counteract the
effects of biotic mortality on workers within a given nest. Such systems are found
in Neotropical species of Polistes. In the Vespinae, another major group, the
queen is much larger in size and is the sole egg producer in the nest. In this group
there is considerable sophistication in the structure of nests, with carton nests
being located in the ground, on branches, or in tree cavities. Tropical wasps are
basically predatory but also exhibit opportunistic foraging for nectar, extraflora
exudates, rotting fruit, and other materials rich in nutrients such as carbohy-
drates.
Wilson(SO) has traced the hypothetical evolutionary steps in the development
of eusociality in wasps, accounting for the various forms of pre social behavior
exhibited by families such as the Sphecidae and Pompilidae. The steps involve a
progressively increasing care for young, food provisioning, and elaboration of
nest structure to house overlapping generations of young and adults. Wilson(SO)
emphasizes that studies of wasps have yielded much of our understanding of the
nutritional control of castes; trophallaxis, the process by which adults exchange
food with larvae; and the mechanisms by which queens regulate the activities of
the colony. The social wasps of the tropics are probably significant predators on a
variety of soft-bodied insects.
Y oung(220) and Young and Moffett(221.222) found that vespids take heavy tolls
of young lepidopterous larvae in a study of the butterfly Mechanitis isthmia (Fig.
3.17), and there is no reason to believe that other species are avoided as prey.
Few studies have been made, however, on the impact of predation by tropical
social wasps on insect populations. Given the usually patchy distribution of
nests, such effects, if significant, are also predicted to be patchy, a function of
nest site relative to optimal prey and other resources and to colony size. Although
some preliminary evidence suggests that under some conditions Polistes popula-
tions can be regulated by predation on larvae and pupae by army ants,093)
 ~:::r:
Z
m
~
o."
m
Z
<
ii
o
z
s:m
~
r-
:J:J
m
C/l
"tI
o
Z
C/l
m
s:m
CO)
:::r:
:to
Z
Cil
s:C/l

FIGURE 3.17. A polybiine wasp taking (in its mandibles) a second-instar caterpillar of the Neotropical ithomiid butterfly M. isthmia off its food plant,
Solanum lancefolium. in northeastern Costa Rica. Note the other larvae present. Despite repeated attacks by the wasps, the larvae were unable to defend them-
selves, and although Ectalomma ants roam over the leaves and larvae, they do not protect them from such predation. But because the ants collect drops of fluid
exuded from the freshly damaged leaf where the larvae are feeding, the interaction could be the initial stage of a coevolved mutualism involving the larvae and ...o
ants in response to intense predation by these wasps. "-I
108 CHAPTER 3

FIGURE 3.18.(a-<l) The effect of human sweat on a species of stinging polybiine wasp in Costa
Rica. The wasps act stunned and lethargic.
MACHINERY OF ENVIRONMENTAL RESPONSE MECHANISMS 109

FIGURE 3.18. (a-d) (Continued)

110 CHAPTER 3

experimental proof is lacking. Army ants also destroy broods of Polybia and
other vespid wasps in the tropics. (92 ) Whether nesting sites for certain kinds of
social insects are limiting resources in the tropics has not been studied. Given the
high density of some taxa such as ant species per unit area of habitat and the
patterns of ecological specialization in some of these, nesting sites may in fact be
limiting resources. However, the study by Matthews(223,224) on the presocial wasp
Microstigmus comes in Costa Rican tropical rain forest suggests a rather high
density of nesting sites per unit area of forest understory.
For most ant species, the grain of the environment with regard to nest sites is
probably far more complex than for wasps in the same habitat. This is an intuitive
impression based on the observation that ant species densities are greater than
densities of wasp species within and between habitats in a region. But such
reasoning is something like comparing bananas with coconuts until more data are
available on resource specificities in various taxa. Preliminary data indicate that
the giant ant Paraponera clavata, a ponerine, prefers to nest beneath the root
systems of only certain tree species. (212) One might expect considerable nest site
limitation in highly specialized nesting taxa such as ant species associated with
hollw twigs, and certain epiphytes.
The studies of Rockwood<225,226) suggest that leaf-cutter ant colonies are
limited by the availability of preferred tree species whose leaves are used as
substrates for fungal gardens harvested for nutrients within the nests. The appar-
ently generalized foraging habits of army ants,(26,27) in which these ants are
predatory on a variety of arthropods, probably resulted from the nomadic condi-
tion of the colonies, in which there is no permanent nest sites. A nomadic
lifestyle allows army ants to become supergeneralists on arthropods in general
and to retrieve energy in small packet from large amounts of habitat. Once again,
however, very little information exists on the impact of army ants as predators in
tropical communities and whether their movement patterns impose a sort of
regulation on social wasp species that tend to nest repeatedly in the same habitats
over a run of years or longer.
Schmidt(227) has reviewed the great diversity of ant venoms found in the
group and relates such biochemical diversity to the tremendous ecological suc-
cess of ants in general. Ants produce some of the most toxic of all animal venoms
known,(227) and the evolution of the sting apparatus in the Hymenoptera(228) has
been a key factor in allowing such large concentrations of insects to escape from
vertebrate predators. Schmidt<227) also suggests that the venoms of other
hymenopterous groups such as bees and wasps might be very different from those
of ants overall. Young(229) reported that human sweat appears to calm workers of
a Polybia wasp species to the point that occupied and thriving nests can be
crushed in a hand coated with human sweat (Fig. 3.18).
We need to relate the level of toxicity of hymenopteran venoms to both
phylogenetic and contemporary ecological factors such as predation rates on
MACHINERY OF ENVIRONMENTAL RESPONSE MECHANISMS 111

nests by vertebrates and invertebrates along with other forms of colony defense.
Even though stingless bees do not sting, they can effectively discourage attack by
vertebrates through behavioral responses. (230) In stingless bees, a dominant group
of social insects in the American tropics and a group of opportunistic pollinators
of many plant species and nectar robbers of others, nests are hidden in tree
cavities and the next entrance is a resinous tube sealed over each night to prevent
other organisms from entering. When a nest is attacked, the bees will swarm and
bite the attacker, and tangle up hair and feathers (A. Wille, personal communica-
tion). Wille and Michenert 230 ) provide a review on the evolution of stingless bees
and their nesting habits.

3.9. DENSITY-RElATED BEHAVIOR IN INSECTS

The density of an insect popUlation influences the environment in several

ways: (1) increasing densities coupled with a lack of dispersal that result in an
overcrowding of the population's membership for the resources the species needs
for survival and reproduction, sometimes resulting in popUlation decline and
extinction; (2) selecting for phenotypes suitable for dispersal to new habitats in
response to increasing densities on limited resource supplies; (3) creating genetic
and nongenetic changes in developmental time or other parameters affecting the
growth rate of the popUlation; (4) alternating existing spacing patterns of the
population's membership. (5) changing various forms of behavior collective or
singly, including mating systems and defense; (6) increasing densities that may
intensify competition with other ecologically similar species in the same habitat.
The laboratory studies of Thomas Park and his students have demonstrated
the deleterious effects of increasing population densities on survival and reprod-
uction,<23P and studies (e.g., Dawson(232); Ayala et ai. (233); Lewontin(84,85» have
shown the operation of density-dependent selection altering the genetic structure
of insect populations. Allee(234) stressed the effects of popUlation density on
developmental time, survival, and other properties of animal species in general.
Many ecology textbooks discuss the effects of density-dependent and density-
independent factors on insect populations. Studies (e.g., Wharton et ai. (59» with
cockroaches have demonstrated that increases in the population density of larval
stages can result in adult cohorts of reduced mean weight of adults and that males
are more sensitive to the deleterious effects of high densities than females. Low
densities promote faster development and higher survival to the adult stage (Fig.
3.19).
Other studies indicate that in some insects, aggregative behavior, in which
several or many individuals of a species stay together, is adaptive for feeding and
defense against predators. Fraser-Rowell(235) demonstrated the role of visual
stimuli arising from individual color patterns in promoting the formation of
112 CHAPTER 3

(500,650)
100
80

1le' 60
CD

tE 40
*

18 21 24 27 30 33 35
Time,wk
FIGURE 3.19. Egg-to-adult development curves for different population densities of nymphs grown
in the same size jars (except for curves IB and 2B, which were grown in smaller containers)
expressed as percentages of all surviving insects, both nymphs and adults. The vertical lines crossing
the curves show limits of the averages for the group of (500,650) and (750, 1000). [From D. R. A.
Wharton, J. E. Lola, and M. L. Wharton, 1. Insect Physioi. 13:699-715. Reprinted with permission.
Copyright 1967, Pergamon Press.]

aggregations in the tropical grasshopper Phymateus purpurascens in Uganda.

Aggregative behavior is most evident in the brightly colored larval stages of this
species and is believed to reinforce noxious properties to predators. The larvae
rest and move as groups. Myers and SmiM236) described the synchronous head-
flicking behavior of tent caterpillars in aggregates as a means of defense against
parasites. Here, the stimulus is the flight sound of a fly parasite that lays its eggs
on the caterpillars. Otte and Joern,(237) also working with grasshopper popula-
tions, developed a model which predicts that the number of females successfully
mated within a population by the most successful males (defined here as those
having optimal territories) will initially increase as density increases, and that as
density increases further territorial defense becomes too costly energetically and
aggressive behavior then decreases. At low densities there are many more suita-
ble territories (suitable bushes) than male grasshoppers, but at high densities
territorial defense of suitable bushes increases and many males are excluded from
entering the bushes and mating by the most aggressive (successful) males. Thus
fewer males mate with more females as density increases, but after a point the
defense of suitable bushes becomes a major energy drain and the system breaks
down.
Because of the tremendous variation in the plant species composition and
structural organization of tropical habitats within a particular region, most nota-
bly in the tropical lowlands, there exists the potential for various strategies of
spacing patterns in insect species inhabiting different trophic levels. Two ex-
tremes within such a spectrum of habitat types might be one in which many
taxonomically unrelated plants are present and in which each taxon occurs in
MACHINERY OF ENVIRONMENTAL RESPONSE MECHANISMS 113

small clumps (patches) or single individuals with considerable distances between

patches. In the second case, few plant taxa are present and the abundances of
each are high. The first habitat therefore is more heterogeneous spatially than the
second one. An example of the first type might be a strip of advanced secondary
forest bordering lowland primary wet forest, and of the second, a strip of very
young secondary growth along a road in the same area. If herbivorous insect
species in both habitats are assumed to be specialists, selection might have
favored the evolution of population growth characteristics favoring the mainte-
nance of population densities at or near carrying capacities for each species. If
such species are regulated by predators and parasites, the populations might have
high growth rate characteristics counteracted by such density-dependent
mortality, and if such mortality is low or absent, the energy drain, so to speak,
has been the tradeoff between parameters such as high clutch size and effective
defense systems against predators and parasites. Under these conditions, and
because there are more insect taxa in the first habitat within the plant-feeding
community, densities of individuals on each host plant are probably low to
moderate but seldom reaching levels for intraspecific competition to be invoked.
The insect predator and parasite popUlations under these conditions are expected
to be composed primarily of generalists capable of exploiting more than one kind
of host, since the density and distribution of each host species make each re-
source rare.
The litter beneath the vegetation in such a habitat is expected to be
heterogeneous physically and chemically, and it may support a greater diversity
of decomposer and scavenger insects and other arthropods, including ants. This
assumes that certain distinctive taste and nutritional features of each plant species
are retained in some form during at least initial stages of the decay cycle. For the
first habitat, therefore, little or no density dependence is expected to be within
the population of specialist herbivore species but perhaps some in predator and
parasite popUlations, since these species (tachinids, various hymenopterans) are
expected to be generalists and therefore overlap in resources.
A different picture is expected for the second habitat. The popUlation of
each herbivorous insect species is expected to be large since resource patches are
large, and within such a habitat predators and parasites might in effect become
specialists on these herbivore popUlations, which can absorb such mortality from
having popUlation parameters promoting rapid population growth associated with
colonizing new resources. Litter communities are expected to be less diverse in
this habitat unless multiple species can exploit the same resources, since patches
are large. Keep in mind that these examples are not necessarily the case for all
situations in the tropics. It is dangerous to oversimplify complex systems or
assemblages of species and other taxa. Habitats with many plant species may
very well support mixtures of specialist and generalist herbivorous insect species,
and most litter organisms may be generalists to a large degree. In general, we
114 CHAPTER 3

will have to examine the population dynamics and general ecology of individual
insect species to determine what sorts of environmental conditions promote den-
sity dependence in the tropics.
In habitats where a species is a resource specialist (low niche breadth) and
popUlation growth is high, density-dependent effects will be operative, perhaps
promoting frequency-dependent selection reSUlting in geneotypes best adapted
for colonizing new habitats. Young(71) has discussed the possible interplay of
forest habitat destruction, host tree specificity, and population density in promot-
ing some cicada species to invade secondary habitats in the tropics. Data from
sweep-sample studies in the tropics(72) suggest some overlap in insect faunas in
different adjacent habitats, and the types of selection pressures affecting each
species might be very different between habitats (see Chapter 8). Following
similar lines as my discussion of the clutch size hypothesis, we can expect
considerable variation in the population structure and dynamics of insect species
within a trophic level, implying the differential role of density dependence,
interspecific competition, and degree of resource specialization as the major
factors affecting populations, particularly in the nonseasonal lowland regions.
Further aspects of popUlation dynamics and other ecological properties of insect
species are discussed in Chapters 5 and 6.

3.10 NOTES ON THE PHYSIOLOGY OF FEEDING IN INSECTS

What follows is a brief introduction to the complex matter of insect feeding,

including food procurement, ingestion, and digestion. My purpose is to highlight
some aspects of feeding in insects that relate to the kinds of generalized and
specialized interrelationships found in the tropics, and to lay a foundation for
discussions of such interrelationships in subsequent chapters. Much of the em-
phasis in the present discussion will be on plant-insect interactions instead of on
feeding by insect predators and parasites on other insects because it is becoming
increasingly evident that much of the structure of tropical terrestrial communities
and whole ecosystems is determined largely by the effects of insects as herbi-
vores and pollinators on plant species and the counterselection of plant species as
resources for insects, and because the bulk of descriptive and experimental data
on the ecology of insects in the tropics concerns the plant-insect ecological
interface. This is not to say there are no other relevant aspects to tropical insect
biology. We have large gaps in our knowledge about the community biology of
predators and parasites, not to mention litter insects and arthropods in general.
And much more autecological or natural historical work is needed on herbivorous
insects and their host plants. Excellent general discussions of feeding in the insects
are found in Brues(238l and Dethier.(239-241l
Using the "feeding categories" of Brues,<238l the tropics support a greater
MACHINERY OF ENVIRONMENTAL RESPONSE MECHANISMS 115

diversity of plant-feeding, predatory, parasitic, and scavenging insects than other

regions of the world. Tropical environments present complex mixtures of food
shapes, colors, odors, and nutrient contents for organisms, and given the rela-
tively small body size of most insects, the grain of food heterogeneity generated
by other organisms in the tropics and by the insects themselves is great. Charles
Darwin emphasized the role of food in the survival of animal species in general.
The implication is, of course, that the behavioral and physiological processes that
shape food procurement and processing form a major, complex sphere of selec-
tion pressures influencing both the evolutionary history of the insects and the
strategies of ecological diversification within genera and species. Within each
level of trophic organization, insects exhibit considerable preferences for certain
kinds of food, based on both nutritive and nonnutritive components of food.
Selection is expected to favor the evolution of a feeding behavior within popUla-
tions of a species that maximize the retrieval of energy necessary for maximal
survival and reproduction. The evolution of sophisticated visual, tactile, and
olfactory sensory organs, mouthparts, digestive tract chemistry, and neuromus-
cular mechanisms associated with food handling and capture have shaped the
evolution of feeding behavior in the insects.
The long evolutionary history of insects in the tropics has provided the
opportunity for considerable ecological diversification as part of the differentia-
tion of whole families, subfamilies, genera, and species. And although insects
feed on a wide variety of animal, plant, and dead organic matter, Fraen-
kel(242,243J was one of the first to demonstrate that most insects have similar nutri-
tional requirements. Most insects require sterols, B vitamins, carbohydrates,
and water to varying degrees.
Food procurement involves being able to recognize and ingest the correct
food items in the habitat. Insects ingest food by chewing it; fluid feeding, in
which enzymes may be injected into the food to assist ingestion; bloodsucking, in
which anticoagulants are injected into the blood; and prey immobilization by the
injection of venom such as through a sting. A few insects, such as some ants,
grow their food, examples being the fungus gardens of tropical leaf-cutting ants
and the' 'honeydew" secretions of certain Homoptera farmed by some species of
ants. In the social insects there is feeding of one individual by another in the form
of a worker caste caring for the young, and such behavior is a major component
of such social systems.
Insects that feed on plants are phytophagous, and if many different plant
taxa are consumed, the species is polyphagous. If food consumption is limited to
plants in one family or closely related families, the insect is monophagous. The
term herbivore is used to connotate the trophic role of the insect in a community.
Phytophagous feeding involves the use of visual, olfactory, and tactile sensory
mechanisms, coupled with gustation. The physiology and structure of
chemoreceptors and eyes in food recognition in insects has been studied exten-
116 CHAPTER 3

sively. Of interest is the food recognition of biting or chewing insects. When a

grasshopper lands on a plant, chemotactile stimulation results in biting, a
nonspecific response. Sustained feeding is due to the presence of certain attrac-
tant substances, some of which, such as sugars and amino acids, may be nutri-
tive. Suitable gustatory stimuli are necessary for continued feeding. Yamamoto
and Fraenkel(244J discuss the roles of olfactory and gustatory stimuli in the ovipo-
sition and feeding of the tobacco hornworm on solanaceous plants, and De-
thier<240,241l gives a general review of the role of secondary substances in plants
on the feeding specificity of insects. Dethier stresses, among other things, the
specificity of certain chemoreceptor sites and secondary substances interacting
with them as the mechanism for host plant selection in phytophagous insects.
Since secondary plant substances apparently have no direct metabolic function in
plants, they are believed to have evolved to protect plants from ingestion by
herbivores (see the recent review of Dethier<241l). Dethier<241l has shown by
behavioral and electrophysiological comparisons that in general the gustatory
receptors of herbivorous insects possess a much wider receptor sensitivity than
the gustatory receptors of nonherbivorous insects. The data suggest that her-
bivorous insects have evolved receptors sensitive to a wide range of secondary
compounds rather than specific deterrent receptors. Dethier reasons that such
generalist receptors provided the central nervous system of these insects with a
large pool of information that formed the basis for the evolution within the CNS
of specific deterrent and acceptance behaviors. It follows that the great secondary
compound diversity present in the lowland tropical floras of the world provide
the response systems for herbivorous insects to act upon, resulting in ecological
diversification in the form of host plant specificities (see Chapter 4).
Visual, olfactory, and tactile stimuli form a major aspect of the evolution of
animal tissue feeding and plant feeding habits in insects. Most insects use some
combination of these types of cues from the environment in the food recognition
process. Studies (e.g., Downes(24SJ) have demonstrated that different sexes may
utilize different kinds of food. Thus female ceratopogonid midges in some genera
and species feed on pollen grains, while males feed on nectar and other sub-
stances rich in carbohydrates. Pollen feeding in insects appears to be related to
females' requiring amino acids from extrinsic sources for the production of eggs,
and some species of Heliconius butterflies use pollen as a source of egg protein
too.(31) I suspect there are many undescribed examples of similar behavior in
tropical insects.
The paucity of female Morpho butterflies and the abundance of males at
baits of rotting fruit (a concentrated source of sugars)(17SJ and the presence of
about equal numbers of male and female butterflies on molds (A. M. Young,
unpublished data) suggests that females may obtain different nutrients than
males. Phytophagous insects appear to have feeding specificity greater than
predatory and parasitic insects. Such an effect is likely the result of prey dis-
MACHINERY OF ENVIRONMENTAL RESPONSE MECHANISMS 117

crimination based largely upon size and not so much on taste, since most soft-
bodied prey should taste about the same (save for noxious substances that could
be detected by biting or sucking predatory insects).
Schoonhoven and Dethier246 ) have reviewed the physiology of host plant
discrimination in the Lepidoptera and have stressed the role of olfactory informa-
tion received by the antennae and maxillae in determining discriminatory
levels. \247) Other forms of specificity in the insects are determined by other
environmental factors. Jaisson(248) discusses the role of early experience in the
determination of nest site preferences in ants, which can be entrained by a plant
stimulus to prefer a particular plant. Heinrich(249) and others have shown how leaf
shape and size affect the feeding behavior of caterpillars, and such studies stress
the role of geometrical and mechanical features of plants in determining how
tissues are consumed.
Janzen(72) has suggested that the mechanisms of feeding specificities in
plant-sucking tropical insects might be considerably different from those regulat-
ing the gustatory preferences of chewing or biting herbivorous insects. Host plant
specificity may not be as well developed in plant-sucking insects as in many
chewing insects, because xylem and phloem fluids may contain fewer secondary
substances. Host plant selection by these forms may be mediated primarily
through the use of physical and mechanical properties of plant tissues rather than
by chemical defense mechanisms. Miles(ZSO) has found that some plant-sucking
Hemiptera secrete a salivary phenolase that may induce the production of a
chemical defense mechanism in plants based upon the oxidation of phenols to
qui nones . Sweet(2SI) argues that phytophagy was the original feeding habit of the
Hemiptera and that selection has favored the expression of feeding on nitrogen-
rich plant tissues (buds, seeds, meristems). Schaefer and 0 'Shea(252l reported
that three unrelated tribes of coreine hemiptems feed on Leguminosae and exploit
those parts of the plants with high nutrient content, and the trait is considered a
symplesiomorphic character. (A symplesiomorphic character refers to the pres-
ence of plesiomorphic character states in different species. A plesiomorphic
character is the character or character state from which transformation begins in
a monophyletic group.) The precise mechanisms responsible for host plant speci-
ficity in both the Hemiptera and Homoptera remain obscure. Young(7J.138,139) has
suggested that the nymphs of many tropical cicadas have preferred primary host
trees in the sense of preferring to feed on the root crowns of these trees, but not
necessarily to the exclusion of root systems of other plants (Fig. 3.20). Adult
tropical cicadas feed on a variety oftree species (Fig. 3.21) in a habitat (A. M.
Young, unpublished data; Young(ZI5», but ovipositing females of some species
prefer to place eggs near particular tree species. (70)
Webb(253) found that spring populations of the South African psyllid Acizzia
russellae were correlated in size with nitrogen concentrations in the leaves of a
host plant, Acacia karroo. Webb also found that Homoptera that feed on the
FIGURE 3.20. Nymphal skins of the cicada Diceroprocta sp. on a Tamarindus indica tree along the
northwest coast of Costa Rica and northern Central America as well. Peak emergence occurs annually
between dry and wet seasons in these strongly seasonal regions of the tropics.
FIGURE 3.21. The cicada F. pronoe (above) feeding from a branch of the secondary habitat shrub
Veronia patens (Verbenaceae) , and Q. gigas (below) feeding from a trunk of the same tree species, in
northeastern Costa Rica. Both cicadas are part of a complex of species widespread from Mexico
through Central America and into South America.
120 CHAPTER 3

cellular contents of plant tissues do not increase greatly in abundance on cut

plants, probably because leaf nitrogen levels are no higher than in uncut plants or
stump sprouts. But phloem-feeding species seem to respond to the increased flow
rate of nutrients in growing tissues of cut plants, promoting increased population
growth. Attempts to make statements about the host specificity of plant-sucking
insects such as aphids and other Homoptera must thus take into account the
physiological condition of the host plant with respect to feeding habit (phloem
feeding, cell feeding). Specificities may change in response to conditions of the
host plant and the population growth requirements of the insect species involved.
Levins and MacArthur<254) believe that a major factor determining the choice of a
host plant by an insect is the effect of such a switch or diet breadth on fitness
within the population. The evolution of monophagy or polyphagy is determined
by the balance between the cost of diet breadth expansion and the cost of search-
ing for new host plant types.
The sequestering of ingested secondary substances from plants by insects
refers to the ability to use these substances for defense. Because most insects are
small, abundant, and full of nutrients such as proteins and lipids, they are a major
food source for many animals. Thus selection within insect species may often
favor the evolution of effective predator-deterring defense mechanisms ,0 14) one
of which is the ability to be unpalatable. One avenue of unpalatability is the
sequestering of plant secondary substances by some insects. In the tropics the
diversity of secondary plant substances is expected to be very high given the high
diversity of plants (although the relationship is not one to one), and predation
rates on insects by other animals can also be high. Thus there might be consider-
able selection for sequestering mechanisms in tropical insects. Alkaloids are
often sequestered, as shown by experiments with danaid butterfly larvae, which
store high concentrations of cardiac glycosides from their asclepiadaceous host
plants.(255) The stored substances, which are bitter, are passed on to the adult
butterfly, which through aposematic coloration reinforces the unpaltable prop-
erties in the predator species with relatively few encounters. Whether or not
flying butterflies are frequently attacked is still somewhat controversial, with
evidence for both points of view .089.204) Studies (e.g., Pliske<l77-179») have shown
that some secondary substances sequestered by butterfly larvae are eventually
incorporated into male pheromones. Thus the pyrrolizidine alkaloids of some
plants are converted to dihydropyrrolizidine derivatives in male pheromones of
danaids. (14 )
The inability to sequester toxic secondary substances from host plants re-
sults in the evolution of cryptic behavior to escape predators, or mimicry to
resemble noxious models. In the tropics (notably the lowland wet regions),
where many groups of phytophagous insects comprise many closely related
genera and species, the pattern of evolutionary acquisition of the ability to
sequester may exist. For heliconiines (Benson(255»), two different genera may
MACHINERY OF ENVIRONMENTAL RESPONSE MECHANISMS 121

feed on the same larval host plant, but only the advanced genus will have the
ability to sequester the secondary substances and become unpalatable. Sequester-
ing appears to be a derived phenotypic trait superimposed upon the detoxification
mechanisms of phytophagous insects. The majority of ithorniid butterflies tend to
possess transparent to translucent wings with reduced markings (Fig. 3.22), and
such genera and species may be palatable even though they feed on the larvae of
Solanaceae, a tropical plant group well known for high alkaloid contents. These
groups may lack the ability to sequester the alkaloids of these plants, even though
many of the same plant species support other aposematically colored insects. A
few ithomiids such as Mechanitis and Hypothyris possess aposematic coloration,
perhaps reflecting unpalatability characteristics derived during the larval stage.
These genera tend to feed on solanaceous species found in open secondary
habitats, whereas the majority of other ithomiids feed primarily in forest under-
stories and utilize different host plant species. In this instance the evolution of
sequestering might have been induced by the transition of some ithomiids to
secondary habitats. The bright coloration of the wings may be more adaptive for

FIGURE 3.22. The ithomiid butterfly Hypoleria cassotis in the shaded primary forest understory in
northeastern Costa Rica. Species such as this one within the Ithomiidae are typical for the shaded
understories of Central and South American rain forests, where adult population densities are gener-
ally low (save for contractions of populations into moist pockets during dry seasons in highly seasonal
areas) and larval food plants (various papery-leaved genera and species of Solanaceae and
Apocyanaceae) are dispersed over large areas. especially within light gaps.
122 CHAPTER 3

courtship in the more exposed habitats in addition to reinforcing unpalatable taste

profiles in these insects.
From the standpoint of insects, plant populations are food resources possess-
ing varying degrees of accessibility of nutrients. Although, as pointed out in the
foregoing discussion, individuals of different plant species may be similar in
terms of the quality and quantity of nutrients available for animals, including
insects, accessibility of nutrients is determined largely by the lines of defense
exhibited by each plant species at a particular place and point in time.
Phytophagous insects are not necessarily seeking out the most nutritive plants to
feed on, but rather those with defense systems that can be most efficiently
overcome by the prevailing genotypic and phenotypic traits of the insect. The
vast botanical literature is filled with examples of insects attacking all stages in
the life cycle of plants and all structures of adult plants and seedlings. As a result
of natural selection, unstable equilibrium is expected between the plant's con-
tinued abilities to defend itself against predators and the predators' abilities to
overcome the defense mechanisms of host plants without appreciable losses in
fitness. In artificial systems such as cultivars and associated crop plantings, the
equilibrium breaks down and usually results in increased predation on the cul-
tivar, because selective breeding often removes the defense mechanisms of the
targeted plant species while enhancing other traits associated with human con-
sumption. Weed species lie at the opposite end of the spectrum from cultivars,
either possessing effective defense mechanisms or saturating the environment
with such a large harvestable biomass that phytophagous insects cannot attack in
a density-dependent fashion. Both strategies result in the successful occupancy of
many habitats.
Levid 256 ) lists the following defense mechanisms in angiosperms: (1) the
texture and composition of stem and leaf surfaces, (2) thorns and resin ducts, (3)
an absence of nutrients, (4) a presence of hormone like substances that alter the
development of insects, and (5) the accumulation of secondary substances in
plant parts subjected to attack. The evolution of secondary substances, as dis-
cussed earlier, is maintained to be a major line of defense by plants against
phytophagous insects. Secondary substances are diverse, being alkaloids, ter-
penes, phenolics, and mustard oil glycosides, and are steroidal and
cyanogenic.(256) The alkaloids enjoy a wide distribution among many taxa of
angiosperms in particular. Secondary compounds can be sequestered in cells or
cellular organelles and released only when the cell is broken, or they may be
stored and secreted by specialized epidermal glands, sometimes located on
trichomes. (256) In terms of general modes of action, terpenes, alkaloids, resins, or
phenolics tend to be glandular exudates and can act as contact poisons or volatile
inhibitors. The object of the game is to discourage insects from eating the plant
tissues.
Levin(256) points out that the production and storage of secondary substances
MACHINERY OF ENVIRONMENTAL RESPONSE MECHANISMS 123

is often an energy drain on plants, not to mention the multitude of other

phenotypic traits that are associated with release of such substances. The propor-
tion of total resources allocated to the production and maintenance of effective
chemical defense systems in plants is expected to vary with many environmental
factors, including season, age of plants, relative allocations to vegetative and
reproductive growth phases, and the degree of attack by insects and the types of
insect feeding habits involved.
Several of the papers in Rosenthal and Janzen(257) stress that the distribution
of secondary compounds or toxins in a plant will vary for different parts of the
plant and with both spatial and temporal properties of the surrounding environ-
ment. Finely tuned timing of the production and release of such substances is also
expected in the individual plant, a condition which may allow a herbivorous
insect to bypass the toxins by feeding on that part of the plant before the sub-
stance is produced. Because of these complexities, and the fact that a plant may
contain a wide spectrum of distinctive secondary compounds interacting to vary-
ing degrees with different species of insects feeding on the plant, it is extremely
challenging for the biologist to confirm the precise relationship between an insect
and a host plant species.
The deleterious effects of plant secondary substances on insects have been
elucidated by experiments (e.g., Feeny(Jo2,258); Feeny et at. (259); Erickson and
Feeny(260); Rehr et at. (261-263». High concentrations of such substances may
retard development even though rates of feeding may not be reduced.(260) Even
though there are many kinds of secondary substances, most of them share the
trait of being effective in deterring attack by insects. The recent tome of Rosen-
thal and Janzen(257) attests to current research interests regarding secondary plant
substances and their effects on herbivorous animals. The field is still embryonic,
with much to be done. Especially with respect to the tropics, one interesting
avenue of research is understanding the long-term relationships between tem-
poral and spatial changes in the secondary substance profiles of plant species in
relation to changes in the insect herbivore guild associated with the plant species,
and changes in the population structure of member species in the guild. How
such interactions change as a function of habitat in the tropics is also of interest,
given the high densities of plant species in some tropical habitats.
Levin(256) discusses the general means of detoxification and sequestering of
secondary substances by herbivores. Detoxification refers to the ability of an
insect or other organism to reduce or remove the toxic features of the substance,
usually by oxidations, reductions, hydrolyses, and conjugations, depending upon
the substance involved. The chemical interactions occur in the digestive tract,
and the detoxified substances are usually voided in the feces following the
retrieval of nutrients and water. Microsomal mixed-function oxidases appear to
be the major class of detoxification enzymes in insects, and there are considera-
ble differences in the level of microsomal oxidase activity among different
124 CHAPTER 3

FIGURE 3.23. The life stages (a-d) of the Neotropical ithomiid butterfly M. isthmia, a species
associated with secondary habitats in Central and South America. The larval food plants are tough-
leaved, and spinose species of Solanum found in open habitats, including pastures and roadsides.
MACHINERY OF ENVIRONMENTAL RESPONSE MECHANISMS 125

7 -
....

FIGURE 3.23. (Continued)

126 CHAPTER 3

species of insects attacking the same plant species in a particular habitat, and
differences within the same insect species in different habitats or utilizing dif-
ferent plants as food. Natural selection is expected to enhance the efficiency of
detoxication in insect species, since the process requires high amounts of
energy. (256) As expected, microsomal oxidase activity is higher in polyphagous
insects than in monophagous insects. (264)
Stamp(265) discussed the existence of cluster oviposition in butterflies, and
A. M. Young (in manuscript) has extended her ideas to the role of communal
feeding in Lepidoptera larvae as an adaptation to host plant species presenting
formidable physical barriers to successful feeding by the individuallarva."Based
on natural history studies of tropical butterflies, Young suggests that cluster
oviposition occurs primarily when a present-day host plant, or host plant in the
evolutionary past of the insect, has necessitated communal or group feeding to
ingest the food. Ghent<266) studies such behavior in sawflies. Young(88) found
communal feeding by the larvae of the egg-clustering ithomiid H. euclea, which
uses a solanaceous host plant with thick, tough leaves. Young(220) and Young and
Moffett(22\,222) reported communal feeding in the larvae of the egg-clustering
ithomiid M. isthmia (Fig. 3.23) and suggest that bites by multiple first-instar
larvae are necessary to reach the soft tissue layer of the leaf. The larvae, particu-
larly the early instars, form tight aggregations on the pilose leaves of the host
plant as a result of cluster oviposition (Fig. 3.24), and subsequent defoliation of
the plant (Fig. 3.25) is a function of group size. Cluster oviposition and gregari-
ous larval behavior are rather rare within the Lepidoptera. Where they do occur,
they may be related to a variety of ecological factors, including host plant patchi-
ness, average patch size of food plants within an area, preferences for shade
versus open areas, and degree of mechanical toughness of plant structures con-
sumed by larvae (for example, when it is high), may select for cooperative com-
munal feeding by larvae. Cluster oviposition and larval aggregations also char-
acterize the life cycles of Neotropical Battus butterflies (Papilionidae) (Fig.
3.26). Such behavior may facilitate the mechanical breakdown of tough plant
tissues during larval feeding.
In an interesting study on Barro Colorado Island, Coley(267) examined the
patterns of leaf damage from herbivorous insects on plant species composing
pioneer and forest succession communities. She found that grazing rates were
high on the mature leaves of pioneer species (shade-intolerant species) and lower
on the shade-tolerant species of forests (Table 3.1). Young leaves in both groups
of tree species had significantly more grazing damage than mature leaves. The
data suggest that mature leaves in general and those of climax forest habitats in
the tropics have better defenses against attack by herbivorous insects than young
leaves. The majority of tropical butterflies deposit eggs on or near young leaf
tissue, and young larvae consume young leaf tissue. In many cases first- and
second-instar cannot feed on mature leaves. These patterns are not universally
FIGURE 3.24. M. isthmia ovipositing on its larval foodplant (a), and gregarious second-instar
larvae (b).
128 CHAPTER 3

FIGURE 3.25. An assay of leaf damage to Solanum hispidum by a group of 20 larvae of M. isthmia
within a 10-day period.

true, since some species (M. peleides, for example) differentially place eggs on
mature leaves and the larvae feed successfully there. But the interactions of
herbivorous insects with foliage of different ages and defensive properties may
be more complex than they seem. While theory predicts that young leaves of
trees should be preferred by herbivores, the Temperate Zone study of Raupp and
Dennd268 ) on the leaf preferences of adult and larval stages of the imported
willow leaf beetle, Plagiodera versicoolora, indicates that preference for young
leaves occurs only in the adult stage . Survivorship of eggs and feeding success of
young larvae were found to be greater on mature leaves. Because adults of this
beetle and many other herbivorous insects prefer young leaves, survival of young
larvae and eggs on young leaves in the wild is very low, owing to damage
effects. In this system selection would appear to favor oviposition on mature
FIGURE 3.26. Egg cluster of the Neotropical swallowtail butterfly Battus belus on the underside of
a mature leaf of the vine Melothria guadelupensis (Curcubitaceae) (a), and first-instar larvae aggre-
gate (b). Although this butterfly is an Aristolochia feeder in the larval stage (Aristolochiaceae), egg
clusters are sometimes placed on Melothria leaves entwined with the correct food plant.
130 CHAPTER 3

TABLE 3.1
The Grazing Rates of Herbivorous Insects Associated with a Variety of
Plant Species in Tropical Rain Forest on Barro Colorado Island, Panama u

Mature Young
Species leaves leaves

Persistent
Faramea occidentalis 0.002 0.069
Virola sebifera 0.002 0.108
Prioria capaifera 0.002 0.014
Swartzia simplex 0.003 2.504
Simarouba amara 0.003 0.026
Trichilia cipo 0.003 0.522
Poulsenia armata 0.004 0.027
Desmopsis panamensis 0.004 0.783
Tachigalia versicolor 0.005 0.775
Tetragastris panamensis 0.005 1.477
Protium tenuifolium 0.008 0.928
Rirtella triandra 0.042 0.120
Quararibea asterolepis 0.114 0.316
Aiseis blackiana 0.136 0.096
Zanthoxylum panamense 0.136 0.701
Cupania sylvatica 0.311 1.151
Mean 0.041 0.631
Standard error 0.024 0.151
Number of species 16 16
Number of plants 120 115
Number of leaves 452 509
Pioneer
Didymoponax morototoni 0.007 0.001
Croton bilbergianus 0.025 0.310
Zanthoxylum belizense 0.081 0.624
Spoldias radlkoferi 0.186 1.492
Miconia argentea 0.189 0.509
Ochroma pyramidale 0.191 0.178
Aichornea costaricensis 0.210 0.818
Luehea seemannii 0.456 1.108
Cecropia insignis 0.797 0.111
Trema micrantha 1.071 0.053
Cecropia obtusifolia 2.267 1.299
Mean 0.466 0.663
Standard error 0.162 0.161
Number of species II II
Number of plants 102 105
Number of leaves 386 398

UData from Coley.(267)

MACHINERY OF ENVIRONMENTAL RESPONSE MECHANISMS 131

leaves, since the loss in fitness due to laying eggs on leaves of reduced nutrient
quality (mature leaves) is less than the loss from the mortality of early stages on
leaves of high nutrient quality (young leaves). The data suggest the importance of
examining the whole assemblages of herbivorous insects associated with both
young and old leaves of host plants, and the interactions among the species. Such
effects may even be more important to examine in tropical herbivorous insect
assemblages associated with particular tree species, since species richness may
be greater in many cases than in comparable Temperate Zone communities.
The ability for an insect to feed successfully on a mature or young leaf
depends upon its ability to overcome the defensive systems of the host plant.
Feeny et at. (259) have demonstrated that some polyphagous species of insects
have a higher activity of detoxification enzymes in their digestive tracts to handle
the chemical defenses from diverse plant groups. Krieger et at. (264) found that the
mixed-function oxidases (detoxification enzymes) of the polyphagous larval ar-
myworm can be induced by a variety of secondary plant substances. Carroll and
Hoffman(269) reported that experimentally damaged leaves of Cucurbita
moschata exhibit marked increases in certain substances at the damage site
within a short period and that these substances stimulate feeding in a chrysomelid
and inhibit it in a coccinelid. But the coccinelid under natural conditions cuts a
trench around the damaged areas, preventing the flow of these defensive sub-
stances to the area, and it can therefore feed. Young(270) reported that later instars
of the tropical butterfly Aeria eurimedea cuts the midrib of leaves of the
apocyanaceous host plant before feeding on the distal portion of the leaf. Such
behavior presumably blocks the flow of secondary substances, known to be very
toxic in the Apocyanaceae,<27lJ to the feeding site.
Herbivorous insects associated with different host plants known to have
toxic secondary substances are unable to feed successfully on one another's plant
species, but a generalist species can to some degree. Gilbert and Ehrlich(272)
discuss the evolutionary affinity ofthe Ithomiidae and Danaidae, and suggest that
the exploitation of closely related plant families by the two groups is further
evidence of their relatedness. Some primitive genera of ithomiids, primarily a
solanaceous feeding group, use the Apocyanaceae as larval host plants, while
danaids use the Asclepiadaceae. Because of the close evolutionary relationship
between the Apocyanaceae and Asclepiadaceae, Young(273) has demonstrated in
a preliminary way partial exchangeability in host plants between the two groups
of butterflies. In the analysis of feeding preferences in herbivorous insects, it is
necessary to determine whether instances of polyphagous feeding or switches to
new host plants result from genetic changes in populations or from physiological
induction of already present adaptations associated with gustation and digestion.
Given the high overall diversity of insects and plants in the tropics, many cases
where shifts in host plants by insects, in response to changes in habitat or competi-
tion pressures or other ecological factors, may not be evolutionary at all. Many
132 CHAPTER 3

such shifts, depending upon the phenotypic flexibility of the insect species, may be
behavioral shifts in which new resources are encountered and found to be accept-
able. Yet good evidence exists for the genetic regulation of feeding preferences
in Temperate Zone Lepidoptera. (274) Young(27S) has discussed the possible role of
genetic and physiological processes in the evolution of polyphagous and
monophagous feeding behavior in tropical butterflies.
The nutritional requirements of insects vary considerably according to the
kind of insect and its habits. Insects require environmental sources of at least 10
amino acids for tissue and enzyme synthesis, and carbohydrates as an immediate
source of energy. Various vitamins, inorganic salts, and small quantities of fats
are also essential for normal development and reproduction. Under the working
hypothesis that many tropical dicotyledonous plants have evolved effective phys-
ical and chemical defense systems against herbivores, it is clear that such de-
fenses insulate or protect the nutritional components of the plants from being
mobilized by herbivores. The concept of coevolution between plants and her-
bivorous insects is examined in the next chapter. The greater elaboration of plant
taxa in the tropics has entailed the diversification of rich arrays of chemical
defense systems, and these processes have been a source of both evolutionary
and short-term behavioral and physiological challenge to insects. The develop-
ment of host-location mechanisms in parasitic and predatory insects also depends
in part upon the elaboration of plant odors associated with defense against her-
bivorous insects. Rhoades(276) has reviewed the evolution of chemical defenses
against herbivores by plants. Toxic secondary plant substances playa major role
in determining the feeding specificities of insects. (277) Tropical lowland habitats
must present a very heterogeneous odor and tactile field for herbivorous insects.
The complexity of such environments is appreciated by a consideration of
oviposition behavior in addition to feeding.
Many butterfly species exhibit precise oviposition for placing eggs carefully
on the correct host plant for the larvae, although(278-280) errors are made in which
eggs are placed on incorrect plants or plants toxic to larvae. Sometimes insect
eggs are placed on host plants that are suboptimal for normal development, and
fitness is reduced (Fig. 3.26). Using McClure's study(28I) of introduced scale
insects in Connecticut and Coley's study(267) of herbivore grazing effects on
tropical foliage, we can argue that selection should favor an oviposition behavior
that places eggs on the correct parts of host plants as well. The studies of
Gilbert,(3!) Benson,(282) and Benson et al. (283) show the great degree of
oviposition-site selectivity in Heliconius butterflies in Central and South
America. In the interaction of these butterflies with their passifloraceous host
plants, the oviposition behavior is defined by availability of meristems, plant
size, presence of other heliconiine species, and predation by ants. The studies of
Young and Muyshondt in Central America have shown that oviposition behavior
can be very specialized in most groups of butterflies. (77,221,222,270,284-288,289-319)
MACHINERY OF ENVIRONMENTAL RESPONSE MECHANISMS 133

adult adult
host plant range host plant range

Equilibrium

larval larval
host plant range host plant range

larval
host plant range

Alt. 3
adult
host plant range
I~-
U>-
o ;;;J
.9-]
>0
o c:
*:~
'i m
> >
~.~
*~
-
larval
host plant range

FIGURE 3.27. Hypothetical model for the coevolution of alternative (alt.) adult and larval food
plant ranges for a butterfly population that has successfully colonized an island containing five
equally abundant larval food plant species. [From C. Wiklund, Oecologia 18:185-197 (1975).
Copyright 1975 by Springer-Verlag.]

Although the larval host plants of ovipositing female butterflies are known to
determine the host plants the butterfly oviposits on for some species, Wik-
lund<32Q,321l found no such correlations in tests with Papilio machaon using four
alternative host plant species. Such data suggest a lack of dependence between
larval host plant preference and that of ovipositing adults.
Whether similar preference patterns exist in tropical butterflies remains to
be studied. When large numbers of closely related genera or species of butterflies
occur in a tropical habitat, as is often the case, intense selection might favor
specialization at both the larval and adult preference levels as a means of lower-
 TABLE 3.2
Some Predicted Strategies of Adaptive Response by Tropical Butterflies to the Spatial Patchiness of Their Larval
and Adult Resources in Tropical Forests"

Spatial Relative
patchiness patch Larval feeding Adult feeding
Habitat of host plants size pattern pattern Optimal adaptive strategy

Forest Very high Small Polyphagous Generalist Population polymorphism;

canopy low phenotypic flexibility
Forest High Small Polyphagous to Generalist Monomorphism; high pheno-
understoryb monophagous typic flexibility; individual
polymorphic
Forest Low Large Monophagous Generalist Monomorphism; low phenotypic
understory (palmaceous) flexibility
Forest High Small Polyphagous Generalist As per forest canopy
light gapC
Forest Low Large Monophagous Generalist Monomorphism; low phenotypic
light gap flexibility
Secondaryd Low Large Monophagous Specialist Monomorphism; low to high
succession phenotypic flexibility

a From Young. (27Sl

"Two cases are recognized here. Woody understory trees and woody vines generally have very patchy distributions, thus converging with the forest canopy strategy for most
butterflies using dicot host plants; some understories are highly palmaceous, selecting for a specialist larval feeding pattern. In both situations, as with the forest canopy, adult
butterflies of the tropical forest tend to be generalized feeders of nectar, pollen, fungus, and decaying fruits.
'The forest light gap represents a highly transitional habitat between the highly patchy forest flora and the less patchy flora of secondary succession. Two cases are recognized
here, although most light gaps very likely are mixtures of both.
"It is believed that reproductively active adults of forest canopy butterfly species are generally prone to high dispersal, facilitating outcrossing and reducing inbreeding. Inbreeding
may be more frequent in butterfly species of secondary succession.
MACHINERY OF ENVIRONMENTAL RESPONSE MECHANISMS 135

ing competition. Resource limitation in the tropics and high cooccurrence of

many ecologically similar species in an area or habitat may select for highly
coadaptive gene complexes and physiological mechanisms that influence both
adult oviposition preferences and larval host plant preferences in the same direc-
tion in individual species.
Based on his studies of the butterfly P. machaon, Wiklund020,3w suggests
that host plant range of adults within a species and of larvae in the same popula-
tion are controlled by independent gene complexes. Wiklund(321l has developed a
model for the coevolution between adult and larval host plant ranges in which it
is assumed that larval mortality resulting from the inability to accept a host plant
as food is the only selective pressure, adult and larval host plant ranges are
determined by separate gene complexes, and the host plant family is large and
contains many biochemically similar species. Given three very different starting
popUlations as alternatives (Fig. 3.27), Wiklund predicts that an equilibrium
eventually will be established between adult and larval preferences in another
population. Host plant preference in butterflies and other insects is determined by
(1) the phenologies of both butterflies and host plants, (2) oviposition preferences
of adults, (3) host plant preferences of the larvae, and (4) the combination of (1)
and (3). The question of how adult oviposition preferences can be linked to,
rather than be independent of, larval host plant preferences has been discussed by
Young (275) for tropical butterflies. Young suggests that the evolution of
polyphagous or monophagous feeding in tropical butterflies is a function of the
patch structure of tropical vegetation in addition to the genetic or physiological
ability to detoxify distantly related plant groups. Young believes that polyphag-
ous tropical butterflies possess genetic systems of mixed morphs and that the
population is polymorphic as a whole. Monophagous butterflies are believed to
have populations of genetically similar individuals and with a high amount of
genic variation within individuals. Monophagy can also result from highly
monomorphic populations possessing various forms of phenotypic flexibility
from behavioral, physiological, or developmental switch mechanisms. Different
strategies are suggested for different kinds of tropical habitats (Table 3.2).
Monophagy is predicted to occur when a host plant family is locally very abun-
dant, i.e., has many large patches in an area. Oviposition "errors" in some
Neotropical butterflies may be related to a general colonization strategy(22) or
hiding eggs from egg predators.(323,324)
 CHAPTER 4

ECOLOGICAL ASPECTS OF
PLANT DEFENSES AGAINST
INSECTS

In this chapter a review of the roles of herbi vorous insects as selection pressure in
the evolution of morphological and chemical systems of defense against herbi-
vores in plants is given, with an emphasis on examples from the tropics. Addi-
tional aspects of this topic are touched upon in later chapters as well. The survey
and analysis is by no means exhaustive, but rather is designed to illustrate some of
the interesting defense strategies of plants against herbivorous insects and the
implications for such interactions on food chains in the tropics.

4.1. PLANT DEFENSES AND COEVOLUTION: AN OVERVIEW

In Temperate Zone habitats there are many well-documented cases of severe

defoliation by insects in certain tree species. Typically an infestation is initiated
in the spring and steadily increases throughout the summer, and the insect species
involved in such decimations of foliage are uni- and multivoltine. The gypsy
moth, tent caterpillar, and spruce budworm are three herbivorous insects in-
volved in severe outbreaks on forest and fruit trees in North America. These
species tend to exploit a wide range of tree species and do not appear to be
specialists. In tropical forests such outbreaks are less frequent although not
entirely absent. The young and mature leaves of a particular plant species may
exhibit damage from several different herbivorous insects, but individual plants
are seldom stripped of their foliage.
The defoliation of a plant may involve a loss of leaves, meristem, buds,
flowers, fruits, and seeds. As with other organisms the individual plant is viewed
as allocating energy to vegetative and reproductive functions. Vegetative
functions include those biochemical and physiological processes that promote the
growth and maintenance of the individual and also the development of
morphological and chemical defenses against herbivores. The simultaneous or
sequential allocation of energy to reproduction promotes the differentiatipn of

137
138 CHAPTER 4

inflorescences and flowers, then seeds and fruits. The individual plant is viewed
as having some sort of dynamic equilibrium with the internal and external envi-
ronment, so that fitness is maximized as much as possible.
An array of phenotypic types in the popUlation promotes different types of
responses to the selection pressures of the environment. Other forms of selection
pressure that also affect the modes of adaptation in plants include soil texture and
nutritive properties, intra- and interspecific competition, position in the habitat
relative to exposure to the sun and shade tolerance, availability of seed or fruit
dispersal agents, seed and fruit dispersal methods, availability of pollinators and
system of pollination, and seasonality. Other, often subtle features of the envi-
ronment, such as the influence of microorganisms, are also operative, and the
above list is by no means complete. The high diversity of tree species in lowland
tropical wet forests (see, for example, Pires et al. (325); Connell(65); Janzen(29))
suggests that each plant species in such a habitat is highly specialized in some or
all of the above adaptations. In the nonseasonal lowland tropics, the interplay of
herbivore pressure associated with foliage, fruits, and seeds, or competition on
the one hand balanced by adequate pollination and fruit or seed dispersal on the
other accounts to a large extent for the observed lack of species dominance in
most habitats. The role of herbivores in structuring the organization of tropical
forests is reviewed in Chapter 8, when the impact of insect populations on
tropical plants is approached from the view of how insects promote diversity in
the tropics; other aspects of this topic, including pollination systems and fruit and
seed dispersal systems, are also discussed.
Out of the many elegant studies conducted in the tropics emerges the follow-
ing delineation of the modes of interactions between insects and plants:
1. Co evolutionary associations between plants and herbivores, exemplified
by butterfly-host plant interactions.
2. Mutualism between ants and plants.
3. Seed and seedling predation by insects.
4. Interactions between extrafloral nectaries and ants.
The means by which a plant species adapts to selection pressure arising from
herbivorous insects consists of four alternative strategies: emigration to new
habitats where the selection is, or is very likely to be, less intense; elaborate
specialized morphological traits such as hairs or spines, designed and strategi-
cally placed to discourage feeding and oviposition; elaborate secondary plant
substances that block feeding; and mutualism with an insect partner (ant) that
helps to defend against herbivores. It should be immediately apparent that these
categories are not necessarily mutually exclusive. But what particular pathway of
defense is expressed in a function of genotypic and phenotypic flexibility coupled
with preadaptations in a population, types and intensity of herbivore pressure,
and potential availability of mutualist ant species in the habitat? Since not all
ECOLOGICAL ASPECTS OF PLANT DEFENSES AGAINST INSECTS 139

parts of a particular plant are equally vulnerable to attack by herbivorous insects,

it stands to reason that the distribution of defense mechanisms within the plant
need not be, and should not be, uniform. Rhoades(276) suggests that the allocation
of secondary substance to a particular tissue or structure is proportional to the
value of that structure to the plant. Yet one can clearly predict that other forms of
physiological constraints within the phenotype of a plant will alter such effects,
and generalizing about such effects and defensive compounds across the board is
difficult.
On the other hand, in encountering a potential host plant species with
effective defense systems the herbivorous insect species can do the following:
individuals within the insect population can adapt biochemically or physiologi-
cally in terms of digestion and assimilation of detoxified secondary substances;
individuals can "learn" new feeding habits to overcome the blocking of favor-
able gustatory cues from the plant, and perhaps feed on other parts of the plant
where cues are more inducive to sustained feeding; and a portion or all of the
population can invade a new habitat where a more accessible host plant popula-
tion is found, thereby circumventing entirely the defensive system of the original
host plant species. The development of the detoxification enzymes necessary for
handling secondary plant substances, through evolutionary modifications in
metabolic pathways, is a requirement for some insects to overcome the chemical
defenses of a particular host plant. Such selection must ultimately be coupled
with the peripheral and centrally controlled sensory physiology of the insect so
that each insect species recognizes the proper host plant to feed on under most
conditions. Thus behavioral diversity, at the level of high specificities in recog-
nizing host plants, is coupled with the diversity of detoxification systems, and
together they account for the expression of considerable feeding diversity, espe-
cially in tropical insects.
In the tropics, where physical conditions of the environment are generally
stable, a greater diversity of coevolved relationships is expected between plants
and insects and other animals. The higher specialization in feeding habits means
that many insects are specialist feeders on plants, one factor contributing to the
absence of outbreaks and severe defoliation. But such specialization within an
insect species may entail the allocation of considerable energy to the metabolic
pathways responsible for the morphological or chemical defenses of the host
plant, removing resources that could potentially be used to increase fecundity or
some other highly selection-sensitive portion of the phenotype concerned with
increasing the rate of population growth. The tradeoff between increasing popu-
lation growth and the acquisition of mechanisms for overcoming plant defenses
therefore precludes outbreak conditions in most tropical insects associated with
species-rich and stable habitats.
Another energy sink operative is that endowed by the patch structure of the
vegetation: resources must be allocated to effective interpatch dispersal in a
140 CHAPTER 4

specialist species. It is therefore the species composItIon of the vegetation,

coupled with the patch structure of the vegetation and the great diversity of
morphological and chemical systems of defense by plants', together with the fact
that the majority of insects are plant feeders, that accounts for the high diversity
of insects in the tropics, notably the nonseasonal lowland tropics. Each plant
species can be viewed as being in some sort of dynamic equilibrium with the
biotic agents affecting it, including herbivores and pollinators. If such an equilib-
rium is shifted drastically by some environmental perturbation and short-term,
rapid responses (of the sort discussed by Slobodkin and Sanders(104») are ineffec-
tive, considerable fitness in that generation may be lost. Rockwood(326) created
such an environmental pertubation on trees in lowland Guanacaste Province,
Costa Rica, where he deliberately hand-defoliated trees to simulate intense her-
bivore activity. He demonstrated a significant loss of fruit and decrease in seed
crop size following the defoliation. He concluded that heavy defoliation in wild
trees virtually eliminates seed production for the year it takes place. Various
studies have shown that growth and reproduction in plants greatly depend upon
leaf surface area since most of the photosynthesis takes place here. Breedlove
and Ehrlich(327) found a drastic reduction in seed set in lupine popUlations as a
result of severe predation by the larvae of a lycaenid butterfly. Heavy defoliation
reduces fitness by postponing the age of first reproduction and lowering the
number of seeds and fruits produced. Varying levels of defoliation of individuals
of a given plant species in a habitat are likely to be regular occurrences, and
during periods of intense defoliation a plant may respond by subsequently reduc-
ing energy allocations to reproductive structures and maximizing the replacement
of those vegetative structures involved in photosynthesis, a physiological shift
that minimizes the depletion of food reserves to nonvegetative functions.(328)
Whether an insect species is a specialist or generalist feeder is a function of
the phenotypic ability to handle more than one kind of plant species. I equate
generalist with the polyphagous habit and specialist with the monophagous habit.
Rosenthal and Janzen(257) provide a collection of papers concerning the many-
faceted aspects of the role of secondary plant substances in the defenses of plants
against herbivores, and the reader is urged to consult that volume for an introduc-
tion to much of the current thinking about this topic. One might expect that a
great deal of host plant selection takes place at the level of how insects perceive
and process secondary plant substances. Other substances, such as tissue water,
may also determine feeding preferences. It is suggested, therefore, that water
content, physical properties such as toughness and nutritive and nonnutritive
substances in plants act together to determine the attractiveness of the plant to
particular insects as a food source. It is also largely a matter of the cost to the
insect to exploit a particular plant species in terms of possible losses in fitness in
other niche components, and the availability of alternative food sources in the
habitat or nearby habitat. Owing to this form of cost-benefit relationship, plant
ECOLOGICAL ASPECTS OF PLANT DEFENSES AGAINST INSECTS 141

species evolve to maximize the energy investment in secondary substances with

defensive roles but do so only until any gain in adapting to herbivores is balanced
by the cost in fitness in doing so. But cost cannot be viewed solely in terms of
current tradeoff, because the eventual penetrance of the ecological barrier created
by defenses of the potential host plant(6) may endow a much greater fitness to the
insect population. Thus in addition to protecting itself against many herbivorous
insect species, the plant allows itself to be penetrated by one or few such species,
and these insect species in tum develop protective devices against their own
predators through becoming unpalatable. Unpalatability, in this context, is the
result of the insect developing the appropriate mechanisms to incorporate the
secondary plant substances, in some form, into its own body tissues. Thus the
coevolution between plants and insects involves not only the processes by which
insects become feeders on certain plants but also the development of unpalatabil-
ity in insects.
Coevolution between plants and insects depends upon the processes of
coadaptation between insect and plant. The concept can be applied at the famil-
ial, subfamilial, generic, and specific levels of taxonomic organization. Benson
et al., (283) for example, discuss both the overall and very fine features of the
coevolved associations between heliconian butterflies and their larval host plants,
the Passifloraceae. They emphasize the joint role of evolutionary history and
contemporary ecological factors in molding the manifold specialized relation-
ships of the butterflies to their host plants and the geographical diversification in
these associations. They found a basic pattern of host plant exploitation in which
restricted primitive taxa of heliconians oviposit on the mature leaves of primitive
passion vines of large biomass and the more evolved species of Heliconius using
leaves and meristems of the Granadilla and Distephana lineages of passion
vines. Other interesting patterns of plant usage and local adaptations of both
butterflies and plants were found. The work of Benson et al. (283) provides an
excellent example of how the general coevolved association between heliconians
and passion vines has become diversified geographically and locally, and also
shows some patterns accountable from the evolutionary history of the groups as
well.
Other studies of butterfly -host plant associations in the tropics (e.g. , Gilbert
and Smiley;(329) Rathcke and Poole~330») suggest that the evolution of physical
defenses such as trichomes follow the evolution of chemical defenses in host
plants. The studies of Rathcke and Poole(33O) and Young and Moffett(221,222) on
the feeding behavior of the larvae of M. isthmia suggest that the larvae of the
tropical butterfly can exhibit behavior patterns associated with avoiding or cir-
cumventing the thick layer of trichome hairs covering the leaves of the host plant.
Such studies and stress the dynamic nature of the coevolutionary process between
insects and plants. The relationships are not static but continually changing in
response to selection pressures promoting specialized adaptations in each or-
142 CHAPTER 4

ganism to counteract the other. In some groups of tropical butterflies there is

considerable ecological specialization to feed on certain plants. A. M. Young
(unpublished data) found that several genera and species of Ithomiidae in a forest
habitat exhibited considerable diversification in larval host plants, with little
overlap among species. Young(301) also found that the little ithomiid Oleria
zelica pagasa utilized a vine growth form of solanaceous host plant, one not used
by other i.thomiids in the same habitat.
Another interesting example comes from the association of some species of
Morpho butterflies with Leguminosae as larval host plants in tropical wet forests.
I have been studying the natural history of two very similar species, M. peleides
and Morpho granadensis at one site in northeastern Costa Rica. In earlier
studies,<31J,316,331l peleides was found to accept a broad range of leguminous
trees and vines as larval host plants (Fig.4.1), and eggs and larvae of this
butterfly have been found on these plants in the wild (Fig. 4.2). Several genera
and species of papilionaceous and nonpapilionaceous legumes are involved.(316)
In northeastern Costa Rica both butterfly species are sometimes found together in
the same habitat. But granadens(s is never as abundant in the adult stage as
peleides where the two co-occur, and granadensis seems very restricted in terms
of habitats while peleides occurs in several kinds of habitats. Philip De Vries has
also gathered data on the geographical distribution of granadensis in Costa Rica,
and his results tentatively show that this species is also restricted geographically,
much more so than peleides (summarized in Young(331a»). Reciprocal feeding
tests with the larvae of both species indicate that granadensis utilizes only one
host plant, Machaerium seemanii, of peleides, and larvae do not accept others
such as Mucuna urens. While preliminary, the data suggest that the coadapted
relationship between these morphos and legumes is not the same in the two
species, perhaps a result of contemporary ecological factors, including habitat
type and breadth of habitat used by each species. The role of habitat diversity in
insect diversity in the tropics is examined further in Chapters 5, 7, 8, and 9.
The evolution of chemical defense systems in tropical plants is also ex-
pected to vary with respect to other forms of environmental factors beyond those
shaped by herbivores. Janzen(332) predicted, for example, that regions in the
tropics with nutrient-poor soils, such as white sand soils, should have plant
species with greater concentrations of secondary substances very effective
against herbivores. This prediction comes from the observation that plant as-
semblages on such soils are vulnerable to extinction since replacement from
severe herbivore pressures would be very difficult. McKey et al. (333) found that
the mature leaves of common tree species in the tropical rain forest on white sand
soils in Cameroon contained about twice the concentration of phenolic com-
pounds found in similar forest vegetation on lateritic soils in Uganda, confirming
Janzen's hypothesis. The implications for herbivorous insects in the tropics is
clear: the evolutionary challenge for insects to exploit plant species growing in
ECOLOGICAL ASPECTS OF PLANT DEFENSES AGAINST INSECTS 143

FIGURE 4.1. Fourth-instar larva of Morpho polyphemus (Morphidae) from El Salvador (a), and
third-instar of M. peleides from Costa Rica (b).
144 CHAPTER 4

FIGURE 4.2. Seedling of Machaerium seemanii (Leguminosae), a larval food plant of M . peleides
in the central highlands of Costa Rica (a) and a fifth-instar caterpillar of this species in typical resting
position on the stem of the food plant (b).

nutrient-poor soils should be greater than for the exploitation of similar plant
species in relatively greater nutrient rich regions of the tropics.
Overall species richness of insects as herbivores on those plant parts with
ECOLOGICAL ASPECTS OF PLANT DEFENSES AGAINST INSECTS 145

high concentrations of secondary substances in communities on nutrient-poor

soils should therefore be lower than in comparable communities on nutrient-rich
soils. The average plant species in the nutrient-poor community cannot afford the
cost of withstanding severe herbivory, the lack of required nutrients for rapid
growth make recovery impossible. The metabolic cost for elaborating effective
chemical defense systems is overshadowed, in this case, by the potential threat of
localized extinctions resulting from severe defoliation or other forms of herbiv-
ory. Tropical plant communities occupying nutrient-rich soils should be buffered
against such effects by the ability to recover with the mobilization of soil nutri-
ents. This is not to say that plant species in such communities would not have
antiherbivore defense compounds. It does say that the concentrations of such
eompounds might be less than for similar species in communities on nutrient-
poor soils. Thus soil quality, in terms of nutrient profile, is a fundamental
geographical factor in shaping not only the composition of vegetation of a region
but also the phenotypic commitment to chemical defenses within populations of
each species, and the composition and species richness of the herbivorous insects
found on these plants.
From studies in the tropics, it is now clear that the coevolutionary process
between plants and animals, especially insects, goes beyond the interactions of
herbivores with their host plants. Frankie(117) has reported on the strategies of
annual flowering and pollinator activity in relation to how tropical communities
are organized in time. H. G. Baker and I. Bakd216 ,217) stress the possible role of
nectar constituents in promoting specific patterns of coevolution between plants
and pollinators. Van der Pijl(34) has outlined the general diversity of pollination
systems in the tropics. Coupled with the data of Frankie(117) and studies of nectar
by H. G. Baker and I. Baker,(216.217) the evidence indicates that the great diversity
of pollinating insect species in the tropics is promoted by the coevolution of plant
species with their pollinators in specific habitats. High diversity of pollinators is
also promoted by allochronic flowering patterns amoung plant species in the
same community or habitat, often regulated by, or timed with, seasonal cycles of
rainfall(335) (see also Chapter 7).
The energetics for maintenance of optimal foraging behavior associated
with thermal and water-stress conditions (reviewed by Heinrich(165) for bees)
represents the combined effects of selection within plant populations for flower-
ing periods and nectar nutrient profiles and selection within pollinating insect
populations in terms of physiological constraints of temperature and water stress,
flight muscle physiology (see Chapter 3), and resource needs for maintenance
and reproduction. See Chapter 8 for discussion of pollination in tropical com-
munities.
The degree to which an abundant plant species becomes exploited by a
particular herbivorous insect form depends, in some cases, on the degree to
which the plant population in question occupies a heterogeneous environment
146 CHAPTER 4

and to which the insect species is capable of maintaining fitness on different kinds
of environmental patches occupied by the plant species. In the tropics, particu-
larly lowland and montane humid or wet areas, considerable topographic
heterogeneity may create different microenvironmental factors influencing both
plant and herbivore distributions, and the two may not always overlap. If the
plant species is capable of occupying patches where the herbivore does poorly,
this may function as an effective escape mechanism for the plant in space, if the
average size of the antiherbivore patches and the frequency of such patches are
great enough so that the plant species maintains a breeding population capable of
withstanding fluctuations in the environment perhaps totally unrelated to the
plant-herbivore interaction. Individual groundnut trees growing on sandy or
gravelly soils in some parts of Nigeria experience severe defoliation by the larvae
of the cos sid moth, Azygophleps albovittata, while individuals on patches of
lateritic soils escape such predation.(336) The physiological explanation of why
groundnut trees on lateritic soils are less attractive to the larvae of this moth than
trees with similar leaf crowns on sandy or gravelly soils is not known. The spatial
patchiness of soil types keeps populations of the herbivore relatively isolated,
reducing the likelihood for an outbreak of the species as a pest of groundnuts. A
dry-season diapause in this herbivore also contributes to the prevention of out-
break conditions in all patches.(336) In the tropics, high percentages of butterfly
taxa in local communities do not feed on flower by-products such as nectar and
pollen. Pliske(t77,178) has shown that various male ithomiine butterflies are at-
tracted to plants containing pyrrolizidine alkaloids by the release of certain highly
volatile esterifying acids released from the rotting plant tissues where the insects
feed.
Using the chemical defense system of the Cruciferae as a model, Feeny002)
has stressed that herbaceous plants exposed to herbivores with only low fre-
quency should evolve chemical defenses in small amounts. Feeny suggests that
herbaceous plants in general should have relatively low concentrations of toxic
secondary substances, whereas the mature foliage of trees, shrubs, and grasses
have well-developed chemical defenses against herbivorous insects. In the
tropics, one might therefore expect that the species richness of forest herbivorous
insects is much lower than that of adjacent secondary habitats dominated by
herbaceous species. The sweep samples of Janzen(72) from tropical lowlands
suggest a greater number of both chewing and sucking herbivorous insect species
in secondary growth, and the pattern is not correlated with a greater absolute
diversity of plant species there. The sweep samples compared include those from
forest understories, and little data are available for insects in the canopy. The
forest environment may contain a number of specialized herbivorous insect
species, since Feeny(!02,258) and others predict that trees should have greater
concentrations of toxic secondary substances in mature leaves, and Coley's
ECOLOGICAL ASPECTS OF PLANT DEFENSES AGAINST INSECTS 147

study(267) in Panama showed that grazing rates were lower for mature leaves of
forest species. Thus the overall species pool of herbivorous insects is lower in
forests, but those present probably include a good many specialist species.
In contrast, secondary habitats in the tropics may select for "patch
specialists" and generalist species of herbivorous insects using the following
reasoning. Patch sizes of plant species may be larger in secondary habitats,
perhaps even to the extent that the popUlation of a particular herbivorous insect
species is confined largely to one or few patches of a host plant. This insect is a
"patch specialist." On the other hand, the presence of a pool of a lower number
of plant spec1es than what is found in an adjacent forest, but also a pool with
species having less effective toxic secondary substances, promotes the develop-
ment of generalist feeding habits in some insects. Many Homoptera may fit this
category. Thus secondary habitats in the tropics may support mixtures of
specialist and generalist insect species as a function of larger patch sizes per host
plant species, generally fewer host plant species, and less toxic secondary sub-
stances found in the average plant species of secondary habitats.
Not all kinds of secondary habitats are expected to be similar in these
properties. The presence in some lowland tropical forests of dominant tree
species, such as P. macroloba in northeastern Costa Rica and elsewhere in
southern Central America, provides an increased uniform biomass providing the
evolutionary opportunity for some specialist insects to exhibit convergence for
feeding on that abundant tree species. In this hypothetical situation, different
insects are specialists on different tree species in the forest environment. But if
one tree species through evolutionary time becomes a dominant, perhaps as a
result of high seed and seedling survivorship coupled with colonization of light
gaps and marginal areas,(337) and if the foliage or phloem fluids of this tree
contain very low concentrations of secondary substances, cost-benefit considera-
tions may promote insects to invade the tree as a new resource that is readily
available in large patches and free of a large energy drain associated with detox-
ifying toxic secondary substances. The striking occurrence of several genera
and species of cicadas on Pentaclethra trees in Costa Rica might be an example
of such convergence, although other explanations also exist. With disjunct patch
structures of plant species in the tropics, the specialist or generalist feeding
strategy in herbivorous insects must be a function of both the physiological
capabilities to feed on certain species and the patch structure of each species.
If the tree species of tropical forests represent very different sets of toxic
defenses to herbivorous insects, that is, each species has high apparency of its
defense substances, the specialist feeding behavior will evolve in the insect. Low
apparency among several plant species in a habitat would promote generalism in
insects. The generalist strategy is therefore most likely to tum up in insects in
secondary habitats in the tropics if in fact most secondary growth plant forms
148 CHAPTER 4

have low apparency of toxic defenses. The patch specialist insect might in fact be
a generalist physiologically that has happened upon an attractive resource with
sufficient biomass to allow the majority of the popUlation to exploit it as a food
source. Further studies on the levels of toxic secondary compounds in plant
species in different habitats in the tropics, as well as detailed natural history
studies of feeding preferences in tropical insects, are needed to examine some of
these ideas.

4.2. EVOLUTION OF UNPALATABILITY AND IMPLICATIONS FOR

FOOD CHAINS

The foregoing discussions have emphasized the general role of the versatil-
ity of response mechanisms of insects to the environment, with the focus being
how one explains the great diversity of insects in the tropics. I use the term
diversity here not in the sense of weighting species numbers by abundances of
individuals in popUlations, but simply to refer to the great number of genera and
species of insects per unit area of tropical habitat. Reference to the high diversity
of insects in the tropics usually means to regions and habitats where insect
diversity is greatest, namely, the nonseasonal to slightly seasonal lowland tropi-
cal rain forests, and to the pattern of environmental heterogeneity imposed by the
existence of different terrestrial habitats.
I have emphasized in a general way that the long evolutionary history of
insects as a group and the increased stability of the physical environment in
tropical regions have together promoted insect diversity. Of course, the impor-
tant roles of biogeographic events such as expansion and contraction of lowland
and mountain forest refugia in South America must also be taken into account in
explaining present-day distributional patterns of tribes, genera, and species. The
studies of Gilbed338l and Bensod282 ,339l stress the interplay of evolutionary his-
tory of a group of tropical butterflies and contemporary ecological factors pro-
moting diversity within the family. In this system, it is the geographical and local
changes in the environment that are continually restructuring the coevolved as-
sociations between the adult butterflies and their food sources, and between the
larvae and their host plants. Simberloff340l points out that there are relatively few
examples of the rates of local population extinctions for any group of organisms,
including insects. It is necessary to document such extinctions or lack thereof,
through detailed field studies, as a means of testing the generality of the equilib-
rium island biogeography model of MacArthur and Wilson.(34l)
In both a broad sense and at the level of specific interactions between insects
and plants, the evolution of plant communities has promoted the diversification
of insects. The following pattern of events is suggested.
ECOLOGICAL ASPECTS OF PLANT DEFENSES AGAINST INSECTS 149

1. Evolution of flowering plants, followed by regional differentiation and

speciation in relation to the physical environment, including land area
size, soil quality, topography, and prevailing regimes of rainfall and
temperature.
2. Overlapping with (1) of the evolution of insects, with an emphasis on
diversification of qualitatively distinct feeding habits resulting in herbi-
vory, pollination, etc.
3. Large-scale diversification of tropical plant communities, including the
establishment of different ecosystems at different altitudes, soil types,
and climatic regimes.
4. Development of high speciation rates in tropical floras in response to
year-round favorable conditions of climate, competition for space, and
continual natural occurrence of habitat differentiation resulting from
floods, mud slides, volcanism, tree falls in forests, and early systems of
agriculture.
5. Induction of high speciation rates in tropical insect faunas in response to
extant floras to reduce competition.
6. Induction of selection pressures for speciation in plants through herbivory
and pollination. Elaboration of specific plant parts and metabolic path-
ways resulting from such forms of environmental pressures. High
specialization in the tropics for seed and seedling predators, herbivores
ofleaves and meristems, seed and fruit dispersal agents, and pollinators.

Events (I) through (6) establish equilibria in the tropical biotas in terms of
species number and types of interrelationships, but superimposed upon such
patterns in evolutionary time are large-scale biogeographical events such as
Pleistocene glaciation that create contractions and expansions of biotas (see also
Chapter 10). The result is the need to interpret present-day distributions of
species and other taxa of insects in the tropics not only in terms of contemporary
and past ecological factors promoting diversity, but also in terms of biogeo-
graphical events causing regional changes in biotas.
Like plants, insects do not adapt to the environment in terms of food recog-
nition and food processing as related to fitness. Other pathways of selection
pressure include that caused by the trophic position of most insects as food for
other organisms. Visually hunting insectivorous vertebrates and many inverte-
brates create intense selection pressure in the tropics, especially where speciation
in these groups has also been high. In insects, therefore, the expression of
phenotypic traits that promote passive escape from predators or make insects less
attractive as food for predators is linked in some ways to the interactions of
insects with their host plants. Part of the great ecological diversification of
insects in tropical habitats is related to the insects' ability to survive being eaten.
150 CHAPTER 4

Yet we have no well-documented examples from the tropics of insect popula-

tions going extinct as the result of intense predation by vertebrates.
The classic examples of aposematic coloration in many tropical insects,
notably butterflies, grasshoppers, beetles, and hemipterans, represent the out-
ward manifestation of unpalatable properties to predators. Such properties are
derived (e.g., Brower et al. (342); Rothschild(343); Rothschild and Kellett(344);
Rothschild and Reichstein<345); Rothschild et al. (346)) from the incorporation of
toxic secondary substances from plants into the bodies of insects. The ability to
incorporate such substances and tum them into defenses against predators is
considered to be a highly evolved characteristic.(339) Thus there are many exam-
ples from tropical biotas where several insect species may feed on a plant species
with toxic secondary. Thus it is not always the case that an insect species feeding
on a toxic plant will be unpalatable and possess aposematic coloration. When a
toxic plant has several insect species feeding on it, individuals belonging to one
or more of these species may be cryptically-colored, suggesting an absence of
unpalatability. Such an insect does not have the ability to incorporate the toxic
substances from the plant into its own body tissues. The ability to improve
survival based upon the evolution of warning coloration depends on being able to
sequester the toxic compound and then assimilate it in some effective form into
the body tissues. Brown and Netd 205 ) have reported that some birds in Brazil
successfully feed on supposedly unpalatable butterflies, so the system is not fool-
proof. Pough and Brower(347) discuss a form of functional palatability in which
males of the butterfly Ascia monusta are attacked more frequently than females
as a result of behavioral differences such as more frequent thermoregulation
in males, even though both sexes are unpalatable. Well-studied systems in-
clude the sequestering of cardiac glycosides from Asclepiadaceae and Apocyan-
aceae, two tropical groups, by several different taxa of insects.
The studies of Brower and Moffitt(348) and Brower et al. (349-352) suggest that
considerable geographic variation exists in both the concentrations of these toxic
secondary substances in host plants and the degree to which insects become
unpalatable as a result of feeding on them. Brower et al.(35!) studied the car-
denolide profiles and palatability in a population of the butterfly Danaus
chrysippus in West Africa and found considerable variation in the concentration
of this substance in a sample of butterflies from one area. They related these
differences to the spectrum of host plant variation and differential abilities of
different butterflies to sequester the substance efficiently in the larval stage.
Largely as the result of the studies of Brower and Moffitt,(348) it is evident that
butterflies and other herbivorous insects whose immature stages or adults feed on
plant groups known to contain toxic secondary substances and sequester these
substances are not always unpalatable to potential predators, because individual
species of host plants may vary considerably in the content of such substances,
within local populations and geographically. The studies of Duffey and Scud-
ECOLOGICAL ASPECTS OF PLANT DEFENSES AGAINST INSECTS 151

det353 ) and Vaughn(354) demonstrate the role of cardiac glycosides in making

selected Hemiptera and Coleoptera unpalatable. Duffey and Scuddet353 ) suggest
that the occurrence of several unrelated insects with the same palatability prop-
erties, small shape and size, and similar patterns of warning coloration provides a
basis for the co-occurrence of these insects throughout much of North America
and the Caribbean. In an important set of experiments, Vaughn(354) demonstrated
that the amount of sequestering of cardiac glycoside from milkweed seeds by the
milkweed bug, O.fasciatus, is inversely proportional to the concentration of the
secondary substance in the seeds (Fig. 4.3).
Some lowland and mid-elevation tropical rain forests support species- and

•
10.0 I
••
•
•
•
I
~ B.O
.s'" •
z •
o •• •
~

..• ) /..
a: •
• ••
~ 6.0
UJ
()
Z •
o
()

I·
UJ
Q 4.0
•
...J
o N=39
Z
UJ

I
o
a:
«
()
2.0

/
2.0 4.0 6.0 B.O

WHOLE SEED CARDENOLIDE CONCENTRATION (mg/g)

FIGURE 4.3. Patterns of increasing concentrations of cardenolides in Oncopeltus milkweed bugs as

a function of increasing concentrations of these toxic secondary substances in whole milkweed seeds.
[From F. A. Vaughan, J. Chern. Ecol. 5:89-100 (1979).]
152 CHAPTER 4

FIGURE 4.4. Young nymphs of the coreid bug, P. blandus. aggregated on the underside of a leaf of
the food plant A . warscewiczii (a); the nymphs are bright orange with black bulbous antennae and
legs. Older nymphs (b) are metallic green and orange. The nymphs aggregate on the food plant
throughout the juvenile period and disperse as adults.

genera-rich assemblages of butterflies associated with a particular microenvi-

ronment within the forest, such as the ground cover and low understory vegeta-
tionY86.189) In such well shaded environments, scent organs rather than strictly
ECOLOGICAL ASPECTS OF PLANT DEFENSES AGAINST INSECTS 153

FIGURE 4.4. (Continued)

visual mechanisms may playa major role in species-specific courtship patterns,

although no inverse relation necessarily exists between color patterns and effec-
tiveness of scent organs in courtship for tropical butterfliesysl) The solution to
understanding the distribution of certain kinds of scent structures in butterflies is
the examination of both phylogenetic and ecological properties of the species
involved. (lSI)
154 CHAPTER 4

In some Neotropical Hemiptera, the adults are migratory, large, and not
aposematically colored, but their nymphal stages are sedentary, gregarious, and
warningly colored. Aldrich and Blum(183) demonstrated the defensive role of
aggregative behavior in the nymphs of the coreid Thasus acutangulus in lowland
Guanacaste Province, Costa Rica, during the dry season. The nymphs form
conspicuous clumps on the host tree and, when disturbed, simultaneously spray
anal fluids and produce a defensive secretion covering the abdominal terga. The
aposematic display is released by mechanical disturbance, and aggregation is
induced by a pheromone. The behavior presumably is adaptive: the large bugs
are active in the dry season and therefore under conditions of reduced foliage.
Without such a defense, the bugs would be sitting targets for birds, lizards, and
some amphibians. The brightly colored nymphs of the hemipteran Paryphes
blandus form conspicuous, tight congregations on the leaves of Anguria (Cur-
cubitaceae) (Fig. 4.4), the adult bugs breed there (Fig. 3.9), and presumably
these insects are unpalatable due to uptake of toxic substances from the host
plant. (355)
In the tropics one encounters many examples of aggregative behavior in
brightly colored insects, but further studies are needed to relate such conditions
to host plant chemistry, de novo synthesis of defensive compounds by insects,
and kinds of predators involved. There is also a need to analyze the spectrum of
individual variation for such aggregating insects in terms of the spectrum of host
plant species utilized. If, for example, the basis for unpalatability is derived from
host plants, and if the insect is polyphagous, how is unpalatability conferred if
the toxic substances are different among the host plant species in the habitat? Is it
a system in which some morphs within the insect population are physiologically
adapted to sequester the compounds from one or a few of the host plants? Or can
any individual in the population sequester the compounds from several different
host plants? These two alternative patterns of interaction imply different genetic
and physiological traits. The popUlation may consist of distinct genetic morphs
each adapted to process some subset of the total food spectrum, or the population
is essentially monomorphic with each individual possessing enough phenotypic
flexibility (in the form of mixed-function oxidases or a similar system) to seques-
ter the compounds from all host plants. The evolutionary implications of un-
palatability in tropical insects also relate to population growth features. Unpalat-
able species may have, on the average, reduced population growth capacities
resulting from energy being allocated to those physiological mechanisms endow-
ing unpalatability. Yet if predation rates are high on immature stages such as
eggs or larvae, new adaptations may arise to offset the loss in fitness from
reduced fecundity or some other characteristic related to population growth. For
example, Gilbert(31) has found pollen feeding to be a specialized behavior of
some species of supposedly unpalatable Heliconius butterflies (Fig. 4.5), and he
ECOLOGICAL ASPECTS OF PLANT DEFENSES AGAINST INSECTS 155

FIGURE 4.5. (a) Fifth- (final) instar caterpillar of Heliconius hecate . a specialized herbivore of
some species of Passiflora (Passifloraceae) in Neotropical rain forests. and (b) the adult butterfly.
clinging to pupal shell.
156 CHAPTER 4

concludes that pollen feeding enables the butterflies to increase egg production in
response to high predation rates on eggs by ants.
Detailed life cycle studies of tropical butterflies (e.g., Muyshondt(284-287);
Young(77,221-222,270,28!f-315); Young and Muyshondt(316-319» that include records of
larval host plants and adult food resources provide the information necessary for
developing conceptual models of how specific groups of insects, such as the
major taxa of butterflies, are distributed locally (habitat wise) and geographically
in the tropics. Such information also provides insights into why some tropical
butterflies adapt well to man-made environments such as gardens.(356)
The double-edged feature of the concept of coevolution between plants and
insects is that, on the one hand, the expression of toxic secondary substances in
plants for defense against herbivores may screen out many potential feeders on that
plant species, while on the other, other species may "penetrate" the ecological
barrier through evolutionary time, overcome the defense system, and feed success-
fully on that plant. The ecological advantage of feeding on a plant with effective
toxic substances for antiherbivore defense is that those species that possess the
ability to detoxify or sequester the toxic compounds enter into an essentially empty
universe in terms of food and absence of competitors. Intense ecological factors
such as interspecific competition or high predation in the insect population as-
sociated with an "original" host plant species popUlation probably promote
selection for switching to a new host plant species in the same habitat or in
adjacent habitats. The ability to make the switch would be a function of many
factors, including (1) taxonomic distance between the original and new host plant
species, (2) the ability of the insect to respond in the short term (behavioral,
physiological) to feed on the new host plant, (3) the long-term adaptive response
capabilities in terms of the genetic structure of the insect population, and (4) the
amount of ecological stress associated with the new host plant species, in terms
of competing species already established on it, and the kinds and intensity of
mortality factors, biotic and microenvironmental abiotic factors, that could lower
the population well below the carrying capacity and so low as to promote extinc-
tions. Ehrlich and Raven(6) discuss some of these aspects of the coevolutionary
process.
In plant populations, considerable variation is predicted among individuals
for the concentrations of toxic secondary substances present in a particular tissue
or organ. Such variation is expected since the populations of most plant species,
particularly in the tropics, where selection pressures from herbivores and
pathogenic microorganisms are expected to be intense, will contain different
genotypes and phenotypes with somewhat different character states of various
traits. Janzen<40) has stressed the adaptive role of maintaining outcrossing in plant
species in lowland tropical forests, to the degree that certain plant species have
evolved temporal patterns of flowering that promote long-distance pollinator
ECOLOGICAL ASPECTS OF PLANT DEFENSES AGAINST INSECTS 157

movement patterns to ensure genetic exchange. Superimposed, so to speak, upon

the residual variation among individuals in plant populations in the tropics is the
pattern of temporal changes in concentrations of toxic secondary sup stances
found in individuals. Feeny, (258) working with oaks in the Northern Temperate
Zone, assessed seasonal changes in the concentration of tannins in leaves and the
impact of these changes on herbivorous insects. One type of pattern is the
difference in concentrations of defensive compounds in young versus older tis-
sues and structures. Shifts in concentration and distribution of defensive com-
pounds may be matched with seasonal patterns of vegetative growth and reprod-
uction in tropical plants. Some Neotropical vertebrate omnivores exhibit consid-
erable shifts in diet as a function of size, sex, and age,(357) with such shifts related
to changing physiological needs and availability of food sources. In tropical
herbivorous insects, similar shifts may occur in terms of progressive reduction in
quality of a food type during the growing season. Some Neotropical Orthoptera
and carabid beetles are opportunistic predators on other insects and plant mate-
rials, respectively. The diet of a herbivorous insect must be viewed in terms of
temporal changes in quality of suitable food sources and quantitative changes in
food abundances. Many tropical butterflies preferentially oviposit on young
leaves and meristem of host plants, and according to general theory about plant-
herbivore coevolution, such tissues may have lower concentrations of toxic sub-
stances, making it easier for young larvae to assimilate nutrients efficiently.
Young,(292) for example, described the preference of Dismorphia butterflies for
placing eggs on the young leaves of Inga. Such studies lead us to believe that her-
bivorous insects exploiting the young tissues of host plants to the exclusion of older
tissues will be subject to considerable resource limitation, which is a function of
the vegetative growth pattern of the host plant species. The effects are expected
to be pronounced for specialist herbivores.
Frankie et al. (118) have described some aspects of the seasonal or temporal
vegetative growth patterns oftrees in both seasonal and nonseasonal lowland tropi-
cal forests. An important implication for coevolutionary theory from this study is
the great variation among tree species in these patterns, which must impose
considerable variation in the population cycles of insects associated with these
trees. Taking this one step further, those groups of butterflies and moths, and
flies and beetles which make up pollinator communities in tropical forests will
also be regulated in the larval stages by the availability of suitable host plant
tissues, if the species are specialists on certain tree species and if they are also
specialists on young tissues. The impact on plant species that require these in-
sects for effective pollination is that limitation in the abundance of pollinators
through time will be a function, in part, of patterns of resource limitation
on larval food sources, which are usually different taxa of plants from those
being pollinated by the adult stages of these insects. Such considerations point
158 CHAPTER 4

to the immense complexity of interactions in the tropical forest habitats, espe-

cially from the viewpoint of different selection pressures influencing adults
and immature stages in insects.

4.3. NOTES ON DEFOLIATION IN THE TROPICS

In the preceding discussions it was generally concluded that the incidence of

severe defoliation in tropical plants ought to be considerably less than for plant
species in comparable Temperate-Zone habitats, primarily because of the in-
ferred role of biotic mortality factors and deleterious effects of competition in
keeping popUlations of herbivorous insects in the tropics well below the carrying
capacity. Yet we must recognize that not all species of herbivorous insects in the
tropics will be regulated by predators and parasites and those that are will not be
regulated to the same degree. Because the essentially nonseasonal lowland wet
forest tends to have the greatest numbers of insect species and plant species,
incidence of defoliation in natural forest habitats may be low in these regions. As
new insect species evolve in these regions, there might be a short period initially
when populations are not exploited by predators and parasites. But because
predators and parasites tend to be generalists, these organisms will rapidly ac-
cumulate on the new species and the species will come into equilibrium with the
rest of the communities. If the new species is a good competitor and has a low
popUlation growth rate as a result of an evolutionary tradeoff with energy allo-
cated to defensive strategies, it may remain "cryptic" as a new resource for
predators and parasites for some time. But if the species consists of one or more
popUlations with individuals programmed for rapid population growth, resulting
in a swamping of a host plant species, the new species becomes a signal for
predators and parasites and the response to the new resource may be rapid.
Another factor in such considerations is the degree to which different
species of predators and parasites that are potentially capable of exploiting the
new resource are limited enough in sufficient resources to make such a facultative
switch. Then again, such "decision making" may not be an issue if the exploita-
tion of the new host is basically a matter of simple encounter and easy access to
it. The reestablishment of an equilibrium between the new herbivore species and
its host plant mediated by regulation from predators and parasites reduces the
likelihood of outbreak conditions and subsequent defoliation. If the new species
enters into a competitive adaptive zone with other species, the tradeoff in energy
allocation associated with niche restructuring to lower competition will also keep
the popUlation below outbreak levels.
Other factors may influence the degree to which predators and parasites can
regulate herbivore populations. Campbell and Duffey<358) have discovered that
tomatine, an alkaloid toxin in tomato plants, is passed on from the body of the
ECOLOGICAL ASPECTS OF PLANT DEFENSES AGAINST INSECTS 159

herbivore, a pest of tomatoes, to the parasitic wasp, and the substance is toxic to
the wasps. Such a mechanism lowers the ability of the wasp to regulate the
herbivore population. The wasp is an endoparasite of the herbivore and therefore
ingests the tissues laden with tomatine. Thus the effectiveness of parasitic wasps
and flies in regulating herbivore populations will be a function of the amount of
toxicity of the secondary substances accumulating in the body tissues of a poten-
tial host species and the ability of the parasite to buffer the toxic effects of the
secondary substance. Given the diversity of plant species and secondary sub-
stances in tropical floras, such selection might be intense in the tropics and
constitute yet another level of coevolutionary associations, namely, between
parasites and predators of herbivores, and the defensive compounds of the host
plants, for those herbivores that sequester these substances in their diets.
Although defoliation seems to be an ecological phenomenon of Temperate
Zone communities, I believe many examples of it will be discovered in high-
altitude communities in the tropics, lowland forest light gaps expanding into
secondary growth habitats, a variety of disturbed habitats including agricultural
systems and horticultural habitats, and very seasonal habitats. Kenyi<360) de-
scribed the defoliation of Capparis tomentosa by the larvae of the butterfly
Belenois aurota in Uganda and found it to be a seasonal event each year with
larvae selectively browsing young leaves. The observed defoliation is thought to
reduce the host plant population to the extent that other plant species in the same
park environment increase in abundance. Defoliation of a host plant species will
also depend upon the amount of defensive chemistry variation among individuals
of the host plant popUlation and the ability of the insect popUlation to handle that
spectrum of variation in host plant defenses. The extent of herbivore attack will
also depend upon the spatial arrangement of individuals in the plant popUlation
(e.g., Vandermed 359 ) . The isolated individuals may have a significantly lower
probability of being attacked than an individual in a large clump. Yet such
patterns will vary considerably with the kind of herbivore, its dispersal ability,
and variation in attractiveness of host plant individuals.
If temporal patterns of growth and overall diversity of plant species are low
in high-altitude regions in the tropics such as parts of altiplano in the Peruvian
Andes, there might be periods of defoliation corresponding to those periods in
which vegetative growth of host plants is high and herbivore or omnivore popula-
tions are somewhat synchronized to the flushes of these resources. Physically
controlled environments are believed to have a greater incidence of insect out-
breaks as a function of synchronized seasonal emergence patterns and reproduc-
tive cycles during the limited growing seasons of such environments. It does not
necessarily follow that highly seasonal lowland tropical areas will also contain
communities characterized by seasonal outbreaks of herbivorous insects and
defoliation. Lowland seasonal environments represent old ecosystems with con-
siderable diversity of organisms and sufficient evolutionary time to permit
160 CHAPTER 4

species in different trophic levels to adjust population structures to seasonal

fluctuations in resources. As with nonseasonal lowland environments in the
tropics, such interactions are finely tuned so that herbivores, predators, and
parasites follow in close synchrony the seasonal changes in the environment. The
major departure from all of this reasoning is, however, the consideration of
agricultural systems in the tropics and their proneness to insect outbreaks and
defoliation. See the discussion of this important topic in Chapter 9.
Ferns make up a major group of tropical plants that are not intensely ex-
ploited by insects as food, and the reasons are not obvious. (361) The evidence
suggests that toxic defensive substances in ferns cause low predation rates by
herbivorous insects, but interestingly, about 75% of all polyphagous insects
feeding on ferns are also associated with woody angiosperms,(361) a condition
suggesting that under-utilization of ferns by insects is largely the result of other
suitable food supplies being readily available in the same habitats. The explana-
tion, therefore, may be an ecological one, related to the local distribution and
abundance of angiosperm host plants versus ferns as hosts.

4.4. NOTES ON SEED PREDATION IN TROPICAL HABITATS

The studies of Janzen(41.362) demonstrate the role of host-specific seed and

seedling predation in determining the species composition of tropical forests.
Working primarily with various species of Neotropical leguminous vines, Jan-
zen(363,364) and Rehr et at. (261,263) have shown that the large seeds of these plants
contain high concentrations of toxic secondary substances that make the seeds
immune to attack by certain seed-eating insects. For example, the seeds of the
common vine Mucuna contain high concentrations of free L-dopa, which when
isolated and combined with a synthetic diet leads to substantial mortality and
deformities in the southern armyworm larvae. (261) Rehr et at. (261) surveyed a
variety of seeds from Central American legume species and conducted similar
feeding tests demonstrating that these seeds contain high concentrations of "in-
secticidal" amino acids that repel or kill armyworm larvae and presumably have
similar effects on many seed-eating insects in tropical forests. The various insect
species that attack such seeds in the wild may be under limiting conditions of
other food sources, and considerable selection might favor the expression of new
coevolved relationships between these insects and leguminous seeds.
The studies of Janzed 365 ) have shown that nonleguminous trees in tropical
forest have similar coevolved systems. From his study it is clear that seeds of
tropical trees may contain toxic secondary substances and other kinds of toxic
substances that through evolutionary time have filtered out a potentially large
assemblage of seed-eating insects, at the same time resulting in a few species'
overcoming the defenses and successfully attacking the seeds in annual seed
ECOLOGICAL ASPECTS OF PLANT DEFENSES AGAINST INSECTS 161

crops. Janzen(29) has shown that seeds are rapidly discovered on the ground
beneath the parent tree or vine and that discovery rate and intensity of attack on
seeds drop off considerably as one moves away from the tree. The system is most
evident for tree individuals within forest communities, but sometimes the entire
trees or seed crop of vines in secondary succession habitats escapes from pred-
ators, presumably because the crops escape detection by insects. Mammals such
as squirrels and monkeys are the primary dispersal agents of seeds, carrying off
pods and then dropping some seeds away from parent trees or vines. The interac-
tion of host-specific seed predators, annual patterns of seed crop production, seed
characteristics such as size, shape, and toxicity, (362) and mammalian dispersal
agents largely determine the distribution of adult and juveniles of tree species in
tropical forests. Other processes, such as differential rates of seed survival and
germination on different soil patches within a forest, inhibitory effects of other
plant species on the germination or seedling growth process, and other microen-
vironmental factors also help determine the spatial distribution of plant species in
tropical forests. Janzen(29) also discusses the impact of predation on seedlings in
causing virtual loss of seed crops beneath parent trees and reviews the theory and
evidence of the ecological processes shaping plant distributions in tropical
forests. The model developed by Connell(65) and Janzen(29) predicts that each tree
species in a tropical forest will have its own set of host-specific herbivores
attacking seeds and seedlings. Different kinds of insects are involved in attacking
seeds and vegetative structures. Herbivorous insects such as some Lepidoptera
may be associated with the canopy of such trees, but larvae will also attack
seedlings growing beneath the tree. The canopy crown is envisioned to act as a
signal to attract herbivores while filtering out many others through defensive
compounds present. Cicadas may constitute another group of insects attacking
forest trees in the tropics and using the canopy crown to search for the appro-
priate root crown necessary for nymphal development.
A dominant tree species of lowland tropical forest in the eastern watershed
of southern Central America is the legume P. macroloba, whose population
structure and dynamics has been studied by Hartshorn.(337) Unlike most trees of
these forests, this tree species is abundant per unit area of habitat, and adults and
seedlings are often found tegether. At least three genera and six species of
cicadas emerge in large numbers beneath individual adult trees of this species
year after year, and the same trees are full of cicada nymphal skins each
year.(70.71) Even though the forest understory is full of emergence sites, the
nymphs appear to emerge directly beneath the crown shadow of Pentaclethra,
suggesting a possible feeding association of nymphs with the root crowns of the
tree. But given the abundance of intermixing root systems from other tree
species, the hypothesis is that Pentaclethra adults, and their cohorts of seedlings
beneath them, act as signals for mated female cicadas to oviposit within the
patch. Young<70,71l reports frequent oviposition by Z. smaragdina in Penta-
162 CHAPTER 4

clethra seedlings beneath adult trees. Pentaclethra xylem or phloem fluids may
have relatively fewer toxic substances than other tree species in the same habitat
and therefore represent an accessible food source for cicada nymphs on its root
crowns. Thus there might be strong selection for cicada species to converge and
exploit the root crowns of this tree species, and the preference is mediated
through the oviposition behavior. White and Strehl(366) discuss the effects of
low-nutrient plant fluids on the evolution of developmental time in cicadas and
other insects, and because of such an effect, one might predict that cicadas prefer
host plant species where nutrient assimilation is most efficient, thus speeding up
developmental time, even if only in small amounts. Pentaclethra trees; because
of a hypothesized lack of toxic defenses in internal fluids, may be an optimal
resource for cicadas in tropical forests. Because patch size is relatively large,
considering the combined root crowns of adults and juveniles, not to mention the
biomass available for oviposition sites, the environment can support populations
of several genera and species.
This is not to imply that Neotropical cicadas in geographical zones where
such a tree species occurs will feed only on this tree species. The opposite is
probably true, namely, that cicada nymphs will attempt to feed on most roots
encountered within a certain stratum of soil. What is being suggested, however,
is that while Pentaclethra may have a primary host root crown. Other tree
species are exploited in direct proportion to the frequency of encounter by bur-
rowing nymphs. Disease and predation associated with egg masses and young
nymphs undoubtedly keep cicada populations below the carrying capacity of the
habitat. Adult cicadas are also attacked frequently by birds such as oropendulas,
which successfully locate singing males.
Janzen(29) discusses the high survivorship of Pentaclethra seeds and
seedlings beneath parent trees based upon the immunity of seeds to attack by
seed-eating insects. It should be pointed out that the patterns discussed above for
cicadas in tropical forests can also be explained by other hypotheses. For exam-
ple, an alternative explanation for the apparent convergent association of cicadas
with Pentaclethra is that this tree species is the most abundant resource type in
the forest habitat and therefore can be exploited by several species without
competition. This hypothesis implies nothing about the toxic compound chemis-
try of internal fluids being ineffective relative to other tree species. The combina-
tion of large patch size and relative ease of exploitation together may account for
associations of cicadas with certain tree species in tropical forests.
A similar pattern seen in advanced secondary growth forests in the same
region is the association of some cicadas such as Z. smaragdina and F. sericans
with large patches of Goethalsia meiantha (Tiliaceae).(71,215) But one must also
examine the evolutionary history of cicadas and broad patterns of host tree
exploitation to account for present-day distributions and associations. Young(137)
reported that the majority of species in Costa Rica are associated today with
ECOLOGICAL ASPECTS OF PLANT DEFENSES AGAINST INSECTS 163

leguminous tree species in general and that legumes might have constituted the
original host plant group during the early adaptive radiation of cicadas in the
tropics. Thus, an overall preadaptation to legumes in general, coupled with
contemporary ecological factors, accounts for the associations of cicadas with
particular habitats in the tropics and for association of nymphs with the root
crowns of certain tree species in these environments. Ecological shifts to dis-
turbed habitats may occur in cicada species best adapted to marginal habitats and
may involve shifts to nymphal feeding on different plant families, especially for
those plant species with large patch sizes and low concentrations of defensive
compounds in xylem or phloem fluids. Some recent Temperate Zone studies with
plant hoppers show that individuals of a single species preferentially select indi-
vidual host plants or parts of host plants containing the highest concentrations of
amino acids in phloem or xylem sap.(367) The feeding on plants containing high
concentrations of nitrogen-containing nutrients increases the fitness of some
sap-feeding Homoptera. (368,369) Such studies reveal that sap-feeding insects in the
Temperate Zone exploit host plants in many different families, although anyone
species may exhibit a coevolved interaction with a single host plant species.
Since the tropics contain a greater array of Leguminosae in a variety of terrestrial
habitats than comparable areas of the Temperate Zone, a greater evolutionary
opportunity exists for some sap-feeding forms such as cicadas to coevolve with
these plants, presumably as the result of the high availability of these plants in the
tropics and the presumed higher concentrations of nitrogen-containing nutrients
in the xylem fluids of roots in these plants. Outside of the Leguminosae, one
might expect to find strong selection for cicadas to exploit plant species with
mycorrhizal associations that provide additional nutrients to the root crowns of
these plants; a large portion of plant species in tropical forests may have such
associations. (370)
Studies by Sgarbieri(37!) have demonstrated the considerable variation in
protein content of legume seeds from many plant species in Brazil, variation both
among and within species. Similar patterns of nutrients and nonnutritive con-
stituents may also occur in phloem fluids of tropical legumes, providing consid-
erable heterogeneity in resource quality for cicadas and other insects. Thus
preadaptation to legume feeding in cicadas in tropical forests is followed in
evolutionary time by the acquisition of abilities to feed successfully on legume
species with different nutritive and nonnutritive profiles. Other phenotypic
characteristics, such as the attractiveness of bark and pith to ovipositing females,
will also influence host selection in cicadas. Our overall knowledge of the natural
history of tropical cicadas is fragmentary at best, but the interactions of these
organisms with tree species in tropical forests may shed some light on the
original plant species composition of tropical forests.
Various examples of the chemical and functional diversity of antiherbivore
defenses of seeds in tropical floras are reviewed by Janzen. (41,42) A noteworthy
164 CHAPTER 4

point about these studies is that fewer data are available for seed defenses in
mountain floras in the tropics. It is difficult to generalize about the antiherbivore
defenses of plants and their role in shaping the species richness of tropical
habitats, but an even greater problem is predicting changes in such interactions in
mountain forests. Whether or not the large seeds of legume vines such as M.
urens are attacked by species of bruchid weevils along similar lines as studied in
lowland habitats remains to be investigated in high-elevation regions. The selec-
tive forces affecting a vine population high up on the fertile slopes of a volcano in
southern Central America might be very different from the lowlands in the same
region. For example, a preliminary examination of seed mortality and vine
recruitment levels in a forest patch containing several adult M. urens vines at
about 1600 m elevation indicates little evidence of attack by seed-eating insects
or high predation on "milk stage" seeds by squirrels. Rehr et ai. (261) predict that
Mucuna seeds in general should be free from insect attack. Data from a highland
population of M. urens seem to confirm this. But the high level of predation by
squirrels in this popUlation results in low recruitment of new vines within this
habitat. Presumably the squirrel popUlation is confined to the small patch of
forest, and foraging for milk stage seeds is unusually high. Such foraging repre-
sents a facultative response by squirrels to restricted habitat and easy access to
seed crops maturing on several vines.

4.5. NOTES ON PLANT-ANT INTERACTIONS IN THE TROPICS

An earlier discussion in Chapter 1, pointed out the great taxonomic and

ecological diversity of ants in the tropics. The evolution of eusociality, tremen-
dous diversification in nesting habits, and diversification in feeding habits figure
as the prime factors accounting for the great ecological success of ants in the
tropics. From the poles to the equator, the number of subfamilies, genera, and
species of ants increases gently. Lowland tropical rain forests tend to have the
most ant taxa within the tropics. The pattern of environmental heterogeneity both
horizontally and vertically in lowland tropical rain forests has provided ants with
considerable opportunities for ecological diversification. I define ecological di-
versification in ants as the specialization within different taxa for particular
nesting sites in different microhabitats within the forest environment and the
associated specialization in feeding habits in many of these taxa. (See Wilson(80)
for a complete review of ecological diversification of ants.) While many tropical
ants are opportunistic scavengers and predators, others combine such behaviors
with specialized foraging activities, while still others exhibit completely
specialized foraging habits. From an initially terrestrial or cryptozoic existence,
ants have been able to exploit different foods and nesting sites arboreally and
otherwise in tropical forests in response to the great diversity of available re-
ECOLOGICAL ASPECTS OF PLANT DEFENSES AGAINST INSECTS 165

sources. Extensive surveys of tropical insects will probably leave undescribed

many taxa of forest canopy ants and of ants exhibiting behaviors adapted to
canopy existence.
Some forms of ecological specialization in tropical ants result from the
structural and organismic complexities of lowland tropical forests. The long
growing season, generally favorable climate, high diversification of plants and
various animals, promote diversification in ants, and one can envision such a
process laying the fundamental pattern of speciation in tropical ant faunas. Dif-
ferent patterns of ant speciation are associated with tropical islands. (91) But just as
taxon cycles in some ant faunas of tropical islands have molded the geographical
distribution of ant species in archipelagos, so could phenomena similar to the
stages of a taxon cycle govern the diversification of ant faunas on tropical
mainlands. The invasion of ant species into secondary successional stages and
disturbed habitats may follow such patterns.
Another form of ecological specialization in tropical ant faunas is charac-
terized by the entering of particular taxa into obligate mutualism with certain taxa
of plants. Such associations do not form a major portion of tropical ant faunas, but
another form in ant-plant interaction, namely, the feeding by ants on exudates of
extrafloral nectaries of plants, constitutes a fairly common form of opportunistic
feeding exhibited in many different ants. In the mutualism between plants and ants,
both partners exhibit specialized physiological, behavioral, and morphological ad-
aptations or phenotypic features designed to promote the interaction. Perhaps the
best known examples come from the classic studies of Janzen(J3) on the obligate
mutualistic actions between certain species of Pseudomyrmex ants and Acacia
shrubs in Central America. Although the association of these ants with acacias
was known for some time, the experiments of Janzen(33) confirmed and proved
the nature of the mutualism. Janzen, for example, was able to demonstrate that
young swollen-thorn acacias deliberately deprived of their Pseudomyrmex found-
ing colonies exhibited a reduced fitness compared with ant-occupied plants in the
same habitats. Janzen also demonstrated the detailed adaptive traits in both ants
and plants for the mutualism. Janzed 372 ) describes the geographical distribution
of swollen-thorn acacia species in Central America, illustrating that the group is
mainly adapted to agricultural habitats such as pastures. The major traits of the
mutualism are summarized in Table 4.1. Keep in mind that these phenotypic
characters have developed in response to selection pressures favoring the evolu-
tion of mutualism between Pseudomyrmex ants and acacias. Presumably a pre-
swollen-thorn acacia popUlation experienced considerable herbivore damage to
vegetative structures and the mutualism evolved as a means of reducing predation
on the plants. Janzen shows that the ants, a complex of several species in Central
America, protect the vegetative parts of acacias by removing herbivorous insects.
The ants possess a powerful sting and are mobilized rapidly on the leaves and
stems of the plant in response to mechanical disturbance. The acacia provides
166 CHAPTER 4

TABLE 4.1
The Major Phenotypic Traits of Ant Acacias a

General features of acacias of importance to Specialized features of swollen-thorn acacias

the interaction (coevolved traits)

Woody shrub or tree life form Woody but with very high growth rate
Reproduce from suckers Rapid and year-round sucker production
Moderate seedling and sucker mortality Very high unoccupied seedling and sucker mor-
tality
Plants of dry areas Plants of moister areas
Ecologically widely distributed Ecologically very widely distributed
Leaves shed during dry season Year-round leaf production'
Shade intolerant, sometimes covered by vines Shade intolerant and free of vines
Stipules often persistent Stipules longer persistent, woody with soft
pith
Bitter-tasting foliage Bland-tasting foliage
Each species with a group of relatively host- Each species with a few host-specific phyto-
specific phytophagous insects, able to feed in phagous insects, able to feed in the presence
the presence of the physical and chemical of the ants
properties of the acacia
Foliar nectaries Very enlarged foliar nectaries'
Compound unmodified leaves Leaflets with tips modified into Beltian bodieS'
Flowers insect pollinated, outcrossing Flowers insect pollinated, outcrossing
Seeds dispersed by water, gravity, and rodents Seeds dispersed by birds
Lengthy seed maturation period Lengthy seed maturation period
Not dependent upon another species for survival Dependent upon another species for survival

aFrom Janzen.(33)
'Essential to the interaction.

domicile in the form of hollow swollen thoms, and food in the form of nutrient-
rich secretory structures (Beltian bodies) on the leaves. Ants actively remove
vegetation surrounding the young acacia, increasing its likelihood of surviving in
the canopy of secondary growth vegetation. Acacias deprived of ant colonies
experience intense attack by herbivorous insects and may die. The swollen stipu-
lar spines and foliar nectaries of ant acacias are adaptations to acacia ants.
Rehr et al. (263) report on the chemical defense in non-ant acacias in which
leaves of these plants were found to be cyanogenic and thereby toxic to herbivor-
ous insects. Leaves of ant acacias are not cyanogenic. An absence of evolution-
ary contact between Pseudomyrmex and some acacias, or the lack of phenotypic
flexibility to develop a mutualism with ants if such contact is made, promotes
selection for the expression of toxic compounds as defense against herbivores.
Janzen(373) also demonstrated that occupancy of Barteria (Passifloraceae) shrubs
by Pachysima ants in Nigerian rain forest enhances the vegetative growth of the
plants. The ants reduce herbivore damage and attack plants adjacent to their host
ECOLOGICAL ASPECTS OF PLANT DEFENSES AGAINST INSECTS 167

shrubs and clean Barteria of various debris. A similar interaction has been
studied in Central America and the Caribbean islands.(374,375) Risch et al. (376)
describe the mutualistic association of Pheidole ants with three species of Piper
(Piperaceae) in lowland tropical rain forest in Costa Rica. The ants occupy
hollowed-out stems and petiolar cavities of the plants and feed on lipid-rich
secretions from secretory bodies in the petiolar cavities. In addition to removing
nearby vines, the ants deposit debris in the hollowed stems which is then ab-
sorbed by the plant as an additional source of nutrients.
Bentley(377) found that individual flowering plants of Bixa orellana in Costa
Rica with higher frequencies of visits by ants to extrafloral nectaries have higher
levels of mature fruit production than other plants with reduced visits by ants.
She concludes that the presence of many ants on a plant reduces the activities of
herbivorous insects and thereby increases fitness by increasing yields of mature
fruit. The daily pattern of nectar production in extrafloral nectaries suggests a
peak in late afternoon; ants also are active at this time (Fig. 4.6). The interaction
is so well tuned that peak nectar production in extrafloral nectaries occurs at the
time mature buds are produced (Fig. 4.7). Studies (e. g., Keelet378 ); Keeler and
Kaul(379); KoptUr<380» indicate that extrafloral nectaries produce nutrients for ants
and are strategically placed on plants to maximize protection against herbivorous
insects. The elaboration of extrafloral nectaries in many plant species is a form of
facultative mutualism in which the plant's phenotype includes ant-attracting sub-
stances to increase the general presence of one or more ant species on leaves and

'v6'1/I,
en
.~
0.8r
.E

.\ /.-."•"".-.-.
e I· 6-- _

..... -.I
.c
I -6.. , M
~ !S • • \ .-. I
\
Q) '"-
i: 1\ / Q)
c.

....
0 .-\
;;:: I
'.-e-e • e 2-
il:J 0.4 ~ i:
0
\ I ;;::
.0

...
c:
Q)
\ I
...
'"-
ctl
(J
Q)
'"-
Q)
C.
z
...
en
c:
<!
06.00
Time of day

FIGURE 4.6. Ant activity (e), as the mean number of ants per 10 buds, flowers, or fruits over
successive hours, and hourly nectar production (6) from extrafloral nectaries of the Neotropical tree
Bixa orellana. Nectar production or flow was measured as the mean volume of nectar secreted in 3 hr
by 10 inflorescences over three 24-hr periods. [From B. L. Bentley, 1. Ecol. 65:27-38 (1977).
Copyright 1977 by the British Ecological Society.J
 168 CHAPTER 4

...'" 2
]

------.
0 ..c
'"....
Q)
4 M
....
~ Q)

a 0.

/I
:;::
3
~

-
OJ
~
.c 2-
....
...c....
Q)
~
<J
Q) 2 Q)
Z
...'"c
0.

«
\ /
I 6/
L -__________ ~~--------~--~~----------~~------------~
10 20 30 40
Time (days)

Young Mature TYoung Fruit Fruit

bud bud I
fruit <0·5 em >0·5 cm
Flower

FIGURE 4.7. Ant activity (e) and nectar flow (L) for the six developmental stages of buds,
flowers, and fruits of the Neotropical tree B. orellana. Ant activity is the mean number of ants per 10
buds, flowers, or fruits averaged over 24 hr, based on hourly observations, and nectar flow is the
mean volume of nectar secreted in 3 hr by 10 buds, flowers, or fruits. [From B. L. Bentley, 1. Ecol.
65:27-38 (1977). Copyright 1977 by the British Ecological Society.]

stems, which in turn makes the plant less attractive to herbivores. Although
extrafloral nectaries are found in plant groups outside the tropics,(379) they are
probably more abundant in tropical floras, where herbivore pressure is great and
where there is a large pool of ant species that could enter into such associations
opportunistically. There might be a correlation between the ability of a plant to
produce toxic secondary compounds as effective antiherbivore deterrents and the
degree of phenotypic commitment to other defense traits such as extrafloral
nectaries and their associated phsyiology: plants incapable of producing such
secondary substances are more likely to develop extrafloral nectaries to attract
ants, and within this group a small percentage may evolve into obligate
mutualisms. But facultative mutualisms between plants and ants are expected to
be more common.
Foraging responses in some tropical ants to the exudates of extrafloral nectar
ies is most likely the result of encounter frequency with such resources for most
cases of facultative associations. Carroll and Janzen(197) discuss shifts in the
ECOLOGICAL ASPECTS OF PLANT DEFENSES AGAINST INSECTS 169

foraging behavior of ants in response to differential needs of colonies in terms of

breeding. Opportunistic foraging coupled with periodic phases of different nutri-
tive requirements within nests determine the exploitation of extrafloral nectaries
by each ant species. The exudates of extrafloral nectaries may contain substances
such as lipids and carbohydrates, not found in sufficient concentrations in other
food resources such as prey or debris animal tissues. Intake of lipids, for in-
stance, might be important for proper nourishment of larvae in nests, although
there are no data on such requirements. The observed patterns of extrafloral
nectary exudate production being timed with the reproductive cycles of bearer
plants suggests that these plant species are vulnerable to intense herbivory that
could lower fitness in terms of seed crop production. The highly timed produc-
tion of exudates within a population of a particular plant species may be so
narrow that it becomes an unpredictable resource for foraging ants. Thus exam-
ples of facultative mutualism are expected to consist of ant species with flexible
foraging habits. In the Neotropics, ant groups such as Ectatomma typify such
opportunistic forms of behavior. The investment return for a plant species to
allocate resources to elaboration of extrafloral nectaries and associated exudates
will depend in part upon the synchrony of production within the population: the
greater the synchrony, the greater the reinforcement for worker ants and the
greater the protective role of the ants. Strong stabilizing selection should be
influencing those traits governing the development and timing of exudate elab-
oration among different individuals within the population. Presumably ants at
extrafloral nectaries reduce attacks by herbivores. The adaptive role of these
structures and their secretions is seen in some Passifloraceae that have evolved
extrafloral nectaries that mimic the eggs of heliconiine butterflies. Females in
search of oviposition sites will avoid placing eggs on meristems with egg
mimics(338) (Fig. 4.8), reducing attacks of these vines by heliconiines.
The presence of ants with or without extrafloral nectaries does not necessar-
ily mean that herbivores will be absent from the plant. Counterselection or
short-term behavioral responses induced by ants is predicted in these insects. The
extremely specialized egg placement behavior in some tropical butterflies, in
which eggs are placed on tips of tendrils and other structures where ants cannot
reach them, may be responses to ant predation. Other factors, including the
proximity of ant colonies and foraging habits of ant species in a habitat, will also
contribute to the probability of ants being present on a plant with extrafloral
nectaries. The evolutionary commitment on the part of plant species to evolve
extrafloral nectaries will be a function of (1) the genetic variation in the popula-
tion for those phsysiological and morphological control mechanisms that would
be involved in the development of these structures and (2) the long-term intensity
of predation by herbivores and the reduction in fitness resulting from such dam-
age.
Because selection pressures arising from herbivory are expected to vary
FIGURE 4.8. Some examples of extrafloral nectaries in the Passifloraceae of Neotropical forests. (a)
The cuplike structure at the base of the folded leaflet is an extrafloral nectary. (b) The light round,
dot/ike areas on the leaf of Passiflora aff. capsularis are extrafloral nectaries.
ECOLOGICAL ASPECTS OF PLANT DEFENSES AGAINST INSECTS 171

considerably within and among plant species in the tropics, the production of
toxic secondary compounds in plants may be facultative responses during periods
of intense predation. Thus once a tree or herb is attacked, the physiological
machinery of the plant produces more of the substance to increase effectiveness
as a feeding deterrent. Such facultative chemical defense systems, as
hypothesized here, can work in concert with other lines of defense of plant
species, such as leaf toughness, water content, and phenological patterns of leaf
production. The possession of facultative response mechanism of this sort, along
with other more stable phenotypic characteristics, may be less of an energy drain
on the plant species in the sense that the machinery is geared up only as needed.
Such a system might be very adaptive for plant species in seasonal regions of the
tropics where facultatively controlled defenses are brought into play during the
periods of peak herbivore activity, such as the first half of the rainy season.
 CHAPTER 5

DISTRIBUTION PATTERNS
OF INSECTS IN TROPICAL
HABITATS

As discussed in Chapters 1 and 2 various authors (e.g., MacArthur(15,16» have

considered the joint effects of within- and between-habitat diversity being the
major determinants of observed patterns of high species richness in the tropics.
This approach implies that much of the adaptation within species occurs as a
result of evolutionary responses to patterns of local environmental heterogeneity,
and such patterns of resource distributions and biotic interactions among or-
ganisms are very diverse in some regions of the tropics. Such an approach
complements considerations of broad-ranging adaptive responses such as ob-
served patterns of geographical variation in phenotypic characters and gene fre-
quencies within a species. It is well known that studies of geographic variation
within single species and for closely related species resulted in the formulation of
the concept of speciation by allopatry or geographical differentiation. But in this
chapter, we are concerned with adaptation in local environments rather than over
a broad geographical range.

5.1. THE CONCEPT OF LOCAL ADAPTATION

One can speak of the adaptive responses of individuals within a single

population of a species as adaptation to the local environment. Many studies have
demonstrated that such local patterns of adaptation within populations may differ
among different populations of a single species and that major determinants of
local adaptation include (I) spectrum of potential resources available for a
species, (2) biotic interactions with other organisms, and (3) local effects of the
physical environment. These three sets of factors impose patterns of environmen-
tal heterogeneity upon organisms, and adaptation is often a function of the
distribution of these ecological factors in time and space. Such distributions are
often patchy, and the collection of patches associated with a particular ecological
factor is said to have a grain pattern, where grain refers to the arrangement of the

173
174 CHAPTER 5

patches as experienced by individuals of the species. (3,381) Furthennore, the

abundance of patches and the relative fitness of individuals in each patch deter-
mine the manner in which a species exploits resources in a particular habitat.
Either within a habitat or between two or more habitats in a region, the distribu-
tion of resources for a species may be coarse grained, resulting in an individual
remaining in a single patch for all of its prereproductive and reproductive life.
The result is that different genotypes will adapt to different patches, and consid-
erable microspatial variation in the patterns of adaptation can result. Such
microspatial variation is expected to be most evident if there is reduced gene flow
among patches. This type of response mechanism to environmental heterogeneity
results in differentiation within a species at the local habitat level; as a means of
evolutionary change in a species, it is therefore a very different mechanism from
geographical changes.
Heterogeneous environments establish a mosaic of different selection pres-
sures that in some cases will favor the maintenance of genetic variation in
popUlations occupying such environments. In such situations, disruptive selec-
tion maintains the polymorphism, with different genetic morphs adapted to dif-
ferent niches or portions of the environmental mosaic. Habitat selection promotes
the maintenance of alleles advantageous to the species in different patches of the
environment, and in some cases deleterious alleles can also be maintained in a
popUlation by this mechanism.(382) The maintenance of such an allele represents
an evolutionary compromise when limited resource patch types are present, and
the population requires the exploitation of all patch types to maintain a popula-
tion size sufficient to prevent local extinction.
In species responding to available resources in a fine-grained fashion, that
is, in which resource patches and different habitats may be indistinguishable in
tenns of differential fitnesses of genotypes, most or all individuals in the popula-
tion will utilize the same resources and therefore be generalists rather than a
mixture of specialists as developed above for coarse-grained environments.
Genetic differentiation among populations or subpopulations, in fine-grained
environments, will be nonexistent. Such considerations apply to the analysis of
niche differentiation, or lack thereof, within a single species and in a particular
habitat, or along a series of habitats in the same region. Considering the tropics,
we would expect the great species richness of plant species within some lowland
habitats in particular to provide the potential for a great deal of niche differentia-
tion within a herbivorous insect species, assuming for the moment an absence of
intense interspecific competition and the availability of patch sizes sufficiently
large to support a breeding popUlation of that species. The same reasoning can be
extended to summed floristic diversity of several adjacent tropical habitats.
Several studies with Drosophila have demonstrated that insect species ex-
hibit considerable microspatial preferences in resource utilization,(383,384) and
revealing along these lines are studies on the specialized oviposition habits of
DISTRIBUTION PATTERNS OF INSECTS 175

these flies. (385) Habitat selection is a major behavioral mechanism generating

within-habitat spatial variation in the distribution of Drosophila species on the
island of Hawaii. (385) The analyses of thermal preferences in Puerto Rican Droso-
phila by Levins(386) illustrate considerable diversification among species for dif-
ferent microhabitats. Other studies (e.g., Powel1 (87 ); McDonald and Ayala(388»
indicate that microenvironmental heterogeneity influences the genetic structure
of Drosophila populations, so that observed phenotypic differences correspond to
genotypic differences within popUlations.
There is now considerable evidence that genetic variation in populations is
promoted by the complexity of the environment. Within biological limits, one
might say that as the environment becomes increasingly complex, genetic varia-
tion in the popUlation of a given species should also be high so that the species
can thrive in the environment. I use the term thrive to mean the capacity to
maintain a breeding popUlation at or near the carrying capacity or to counteract
severe biotic or abiotic mortality so that the population can continue to increase
even though well below the carrying capacity. In complex environments, it is
believed that one kind of allele for a given locus is favored in one kind of
microhabitat, another allele in a different microhabitat, and so on. The utilization
of many different microhabitats within a particular habitat is maintained by
natural selection, since such a pattern of resource exploitation allows the species
popUlation to lower the possible detrimental effects of both intra- and interspe-
cific competition.
Field studies with Drosophila have demonstrated, in some cases, a positive
relationship between levels of genetic variation and ecological complexity of the
environment. DaCunha and Dobzhansky(75) demonstrated, for example, a clear
relation between the incidences of chromosomal polymorphisms and floral diver-
sity in D. willistoni. Taylor and Powell(389) showed that morphs of D. persimilis
collected over a region of field and forest habitats in the Northern Temperate
Zones retain habitat associations when marked, released, and subsequently re-
captured. But other studies of different species do not show a clear relationship
between the kinds of chromosomal polymorphisms present in a population and
habitat affinities.(390) On the basis of various kinds of electrophoresis, there have
been attempts to determine if different enzyme-genotypes within a popUlation are
associated with different microhabitats, the basis of selection being suitability of
microhabitat occupancy at the biochemical level.
Given such approaches, the question of habitat selection distills down to the
determination of whether there is a pattern of genetic variation in a population
and the degree to which different genotypes exhibit preferences for certain mi-
crohabitats present in the habitat. The popUlation is seen as consisting of several
subsets of different genotypes, each exhibiting a preference for different mi-
crohabitats, with breeding largely confined to those genotypes associated with a
particular microhabitat. Thus many subdemes may exist, and the result is a
176 CHAPTER 5

preservation, along with the effects of selection, of the microhabitat affinities of

each genotype.
In the tropics, where there is often a great deal of host plant specificity in
many phytophagous insects, selection might result in different genotypes exploit-
ing different host plant species in the same habitat, or over several habitats if the
entire breeding population is distributed over more than one habitat, and breeding
occurs among individuals in different habitats. The selection for ecological di-
vergence in such insect popUlations can arise from interspecific competition for
the same plant species as resources or during periods following colonizing epi-
sodes in which population density might be very high, resulting in considerable
intraspecific competition. Both sorts of situations can create selection favoring
microhabitat divergence where each microhabitat type is a different host plant
species.
The studies of Bush(391) on Temperate Zone, fruit-inhabiting flies in Wis-
consin indicate that such divergence may lead to partial reproductive isolation
and the beginnings of speciation. It is tempting to speculate that the high diver-
sity of insects in the tropics is the result of considerable microhabitat divergence
on resources with high harvestable productivity, initially giving rise to different
genotype sets associated with certain resources, then leading to speciation. Such
a system would fit the lowland tropics well, where resource diversity for insects
is high. Thus one can envision habitat selection for a given species of insects in a
lowland tropical field as being the association of the breeding population with
that habitat, and also on a fine scale, in that microhabitat selection within the
habitat promotes ecological divergence, i.e., different insect species using dif-
ferent plant species. Depending upon the genetic systems associated with feeding
behavior and physiology, and the amount of genetic variation in the population
for alleles involved in the regulation of feeding, polyphagy or highly flexible
monophagy can result.(27S) Which type of expanded feeding evolves, assuming the
initial condition to be traditional monophagy, would also be a function of the grain
of the environment in terms of plant species diversity and the patch structure of the
plant species found in that habitat? Given the great taxomonic diversity of low-
land tropical floras, high harvestable productivity, and great structural diversity
of plant species in the tropics, the response surface available to insects is ex-
pected to be more heterogeneous and textured in the tropics than for most com-
parahle habitats in Temperate Zone regions.
Habitat selection can be viewed as the pattern of niche differentiation exhib-
ited by individuals of a particular species and in which the resulting spatial
isolation works in concert with geographical isolation in causing the local dif-
ferentiation of populations. It is the mechanism by which the individuals compos-
ing the breeding population of a particular species come to exploit different
resources within or between habitats. The adaptive significance of such behavior
is that different segments of a popUlation essentially nonoverlap in terms of
DISTRIBUTION PATTERNS OF INSECTS 177

resource utilization, lowering the effects of intraspecific competition or allowing

a population to become so scarce in terms of density per patch that the deleterious
effects of interspecific competition can be avoided. The problem, however, is to
recognize and detect those cases of habitat selection in which a genetic dif-
ferentiation has occurred within the population so that different genotypes are
exhibiting different preferences in the habitat(s). This is important because many
cases of habitat selection may reflect short-term behavioral responses without
genetic differentiation, possibly precluding the evolutionary response altogether.
Thus not all cases of habitat selection involve shifts in genetic structure within
populations. Maynard Smith(111) has shown that habitat selection can act to
preserve genetic variation in populations. In Drosophila populations pronounced
differences in allozyme gene frequencies and chromosome inversion frequencies
often correspond to differential usage of the environment.
The inverse relationship between allozyme diversity and species diversity in
species of Drosophila in Hawaii (392) suggests that areas of high species rich-
ness are relatively unpredictable environments so that genetic variation in local
populations is unnecessary. Such data complement the notion that high
amounts of some kinds of genetic polymorphisms in populations function to
allow the species to occupy a range of microenvironments or exploit a range of
resources, notably in areas where the environment fluctuates unpredictably and
niches and resources are both changing in frequency and distribution. The role of
environmental fluctuations on the genetic structure of populations has been dis-
cussed by Levins.(3,381) In groups of tropical butterflies exploiting a range of
larval host plants locally, under some conditions polymorphisms may regulate
the ecological flexibility to exploit a range of such resources, although under
other circumstances high phenotypic flexibility per individual may be the optimal
strategy.(27o) The selective advantage to polymorphic individuals in tropical but-
terflies that participate in mimicry associations has been studied. (393)
The relevance of a discussion of the distributional patterns of insects in
tropical habitats to the general subject of habitat selection is the potential applica-
tion of the concept to explaining the high diversity of insect species in tropical
habitats. The individuals forming the population of a particular species are sub-
ject to a variety of ecological factors, including competition, predation and
parasitism, and subtle changes in the physical environment. Habitat preference
largely becomes a matter of the species positioning itself in some optimal way
within the existing environment so as to maintain a breeding population at or near
the carrying capacity. Explaining the richness of insect species encountered in
various sweep samples (e.g., those of Janzen(72») is the task of explaining patterns
of resource partitioning within and between habitats by insect species in each
trophic level. Taxa capable of existing in more than one kind of habitat might
exist, while others are habitat specialists. Specialization, or lack of it, is a
function of the behavioral plasticity available among individuals in the popula-
178 CHAPTER 5

tion, genetic variation in the population for long-term responses, and the ability
to use either the same resource among habitats or different resources in different
habitats. Given the observed high levels of insect taxa in tropical habitats and
over much of the tropics in general, the explanation for this diversity must be
accounted for in terms of broad biogeographical phenomena promoting regional
differences in species composition of faunas and also the niche-differentiating
effects of local patterns of environmental heterogeneity such as within- and
between-habitat resource diversity. The issue involves examining differentiation
in ecological habits within a single population of a species and relating patterns to
other species in the same habitats.

5.2. EXAMPLES OF HABITAT DIFFERENTIATION IN THE TROPICS

It is safe to say that the lowland tropical wet regions of the world probably
contain the greatest diversity of habitats of any region, tropical or extratropical.
Much of this diversity is the result of local instability in forest environments,
creating light gaps, a variety of secondary habitats, and ephemeral habitats
associated with the edges of rivers and streams. Connell(65) has suggested that the
lowland tropical rain forest ecosystem is unstable in the sense that there is
considerable replacement of plant species due to ecological processes such as tree
falls, mortality of seeds and seedlings, and floodings. In many localities a series
of disturbed and undisturbed habitats is expected to be found, some a result of
natural perturbations in the ecosystem, others the result of man's activities such
as agriculture and lumbering. Thus the lowland tropics is actually a mixture of
many different kinds of habitats, some characterized by a wealth of plant and
insect species, others by considerably lower biotic diversity. In these contexts I
define habitat as a distinct assemblage of plant species occupying an area with
more or less recognizable boundaries and probably occupied by a distinct as-
semblage of other organisms. The habitat may comprise several distinct mi-
crohabitats occupied by smaller subsets of the species characterizing the whole
habitat.
Tropical habitats, particularly those in the nonseasonal or mildly seasonal
lowlands and mid-elevations, are characterized by complex assemblages of dif-
ferent communities of plants and animals, reflecting the complex trophic organi-
zation of these ecosystems. The combination of relatively long stable environ-
mental conditions and the dynamics of coevolved interactions between groups
such as plants and insects have paved the way through evolutionary time for the
evolution of complex communities. Many communities in tropical terrestrial
habitats are relatively more complex than similar Temperate Zone communities
in terms of number of species and the degree to which a portion of the commu-
nity's membership are specialist species in terms of exploiting resources. In
DISTRIBUTION PATTERNS OF INSECTS 179

general, the species making up a tropical community may be more in equilibrium

with their resources than species in a Temperate Zone community. In either the
tropical or temperate region, different communities exhibit varying degrees of
species complexity, specialization of member species, and degree of coevolved
interactions. Yet on the average, the lowland and mid-elevational regions of the
tropics contain more complex communities. The greater such complexity, in
terms of numbers of species and ecological specialization, the greater the vul-
nerability of the community to changes in the environment in which it thrives.
Relatively small changes in physical factors in a lowland or mid-elevation tropi-
cal habitat may have large, drastic changes on the community: the equilibrium
between species and their resources can be shifted in such a way as to either
promote the loss of one or more species from the community or result in a
replacement of one or more species with invading species. Ecological specializa-
tion in some communities, and high species number, promotes conditions favor-
ing an unstable equilibrium, and sustained perturbations, even if small in mag-
nitude or intensity, may lower or shift diversity. (65) Such effects of environmental
changes may impact less in many Temperate Zone communities, but because a
spectrum of community complexity patterns is found at any latitude, this is not
universal. Vulnerability to environmental perturbations will also vary with
trophic status and the degree to which member species are specialists within the
trophic level.
Habitats and their component microhabitats are both open systems, with
energy flow within them and with other habitats. Thus epiphytic tank bromeliads
are one kind of microhabitat within the primary tropical rain forest habitat, and
there is certainly an exchange of energy with other components of that habitat,
mediated by insects and other animals that may breed or develop within this
microhabitat but also interact with other microhabitats. The death of such a
bromeliad makes it a component of the ground litter microhabitat. Depending
primarily upon the permanency of the habitat and the food chain complexity it
contains, resident species will be specialists or generalists, or some mixture, for a
given trophic level.
Richards(394) provides detailed descriptions of the floristic composition and
structural organization of tropical forests, with an emphasis on the lowland
tropical rain forest. Holdridge(395) and Holdridge et al. (396) discuss in detail the
relationships between climate and other features of the physical environment as
determinants of the flora of different tropical forests. It is easy to conceptualize
the tropical forest as a single phenomenon with three distinct forms: rain forest,
cloud forest, and semideciduous forest. Given these distinctions, tropical forests
are most developed in the equatorial regions of Central and South America,
Africa, and Indo-Malaysia. The vertical stratification of the rain forest is stressed,
and the fact that groups of invertebrates associated with the soil in Temperate
Zone forests are associated primarily with the vegetation in the tropical forest.
180 CHAPTER 5

The ground fauna of the lowland tropical rain forest appears to be somewhat
impoverished (Table 5.1). In semideciduous forests there are seasonal shifts in
the density and species composition in some groups of invertebrates, with con-
tractions of faunas in the dry season (Fig. 5.1).
Most of our knowledge of the distribution of insect species in tropical
forests comes from studies of the faunas associated with the lower strata of
primary forest or the low canopy faunas associated with relatively young secon-
dary succession. Explorations and systematic surveys of the canopy faunas of
primary forest will undoubtedly reveal not only new taxa, but also different types
of ecological interactions among species. For example, in some tropical forests,
substantial accumulation of debris is expected in the canopy or just beneath it,
and these materials may function as food or domicile for various insects as-
sociated with the canopy. In the realm of risky speculation, insect groups less
abundant in the ground litter and lower vegetational strata of primary-growth rain
forest may reappear in great abundance in the canopy communities. Thus, rather
than some insect taxa of the lowland tropics experiencing range expansions
through horizontal movements, some may filter through the canopy via vertical
movement, experience considerable ecological diversification there, undergo
horizontal shifts along elevational gradients, and so forth. Such geographical
shifts may involve a return to a ground-dwelling habit, especially if habitats at
high elevations are less structurally complex.
Clearly, insect taxa also have distributed themselves across all the major
vertical strata of a lowland tropical rain forest as well as among all habitats in a

TABLE 5.1
A Comparison of Mite and Collembola Densities in Temperate and
Tropical Forest Soils a

Mean density
per square
Habitat centimeter Reference"

Temperate
Beech mar 8.61 van der Drift (1951)
Spruce mor 14.6 Farss1und (1948)
Fen soil 3.19 Macfadyen (1952)
Hemlock mor 2.92 Wallwork (1957)
Tropical
Rain forest (Brazil) 1.0-2.6 Beck (1967)
Sclerophyl forest 1.0
(Chile)

aFrom Wallwork.'78.)
"References can be found in Wallwork."'·)
DISTRIBUTION PATTERNS OF INSECTS 181

(a) Dr

ft--------.
Litter
0

Iii : I
oI
10I 20
I

Species/sample

(b)

Litter
0

:=: 20 E
s
.... .s::
:::"
'.'
:.:
.,
40 0Ci.
:~:
:.: 60
:::
o, 100
, 200
, 300
,
Individuals/sample

FIGURE 5.1. The vertical distribution patterns of cryptostigmatid mites in wet and dry seasons in EI
Salvador rain forest soil. (a) Number of species per sample; (b) number of individuals per sample.
[From J. A. Wallwork, The Distribution and Diversity of Soil Fauna, Academic Press, London.
Reprinted with permission.J

region. Ehrlich and Raven<6) discuss the patterns of coevolution between but-
terflies and their host plants, indicating that the great physical and chemical
diversity of angiosperms in the tropics is a major determinant of the great diver-
sity of butterflies found there. They argue that it has been primarily the evolution
of secondary plant substances, especially in the tropics, that has driven the
evolutionary radiation in many groups of phytophagous insects, including but-
terflies. In not all cases is oviposition preference in butterflies governed solely by
chemical and mechanical stimuli from plants. Wolfson(397) found that the oviposi-
tion preferences of Pieris rapae were determined largely by leaf water content, with
variations in leaf water content being the source of selectivity in ovipositing
butterflies. Richards(394) has stressed the great diversity of leaf structure in tropi-
cal vegetation, pointing out considerable diversity in terms of leaf toughness,
thickness, pilosity, etc. Some of these differences may reflect differences in leaf
water content, and the flora of a lowland tropical forest is expected to provide a
182 CHAPTER 5

rich spectrum of leaf water content patterns among the plant species found there.
Such a pattern only increases the environmental heterogeneity in terms of pro-
moting specialized or generalized oviposition habits in tropical plant-feeding
insects, including butterflies. Such complexity adds an additional layer of re-
source heterogeneity for these insects over the existing spectrum of chemical
defenses among plant species in a habitat.
Insects, being small poikilotherms, are believed to be successful ecologi-
cally in terms of coping with physical features of the environment such as
temperature, humidity, and moisture. But most of the selection pressures leading
to diversification through evolutionary time have been the adaptations of
phytophagous insects to the flora. Recent studies on the natural history of tropical
butterflies support this view in the sense that not only are certain higher taxa
associated with specific groups of plants, but considerable host plant specializa-
tion exists within certain genera and species of butterflies. A species, through the
medium of the individuals in the popUlation, and through the population as a
functional unit in a community, adapts to the local environment by adapting to
the physical environment, to potential or actual competitors, to a particular
spectrum of resources (including food), and to the removal of genotypes from
predation and parasitism. The joint phenotypic responses to these environmental
factors molds the strategy of reproductive isolation in a species.
The view of Ehrlich and Raven(6) gives us a handle on the broad patterns of
ecological relationships between phytophagous insects and their host plants.
However, in trying to explain why a particular insect feeds on a particular host
plant species it is necessary to examine the character states for the four classes of
environmental effects mentioned here. The interplay of these factors and the
constraints of genotype and phenotype frequencies within a population are what
determine patterns of habitat selection in species of insects in the tropics and
elsewhere. Where closely related species exhibit distinct patterns of host plant
utilization within a habitat or occupy similar resources or patches in different
habitats, we must examine these factors in determining the selection pressures for
the observed character displacement. Spatial character displacement over rela-
tively short distances in and among habitats within a region is one mechanism of
accumulating a high number of species in a region. See Chapter 7 for a considera-
tion of patterns of seasonality in tropical insects, a form of temporal character
displacement.
Patterns of local character displacement among closely related species in
tropical habitats may involve not only phenotypic traits such as host plant prefer-
ence and shade tolerance, but also the ecological factors underlying population
growth. A species of leaf hopper attacking trees in a forest understory may
experience the environment very differently from an allied species attacking
herbaceous plants in an adjacent disturbed habitat. The ease of nutrient retrieval
may be different: the leaf hopper in the forest habitat allocates more energy to
DISTRIBUTION PATTERNS OF INSECTS 183

feeding and therefore has less energy for evolutionary increases in fecundity.
These effects, coupled with the existence of larger host plant patches in the
disturbed habitat, may result in a buildup of larger populations of the leaf hopper
species in the disturbed habitat. Predation intensities may also vary, as well as
the number and density of competing species on the same host plants. Thus
evolutionary changes in the population growth parameters (see Chapter 6) in
different directions in allied species in different habitats are phenotypic traits
reflecting character displacement and habitat selection.
Some studies of aquatic insects in the tropics, such as the study by
Hynes(398) of aquatic Heteroptera in Trinidad and Tobago, also indicate consider-
able niche diversification among species. But this particular survey includes both
geographical and habitat collections, combining the effects of both factors in
determining niche divergence within a habitat. Geographical differentiation
within a particular group is expected, sometimes to the degree that selection
pressures within a particular region will be markedly different than elsewhere in
the range. The example of large, flightless weevils with plant growth on their
elytra in the moss forests of Papua, New Guinea (Gressitt et al. (399») exemplifies
the effects of local adaptation to a set of environmental conditions not found
elsewhere in the range of this subfamily of weevils. The phenotypic traits present
on Papuan weevils indicate strong selection pressure favoring well-developed
crypsis, probably as a result of the flightless conditions of the insects. Sedentary
insects in tropical forests may be more vulnerable to predators than more mobile
species because they are deprived of escape behaviors associated with move-
ment. Silberglied and Aielld 4oO ) studied the wettable body surfaces of some
bark-inhabiting bugs on Barro Colorado Island in Panama; wettability conceals
the bugs from predators by making the bugs match wet bark in the forests. These
insects are relatively sedentary, spending most of their time on wet bark.
Jiron and Zeledon(4ol) report on the distribution of Anastrepha (tephritid)
flies in various kinds of fruits in Costa Rica and demonstrated considerable
overlap among three species of the flies for several fruits examined. The effect is
regional: the species occur in different regions of the country but exploit some of
the same fruits. Many species of drosophilid flies are commonly seen on rotting
and fleshy fungal growths in tropical wet forests, including species that breed
there and those whose adults only feed there. The occurrence of such food
sources in tropical forests provide a microhabitat for considerable co-occurrence
in drosophilids. Drosophilids are also found on rotting fruits in such habitats.
Buruga and Olembo,(402) for example, found that large, fleshy fruits supported
the greatest number of species of Drosophila in forest habitats in Uganda.
Experiments in progress by Young in tropical wet forest in Costa Rica indicate
that considerable habitat selection occurs in some groups of fruit-eating but-
terflies. In one study, a transect of rotting banana baits was set up in both primary
forest and relatively open second growth and the numbers of Morphidae,
184 CHAPTER 5

Satyridae, and Brassolidae were scored for as long as the baits lasted. Very little
overlap between habitats was found, and the overall number of species of these
butterflies in the second growth habitat was low compared with the number in the
forest habitat. Three species of satyrids were seen at the second growth baits, as
were a few individuals of M. peleides. In the forest habitat, three species of
Morpho were found on the baits and at least three genera of satyrids and two
genera of brassolids. Such data simply confirm something that most tropical
biologists have known for some time, namely, that different groups of butterflies
are associated with different habitats, primarily because their resources are not
distributed evenly among all habitats in a region. In tropical communities with a
high number of ecologically similar taxa, some pattern of resource partitioning is
expected if some resources are in limited supply. Sometimes the partitioning is
evident within a particular habitat or is more diffuse at the regional level. Thus
Waltd403 ) discovered considerable evidence of resource partitioning among 28
species of dung beetles in Zaire, and these ecological differences were exhibited
in terms of nest sites, flight patterns, and other phenotypic traits related to
feeding and breeding. The observed patterns of differentiation were most evident
on a regional basis where the species overlapped considerably.
In the central highlands of Costa Rica, some species of cicadas partition the
local environment along rather distinctive boundaries. For example, F. pronoe
emerges among coffee bushes adjacent to a forest-lined stream, whereas two
others, Q. gigas and Z. smaragdula emerge on the trees along the stream, only a
few meters away (Fig. 5.2). The mechanisms responsible for such patterns have
not been determined, but similar surveys of cicada emergences in premontane
tropical wet forest by Young(71) suggest that cicadas exhibit considerable di-
vergence in host tree preferences; the nymphs may associate with the root crowns
of different trees. Young was able to recognize those cicada species that emerged
in secondary growth and those emerging in primary growth habitats in that
region. Young also suggested that the creation of secondary habitats in the wet
tropics may provide the opportunity for some cicada species, and probably many
other insects, to invade new environments as they become available as a conse-
quence of land-clearing activities of man and natural events such as major flood-
ings of rivers.
Schustd404 ) discovered that flower-visiting Orthoptera in the tropics often

FIGURE 5.2. Cicada habitats in the central highlands of Costa Rica. Above: strips of Zygia latifolia
trees (Leguminosae) that border streams (to the left), with which cicada emergences are associated.
Below: discarded nymphal skins of the large cicada Q. gigas clinging to a Zygia tree. In heavily
disturbed habitats in Central America, cicada emergences are associated primarily with forest remnants
as shown here, although some species have secondarily become associated with leguminous shade
trees over crops such as coffee. The association is probably defined in terms of suitable oviposition
sites (branches) and root crowns for suitable nymphal development.
186 CHAPTER 5

exhibit considerable preferences for certain plant families. Working in the tropi-
cal wet regions of Peru and the Canal Zone, he concluded that orthopterans may
seriously reduce the breeding capacity of the plants visited because the insects eat
the anthers, probably as a source of protein. The study of Fraser Rowen<404a)
shows that acrid ids (Orthoptera) in Costa Rican lowland tropical rain forest are
specialist feeders on plants. Fraser Rowell suggests that great differences in
secondary compounds between different plant species in such environments pre-
cludes the evolution of a generalist feeding strategy in the acridid species found
in this habitat. I have observed that some orthopterans are opportunistic predators
on insects at lights in Costa Rican wet forest regions, and one gets the impression
that associations of orthopterans with flowers are not specialized associations but
rather ones that occur opportunistically. Schustet404 ) does make an interesting
point about the activity possibly reducing the pollination of the attacked flowers
by bees, since structures are destroyed and pollen is taken. A similar phenome-
non with Trigona bees as the bandits, may be occurring in some cacao planta-
tions in Central America. Young(40S) found that Trigona jaty collects large amounts
of pollen from cacao flowers but does not pollinate the flowers. The pollination
of cacao is done primarily by ceratopogonid midges whose behavior of visiting
the flowers is such that pollination can result. (406) Yet the high densities of T. jaty
in some cacao plantations suggests that the bees are actually robbing pollen that
would otherwise be available for reproduction. (See Chapter 9 for a more in-
depth discussion of this complex matter.) An interesting further consideration
concerns the kinds of habitats adjacent to cacao plantations that may alter the
foraging habits of T. jaty on cacao. For example, I discovered a plantation
bordered by dense flowering patches of the composite Schistocarpha oppo-
sitifolia in which very large numbers of T. jaty bees were foraging. On the
adjacent cacao, the bees were virtually absent. The tentative conclusion of such
descriptive studies is that the availability of alternative resources in one habitat
will influence the foraging habits for another resource in an adjacent habitat. Our
knowledge of insects in tropical habitats must include an attempt to understand
how different sets of habitats, with differing resources, influence the patterns of
exploitation of particular resources within each component habitat. Few studies
are designed around such a goal.
Some kinds of agricultural habitats in the wet tropics may support a diver-
sity of microhabitats, as suggested by Laessle's study(407) of the bromeliad flora
of several cacao plantations in Trinidad. In that study, the cacao plantations were
similar in design and the composition of the bromeliad flora and shade charac-
teristics of habitats were studied for caco plantations and adjacent strips of forest
habitat. The data obtained (Table 5.2) indicate considerable heterogeneity in the
response patterns of different bromeliads to the three kinds of habitats: high-
shade forest, low-shade forest, and very low shade cacao plantation. The author
recognizes three different ecological groups of bromeliads on the basis of tol-
 TABLE 5.2
The Density of Bromeliads per Five Trees of Eyrthrina micropteryx in Cacao Plantations in Four Distinct Rainfall
Zones of Trinidad a

Density

p values of indicated
Bromeliad A B C D density changes b

Catopsis sessiliflora 2.1 ± 0.5 0.5 ± 0.5 0.6 ± 0.4 0.2 ± 0.2 C to F; p < 0.05
Vriesia amazonica 2.5 ± 0.5 0.0 0.1 ± 0.1 0.0 C to D, E, F; P negligible
Vriesia procera 7.8 ± 0.7 6.2 ± 1.6 7.3 ± 0.8 3.7 ± 1.8 No significant change
Guzmania monostachia 6.9 ± 0.8 8.8 ± 0.5 8.5 ± 0.4 7.1 ± 1.3 No significant change
Gravisia aquilega 8.0 ± 0.7 9.3 ± 1.0 8.5 ± 1.0 5.0 ± 2.2 No significant change
Wittmackia lingulata (1.5 ± 1.0) 8.9 ± 4.3 6.9 ± 1.7 1.2 ± 1.2 E to F; p < 0.05
Aechmea nudicaulis 8.3 ± 0.8 13.0 ± 2.3 20.1 ± 2.3 8.3 ± 2.0 E to C or F; p < 0.02
Tillandsia Jasciculata 5.6±0.7 8.2 ± 1.2 7.5 ± 0.7 10.6 ± 1.3 C to F; p < 0.05
Tillandsia utriculata 0.0 0.0 2.1 ± 0.5 3.9 ± 0.8 C & D to E or F; p negligible
Tillandsia subimbricata 0.0 0.0 1.3 ± 0.6 6.4 ± 0.5 C & D to E or F; p negligible
Number of samples 12 6 10 6
(of five trees)

aprom Laessie.(407)
'The values given are means of square-root transfonnations of actual frequencies per five trees.
188 CHAPTER 5

TABLE 5.3
The Distribution and Abundance of Bromeliads in Two Forest Strips
Differing in Light Intensity and in an Adjacent Cacao Plantation
in Trinidad a

Plantation
Forest I Forest 2
Strongest
Dark due Well lit by illumination
to lianes river at side also drier

Exposure group
C. sessiliflora 0 4 49
C. floribunda 0 8 3
V. procera 8 28 144
V. amazonica 3 5 5
Sun group
T. fasciculata 0 I 61
G. monostachia 8 60 101
Aechmea Mertensii 1 5 0
Tillandsia bulbosa 8 10 2
G. aquilega 0 28 142
A. nudicaulis 6 79 201
Guzmania sanguinea 0 0
Hohenbergia stellata 0 4 0
Shade-tolerant group
Tillandsia anceps 4 34 0
Tillandsia monadelpha I 26 0
Vriesia albiflora 0 0
Vriesia simplex 10 0
Guzmania lingulata 48 78 0

aprom Laessie.(407)

erance to shade (Table 5.3). The implications of such patterns for the insect
species associated with the bromeliads are also apparent: tank bromeliads in
shaded forest habitats will have larger communities of aquatic and arboreally
cryptozoic insects than bromeliads associated with exposed or dry habitats in the
same region. But the bromeliads found in cacao plantations may support damp-
to-wet leaf litter communities as well as herbivores associated with leaves and
flowers. Cacao plantations can support a high density of different bromeliads in
shade trees such as Erythrina (Table 5.2). The differences in distribution of
various bromeliads between different kinds of forest habitats and cacao planta-
tions impose a pattern of environmental heterogeneity for insects and other
organisms exploiting these microhabitats.
DISTRIBUTION PATTERNS OF INSECTS 189

The studies of M. H. Robinson and B. Robinson(408) described another form

of specialized microhabitat in tropical wet forests, the association of
Drosophilidae with certain orb-weaving in forest understories. These are
commensal associations in which the flies clean the mouthparts and anal areas of
the spider and the flies take food from the spider as it is being eaten. Thus the
spider and its web form a sort of specialized microhabitat for the flies. Young(408a)
and some earlier workers, notably P. Calvert, described the association of certain
bromeliad-breeding giant damselflies with the same orb-weaving spiders in-
volved in the commensal associations with flies. Young observed that certain
kleptoparasitic spiders in the webs of the orb weavers are taken as prey by these
damselflies. Thus these damselflies essentially glean the web microhabitat of
kleptoparasites.
Such studies imply that some types of resources are very limited in abun-
dance, suggesting ecological constraints on species specialized to exploit them.
Matthews(223,224) studied the nesting behavior of the social wasp Microstigmus
comes in a tropical wet forest in Costa Rica and found that the wasp selectively
nests on the undersides of leaves of certain palms in the forest understory. The
wasps require scrapings of tissue from the undersides of these palm leaves to
make their nests. Within a patch of approximately 75 palm plants, about 60% had
no nests and the remainder one to five nests apiece. The data suggest that the
wasp popUlation is spatially distributed within a relatively small area of forest
habitat (about 35 X 40 m) where a required nesting resource is located. Nesting
site selection in some social insects may also depend upon the degree to which
the species can defend itself against enemies.
In a very interesting study, Windsot409 ) found that nest site selection in a
docile tropical wasp Mischocyttarus immarginatus is determined largely by the
locations of nests of another polybiine wasp, Polybia occidentalis, which is a
fierce attacker. Such interspecific nesting relationships show that a docile as-
sociate nests immediately adjacent to an aggressive associate species, presum-
ably as a means of protection against predators such as small vertebrates. In ant
species with specialized nesting requirements, the distribution of nests in a
habitat will be determined in part by the distribution of preferred nesting sites.
For example, Young and Hermann!212) found that mature Pentaclethra trees in
Costa Rican tropical wet forest tend to have a greater number of nests of the
giant, aggressive ant Paraponera clavata associated with the roots of these trees.
But that study did not include an experiment to demonstrate the apparent prefer-
ence of this tree species over others as a nesting site in this ant. This ant is
capable of various kinds of feeding, including expolitation of extrafloral nec-
taries,(211) and it would be interesting to determine if its nesting habits include
some specialized relationship to the root systems of certain tree species in forest
habitats. The ant is a subterraneous nester that forages primarily in the forest
canopy, thus exploiting a broad range of microhabitats within the forest habitat.
190 CHAPTER 5

Tropical forest and secondary-growth habitats may contain a wealth of

microhabitats in the form of both nesting sites and food for the great many ant
species found in these places. Jeanne(410) demonstrated that rates of predation by
ants are greater in the tropics than in Temperate Zone areas and that predation
rates in dissimilar tropical habitats, such as forest understory and open field, tend
to be similar. He found more pronounced habitat differences in predation rates
for ants in different Temperate Zone habitats than in tropical habitats (Table 5.4).
The greater pools of ant species associated with both forest and secondary-
growth habitats in the tropics result in high removal of prey under both condi-
tions, whereas in the Temperate Zone, more of ant species are in the field habitat
and therefore predation rates are higher. The availability of many kinds of prey
and other food sources and a diversity of nesting sites account for the high
species richness of ants in both tropical forest and field habitats. In Jeanne's
study, (410) the number of ant species found in the tropical habitats was considera-
bly higher than for the other regions examined (Table 5.5).
Ant species which depend upon the exudates of homopterous insects may be
an example of a hypercontingent species as defined by Colwell.(411) A hypercon-
tingent species depends directly upon a relationship between at least two other

TABLE 5.4
Degree of Habitat and Site Specificity among Foraging Ant Speciesa

Shared habitats or New

microhabitats Hampshire Florida Mexico Costa Rica Brazil

Forest microhabitats·
17 57 58 59 57
2 17 13 20 16 15
3 8 10 16 2 17
4 25 20 4 18 7
5 33 0 2 5 4
Field microhabitats·
I 38 30 46 56 46
2 31 40 46 18 27
3 31 40 8 26 27
Field and forest C 27 31 16 28 3
Canopy and forest understoryd 100 50 36 44 44

aFrom Jeanne.(410)
• "Forest" and "Field" entries give the number of ant species collected from I, 2, ... , in microhabitats, as a
percent of all species collected from that habitat at one locality.
C "Field and forest" entries give the number of species collected in both habitats, as a percent of the sum of all

species collected in the two habitats at one locality.

'''Canopy and forest understory" entries give the number of species collected' in the two set of microhabitats,
as a percent of all species found in the canopy at one locality. ("Canopy" here does not include canopy
trees at 15 m.)
DISTRIBUTION PATTERNS OF INSECTS 191

TABLE 5.5
Ant Species Richness at Five Localitiesu,b

New
Hampshire Florida Mexico Costa Rica Brazil

Forest Habitat
Ground 12(70) 16(72) 23(54) 26(71) 24(78)
Twigs 8(36) 11(42) 15(34) 16(58) 19(66)
Leaves 9(30) 12(49) 19(39) 14(60) 12(71)
Ground + twigs + leaves 12(136) 24(163) 40(127) 33(189) 35(215)
Trunks 5(33) 7(35) 18(44) 13(68) 15(77)
Canopy above 15 m 7(25) 12(42) 11 (19) 16(37) 16(22)
Canopy at 15 m 14(29) 17(23)
Entire forest habitat 12(194) 30(240) 59(190) 50(323) 50(337)
Field Habitat
Ground 8(45) 15(99) 12(67) 18(72) 12(71)
Twigs 11 (55) 13(59) 13(53) 20(76) 18(69)
Leaves 12(58) 13(52) 14(53) 20(77) 17(47)
Ground + twigs + leaves 16(158) 20(210) 24(173) 34(225) 26(187)
Buildings 2(2) 7(39) 2(3) 13(60) 7(32)
Total 22(354) 40(489) 66(366) 68(608) 74(556)

a From Jeanne.(410)
"Data are numbers of species. Number of baits at which ants were collected at each microhabitat, habitat, or
locality is given in parentheses.

species. Thus phytophagous Homoptera depend upon plants for food, and the
ants obtain exudates from the bugs as a food source. As pointed out by Colwell,
the hypercontingent strategy may involve a dependence for food, dispersal,
defense, reproduction, or support. Thus the palatable insect species, such as
many butterflies in the tropics (Fig. 5.3), some of which mimic highly unpalata-
ble models, are examples of hypercontingent species, since survivorship of indi-
viduals in their populations depends on the interaction between host plant and
larvae of the model species as the mechanism of unpalatability enforcing the
relationship.
Colwell suggests that the interplay of various adaptive strategies in or-
ganisms accounts for part of the high tropical species diversity noted in some
regions and habitats. He suggests a mixture of sequential specialists, namely,
species that differentially exploit resources through time such as seasonal dis-
placement (Chapter 7); interstitial species, which depend upon several different
widely dispersed resources; grain specialists, which exploit specific resource
patches of one or few kinds in a habitat; and hypercontingent species. Colwell
developed these definitions of adaptive strategies from his studies of the interac-
tions of hummingbird species with seasonally available flowers, the association
of nectarivorous mites with flowers and hummingbirds, and the influence of a
192 CHAPTER 5
DISTRIBUTION PATTERNS OF INSECTS 193

FIGURE 5.3. The fourth-instar caterpillar of the Neotropical nymphalid butterfly Hamadryas februa
actively feeding on a mature leaf of Delachampia heteromorpha (Euphorbiaceae) in the lowland
tropical dry forest zone of northwestern Costa Rica (a); dorsal and ventral sides of the adult (b).
194 CHAPTER 5

nectar thief on the system. The definitions are useful in examining a variety of
interactions among species in different trophic levels within tropical habitats.
Tropical communities are believed to consist of species with a variety of
different adaptive strategies and in which most of the adaptation is a response to
biotic associations rather than physical environment. For example, one might
hypothesize that some kinds of endoparasites of phytophagous insects are grain
specialists, in that closely related species exploit different hosts and in proportion
to their abundances in the habitat. And the abundance and distribution of the
insect hosts are a function of the distributions of the various host plant species
exploited. Some phytophagous insects in the same habitat, rather than being
specialists on one or a few resource patches in the habitat, may be interstitial
species in the sense that they feed on a variety of plant species dispersed in the
habitat. The coexistence of two possibly competing endoparasites therefore de-
pends, in this hypothetical example (and borrowing from Colwell(411l), upon the
grain structure of the environment, determined in this case by the feeding pat-
terns of the phytophagous insect species. The assumption is made, of course, that
each species of endoparasite is a specialist on a particular insect host, and this
may not be the case if most endoparasites of phytophagous insects in the tropics
are generalists. Few data in either direction are available. In the first instance,
specialist endoparasites would be considered hypercontingent species because
their existence in the habitat depends upon the successful interaction of their
hosts with their food plants.
As pointed out in Section 5.1, a major logistical problem with determining
the degree of niche overlap among ecologically similar species in a habitat is the
level of empirical resolution to detect differences. Several studies of resource
partitioning in natural popUlations of microsympatric Drosophila species have
emphasized this matter. Two or more closely related and ecologically similar
insects may overlap tremendously in terms of diet, and the populations of each
species may be large. The question of co-occurrence necessitates the examination
of other components of comparative natural history to determine possible
mechanisms of resource partitioning, assuming resources of limited supply. Thus
two closely related chrysomelid beetles in a tropical field may coexist on the
same host plant species with complete overlap of diet in terms of host plant
selection. But resource partitioning may exist in terms of the spatial exploitation
of the host plant population. Most individuals of one beetle species prefer to feed
on host plant individuals with a certain degree of exposure, while a second beetle
species prefers the opposite level of shade on the plant. Or each species may feed
on different parts of the plant, in which case considerable nonoverlap of niches
may exist at the level of secondary substance types and concentrations experi-
enced by feeding beetles, and perhaps also in terms of nutrients received.
Other forms of spatial separation of the potentially competing species could
include differing preferences for young versus older individual plants. If it can be
DISTRIBUTION PATTERNS OF INSECTS 195

demonstrated that the host plant is a limited resource to both species of beetles in
that particular habitat, the observed patterns of resource partitioning can be
assumed to result from competition between the species. We would also have to
demonstrate absence of feeding on other plants and the degree of movement of
beetles in other habitats, where the relationships of each species to the flora may
be very different. Once the pattern of resource partitioning was determined, we
would need to examine phenotypes of each species to determine whether
morphological, behavioral, or physiological differences are correlating well with
the observed spatial partitioning. The latter is not easy to do in most cases. Both
species may respond to temporal changes in the habitat in the same way, adjust-
ing to daily influences of thermal gradients and shade gradients, and doing so in
such a way as to maintain an optimal spacing pattern of individuals within the
popUlation.
Good data are available on the overall species compositions of the insect
faunas associated with various terrestrial habitats in tropical regions, including a
few studies comparing faunas over altitudinal gradients. I am referring here
primarily to the sweep sample studies of Janzen,(72) Janzen and Schoener,(134) and
Janzen et at .. (412) There are also data on the specific host associations of some
insect taxa such as seed-eating beetles and bugs, and larval host plants of
Lepidoptera. We still need to determine the niche breadth of selected insect
species in a habitat and to relate these measurements to competition, predation,
and spatial and temporal availability of resources. Where very different kinds of
habitats occur together in an area, it is also necessary to determine the amounts of
species overlap among habitats, movement patterns of individuals among the
habitats, and changes in the natural history of individual species due to dif-
ferences in habitats for species that occur in more than one habitat. Such mea-
surements should be incorporated into some of the major theories of maintenance
of high species diversity in the tropics and the lack of high species diversity in
some areas of the tropics and temperate zones. The plant-insect interface seems
to be a good starting point, using the approach of comparative natural history.

5.3. INSECT COMMUNITIES IN TROPICAL HABITATS

The studies of Futuyma and Gould(413) indicate that few plant species in the
Temperate Zone habitat studied have distinct assemblages of leaf-chewing in-
sects associated with them. Many insects in such samples tend to be generalists
on many plant species. Typically a Temperate Zone deciduous forest has consid-
erably fewer species of plants than most lowland tropical forests, and a partial
explanation for a lack of correspondence between leaf-chewing insects and host
plant species is that the average insect species in such a habitat must generalize to
feed on several different plant species in order to maintain a breeding population.
196 CHAPTER 5

Although such a habitat does have a complement of specialist species,(413)

ecological specialization in most phytophagous insects under such conditions
may be more difficult to evolve than in a tropical forest habitat, where a greater
number of plant species per family occurs and the balance might be partially
tipped toward specialization as reflected in monophagy and limited polyphagy
(defined here as the exploitation of two closely allied plant families). Specializa-
tion in tropical forests also depends upon the patch structure of each plant species
relative to the population growth requirements for maintaining breeding popula-
tions.
Murdoch et al. (414) found that insect diversity in a Michigan old field corre-
lated very well with plant diversity and plant structure. There is reason to expect
similar correlations in tropical communities: both parameters of plant com-
munities are higher in most habitats.
With the exception, perhaps, of some defogging studies of the Orthoptera
associated with the canopy of Costa Rican and Amazonian tropical rain forests,
the sweep-sample studies of Janzen(72) in Costa Rica and the Caribbean provide a
picture of insect communities associated with different types of foliage in dif-
ferent tropical habitats. The reader is referred to Janzen,(72) Janzen and
Schoener,(134) and Janzen et al. (412) for in-depth analyses of the extensive data
obtained with the sweep-sample approach. The findings of these studies indicate
the effects of seasonality, habitat type, elevation, and island effects on the
composition of insect communities in the tropics. Some of the findings are
discussed here.
A basic comparison in terms of distinct habitats is that between insects in
secondary vegetation and in primary forest understory at a particular site. For
example, in tropical wet forest in Costa Rica there is a drastic reduction in
numbers of both bug species and individuals of each species for adults going
from secondary to primary forest understory; the transition is far less noticeable
for adult beetles.(72) The greater abundance of non woody stems in secondary
habitats promotes a greater diversity of plant-sucking insects such as adult
Homoptera, whereas leaf-chewing adult beetles have leaves in high abundances
in both habitats. Janzen found little overlap in species composition between the
two habitats. Interestingly, Janzen found the greatest di versity of insect species at
intermediate-elevation habitats. He explains that there is a higher harvestable
productivity in plant communities at intermediate elevations due to reduced plant
maintenance costs on cool nights. Above intermediate elevations (near 3400 m),
the physical factors are more rigorous and there is a drastic reduction in the
number of species and individuals and responses to decreased diversity of flora.
Lowland tropical areas with a long and severe dry season show a great decline in
insects during the dry season, with many insects moving horizontally into moist
forest refugia as the dry season advances. Areas with a mild dry season show
increases in insects during the dry season, suggesting that the wet conditions of
DISTRIBUTION PATTERNS OF INSECTS 197

the wet season tend to suppress insect populations, and many species may exhibit
greatest reproduction, survivorship from biotic factors, and flushes of population
growth in the dry season.
Areas with pronounced dry seasons are characterized by insects passing the
dry period in the adult stage and in reproductive diapauseY16) Because vegetative
growth in most plants ceases during the dry season, insect populations decline or
move to other habitats where the vegetation remains lush to some degree. Jan-
zed 72 ) found reduced numbers of insect species and numbers of individuals per
species on islands in the Caribbean and a greater percentage of predatory insects,
principally ladybird beetles. The reduced floral diversity of these islands cannot
support a phytophagous insect fauna comparable in complexity to those found in
mainland tropical habitats; thus the proportion of predatory species increases.
The greater ecological specialization in tropical herbivorous insects makes it
more difficult for these species to establish and maintain populations in island
habitats, much more so than for generalist predatory species.(415) The relatively

TABOGA PAS lURE WE T

4 I
I
.-'-. ........ _., ·-·fl!l~~~·-·-·-·-·-·-· _0"'" .-'-'-'-'-"
.,/ I
I .-.
1 -.
I TABOGA PASTURE. DRY 1
~ ____ ~~---~_~~llU~~~~____ _

.... ~_ /'~ ----t-·~~ . ....................... t·PRov·lliEN·c·IA····· .............. .

.... '" - I"
,/i' ·····················1· I

!IlO 1600 2400

CUMULAT IVE NUMBER OF SIIEEPS

FIGURE 5.4. The cumulative species diversity index, H', for species of adult bugs in 100-sweep
increments for several localities in Costa Rica. See E. C. Pielou. The measurement of diversity in
different types of biological collections, J. Theor. Bioi. 13:131-144, (1966), for a discussion of H'
and how it is calculated. [From D. H. Janzen, Ecology 54:687-708 (1973). Copyright 1973 by the
Ecological Society of America.]
 198 CHAPTER 5

greater flexibility in feeding habits may provide predatory insects with the resili-
ence needed to survive unfavorable periods associated with colonization of is-
lands(415J and may preadapt them to surviving dry seasons in such habitats.
The sweep-sample studies demonstrate that not all tropical terrestrial
habitats will support the same kinds of insect communities associated with vege-
tation (Figs. 5.4 and 5.5). Trophic organization and insect diversity appear to be
greatly influenced by the composition and structure of the vegetation in a habitat
and the degree to which the harvestable productivity of the habitat changes with
seasons. Insect species composition and population sizes within a particular
habitat will also be influenced by the size distributions of adult insects present.
Mean body size, and the variation of this statistic within an age class of a certain
species in one habitat, influences how the population's membership can distrib-
ute itself upon the available resources and also affects interactions with individu-
als of other species. Schoener and Janzen(416J report that the range of size varia-
tion in insects is greatest in communities from tropical habitats. They found that
spatial heterogeneity in moisture and fluctuations in the abundance of resources
in time are two features of habitats that correlate well with observed differences
in the size distributions of insects. Insect species from drier areas or from areas
with longer growing seasons tend to have larger body sizes, whereas those from
areas with more uniform moisture conditions and less distinct growing seasons
(e.g., lowland tropical rain forest) exhibit considerably less variation in size (Fig.
5.6). Drier habitats promote larger body sizes in insects to circumvent dessica-
tion limits associated with small body sizes, and larger body sizes also result
when growing seasons are long in some habitats; an evolutionary adjustment in

TABOGA PASTURE. WET

,._0_._.- 0_.
_._._._._0 _
LA SELVA
SAN VITO BEAN FIELD
S ..,.'" TABOGA PASTURE. DRY
~ .•...•................. ........ ._~ ....................P.~QY.lllE~C.VI .. ·I .......... •............................·
u

SOO 1600 2400

CUMULATIVE NUMBER OF SWEEPS

Figure 5.5. Cumulative number of adult bug species in 100-sweep increments for several localities
in Costa Rica. [From D. H. Janzen, Ecology 54:687-708. (1973). Copyright 1973 by the Ecological
Society of America.]
DISTRIBUTION PATTERNS OF INSECTS 199

PROPORTION AREAS WITH LONG GROWING SEASON _

FIGURE 5.6. The effect of length of growing season and proportion of dry areas on the distributions
of insect lengths. (---) , Means; (- - -), standard deviations). [Reprinted from "Notes on en-
vironmental detenninants of tropical versus temperate insect size patterns" by T. W. Schoener and
D. H. Janzen, American Naturalist 102:207-224 (1968). Reprinted by pennission of the University
of Chicago Press.]

developmental time can be positively correlated with the length of time habitat
conditions are favorable for growth.
The greater variance in body size expected for insects in tropical rain forest
regions can be examined from the standpoint of size variation within one group,
such as a family. Take Neotropical cicadas as an example (Fig. 5.7). At both
lowland and premontane wet forest sites in northeastern Costa Rica, five genera
and 11 common species of cicadas represent a body length range of 10-55 mm
with a mean of 23.5 ± 20.1 (S.D.). In the northwestern dry lowlands, four
genera and 7 common species represent a body length range of 1O-~5 mm and a
mean of 18.4 ± 12.4 (S.D.). There appears to be a greater range of body lengths
for the common and characteristic cicadas in lowland tropical rain forest.
Of the mosaic of habitat types in the lowland tropical rain forest regions,
some habitats are expected to comprise communities with many specialist
 N
CI
CI

. .:

3 4 5 6 7 8 9 (")
:I:
»
FIGURE 5.7. Shown here. from left to right, for both adults and final-instar nymphal skins , are F. mannifera, F. sericans, F. amoena," ~
F. pronoe, Z. smaragdina, Proarna sallei, and Proarna sp. Interest centers on the determination of patterns of temporal and spatial m
:Jl
utilization of the environment in the tropics, using these insects as a model system. UI
DISTRIBUTION PATTERNS OF INSECTS 201

species, while others may contain major representation by generalist species.

There is no reason to predict that all habitats in such a region will be dominated
by specialist species, including phytophagous insects. For example, the edges of
streams in the lowlands may favor generalist species of carabid beetles, which
are then capable of becoming specialists in other habitats.(417) The stream edge
habitat supports generalist carabids, which then can diversify into other habitats
and undergo radiative adaptation, exhibiting both morphological and behavioral
specializations to new selection pressures. Species of marginal habitats may be
preadapted for such radiations into other habitats. Of interest is how the
generalist species arrived in these marginal habitats in the first place. Stream-
edge habitats are probably very old in the sense that they have existed along with
the forests for long periods, giving time for the accumulation of generalists in
them. These species might have been derived originally from the undisturbed
forest faunas, responding to directional selection favoring invasion into these
marginal habitats.
Young(71) has noted that some species of cicadas in lowland and premontane
tropical wet forests in Costa Rica (and presumably elsewhere) occupy stream-
edge habitats, and some of these species are also invading disturbed habitats
resulting from the clearing away of forests for agriculture and lumbering. In these
habitats, population densities of these cicadas may be very high, indicating
successful colonization. The implication from such observations is that species
adapted to marginal habitats such as stream edges in the lowland wet tropics also
may be preadapted to colonize other kinds of disturbed habitats in addition to
undergoing radiative adaptation to more undisturbed habitats such as primary
forests. But apparently such invasions into other disturbed habitats do not neces-
sarily result in ecological specialization and divergence of the kind expected for
transitions to undisturbed habitats, as disucssed for carabids by Erwin. (417)

5.4. PLANT RESOURCES AND HABITAT SELECTION IN TROPICAL

INSECTS

Plant species in different tropical habitats exhibit different growth profiles,

different reproductive structures, and varying degrees of response to seasonality.
Within a lowland tropical rain forest region, for example, the degree of species
nonoverlap may be high between different habitats and considerable structural
differences should exist among the plant communities in each habitat. Thus both
the upper strata and the understories of primary forest habitats tend to be domi-
nated by woody plants, trees, and vines of various sizes and patch sizes. Young,
disturbed habitats less than 5 years old are dominated by herbaceous plant
species, while various kinds of older secondary habitats are dominated by mix-
tures of herbaceous and woody trees and shrubs.(396)
202 CHAPTER 5

Not only will differences among habitats in a region be expressed in terms

of the collection of the chemical defense systems and mutualisms between ants
and plants that affect the taxonomic composition of phytophagous insect com-
munities in the tropics, but the spectrum of physical differences among plants
will also influence habitat selection in these insects. Janzen,(72) for example, has
suggested that the marked decrease in adult bugs in the forest understories of
lowland wet forest is due to an absence of herbaceous stems, which represent a
readily penetrable resource for these xylomycetophagous and phloeophagous
insects. How the diversity of plant species in a particular tropical habitat will
determine the diversity of phytophagous insects is expected to vary for different
groups of insects, primarily on the basis of differences in feeding habits. The
combined effects of less restrictive physical properties of plants such as woodi-
ness, a large pool of taxonomic and defensive chemistry diversity, and a
heterogeneous patch structure are expected to result in a high diversity of
phytophagous insects in secondary habitats in the tropics. The harvestable pro-
ductivity of secondary habitats is greater than that of forest understories, providing
a larger food base for insects. Whether the canopies of lowland forests are also
characterized by higher harvestable productivity similar to that of some secon-
dary habitats remains to be analyzed, and the insect communities associated with
canopies warrant measurement.
The data of Beaver418 ) indicate that for the beetle families Scolytidae and
Platypodidae the tropics contain a relatively higher number of polyphagous insect
species than do Temperate Zone regions. The greatest increases are in two groups
of these beetles, defined by the feeding habits: the xylophagous species, which
live and feed on wood; and the xylomycetophagous species, which usually live in
wood but cultivate and feed on ambrosia fungi growing in the galleries of these
insects. The mechanism of specificity is therefore determined by the host specificity
of the fungi in the case of xylomycetophagous species but can also be determined
by the oviposition behavior of the beetle, which molds the pattern of host tree
selection in both groups. The same study revealed that phloeophagous species,
those that construct galleries and breed in the phloem and inner bark of trees, are
more diverse in temperate zones, and this pattern is explained in that paper by the
differences in host tree patch structures in temperate and tropical forests.
The hyperdispersion of tropical tree species selects for the generalist
strategy since distances between individuals of a particular tree species (of right
age, etc.) are too great for specialization. Yet patch sizes of tree species in
temperate-zone forests, which have fewer tree species per unit area, are large and
provide larger segments of biomass for several species to utilize. The gradient
appears to be opposite what one would expect from the tree species-rich tropics,
where selection should favor ecological specialization in phytophagous insects.
But the key factor in such considerations is the patch structure and relative
patches sizes for each tree species, relative to the dispersal abilities and reproduc-
DISTRIBUTION PATTERNS OF INSECTS 203

tive cycles of the beetle species involved in these interactions. These beetles may
not come into contact with the diversity of secondary substances generally pre-
sent in leaf and meristem tissues of plants and are precluded from the same sorts
of host plant selection mechanisms affecting the majority of phytophagous in-
sects. Put differently, a forest patch of known size containing 10 tree species in a
tropical region and forest patch of the same size but containing tree species in a
Temperate Zone (and comparable numbers of individual trees in both patches)
are not viewed differently by scolytid or platypodid beetle species present in each
habitat.
The host tree selection mechanism is based on other signals in the trees
rather than based on secondary substance chemistry associated with leaves and
meristems, the kind of selection mechanism predicted. Although secondary sub-
stances might be involved, they are likely to be different from those of other
vegetative structures, perhaps of kinds found in the phloem tissues of most tree
species. Yet in considering the likelihood for specialization in phytophagous
insects associated with leaves and meristems, it is necessary to assess dispersal
and life table characteristics in order to determine the evolutionary advantage or
disadvantage for specialization on a host plant species with a certain patch
structure. Low population turnover coupled with low dispersal ability may act to
favor the generalist strategy in many phytophagous insect species associated with
tropical habitats characterized by a high diversity of plant species. Under such
conditions the generalist strategy would include the physiological ability to feed
on a variety of unrelated plant species in a small area of habitat whose boundaries
for the insect species in question are set by the dispersal abilities of adults and the
popUlation growth parameters of the species in that particular habitat.
Clearly some studies of Temperate Zone insects such as butterflies have
revealed few if any patterns of habitat selection based upon distributions of
resources such as adult and larval host plants.(419) Rausher,(420) however, study-
ing habitat selection in three species of Aristolochia-feeding butterflies in
Mexico, found significant differences in the survival of eggs and larvae of these
species between different habitats. Survival of eggs and larvae was higher in
shady areas than in exposed areas (Figs. 5.8 and 5.9), and selection should favor
oviposition in shady habitats. But Rausher found that two of the species actually
preferred to lay eggs in exposed habitats (Fig. 5.10), presumably locked into host
plant cues having little or no relation to survival of eggs and larvae.
Breedlove and Ehrlich,(327) studying a Temperate Zone butterfly population,
concluded that in some cases the intensity of predation on the plant species in a
region may be so high as to be a source of selection regulating the distribution of
the plant popUlations among habitats. But the pattern may be heterogeneous
(Dolinger et al. (421») in that if the butterfly is not "tracking" the plant popula-
tions, there will be considerable variation in the phenotypic commitment within
different plant populations in the expression of secondary substances used in
204 CHAPTER 5

1.0 PH

I
A -
-
-
PO -
.5

~
<II MO
- iii!!

-...
<II
1.0
II>

I
<II
B
.2
u FIGURE5.S. The distribution of egg clusters
CI .5
CI in sunny and shady habitats for three Neo-
II>
tropical papilionid butterflies, Parides mon-
'0 tezuma (MO) (shaded bars), Battus philenor

-
c: - (PH) (stippled bars), and Battus polydamas
0

u
1.0 (PO) (black bars) at two localities, El Encino
...
<II C (A, C; two different study periods) and Pico
"- de Oro (B) in Mexico. There is a definite
il!l!
.5 ~ml preference for B. polydamas and B. philenor
to oviposit in sunny habitats and for P. mon-
-
- - tezuma to oviposit in shaded habitats. [From
- 11III - M. D. Rauscher, Ecology 60:503-511
Sunny Shady (1979). Copyright 1979 by the Ecological
Habitats Habitats Society of America.)

1.0
)(

II>
>

-
iV

~
...
0
0
U
FIGURE 5.9. Patterns of egg and larval
<II .5 survivorship for P. montezuma, B. philenor,
c: and B. polydamas for all combinations of

-
.;:
CI
these species and habitat patches studied .
0 Solid lines represent curves for sunny habi-
0 tats and dashed lines for shady habitats.

-
c: Day zero is the day the egg cluster was dis-
.2
u
covered in the wild. The "(4)" in thefigure
<II indicates the point of convergence for all
J: four curves, thereafter represented as one
line. [From M. D. Rauscher, Ecology
10 15 20 60:503-511 (1979). Copyright 1979 by the
Days Ecological Society of America.)
DISTRIBUTION PATTERNS OF INSECTS 205

CI I 0 PO
c PH
.;:;
.~
tt
a.
.;;
o
.,c
5:
~

.!!
"'
E
~
'0
c
.Q
u

.t"' "W •
~:::::1

Sunny Shady
Hab itats Habitats

FIGURE 5.10. The proportions of females of P. montezuma (MO), B. philenor (PH), and B.
polydamas (PO) searching for oviposition sites in sunny and shady habitats over all observation
periods at two localities in Mexico. [From M. D. Rauscher, Ecology 60:503-511 (1979) . Copyright
1979 by the Ecological Society of America.]

herbivore defense. Thus populations where the butterfly is absent or scarce have
low alkaloid profiles, and where abundant, high profiles.(421) The plant species
might be able to occupy different habitats and the butterfly only a subset of these.
Another factor to consider is the availability of alternative host plants in a
habitat. Chew,(422) working with montane Pieris butterfly species in the Northern
Temperate Zone, found imprecise correlation between oviposition behavior and
the suitability of host plants for larval growth. She concludes that the patchy
distribution of some crucifer species used by the butterflies promotes the use of
less suitable crucifers. Habitat hereogeneity , in the form of both suitable and less
suitable host plant species having patchy distributions within the habitats studied,
promotes the exploitation of all the host plants by the average female in each
species. It is almost as if the environment is providing too much information for
the female butterfly to process per unit time during the oviposition periods.
Young 0l3 ) found some evidence of habitat separation in heliconine butter-
fly species in Costa Rica. For example, although Eueides lybia and Eueides
aliphera are sympatric and utilize the same larval host plant, Passiflora viti-
folia, in the premontane tropical wet forest region of northeastern Costa Rica,
the former species oviposits primarily along shaded forest edges while the latter
does so primarily in open secondary habitats . In this case the larval host plant
occurs in more than one kind of habitat, while the two butterfly species, probably
potential competitors, have different habitat preferences, presumably mediated
by tolerance to sunlight and related factors. Other patterns for heliconiines (Fig.
5.11) have also been noted utilizing other Passiflora species. (313)
206 CHAPTER 5

FIGURE 5.11. The heliconiine butterflies Dryas iulia (above) and Eueides aliphera (below), which
occur together regionally in Central American rain forests but utilize different larval food plants
(Passifloraceae) .

Slater423 ,424) analyzed the vertical distribution of seed-eating lygaeid bugs in

South Africa and the West Indies and found four ecologically distinct groups
based on feeding behavior: (1) arboreal seed predators, (2) obligatory terrestrial
seed predators, (3) frequent facultative seed predators, and (4) accidental terres-
trial seed predators. He noted that most of the species in these groupings belong
to one subfamily. The evolutionary divergence of these bugs into these ecologi-
cal groups may have resulted from competition among the bugs for the same kind
of resources, and the groups reflect differing amounts of ecological specializa-
tion, with accidental terrestrial seed predators being the most flexible group in
terms of utilizing resources other than fig seeds. An important component of such
interactions involving many insect species is the area of habitat available for the
species to thrive in and the amount of time available for ecological divergence.
. Janzen<37S) has shown that the dissolution of mutualism between Azteca ants
and Cecropia trees in Puerto Rico, compared with the well-developed mutualism
present in many mainland Central American localities, is an adaptive response to
reduced competition among plant species in Cecropia habitats and reduced herbi-
vore pressure on the trees on islands. Some studies, (e.g. , Lawton(42S» indicate
DISTRIBUTION PATTERNS OF INSECTS 207

that vegetation complexity, in terms of plant species composition and the structural
diversity of plants present in the habitat, is a good indicator of the diversity of
phytophagous insects present. Habitats with high structural diversity and species
composition may accumulate a greater diversity of insects in certain trophic
classes.
Young(215) found that many species of Neotropical cicadas will form dense
aggregations of adults in certain trees in disturbed habitats, while such aggrega-
tions are far less numerous in forest habitats (Table 5.6). It was suggested that
cicadas in disturbed habitats select on such trees for optimal mating sites and
feeding and that the absence of aggregates in undisturbed forest is due to the
patchy distribution of each tree species found there, making it difficult so to
speak for cicadas to discover and congregate in those trees that optimize feeding
and mating. Thus certain kinds of behavior in insects will also influence how
populations are superimposed upon available resources, again with the patch
structure of the vegetation in each habitat being a major determining factor.

5.5. DISTURBED AND UNDISTURBED HABITATS: SOME GENERAL

EFFECTS

Temporary or rapidly changing terrestrial habitats in the tropics will present

very different types of selection pressures from more stable habitats. For exam-
ple, wide oscillations in the abundance of resources for insects might result in
large fluctuations in abundance as a consequence of such temporal environmental
heterogeneity. The study of Irmler and Furch (426) on cockroach popUlations in an
Amazonian inundation forest illustrates the dependence of scavenger insects
occupying such changing habitats on more than one kind of food source, namely,
decomposing leaves and decomposing animal tissues. Both kinds of resources are
necessary for juvenile roaches to receive the correct balance of nutrients, but both
are expected to fluctuate in abundance with flooding cycles.
In some temporary habitats such as a crop field, large popUlations of rela-
tively few species of phytophagous insects might be exploited by several species of
parasitic insects. The insects feeding on the crops are derived from the local fauna
associated with either other kinds of disturbed habitats or undisturbed habitats.
The degree of parasitism present in populations of a particular species will depend
in part upon the kind of habitat where the population occurs. For example, the
degree of parasitism within a population in an undisturbed habitat may be consid-
erably lower since a parasite species such as a tachinid fly is expected to be a gen-
eralist on many different host species within relatively small patches offorest. The
biomass per host species is low per unit area of habitat because each host plant
species is hyperdispersed; the result is a great diversity of potential host insects
accumulating over small patches offorest. The greater the disruption of the habitat,
 TABLE 5.6
Abundances of Adult Cicadas on Isolated and Aggregated Individuals of Various Tree Species in
Costa Rican Tropical Forests a ••

Number of Range of
Season of Number of Number of Total Mean number tree number of Total Mean number
peak adult single days of number of ofcicadas aggregates trees in Number of days number of ofcicadas
Tree activity trees observation cicadas (X ± S.D.) observed aggregates of observation cicadas (X ± S.D.)

F. sericans
Pourouma aspera
(Moraceae) Dry 37 13 103 2 ± 2.4 12 5-20 10 650 51 ± 10.5
Goethalsia meiantha
(Tiliaceae) Dry 10 18 2 ± 1.6 54 23 735 30 ± 8.6

Z. smaragdina
Wet 10 8 27 3 ± 0.6 54 52 1483 31 ± 11. 3
Pentaclethra macro/aba
(Leguminosae) Wet 32 27 138 5 ± 7.4 6-11 20 753 35 ± 9.6

F. pranoe
Vernonia patens
(Compositae) Dry 12 10 23 2 ± 3.2 4 5- 30 10 420 38 ± 11.4
 Q. gigas
Dry 12 10 11 1 ± 0.5 4 5-30 10 184 17 ± 8.3

F. pronoe
Gliridicia sp.
(Leguminosae) Dry 19 8 15 1 ± 1.4 35 8 512 60 ± 7.5

Diceroprocta sp.
Tamarindus indica
(Leguminosae) Wet 10 6 10 1 ± 0.4 28 5 105 19 ± 5.3

Q. gigas
Zygia latifolia
(Leguminosae) Dry 4 14 0 0 22 17 2530 124 ± 15.6

Z. smaragdula
Wet 4 12 0 0 22 15 568 37 ± 10.2

aFrom Young.'21"
·With the exception of the last two tree species in this table, all observations were made at "Finca La Tirimbina," near La Virgen de Sarapiqui, Heredia Province, in the premontane
tropical wet forest zone. The observations for T. indica were made at Playas del Coco, Guanacaste Province, and for Z. lati/olia at San Rafael de Ojo de Agua, San Jose Province.
210 CHAPTER 5

going from advanced secondary to monoculture or simple polyculture (two to

three) crop arrangements, the greater the number of parasite species the habitat
can support per unit area. In disturbed habitats, parasites are therefore expected
to be specialists, since the biomass (harvestable productivity) of the few host
insect species present is large enough to support breeding populations of several
parasite species. There might be considerable niche overlap among parasite
species under such conditions since the resource base is large enough to be ex-
ploited by several parasites. This argument is developed for spatial changes in
the environment, in that different kinds of habitats will have different species
composition and abundance of phytophagous insects and such variations will
also be manifested in the diversity and abundances of parasitic insects attacking
these herbivores. Miller(427) developed a similar argument based upon temporal
changes in habitats in a Temperate Zone environment.
Depending upon the trophic level considered, highly disturbed habitats in
the tropics may contain specialist or generalist insect species. As discussed
above, the lower abundance of some plant groups and larger patch sizes per plant
species expected in some disturbed habitats may select for· specialist feeding in
many phytophagous insects, assuming the ability to locate suitable patches and
the ability to adjust evolutionarily population growth parameters (see Chapter 6)
to prevent overexploitation of harvestable productivity while maintaining breed-
ing populations. But if other kinds of resources are patchy and patch sizes are
small, the generalist strategy may be optimal for some insects. For example,
predatory insects and opportunistically omnivorous insects such as carabid bee-
tles may be generalists in some disturbed habitats. In terms of prey actively
sought by predatory insects, the patch structure of these resources is defined
largely by size ranges of the prey and their distributions in the habitat. It is
therefore difficult to generalize about the relative frequencies of generalist and
specialist insects in tropical habitats without considering trophic level and distri-
bution and abundance of preferred resources.
The degree of specialization, of course, will set the limits on the amount of
flexibility there is among individuals in a popUlation of a particular species for
the occupancy of more than one kind of habitat. The foraging habits of many
social Hymenoptera probably involve the retrieval of prey from more than one
kind of habitat. It was discussed in Section 5.2 that the great diversity of ant
species in the tropics is due in part to the exploitation of arboreal nesting sites,
not to mention the great horizontal diversity of available microhabitats in some
habitats. Ant species that have entered into specialized mutualisms with other
orgnisms are expected to have definite patterns of habitat selection, while ant
species with flexible nesting and foraging requirements may occupy more than
one kind of habitat in an area. Leaf-cutting ants may nest in one kind of habitat,
forage there, and also forage in adjacent habitats. The extreme in habitat flexibil-
ity is attained in some species of army ants, whose raiding columns and bivouac
DISTRIBUTION PATTERNS OF INSECTS 211

locations can be found in different habitats. The degree of raiding in army ant
colonies varies directly with the amount of food intake, but emigration rate does
not appear to fluctuate under different food intake regimes. The food-gathering
component of emigration behavior in army ants may not be the major functional
role of such movement patterns in these insects.
Foraging leaf-cutter ants retrieve leaf material from many tree species over a
large area encompassing more than one habitat type (Table 5.7). Foraging in
leaf-cutter ants also has a strong diurnal component (Fig. 5.12), workers leave
nests late in the day and return later at night with booty. Young and Hermann(212) \
found a similar foraging pattern for the giant ponerine ant, Paraponera clavata,
involving both prey and various plant materials as booty. The marked vertical
stratification of the tropical wet forest habitat may result in different resources
being spatially separated for a particular species. A good example is P. clavata,

TABLE 5.7
The Distribution and Abundance of Plant Species as Indicated by Leaf
Sampling, Encompassing an Areas of about 1200 m 2 around the
Leaf-Cutter Ant Nest Studied U

Number Percent
Species recorded frequency

Marlierea montana (Aubl.) Amsh. 87 21.2

Terminalia amazonica (1. F. Gmell) Exell. 26 6.3
Eperua falcata Aubl. 16 3.9
Pithecellobium eperuelorum Sandw. 16 3.9
Dcolea schomburgkiana (Nees) Bth. 13 3.2
Calycolpus glaber (Bth.) Berg. 10 2.4
Ecclinusa psilophylla Sandw. JO 2.4
Licania buxifolia Sandw. 10 2.4
Chondodendron candicans (Rich) Sandw. 9 2.2
Catostemma fragrans Bth. 8 1.9
lschnosiphon gracilis Koern 8 1.9
Emmotum fagifolium Desv. 7 1.7
Cassia pteridophylla Sandw. 6 1.5
Eperua grandijlora (Aubl.) Bth. 6 1.5
Anaxocorea peliolata R. E. Fr. 5 1.2
Myrcia silvatica (Mey.) D.C. 5 1.2
Tovomila cephalosligma Vesque 5 1.2
Calycolpus pyrifer Riley 4 1.0
Dugelia inconspicua Sage!. 4 1.0
Pouleria filipes Eyma. 4 1.0
Tapirira guianensis Aubl. 4 1.0
Simaba cedron Planch. 3 0.7

( continued)
212 CHAPTER 5

TABLE 5.7 (Continued)

Number Percent
Species recorded frequency

Talisia squarrosa Radlk. 3 0.7

Vismia guianensis DC. 3 0.7
Alexa leiopetala Sandw. 2 0.5
Aspidosperma excelsum Bth. 2 0.5
Casearia densijlora Bth. 2 0.5
Catostemma altsoni Sandw. 2 0.5
Connarus lambertii (D.C.) Sagot. 2 0.5
Duroia eriopila L.F. 2 0.5
Eschweilera corrugata (Poit.) Miers. 2 0.5
Matayba opaca Radlk. 2 0.5
Neea constricta Spruce ex Schmidt 2 0.5
Oxandra asbecki (Pulle) R.E.Fr. 2 0.5
Ryania speciosa var. subulijlora (Sandw.) 2 0.5
Monachino.
Simaba multijlora A. Juss. 2 0.5
Tabebuia capitata (Bur. & K. Schum) Sandw. 2 0.5
Tachigalia rusbyi Harms. 2 0.5
Unonopsis guatterioides (A.DC.) R.E.Fr. 2 0.5
Xylopia benthamii R.E.Fr. 2 0.5
Byrsonima incarnata Sand. 0.2
Clusiafockeana Mig. 0.2
Cupania scrobiculata L.C.Rich 0.2
Cybianthus brownii Gleas. 0.2
Eugenia essequiboensis Sandw. 0.2
Guatteria schomburg kiana Mart. 0.2
Himatanthus bracteatus (D.C.) Woodson 0.2
Hirtella davisii Sandw. 0.2
Maprounea guianensis Aubl. 0.2
Marlierea schomburgkiana Berg. 0.2
Miconia guianensis (Aubl.) Cogn. 0.2
Nectandra grandis (Mez.) Kosterm. 0.2
Pouteria cladantha Sandw. 0.2
Pouteria egregia Sandw. 0.2
Pouteria engleri Eyma 0.2
Pouteria jenmani (Pittier) Sandw. 0.2
Rinorea f!avescens (Aubl.) Kze. 0.2
Smilax schomburgkiana Kth. 0.2
Tapirira marchandii Engl. 0.2
Toulicia pulvinata Radlk. 0.2
Weigeltia surinamensis (Spr. f.) Mez. 0.2
Lianes (unidentified) 51 12.4
Unidentified 34 8.3
61 species TOTAL 410

aFram Cherrett.(S81)
DISTRIBUTION PATTERNS OF INSECTS 213

--
Light intensity (lumens/ft 2 approximately)
>400 40-400 08-40 08 or less No reading-dark
Air temperature % relative humidity
'1.- - --I( o C; () a •
_'1.- -'1.--'1.- -x- -'1.- -x- -)(- -'1.100
_x- _x- _x- _x- >-
90 ~
E
80 E
m
70 .~
60 ~
"if'
50

.~ 500
M
c
-- In laden Out
x- - --x
In empty
0········0

-1 400
"~ 300

I
I
I
x
/'

!p~.... ....0·

OL-L-~-L~~~~~-L~~__L-~-L~~~~~-L~~~~~~~
7 8 9 10 11 12 1 5 6 7 8 9 10 11 12 1 2 3 4 5 6 7
Noon Midnight
TIme

FIGURE 5.12. The daily cycle of foraging activity in a Neotropicalleaf-cutter ant, Alta cephalotes.
and associated climatic data. [From J. M. Cherrett, J. Anim. Ecol. 37:387-403 (1968). Copyright
1968 by the British Ecological Society.J

a species that nests beneath certain species of canopy-size trees but which con-
centrates most of its foraging in the canopy and does so with a marked diurnal
cycle of activity. (212) In a study of the nesting and foraging of the Neotropical ant
Zacryptocerus varians, Wilson(428) hypothesized that arboreal nesting resulted
from competition with ground-dwelling ants and scavenging resulted from com-
petition with other arboreal predaceous ants. Wilson was able to associate other
behavioral (and specialized) traits with these aspects of habitat utilization. Got-
wald(429) showed that the diet composition in three species of African driver ants
was determined primarily by habitat, with the major clases of prey being insects,
arachnids, and earthworms. The relative abundance and accessibility of particu-
lar prey types also determined diets(430,431) (Fig. 5.13).
Carroll and Janzed l97 ) report that the abundance of soft wood stems is a
limiting factor in the abundance of arboreal nesting ant species in tropical de-
ciduous forests. They suggest that the abundance of ant nests in tropical forests
 214 CHAPTER 5

70
GHANA: HABITATS KENYA: HABITATS
60
'"f-Z _ Coastal scrub and
grassland
'"f-Z _ Central highlands (woodland)
::> 50 _ . Kakamega Forest
nMlllllnu Guinea savanna- ::>
>- . . Savanna
w woodland
>-
w
a: 40 a:
...0
Q.

30
...o
Q.

f- f-
Z Z
w w
u 20 U
a: a:
w w
Q. 10 Q.

0
~ l1!

i
~ 0
.c -l3 -l30 -l30
:>
~ ~ ·c e
Ii>=~ :> 0;
c
-" "
.2 "
0
'is :.c 0. ~
-"
f-
0
U :;; :t ~ u U is E

"Preyless"
60 70.61 workers
'"Z
f-
50
KENYA: HABITATS
_ Central highlands ,.
-~
::> (woodland) I ~
z
>-
w 40
a:
_ _ Kakamega Forest
Savanna I ...«o
...0 I
Q.
r'I'.I~
30 a:
w
al
f- :;;
z 20 ::>
w
U z
a: 10
w
Q.

] E.E E
E
i !lE !lE E. E
E
.!! i E. "E "e "g .c. !l
0 0 0
"C

8 "€0 .§l" :t :t Z" "0U j is"

.~ :>
0
HEAD WIDTH, mm

FIGURE 5.13. Taxonomic composition of prey taken by African driver ants, Dorylus (Anomma)
species in Ghana and Kenya habitats, and the frequency distributions of head widths of "preyless"
workers and prey-carrying workers of D. (A.) nigricans var. molesta. [From W. H. Gotwald, Jr.,
Ann. Entomol. Soc. Am. 67:877-886 (1974). Copyright 1974 by the Entomological Society of
America.]

will vary directly with the harvestable productivity of the forest patch available
for colonization. Harvestable productivity will be a function of the kinds and
distributions of potential resources in the patch and the amount of competition for
the resources with other ant species. Carroll and Janzen(l97) suggest that small
colony size in tropical ants may result not only as an adaptation for nesting in
small cavities but also as a niche contraction resulting in food specialization
under selection pressure from competitors. Competition for food is believed to be
a major ecological factor determining the local ant diversity in tropical forests.
Studies of ant nest distributions in Temperate Zone grasslands (e.g., YasunO<432»
indicate that intraspecific competition for nesting sites accounts for considerable
spacing in a rather uniform habitat. Similar patterns may exist for ant species in
DISTRIBUTION PATTERNS OF INSECTS 215

tropical grasslands if colonies are very large. But the more heterogeneous nature
of most other terrestrial habitats in the tropics probably reduces the possible role
of intraspecific competition for both nesting sites and food to nothing.
Other social Hymenoptera may exhibit specific habitat associations based
largely upon the kinds of food found there. Wille(21) discusses the specialized
foraging behavior of certain species of Trigona bees for gathering pollen from
Cassia trees. The distribution and abundance of Cassia trees among habitats at a
locality will contribute a major portion of the grain structure of the food envi-
ronment for these bees. Most species of Trigona are fairly opportunistic foragers
at many different plant species in tropical forests, and collectively these bees
probably acocunt for a very large percentage of the pollination in lowland tropi-
cal rain forests such as found on the Osa Peninsula of Costa Rica (A. Wille,
personal communication). Trigona bees probably fit Colwell's definition of in-
terstitial species in the sense that they forage on a variety of plant species in
habitats where patch sizes are relatively small and each plant species is highly
dispersed in the habitat. Many species of Trigona bees collect pollen from
tubular anthers of many different plant species by biting through the anther and
releasing the pollen grains. The grains are picked up with the proboscis, then
gathered on the thoracic sternum, from which it is transferred to the corbicula. A.
M. Young(677a) has found that at least two species of Trigona exhibit this form of
pollen-collecting for the introduced ornamental vine Thunbergia grandiflora
(Acanthaceae) in Costa Rica. The bees also rob nectar from this vine by biting
through the base of the tubular corollas before hummingbirds visit near dusk, an
activity that probably reduces pollination frequency considerably. In Costa Rica
the vines are propogated only by cutting, as there is no fruit set.
Different species or groups of species of Trigona bees exhibit various forms
of pollen-collecting behavior, and some species are capable of changing this
behavior for the exploitation of different flowers with different anther structures.
It is perhaps this very flexibility in feeding that has resulted in the evolution of
aggressive behavior among species of Trigona occupying the same habitats.
Johnson and Hubbe1J<433) found that several species of these bees exhibit both
intraspecific and interspecific aggressive behavior toward flowers and artificial
baits that result in considerable lowering of the amount of time bees spend on
flowers. Johnson and Hubbell(433) found that aggression increases directly with
increases in the sugar concentrations of artificial baits. Such behavior reduces the
amount of nectar or pollen that bees gather on individual visits to flowers in the
wild, and results in a spatial partitioning of the habitat so as to reduce aggres-
sive encounters. The ability of many species to exploit the same species of
flowers promotes the aggressive-behavior syndromes as a means of enforcing
spatial separation of foraging areas among different species or different colonies
of a single species in a habitat. Such aggression is expected to increase in habitats
216 CHAPTER 5

where many bee species occur and where the diversity of flowers is relatively low
through time, or where the few flowering species available at anyone time have
small patch structures and therefore low, harvestable productivity for bees.
Because pollen is presumably rich in a variety of nutrients whose concen-
trations and types vary considerably among different plant taxa in the tropics,
pollen itself is a critical resource for many insects. As studies of flower-visiting
insects in the tropics continue, elucidating the patterns of exploitation of pollen
by many insects other than bees will be possible. Gilbert(31) has determined that
some species of Heliconius butterflies feed on pollen, which in tum acts as a
source of amino acids and proteins used in the production of eggs by the but-
terflies. That female ceratopogonid midges visit cacao flowers suggests that these
insects also use pollen as a source of amino acids for eggs . We must also examine
the kinds of selection pressures operating in such species to promote the expres-
sion of pollen feeding and the accompanying physiological mechanisms involved
in the assimilation of pollen amino acids into egg proteins in these insects.
Gilbert,(31) for example, argues that pollen feeding evolved in some Heliconius
to counteract the heavy predation on immature stages in populations, the net
result being that fecundity is increased evolutionarily by such a mechanism. H.
G. Baker and I. Baket216 ,217) have emphasized the coevolution of pollinators and
plants on the basis of pollen and nectar nutrient concentrations, and it is neces-
sary to examine such relationships for a variety of pollinating insects, including
bees, lepidopterans, beetles, and flies. Specialized forms of pollen feeding may
involve insects utilizing only certain plant species, and such interactions may
result in considerable habitat selection. If a particular pollen source is found only
in a certain kind of habitat, the insect exhibits a preference for that habitat. One
might expect that specialized pollen-feeding associations would be most frequent
in rather stable habitats where such resources may be most predictable over long
periods of time. The strategies of flower foraging developed by J anzen(40) are the
kind that might involve specialized pollen feeders. Less stable or changing
habitats may select for flower visitors that are not specialists on certain species.
Temporal and spatial predictability of flowers for each plant species seems to be
a requirement for the evolution of specialized flower-visiting strategies in in-
sects, and some habitats might be better for this than others.
 CHAPTER 6

POPULATION RESPONSES
TO THE ENVIRONMENT IN
TROPICAL INSECTS

The association of an insect species with a particular habitat means that the
species maintains a breeding population in that habitat. We therefore need to
distinguish between resident and nonresident species associated with different
habitats. The great differences in the taxonomic composition and structure of
vegetation in disturbed and undisturbed habitats in tropical regions, discussed in
detail by Holdridge et al., (396) indicate that the composition of local insect faunas
may differ greatly between habitats in the same region, a phenomenon well
documented by the studies of Janzen and Schoener(134) and Janzen.(72) Thus
different collections or assemblages of insects and other vegetation-dwelling
arthropods are expected for different kinds of terrestrial habitats in tropical re-
gions.

6.1. ADAPTATION TO HABITATS

In order to understand how different habitats may support species-rich as-

semblages of insects in the tropics, and the differences in species abundances
among different habitats in a tropical region, it is necessary to consider how the
breeding popUlations of individual species "fit in" to the habitat. Natural selec-
tion determines various kinds of so-called life history patterns of organisms as a
function of varying environments. Life history refers to the age-specific fecun-
dity and mortality schedules and other parameters that go into the formulation of
the growth rate of popUlations. Natural selection is expected to mold these
ecological parameters within popUlations in such a manner as to increase the
likelihood that the species will survive in a particular habitat or environment.
Because the phenotype is a compromise, there is always a tradeoff in the alloca-
tions of resources to different traits such as vegetative maintenance, reproduc-
tion, feeding, and defense against biotic mortality factors. The patterns of re-
source allocation among these and other phenotypic traits affecting population

217
218 CHAPTER 6

growth will vary in different kinds of environments with different forms of

selection pressure. (434) Natural selection favors increases in fecundity, defined
here either as the age-specific schedule of births (eggs) or the total number of
births in an individual's lifetime, adult survival (from biotic and abiotic environ-
mental factors), and juvenile survival (egg to age of first reproduction for most
species). But there might be compensatory tradeoffs between these ecological
parameters, with an increase in one through evolutionary time being accom-
panied by a decrease in another during the same period. Again, the phenotype is
a compromise in terms of total resources available for various functions related
directly to fitness.
Shifts in the environment may change selection pressures, which in turn
alter the resource allocation profiles for individual genotypes in the popUlation.
Selection pressures affecting fitness parameters associated with population main-
tenance and growth will include fluctuations in food availability and accessibility,
fluctuations in the physical environment, shifts in the abundance and density of
competing organisms, and shifts in the intensity of predation and parasitism.
Shifts in genotype frequency resulting from emigrations and large fluctuations in
population size will at the same time alter the degree of genetic and phenotypic
resiliency in the popUlation to respond to these selection pressures. In both the
lowland seasonal and nonseasonal tropics, changes in food availability, abun-
dance of competing species, and changes in the intensity of predation or
parasitism are probably the major classes of selection pressures affecting the
population structures and dynamics of insect species in disturbed and undis-
turbed habitats. At the high altitude (above mid-elevation) regions ofthe tropics,
fluctuations in the physical environment, with manifold effects on resource
availability, may be the major form of selection pressure. Changes in fecun-
dity, developmental· time, and longevity in response to severe fluctuations
in the physical environment, including long-term unfavorable growing condi-
tions, may be very different from the changes needed to maintain a population
due to shifts in competition or predation. I use the term maintain in this context
to mean "keep growing or remaining stable in size in a particular habitat. "
Lewontid 87 ) discussed the differences in various kinds of tradeoffs arising
from shifts in certain populational parameters over others for species exhibiting
colonizing episodes. Thus a tradeoff between adult survival and increased fecun-
dity may depend on the gain in fecundity per unit of adult survival lost, coupled
with the direction of selection. Lewontin(87) showed, using data from
Drosophila, that about a 10% increase in developmental time is the same as a
100% increase in fecundity, implying that selection pressures would favor the
former phenotypic change in increasing the growth rate of the population in a
new environment. Selection is expected to favor earlier reproduction for a
species invading a new habitat, but eventually popUlation growth rate might
POPULATION RESPONSES TO THE ENVIRONMENT 219

decrease as the species becomes integrated into the community, and the shift
might then be toward postponing reproduction. MacArthur and Wilson(341)
pointed out that the relative value (increase in fitness) of decreasing developmen-
tal time and increasing fecundity depended upon the net reproductive rate (Ro),
since that latter parameter is an estimate of the total average fecundity of indi-
viduals in the population. For example, if Ro is relatively low within a certain
population, the advantage for increasing fecundity would not be as great as for a
higher value of Ro in another population, or within the same population at
different time. But in some stable populations increases in adult survival are the
optimal solution to increasing fitness.
Increasing fecundity or decreasing developmental time appear to be most
adaptive in populations in the early phases of population growth. Thus, depend-
ing upon the stability of the population, selection will act to increase fitness by
affecting different parameters underlying population growth. The avenue of re-
sponse embarked upon by the species will be a function of the type of habitat
occupied. We need to distinguish between relatively stable and unstable tropical
habitats and the unique assemblages of insect species found in each of them. It is
not at all clear whether primary forests in the lowland tropical regions are stable
from the standpoint of turnover in floras at the local (habitat) level. The data
seem to indicate the opposite.(60.435) Yet certain kinds of secondary tropical
habitats may be incredibly stable, with little or no turnover in plant species
composition.
From the viewpoint of phytophagous insects and the insects that feed on
them, shifts in plant species in some tropical habitats are a form of selec-
tion pressure perhaps favoring colonization to more suitable areas. The total
insect faunas may change under such conditions, but so may the population
dynamics of individual component species. Some tropical insects may in fact
consist of shifting mosaics of plant species and associated insect species (and
their parasitoids), while others may consist of stable elements. Shifting condi-
tions are expected to result in shifts in the direction of selection affecting the
parameters of popUlation growth.

6.2. DEMOGRAPHY AND FITNESS IN POPULATIONS:

IMPLICATIONS FOR TROPICAL INSECTS

The basic theory of individual selection is Fisher's Fundamental Theorem of

Natural Selection, which states that the rate of change of fitness within a popula-
tion is equal to the genetic variance in fitness. The rate of evolution is therefore
limited by the variation in fitness of the units being selected. The rapidity of
response to selection within a popUlation of a species depends upon the her-
220 CHAPTER 6

itability of differences in fitness between the units selected. The rate of evolution
depends upon the rate of reproduction and mortality. Fitness, defined for the
individual in a population, has units of time, as seen from Fisher's Theorem:

dW
dt = -
Var (W), so that dt (dW)
fV2 and t (1)
W

where W is the mean fitness and Var (W) is the variance in fitness.
A species with a very long life cycle, such as some cicadas, cannot evolve
as fast as one with a short life cycle, and the intensity of selection for a unit is
directly related to the cycle time for the unit. Thus populational selection must be
generally weaker than individual selection since births and deaths of whole
populations encompass many individual lifetimes. But it is possible that popula-
tions and communities are also units of selection. The interactions of individuals
within a population of a single species, and the interactions among individuals of
different species and with the physical environment, result in forms of selection
pressure not found at the level of the individual. In the tropics, where biotic
interactions, often highly specialized, are common in communities with large
numbers of genera and species, selection at the popUlation and community levels
of ecological organization might be particularly important in determining the
structure of communities associated with different habitats in a region. The
present discussion, however, is limited to a consideration of life history traits in
populations.
As discussed in Chapter 2, natural selection and other factors such as drift,
mutation, and recombination can alter the genetic composition of a population.
The frequency distribution of genotypes is changed to maximize the average
fitness in the popUlation, as the result of individual selection. Each genotype in
the population has a fitness value that contributes to the overall growth rate of the
population. The measure of heritability provides an estimate of the amount of
phenotypic variation in a popUlation that arises from genetic variation and from
the distribution of genotypes and their fitnesses. Phenotypic variance is the sum
of genetic variance (V G) and environmental variance (V E). Phenotypic variance
is measured for a given trait in the popUlation. The measure of heritability is
given by

h 2 = VG = VG
Vp VG + VE
and it varies from 1.0 to 0.0. An h 2 value of 1.0 results when all of the variation
in the popUlation for a given trait is caused by differences among genotypes and
no differences result from environmental effects. Clearly, most h 2 values will
reflect a mixture of both effects for many traits. The genetic variance in tum
result from both additive and nonadditive genetic effects, the latter including
POPULATION RESPONSES TO THE ENVIRONMENT 221

dominance and epistasis. Populations are characterized by population growth

parameters that may have a strong genetic component:
Lx the age-specific survivorship schedule of individuals in the population,
from birth until death.
mx the age-specific fecundity schedule of individuals in the popUlation,
from birth until death.
r the intrinsic rate of increase of the population, the growth rate itself,
computed from the Lx, mx tables, initial size of the population, and
developmental time.
Ro the net reproductive rate of the population, the total number of off-
spring produced by the average female during her lifetime.
T the generation time of the population.
The size, N, of the population is then described by the following formulation,
which brings together some of the above ecological parameters:

1 = ~e-r.x·Lxmx
x=o

It is important to realize that these parameters are determined in part by the

distribution of genotypes in the popUlation and that an optimal strategy, based on
the assumption of individual selection, is to maximize the average fitness of the
popUlation. This should be reflected in the maximization of growth rate of the
population. The differential fitness of genotypes within a population and the
variations in environments account for differences in population growth rates.
Lewontin and White(S4) examined these considerations for populations of the
Australian grasshopper Maraba. Different populations of the same species will
maximize fitness in different ways, depending upon the initial genetic composi-
tion of the population and differential selection pressures affecting components
of fitness such as popUlation growth parameters.
The evolutionary responses made by species to the environment occur by
the following mechanisms: (1) adaptive responses within the population; (2)
differences in adaptive responses between populations of the same species; (3)
differences in adaptive responses among closely related species in the same
habitat or region, or in different habitats or regions; (4) adaptive responses
arising from interactions among unrelated species; (5) responses to changes in the
physical features of the environment. Depending upon the kind and direction of
selection pressures in the environment, a population will either increase, de-
crease, or remain the same. As the result of both laboratory and field censuses of
animal populations, populations exhibit different kinds of census histories, re-
222 CHAPTER 6

fleeting different causal mechanisms (Fig. 6.1). No single population is expected

to be governed by one of these mechanisms; rather the interplay of different
factors usually determine the census history of the popUlation. But census his-
tories are expected to vary considerably with habitat, region, and period for a
given species.
The expression dNIdt = rN describes optimal growth of a population with

I. INSECTS
Suppression by
disease and
emigration
02 Carrying capacity of environment
_._._._._._._._.
determined by food Milne's 'Zone III'

! .. 1
.
1 ... . ~ !1 Suppression by
predator and
!
01 Upper limit of
normal density

Emigration

_
03 Density level where social tension and 'shock-

-..... - ...............................................................................
causes sudden decrease of animal

Epidemic of
- .. ......... .
Fail of
Infectious pathogene
02 Carrying capacity disease

III. RODENTS IN 'CULTIVATION STEPPE'

03 Density level where social tension

causes sudden decrease of animal

FIGURE 6.1. A summary of the various mechanisms of census history fluctuations in three distinct
types of animal popUlations, including insects. Solid lines show changes in population density, and
broken lines indicate census histories of predators (natural enemies). The solid arrows indicate
climatic changes favoring increase in the animal (host) populations, and the dashed arrows show the
effects of changes decreasing population densities. [From Y. Ito, Bull. Natl. Inst. Agric. Sci. Japan
Ser. C, No. 13 (1961).]
POPULATION RESPONSES TO THE ENVIRONMENT 223

unlimited resources. Yet we know that few if any populations can grow for any
appreciable length of time under such ideal conditions. If we were to examine
one parameter of population growth, such as the age-specific survivorship
schedule I x, various kinds of coactions (defined here as interactions between two
or more unrelated organisms) alter the I x schedule from what has been an ideal or
"physiological" condition, reflecting the intrinsic capacity of the species to
survive under ideal environmental conditions (in the absence of coactions and
deleterious effects of the physical environment). The five major kinds of coac-
tions, namely, plant-herbivore, prey-predator, host-parasite, host-pathogen,
and mutualist-mutualist, endow the population with a pattern of "ecological
longevity" considerably lower than that resulting from the physiological Ix
schedule (Fig. 6.2).
It is important to realize that the length of time a population of a particular
species has occupied a certain habitat will determine to a large degree the level of
adaptation to the environment. For example, a species in the early stages of
arrival into a new habitat may still possess ecological traits that adapted the
species to its previous environment. Certain phenotypic traits, evolved in re-
sponse to the previous environment, can influence the trajectory of adaptation in
the new environment. An herbivorous insect that feeds on certain Solanaceae in

I
/
I
/ /
I I
I I I

I
I
I
/ I
I
I
I
/ / /
I
/
/"
I
I
I I
I
/ I
I ~ PHYSIOLOGICAL
I I I
I LONGEVITY
/
I I
W I

/
I I
> I I
I
I

" / "
I I
::::i I I I
< I I I
I I I
/
I I
/ I I
I /

" /" "

I I
I I
Iii I
I
I
I
I
I

~
I I
I I

""zw
I
/
ECOLOGICAL ~ I
I
I
/

"
I
0 LONGEVITY
a:
w
D.

TIME

FIGURE 6.2. A hypothetical representation of the effects of ecological processes on the survivor-
ship curve of an insect species in a particular habitat. Individuals in such a popUlation are believed to
be genetically and physiologically programmed for some optimal pattern of survival (upper curve),
but various ecological interactions (e.g., coactions) in nature actually lower survival in different age
classes (lower curve). The area between the two curves approximates the actual losses in harvestable
biomass from the population as the result of ecological processes (e.g., density-related intraspecific
competition, predation, etc.).
224 CHAPTER 6

the present environment might in fact have developed a feeding relationship with
these plants in the previous environment. Thus the coevolved association is
uncharacteristic of the present environment. (436) Similarly, parameter character
states determining the rate of population growth in the present environment may
still be those carried over from the previous environment.
Species involved in colonizing episodes may have genotypes in the previous
environment that allowed the colonizing episode to occur, and some of these
traits may still be present in the new habitat. In the new habitat the species might
not have occupied all resource patches for the spectrum of resources it is capable
of exploiting. A reason for this could be insufficient time to fulfill population
increase to the point where all resource patches are occupied. When this occurs,
the popUlation will no longer be able to increase, and strong selection might favor
an alteration in the parameters underlying popUlation growth. There might be an
evolutionary reduction in r. Other adjustments in r may result from resources
being in short supply, or being inaccessible as the result of competition being
intense, or being different in some way that precludes their penetration by a
newly arrived species. For example, the chemical defenses of the host plants may
be sufficiently different from the phenotypes of the same host plant species in the
previous environment so as to block attack by the herbivorous insect associated
with plant in the previous environment. In the lowland tropics, insect popUlations
are believed to be regulated largely by various biotic factors, most notably
(1) limited supplies of resources, (2) intraspecific competition resulting from (1),
and (3) intense predation and parasitism that keeps populations below the carry-
ing capacities.
The ecological processes under (1) and (2) above are expected to be density
dependent. Insect species in high-altitude regions are expected to be largely
regulated by density-independent factors such as large fluctuations in the physi-
cal environment. But to some extent all popUlations will be affected by both
classes of factors in the environment. For example, most insect species in the
temperate zones are known to experience rather high mortality in the juvenile
stages (Fig. 6.3), so that survivorship curves for many insects reflect generally
low recruitment potentials for adults. Similar patterns are expected for the
tropics, although far fewer data are available. The high density and diversity of
predatory and parasitic insects and other arthropods in many tropical habitats,
particularly in lowland regions, suggest that mortality on juvenile stages is prob-
ably very high in many species of insects.
Coactions such as prey-predator and host-parasite are expected to be fre-
quent in the tropical lowlands, where the abundance and diversity of these insects
are high. Thus mortality and natality, in the forms of the Ix and mx schedules for
new cohorts in the population of a species in a particular habitat, will determine
to a large degree the size and density of the popUlation in that habitat. But
dispersal, the shift in popUlation numbers and distribution through movement of
POPULATION RESPONSES TO THE ENVIRONMENT 225

1000
f ....Jl-
I ""'IVV~ E L p. EL P ~ ELC PP-P~
MOROCCAN GREEN
LOCUST TORTRIX
Tort;rix
vlridana

HONEYBEE
___ Apis
~melifera
.WOrker)
,,
,,,,
,
I
II

-
I

Pt E L PP-PtA ELPt 20304050

days

FIGURE 6.3. Examples of survivorship curves determined empirically for a variety of insects, with
analysis by stages of the life cycle. In the case of Ips bark beetle, data are given for two different
populations. The letters on the abscissa refer to immature stages (E=egg, N-nymphal instars,
L=larva, PP=prepupa, P=pupa, and C=cocoon); the ordinate is in number of individuals. [Redrawn
and modified from Y. Ito, Bull. Nat!. Inst. Agric. Sci. Japan, Ser. C, No. 14 (1961).]

the population's membership, will also contribute to the size of the population at
anyone point in time. Populations may select for dispersal-prone genotypes in
response to increased population densities, changes in the habitat resulting in
reductions of resources, and the sudden availability of new habitats suitable for
occupancy by the species. Such dispersal functions as a flush mechanism that
promotes escape in space to other suitable habitats, either permanently or tem-
porarily (as in the case of some butterflies and dragonflies). Dispersion refers to
the distribution of individuals in the population at anyone point in time, and
some forms of dispersal concern adjustments in the pattern of dispersion within
that population rather than a shift of the population, or a portion of it, to a new
habitat.
Terrestrial tropical habitats with many plant species and small average patch
size for each species will result in specialized phytophagous insect species also
having patchy distributions, which is a pattern of dispersion resulting from
within-habitat dispersal to certain plant patches. The sudden appearance of a
226 CHAPTER 6

forest light gap, or regeneration of secondary forest, may select for dispersal to
new habitats in two directions: insect species arriving to colonize new resources
in the new habitat, or species exiting in response to a reduction in suitable
resources for maintaining breeding popUlations. Such events point to the
dynamic state of affairs of tropical environments.
The parameter r requires little introduction, as it is the intrinsic growth rate
of a population and represents a balance between birth and death rates (deter-
mined by the Ix and mx schedule). The parameter K is the carrying capacity of
the environment for the population of a given species, and both rand K are
determined by the genetic composition of the population for a particular envi-
ronment. Different kinds of environments, including many in the tropics, will
select for different patterns of resources allocated to levels of rand K in popula-
tions. The concept of rand K selection was introduced by MacArthur and
Wilson(341) to specify the populational characteristics best suited for species
occupying temporary and permanent habitats and for communities varying in
different levels of complexity (in terms of species composition and trophic struc-
ture). For example, species undergoing colonization episodes into new environ-
ments should allocate resources so as to maximize r, and the genotypes with high
r values are expected to increase in frequency in such a population. High r allows
a population to grow quickly and swamp the habitat in the face of certain levels
of environment uncertainty associated with new habitats. Environmental uncer-
tainty usually results in considerable density-independent mortality of individuals
in the population. The tradeoff for high K and therefore low r adjusts the species
to occupying a position in a community at the time the species becomes inte-
grated into the community. Permanent and somewhat stable communities are
generally associated with permanent environments. In a temporal sense a col-
onizing species passes through a temporary phase of population growth in the
new habitat, but to avoid extinction the species eventually becomes integrated
into the existing communities in the environment.
The "r strategy' '(92) is to also make use of temporary habitats such as a
weedy mud flat in a slow-moving stream. Both in this instance and in colonizing
episodes involving the invasion of permanent habitats, genotypes with high r
capacities are favored by selection. There is less selection for genotypes with
good competitive ability, for both intraspecific and interspecific competition.
Competitive ability is conferred primarily by genotypes in the "K strategy. ' '(92)
This strategy allows the species to adjust to a portion of the habitat and to become
a member of the community. By relaxing selection for high r, energy is freed to
invoke mechanisms that relate to competition and escape from predation and
parasitism. The result is that the population approaches the carrying capacity of
the habitat, whereas the population experiencing selection for high r tends to
remain well below the carrying capacity if both abiotic and some forms of biotic
POPULATION RESPONSES TO THE ENVIRONMENT 227

mortality are intense, which is expected in colonizing episodes and the exploita-
tion of ephemeral habitats.
For plant-feeding tropical insects, the r strategy is predicted to occur as new
habitats are created by natural catastrophes such as mudslides, floodings, tree
falls in forests, and volcanic eruptions, and biotic changes in existing habitats
select for colonizing episodes in some resident species. Some tropical habitats
may be characterized by a flux of existing species being replaced by invading
species, even though the numbers leaving and arriving may not be the same.
Such effects may cause herbivore guilds associated with some plant species in a
particular habitat to shift in composition.
Intense K selection may lead to considerable ecological specialization in
terms of an insect species exploiting certain host plants, but some residual flexi-
bility is expected in order to buffer against selection resulting in niche shifts of
potentially competing species. One might expect essentially K-selected popula-
tions to be a mixture of predominantly K -adjusted genotypes but also some
r-adjusted genotypes.
If a species of lepidopteran is a specialist on a few host plant species in a
particular habitat, and its popUlation is regulated largely by density-dependent·
factors such as predation or parasitism, the popUlation appears to be K -selected.
If, however, the intensity of predation or parasitism shifts away from this insect
species, the butterfly population may increase in size and density as selection
increasingly favors the r-selected genotypes through successive generations.
These prevailing genotypes will be most adapted to density-independent factors,
and as the popUlation increases, the specialist feeding habit results in a rapid
depletion of the host plant popUlation and either the butterfly goes extinct locally
(the population crashes) or the species experiences a colonizing episode to
another habitat or region. Other consequences may also result under such a
scenario. As the butterfly becomes very abundant in the original habitat, it may
become a favorite prey for other predators, perhaps in a different stage of the life
cycle, and a new form of density-dependent regulation may result.
If the butterfly undergoes a colonizing episode to another habitat, the
emptied niche left behind may be filled by a competing species in that commu-
nity, or a new species may arrive from another habitat through colonization. The
result is that the community may continue to have the same number of species, or
one less. Multiplying such effects for many species in tropical communities
illustrates that such assemblages may be continually changing in composition.
From this discussion, it is evident that the basic issue in the population
biology of insects in the tropics distills down to the analysis of how resources and
energy are allocated to different vegetative and reproductive functions in the
individuals making up a breeding population of a species. The problem is com-
plex because the community in which the species is found is continually chang-
228 CHAPTER 6

ing. On the one hand, selection within the population will favor the evolutionary
maximization of fitness either through increasing the capacity for the population
to increase (r selection) or through adjusting the population growth parameters to
maximize adjustments to competitors (K selection). These are two ways of
optimizing the use of the environment. Borrowing from researchers such as
Lewontin,(87l I suggest the two different joint Ix and mx functions shown in Fig.
6.4. For insects in terrestrial tropical communities, the local faunsitic composi-
tion will include species with either strategy. Which one is used will depend
largely upon the degree of genetic and physiological specialization on a set of
resources, the intensity of biotic mortality factors, the intensity of competition,
and the intensity of abiotic mortality factors. If most parasitic flies and wasps are
generalists in the tropics, their impact on phytophagous insect communities may
be patchy; certain species are repeatedly attacked while others escape repeatedly.
This is expected to be the case especially if the parasite uses the host's food plant
as a means of locating the host. (437)
Thus some phytophagous insect species may be more K selected than others
in the same habitat: species escaping from such predation will build up larger
popUlations and therefore be more likely to compete with other species for the
same resources. Intense predation favors the r-selected strategy, since popula-
tions are cropped severely, precluding the' activation of density-dependent and
competitive effects. We clearly need to explore the comparative natural history
of insect species exploiting particular plant species in a habitat and to analyze life
table parameters and mortality factors. We must also examine to what extent the

AGE CLASSES

FIGURE 6.4. Reproductive effort as a function of fecundity (age specific m,r curves) weighted by
age specific survivorship schedules (Ix). The function assumes a triangular shape over all age classes
in the population considered, Some environments select for l,rmx curves to be skewed toward
maximal reproduction in young age classes (upper graph), while others select for a postponement of
reproduction until later (lower graph),
POPULATION RESPONSES TO THE ENVIRONMENT 229

population of each host plant species is utilized and to what degree by each
member of the herbivore guild found on each plant species.
When reproductive effort is spread out over many age classes of genotypes
within a population (Fig. 6.4), natural selection acts to adjust the reproductive
effort at each age in such a manner as to maximize the overall fitness of the
population. Colonizing episodes for an insect species into a new environment are
believed to be characterized by genotypes maximizing rate of development to the
adult stage, and higher levels of egg production in younger age-classes within the
adult stage. More predictable environments, with species well integrated into
stable communities, may allocate less energy to early breeding and disperse
reproduction over a greater number of age-classes. Of course, one might expect
selection to favor high reproductive efforts at all age classes capable of reproduc-
ing. But such a pattern of resource allocation may also diminish other compo-
nents of fitness in other stages of an insect's life cycle. Natural selection
functions to balance these cost-benefit effects in the attempt to maximize overall
fitness. Gadgil and Bossert<92) discuss the ramifications of these considerations in
terms of when an organism should maximize reproduction at one point or else
spread it out.
Natural history studies are needed in the tropics to determine the factors
affecting the recmitment of reproductively competent adults to the popUlation of
a species. Populations continually change in space and time, and shifts in
genotype frequency are expected as the result of selection. Different forms of
selection pressure affect different stages in the complex life cycles of insects. We
know little about the impact of mortality factors, resource quality, and impact of
abiotic mortality factors on the population dynamics of any tropical insect. Some
census histories are discussed in a later section of this chapter. For the present,
however, it is necessary to realize that different stages in the life cycle may be
affected by predominantly density-independent or density-dependent factors.
This is true for insect populations in temperate regions as well. For example, Ito
et al. (438) found that fluctuations in the population densities of the moth Hyphan-
tria cunea in Japan were the result of the interplay of both density-independent
and density-dependent factors affecting eggs, larvae, pupae, and adults (Fig. 6.3).
In the insect communities of the lowland tropical regions, popUlation out-
breaks tend to be rather scarce. From studies in the Southern Temperate Zone,
Andrewartha and Birch(439) concluded that popUlation outbreaks were generally
caused by changes in prevailing weather conditions, having little to do with
density-dependent effects of predators, parasites, or competitors. In the lowland
tropics, the annual patterns of rainfall and temperature do tend to be similar over
long periods, suggesting that density-independent regulation of insect popula-
tions may not be prevalent. A very different picture might emerge from long-
term studies in high-altitude regions of the tropics. But cooler temperatures at
high altitudes generally reduce insect diversity too. Population regulation of
230 CHAPTER 6

insects in the lowlands, if it exists, is expected to result primarily from (1)

restrictions in accessibility of resources either because they are very patchy or
because competitors mobilize portions of them (i.e., resource limitation), or (2)
intense predation and parasitism. Both effects may be strongly density dependent.
The predation-parasitism model would involve host species being r-selected,
since their populations would buffer the effects of predators and parasites through
increased reproduction. Such an effect is brought about most efficiently by an
evolutionary reduction in developmental time or secondarily so by increasing
fecundity.(87) The assumption is made here that the predation or parasitism is
indeed intense to the point that increasing the host populations increases the
mortality.
Evolutionarily increasing r may result in a satiation of predators and para-
sites so that a breeding population of the host species can be maintained in that
habitat. If, however, predation or parasitism is low or weak in intensity, its effect
may be density independent and have little regulatory impact on the insect
population. Under such conditions, other forms of density-dependent regulation
such as resource limitation and competition may take over in regulating the
population. The latter situation, of course, promotes the K-selection strategy.
Few data exist on the impact of predator and parasites on insect populations in the
tropics. The concept of biotic control of host popUlations in the tropics, as
opposed to control from resource limitation and competition, comes from the
studies on seeds and seed predators. (29,65,435) In lowland seasonal regions of the
tropics, the dry season may be a time of considerably density-independent
mortality on insect populations, especially if the beginning ofthe dry season, and
first few months, is an unpredictable event in terms of the synchronization of life
cycles. The sweep-sample studies of Janzen and Schoener l34 ) indicate that many
insect species drop out of the field communities in lowland Guanacaste Province,
Costa Rica, during the dry season. Such an effect may result in considerable
mortality for the adults of many small insects and cause many small-bodied
parasitic flies and wasps to drop out of the communities for this period. The latter
could relax predation and parasitism on larger insects that remain active in some
habitats during the dry season.

6.3. FURTHER NOTES ON POPULATION REGULATION

We can generalize about the ecological genetics of popUlations by echoing

Cole,(434) who predicted that decreasing the age of first reproduction of indi-
viduals in a popUlation results in increased growth rate, r, of the population. Or
we can be cognizant that populations are affected positively or adversely by
many factors in the environment. Cole's early formulations concerned the statis-
tic rmax, which is of course the familiar intrinsic rate of natural increase or the
POPULATION RESPONSES TO THE ENVIRONMENT 231

popUlation growth rate when all conditions of the environment are ideal. Under
such idealized conditions fitness is maximized through the realization of rmax.
Yet we know that populations are characterized by a realized optimal fitness,
which is generally lower than rmax, if we equate fitness with r. In nature,
popUlations are regulated by various factors, including resource limitation and
predation.(440) In fact, the interplay of factors such as limited food supply and
predation may effectively regulate populations. (440)
In a lowland tropical field or secondary habitat, many dicotyledenous plant
species often show signs of heavy damage by a variety of herbivores, although
seldom is a plant completely defoliated by a single species of herbivorous insect.
Such observations imply that a plant species may be attacked by several unrelated
insects, although densities of the individual species on each host plant individual
are low. Presumably the popUlation of anyone insect species is distributed
among many individuals of that plant species or, in the case of polyphagy,
among individuals of different plant species in the same habitat.
The usual absence of complete defoliation by a single species is the result of
the dispersion of the insect population among different resource (plant) patches
so as to lower the density per patch and predation or parasitism on the insect
population that further lowers the abundance of individuals in each patch. Such a
popUlation system is thought to be r selected in the sense that the insect popula-
tion is regulated primarily by biotic mortality. Depending in part upon the egg
placement habits of the insect species, there may also be some form of intrinsic
control of the population, such as cannibalism among young larvae on each other
or unhatched eggs in a species that deposits eggs in large clusters. Although female
Mechanitis isthmia butterflies carefully place large clusters of eggs on the larval
host plants, the resulting clutches of first-instar larvae are usually reduced by as
much as 20%, even when egg predators other than Mechanitis larvae are pre-
vented from reaching them (A. M. Young, personal observation). In laboratory
popUlations of some Tribolium flour beetles, egg cannibalism by adults is some-
times a form of regulation for the populations.(441) However, in other tropical
insects, such as the butterfly Perrhybris lypera, egg cannibalism is absent even
though the eggs are laid in tight clusters.(77·291) If there is strong selection for egg
clustering in response to some phenotypic feature of the host plant requiring
coordinated group feeding, there should also be selection for ensuring that poten-
tiallosses in fitness arising from cannibalistic activity of early hatchlings or other
life stages present on the plant are minimized. Such a form of population self-
regulation is probably uncommon in insect populations, despite the many exam-
ples of egg clustering and communal larval (or nymphal) activities in the tropics.
Given the high densities of predatory and parasitic insects in the tropics, such as
ants, wasps, and tachinid flies, the early stages in particular of many insects are
probably cropped down by these predators and parasites; such effects are a form
of popUlation regulation. At high altitudes many small predatory insects drop out
232 CHAPTER 6

of the communities, and scavenger, herbivore, and omnivore assemblages of

such regions are probably regulated by limitations in the food supply and by the
rigors of the physical environment.
Comprehensive theories of population dynamics involve both density-
dependent and density-independent factors in the regulation of insect popula-
tions. Some emphasize density dependence in the regulation of populations,
and still others (e.g., Milne(442); Andrewartha and Birch(439) stress that insect
populations were regulated primarily by density-independent factors such as
weather and climate. The favorable physical conditions of the lowland tropical
rain forests may preclude density-independent regulation for insects in such
regions. Density-dependent factors include intraspecific competition, predators,
parasites, and pathogens. While many biotic factors may affect insect popula-
tions in a density-dependent fashion, others, such as generalized predators
and parasites with many alternative hosts, may be density independent. Den-
sity-dependent factors provide a form of natural control on some insect popu-
lations in the sense of being extrinsic mechanisms of population regulation.
In the lowland tropics, the interplay of several different kinds of biotic factors on
insect popUlations, including both density-dependent and density-independent
factors such as intra- and interspecific competition, predation, or parasitism,
probably keep many insect species well below the carrying capacity of the
habitat. Such effects are examples of intrinsic (intraspecific) and extrinsic natural
control of insect popUlations. These biotic factors are responsible for removing
individuals from a population in various stages of the life cycle, generating rather
concave survivorship curves as found in the populations of many different insects
(Fig. 6.3). Because biotic control is expected to operate on many insect species
in different trophic levels in the tropics, outbreaks of insects are not expected to
be frequent in relatively undisturbed to lightly disturbed habitats. In this context,
popUlation out!:>reaks occur either when the effectiveness of the existing biotic
control factors breaks down or when insect species occupy physically controlled
environments(104) where biotic control mechanisms are not prevalent. Much of
the basic research in tropical biology should focus upon the natural history of
control mechanisms in insect populations as related to the kind of habitat and
degree of habitat complexity, community complexity, and fluctuation patterns of
abiotic factors such as temperature. We need to know something about the range
of hosts, for example, of the average species of braconid wasp or tachinid fly in
different habitats. We also need information on the relative roles of different
biotic factors in the regulation of an insect population in different habitats. Such
data are lacking for tropical insects. When we have information on the composi-
tion of a community found in a particular habitat in the tropics, we will need also
to gather data on the following parameters:

1. Resource abundance and accessibility and the degree to which resources

are a limiting factor for a particular insect species being scrutinized.
POPULATION RESPONSES TO THE ENVIRONMENT 233

2. The degree to which the population of that insect is affected by predators

or parasites (and pathogens).
3. The degree to which effects of overcrowding induce mortality or disper-
sal to new habitats.
4. The degree to which competition among ecologically similar organisms
lowers the carrying capacity of the insect population in that habitat.
5. To what extent mortality is induced by fluctuations in parameters of the
physical environment, and over what time frame.

When discussing the impact of density-dependent and density-independent

mechanisms or factors on the population dynamics of an individual insect species
in a particular habitat, it is important to realize that any single individual in the
population can be subjected to both kinds of mechanisms and that "population
regulation" is essentially the combined effects of both mechanisms. In many
terrestrial tropical habitats, insects will be subject to subtle differences in en-
vironmental factors affecting survival and reproduction, some of which may be
density-dependent factors and others density independent. Thus the collection of
individuals of a grasshopper species occupying a premontane tropical rain forest
light gap will be differentially influenced by a wide range of such subtle factors
so that understanding survival and reproduction become a matter of individual-
to-individual variation in responses to such factors. Collectively, the sum of such
differences in response will determine the tenacity of that population for that light
gap. Conceptually it is easy to examine issues of population regulation in terms
of density dependence and density independence. But the reality of tropical
insects in their habitats is that a wide range of often subtle environmental factors
creates a complex system of population regulation, a system that will vary within
a species over different communities in the same region, among regions (geo-
graphical), and between seasonal and nonseasonal environments.
For lowland tropical regions, it is necessary to assess the role of periodic
catastrophes in the same way that Itd443 ) analyzed typhoon effects on populations
of the beetle Ips typographus on the subsequent potential for population regula-
tion in the region for that species and the likelihood of periods of overpopulation
as part of a general return to equilibrium conditions in the community. It is easy
to imagine and determine that fluctuations in insect popUlations come from a
variety of sources, even in tropical regions (Fig. 6.5), and we must not generalize
about such effects for all species in all regions. Insect populations in the tropical
regions of the world are believed to exhibit diverse forms of census histories
reflecting mechanisms of food limitation, predator limitation, competition limita-
tion, and limitations from the physical environment. We do not necessarily
expect one of A. M. Shapiro's high-altitude pierid butterflies in Ecuador to
possess the same kinds and intensities of regulatory agents as a related species in
the lowlands on the eastern slopes of the Andes below 200 m.(760) The key, it
seems, is to determine the relative role of various types of environmental factors
 234 CHAPTER 6

3
>-
t::
(1) 2
z
w
0 1

I 0
b
~ 3

1910 1915 1920 1925 1930 1935 1940 1945

YEARS

FIGURE 6.5. Insect population census history as related to fluctuations in the physical environment.
The pattern of census history depicted is for the paddy borer moth, Schoenobius incertellus, in Japan.
Redrawn and modified from Y. Ito, Bull. Nat. [nst. Agric. Sci. Japan Ser. C, No. 13 (1961).

(biotic and abiotic) on insect populations in different kinds of regions in the

tropics and then to relate these effects to community structure, habitats, and
patterns of fluctuation in physical parameters. We also need to consider the
effects of trophic status on the type of regulation, if any, operative on the
population of a particular insect species occupying a particular habitat.
This is really a discussion of the differences between rand K selection:
species in physically controlled environments are believed to be ecological
generalists exhibiting large populations and with few if any competing species
and a lack of biotic control from predators and parasites. This is an example of r
selection of the type expected for most high altitude regions in the tropics and
tropical deserts. K -selected species have smaller populations, more specialized
patterns of resource utilization, biotic control, and perhaps competition from
other species. K-selected species are believed to be more prevalent in some kinds
of tropical wet lowland environments. But one must keep in mind that we are
POPULATION RESPONSES TO THE ENVIRONMENT 235

referring to relative effects, and that at any elevation in the tropics we expect to
find mixtures of r- and K-selected species, depending upon the habitat. Given
the diversity and patchiness of plant species in some lowland tropical terrestrial
habitats, one might expect that many phytophagous insect species are K-selected
in the sense of being specialists on certain plant species, with low average food
niche breadth. The comparatively benign physical features of the lowland tropi-
cal wet and dry forests (with the exception of certain coastal habitats) contribute
to the accumulation of plant species in some habitats, which in tum promote
ecological diversification among phytophagous insects. K selection is also man-
ifested in terms of competition between these insects, as evidenced indirectly by
the accumulation of species-rich insect guilds on some plant species, with low
numbers of individuals per species, and in some instances with each component
species specializing on a different part of the plant species. (93) Yet some of these
species may be kept scarce by predation and parasitism. But tropical predators
and parasites are believed to be generalists with some tendency toward r selec-
tion attributes in order to take advantage of diverse resources per unit area of the
habitat. I am referring here primarily to tachinid flies and various hymenopterous
groups such as the Braconidae. If each host patch is small, a predator and parasite
both need to visit many patches within the habitat, and selection favors the ability
to exploit patches of different host species, within some limits. Those limits are
set largely by the searching and feeding physiology of the predators and para-
sites. The obvious implication for such considerations is that some of these
insects become ideal candidates for biological control programs, even in Tem-
perate Zone countries. (48)
While lowland tropical regions may be relatively benign in terms of temp-
erature fluctuations and excessive dryness throughout the year (i.e., nondesert),
some tropical butterflies such as Parides are killed as adults (Fig. 6.6) by severe
rainfall.(444l Such effects produce considerable fluctuations in the abundance of
adult butterflies during the rainy season (Fig. 6.7). In some cases the resulting
survivorship of adults is comparable to that of many Temperate Zone species
(Table 6.1). Adult butterflies sometimes shift their patterns of dispersion mark-
edly in response to seasonality, increasing their population densities and accessi-
bility to some vertebrate predators. In some instances highly palatable butterfly
species are involved, and avian predators expertly remove the more palatable
abdominal contents and discard the exoskeletons.(205) Such predation is density
dependent.
Predation may be high in tropical butterflies that are supposedly palatable
but have other means of defense such as disruptive wing coloration pattems(445l
and in which the disruptive coloration does not reduce predation. Butterflies in
lowland tropical regions may be subject to certain kinds of predation from verte-
brates(186,188,189,207) while both feeding and flying. We need to reexamine some of
the classic concepts of protective coloration and mimicry with more data on
236 CHAPTER 6

FIGURE 6 .6. Adult female of the Neotropical buttert1y Parides iphidamas, resting on a melasto-
maceous plant at the edge of plimary growth tropical rain forest in Costa Rica.

50

40

..,g
~

~
30
~

. 20
'0

~
D

~
or.
10

July August

FIGURE 6.7. Changes in adult population size for the Neotropical buttert1ies Parides neophilus and
P. anchises in Trinidad during the 1965 wet season. Daily population size (both species) is smoothed
by application of three-point moving average. [From L. M. Cook, K. Frank, and L. P. Brower,
Biotropica 3:17-20 (1971).)
POPULATION RESPONSES TO THE ENVIRONMENT 237

TABLE 6.1
Estimates of Daily Survival Rate of Parides on Trinidad Compared with
That of British Lepidoptera a

Species Survival rate Reference b

Parides spp. 0.84 This paper

Polyommatus icarus (von 0.83 Dowdeswell, Fisher, and Ford
Rottenburg) (1940)
P. icarus 0.74 Parr, Gaskell, and George (1968)
Parage megera (Linnaeus) 0.79 Parr, Gaskell, and George (1968)
Maniola jurtina 0.89 Dowdeswell, Fisher, and Ford
(1949)
Coenonympha tullia (Miiller) 0.70 Turner (1963)
Lysandra bellargus (von Rotten- June 0.79, 0.92 Dayis, Frazer, and Tynan (1958)
burg)
September 0.91
Lysandra coridon (Poda) 0.84 Davis, Frazer, and Tynan (1958)
Panaxia dominula 0.84 Fisher and Ford (1947)
P. dominula 0.78 Sheppard (1951)
P. dominula 0.62 Cook and Kettlewell (1960)
Zygaena Jilipendulae 0.71 Manly and Parr (1968)

a From Cook et al. (444)

"References can be found in Cook et at.'444)

actual surivorship in natural populations. As with populations of other insects,

tropical butterfly popUlations may be regulated by a variety of biotic control
agents. (221 ,222)
Whether or not such patterns of mortality behave in a compensatory manner
through time in a fashion similar to that described by Murdoch for Temperate
Zone Carabidae,(446) long-term studies in the tropics are warranted. We also need
to recognize cases in which shifts in the breadth of food resources utilized in a
habitat by a single insect species result from changes in population density: under
conditions of low density, food niche breadth might be lower than for the same
species and popUlation at higher densities. There might be some cases of "facul-
tative broad-base or flexible monophagy or even polyphagy" when popUlation
density becomes unusually high. Such effects will result from changes in the
habitat affecting competition and biotic control agents: periods of relaxed control
or competition may be the prelude to periods of increased population density and
facultative shifts in feeding spectra within a species. Some habitats may exhibit
more changes through contemporary time than others in the tropics. If the habitat
changes in such a way as to relax some forms of biotic control on some insect
species, and if the response to selection is rapid by the insects, there could be a
shift from K selection to r selection in terms of the population dynamics.
238 CHAPTER 6

Given sufficient generations, directional selection may result in an adaptive

intrapopulational shift toward genotypes capable of increasing population growth
rate. In some tropical habitats such shifts may occur regularly for insect species,
depending upon the tenacity of the species for that habitat and the degree to
which gene flow from populations in nearby habitats delays or promotes such
responses. The resulting "outbreak" is successful if the increasing population of
the insect species is able to find and utilize suitable resources. If the habitat
contains large patches of the single preferred resource or consists of a mosaic of
several suitable alternative resources, the species will colonize the resource base
to a degree not realized under K selection. It may be easier in terms of energetic
considerations for the species to exploit a single suitable resource if differences
among resource types are sufficiently different to reduce fitness additively. Out-
breaks of phytophagous insects are known from single-species plantings in the
Temperate Zone, (447) and some regions of the tropics may have regular or occa-
sional "mini-sagas" of such effects as habitats change in cyclic or acyclic
manners.
Depending upon the degree to which scavenger insects and other arthropods
are resource specialists, occupants of benign or harsh regions, and regulated by
biotic control agents, these insects will also respond to habitat changes in a
manner implied above for phytophagous insects in particular. Generalist pred-
ators and parasites of phytophagous and scavenger insect species may respond
less to such perturbations in the environment since they are already generalized to
some degree. But the direction of selection pressures may be sufficiently dif-
ferent so that some portion of the resources of endoparasitic forms and predators
are shifted in a significant manner: if some phytophagous species drop out of the
community as a result of a habitat change, the food base for predators and
parasites may contract, changing the population dynamics of these species. Such
species may increase exploitation on some or all remaining species suitable for
exploitation or experience a reduction in population growth since fewer resources
are available. The generalist tendencies of these species, however, probably
precludes their dropping out of the community: they are tenacious enough to
adjust to the change accordingly. Such predictions warrant long-term studies for
verification.
The previous discussion stresses the great heterogeneity of the tropics in
terms of how component species of specified communities respond to perturba-
tions in the environment. Relatively benign habitats and regions of the tropics,
that is, those where physical conditions are not severe, may be more vulnerable
to such perturbations (floods, severe periods of rainfall, mudslides, etc.) than
other habitats where the environment is generally harsh most or all of the year
(high elevations, deserts, certain coastal formations, etc.). Organisms occupying
the latter regions are already evolutionarily adapted to conditions favoring r
selection in response to abiotic mortality factors and an absence of density-related
POPULATION RESPONSES TO THE ENVIRONMENT 239

biotic control agents. Any conclusions about the kind of regulation of insect
populations in the tropics must therefore be qualified with a consideration of
the regional and habitat effects on individual species associated with each
trophic level. We anticipate a spectrum of different habitats at various intervals
along elevational gradients in the tropics; there is also the effect of increasing or
decreasing elevation on the physical environment, in turn affecting organisms
found there, and the ability of individual species to move in and out of com-
munities at different elevations. Physically controlled communities are composed
of generalist species, and there is room for the addition of additional species
through the effects of migration and range expansions. Biologically accommodated
communities associated with some lowland tropical regions may be saturated
with species, many of which are specialists, and the addition of new species is a
gradual process made possible by the increase in harvestable biomass of such
environments (see Chapter 1).

6.4. SOME EXAMPLES OF CENSUS HISTORIES IN TROPICAL

INSECTS

The modes of adaptation to the environment are largely determined by the

properties of breeding populations of a particular species. The pioneering studies
of Sewall Wright elucidated rigorously the effects of popUlation size on the
evolution of adaptations in popUlations. (448,449) When popUlations are small on a
relative scale, certain events affecting gene and genotype frequencies, such as
drift, may override the effects of selection and gene flow as major determinants
of evolutionary history within species. On the other hand, large populations may
be characterized by considerable genetic variation of various kinds,(1) providing
greater opportunities for new adaptations to appear and stabilize. Populations of
individual species may pass through evolutionary filters of various kinds, charac-
terized by periodic contractions and subsequent expansions, thereby clouding the
effects of what might have been a direct selective pathway to optimal use of the
environment. Initial gene frequencies may vary considerably in different popula-
tions of the same species in a region and the differential effects of selection in
different areas within the region result in different genomes being stabilized,(54)
even if only temporarily. The rate of stabilization of characters through stabiliz-
ing selection largely depends upon the amount of genetic variation in the popula-
tion and the size of the population. Different populations of a species in more or
less continuous contact may result in a postponement of local adaptations being
expressed within each population because the effects of local selection are some-
times swamped by shifts in genetic composition arising from gene flow.
Any attempt to understand the interplay of regional and habitat diversity on
the development and maintenance of faunas in the tropics must also take into
240 CHAPTER 6

account the discreteness and sizes of breeding populations for a species. The
problem for the tropical biogeographer is to confidently recognize the boundaries
of populations in making analyses of historical versus contemporary selection
pressures acting on species; the tropical population biologist requires the same
data for understanding how different species fit into a community and where the
boundaries of communities occur. Yet one major obstacle remains the lack of
long-term studies on the population dynamics of any tropical organism(42) for
periods of at least 10 consecutive years. Some short-term studies of tropical
butterflies, (e.g., on such as that for M. peleides by Young and Thomason(t90»)
provide information on the effects of habitat boundaries on the size of an insect
popUlation and the dynamics of dispersal within it. In that study it was possible to
estimate popUlation size from various methods, something greatly facilitated by
the confidence in actual boundaries of the butterfly popUlation (Table 6.2). Daily
survival rate was calculated from the same capture-mark-release data from the
expression

(1- S) - - -
l-R
tNt + NR
where S is the survival rate per day, N the estimated daily population size, and R
the last day of collecting. (450) Such studie.s also reveal interesting data on sex

TABLE 6.2
Comparative a Daily Population Size Estimates and Survival Rates for
M. pe/eides in the Barranca Study Area during the 1973 Dry Season b

Population size estimates

Survival rates
Fisher and Ford Jolly Manly and Parr [Fisher and Ford
Date (1947) (1965) (1968) (1947)]

23 January 0.56
24 99 84 84 0.60
25 72 72 73 0.63
26 105 103 105 0.67
27 79 66 64 0.71
28 78 73 75 0.75
29 81 97 93 0.79
30 82 80 78 0.84
31 68 66 66 0.89
1 February 55 58 58 0.94
2 61

aPopulation size estimates are calculated by three methods.

bFrom Young and Thomason."90)
POPULATION RESPONSES TO THE ENVIRONMENT 241

ratios, which in the case of Morpho butterflies are generally skewed toward
males, are not due to aberrant sex ratios at birth, but rather result from different
behaviors of the sexes biasing the probabilities of reca~ture toward one sex or the
other.
Although the study of populations may be more challenging in areas where
habitats are not as restricting as in this study, Ehrlich and Gilbert<451) found
convincing evidence that a population of the tropical butterfly Heliconius ethilla
remained stable in size for more than 2 years (Fig. 6.8). That population was
composed of two colonies with virtually no exchange of butterflies between
them. The stability of the population is explained in terms of density-related
forms of predation on early stages and the continual supply of adult resources
found in the habitat. Heliconius butterflies are relatively long-lived as
adults, (451,452) another factor contributing to the observed stability of populations.
In some species of Heliconius such as H. charitonius, communal roosting (Figs.
6.9, 6.10) is a major mechanism contributing to the spatial structuring of the
adult butterfly population.(452--455) Such behavior contributes to observed patterns
of population stability and cohesiveness within the habitat in Heliconius.(452)
Adult butterflies form conspicuous aggregations on the same clump of dead vines
or branches (Fig. 6.9) month after month. (452--454) Capture-mark-release studies of
species such as H. charitonius reveal that the same individuals return to a roost
site over long periods (Fig. 6.10). Longevity in Heliconius is also promoted by
specialized feeding on pollen that is used as a protein source for eggs by the
butterflies. (31)
Another feature of such populations is a general tendency for age classes to

..
V>
...J

:>
0
400
•• AREA A
>
0 o. AREA B
~
"- 300
0

",,,,
a: . 1970 1971 1972

d
<Xl'; DRY SEASON DRY SEASON
~ +1 200
Z

+1-- _____________ f
100

PERIODOO~~~2~~74~~6~~~8~~1*0~~12'-~1~4~~1~6~~18~~2~0~~~
... L···-·4~1~
1 --I
DAY 0 40 80 120 160 200 240 280 320 360 400 440 ... ·820
DATE DEC '69 MAy'70/JULy'70/AUG'70 OCT '70 DEC '70 FEB'71 MAR '72

FIGURE 6.B. Estimates of adult population size for the Neotropical heliconiine butterfly H. ethilla
at St. Andrew's Trace, Trinidad, for different areas (A and B) during two successive dry seasons.
[From P. R. Ehrlich and L. E. Gilbert, Biotropica 5:69-82. (1973).]
242 CHAPTER 6

FIGURE 6.9. H. Beth Braker looking at a roosting site for the Neotropical heliconiine butterfly H.
chariton ius along a shady footpath near the edge of mixed primary and advanced secondary pre-
montane tropical wet forest in northeastern Costa Rica. Several species of Heliconius form nightly
communal roosts in clumps of dead vines such as the ones shown here. Communal roosting is a major
component of the population dynamics of these butterflies.

remain rather similar in distribution over long periods (Fig. 6.11). The regulatory
mechanism in such an insect population consists of density-dependent mortality
in early stages coupled with high longevity. Predation rates on H eliconius are not
well documented, although one study(339) indicated that survival is high in the
adult stage due to specific types of defensive behavior. In tropical butterflies such
as Acraea encedon in Uganda, some populations are predominantly female and
the abundance of males is inversely correlated with overall population size. (456)
Thus the abundance of males in this instance plays a self-regulating role for the
population: mating frequency increases as population size diminishes. Tropical
mosquito populations are usually characterized by considerable fluctuations in
relatively short periods, the result of short adult life-spans (4 days or lesS).(457)
Fluctuations in the abundance of an insect species in an area are expected if
a predator or parasite of that species interacts with it in a patchy manner, as
POPULATION RESPONSES TO THE ENVIRONMENT 243

FIGURE 6.10. A communal roost of the butterfly H . charitonius in Costa Rica. Note marked
butterflies. Marking studies reveal considerable individual fidelity to roosts in these insects. a factor
contributing to population cohesiveness.

expected for generalists exploiting several different resources in a habitat. If the

predator or parasite is a specialist on a few host types , host populations may not
exhibit fluctuations through time since the intensity of the mortality generated
may be more or less constant through successive generations. For insects with
effective defensive mechanisms, particularly in the adult stage, adult survival
may be high and adult population size stays rather constant. Strongly seasonal
environments in the tropics add another dimension, in that periods of uneven
abundances may result from effects of shifting physical conditions on biotic
control agents such as small-bodied parasitic wasps. For example, a large-bodied
lepidopteran may continue to breed during the dry season if larval host plants
 244 CHAPTER 6

200 Total marked 0

160 Young I
Middle rn
Old I
'"
.~
t
....'":l
....
.0
'0
t 80
.0
E
:l
Z 60

40

20

o I
12 !2 12 I 1 22223333" ,,'555
Interval
Sampling periods

FIGURE 6.11. Number of adult butterflies, and abundances within three approximate age classes,
for H. charitonius in Costa Rican mountain rain forest (above), and the numbers of newly marked
butterflies (below). Units on the abscissa are days above and months of the year below. [From L. M.
Cook, E. W. Thomason, and A. M. Young, J. Anim. Eco!. 45:851-863 (1976). Copyright 1976 by
the British Ecological Society.J

remain lush (as is expected for some riparian forest habitats), but the relatively
small-bodied endoparasitic braconids that parasitize the larvae of such a species
may disappear temporarily in the dry season, relaxing one type of biotic control
on the herbivore population. Regulatory interactions of this kind may return
during the following rainy season. At the same time, other forms of biotic
control, such as increased predation on adult butterflies in dry season forest
refugia, may occur when populations of individual species become compressed
into relatively small areas of moist habitats. We therefore need to examine census
histories of insect populations in seasonal and nonseasonal tropical regions
over several cycles of seasons to determine the patterns of fluctuations in adult
recruitment and the kinds of environmental regulatory effects operative at differ-
ent times. Shifts in the kinds of resources available to the target insect species
POPULATION RESPONSES TO THE ENVIRONMENT 245

being studied must also be considered, especially for species that are herbivores
and pollinators.

6.5. POPULATION RESPONSES TO ENVIRONMENTAL

HETEROGENEITY IN TROPICAL INSECTS

The challenge to any species, under the assumption of individual selection,

is to maximize fitness. The evolutionary maximization of fitness in populations is
reflected in the tendency for such populations to increase. A useful conceptual
framework for population increase is provided by the familiar formulation
L e-rxlxmx
OX>

1=
x=o
with parameters as defined earlier. Selection should favor the maintenance of an
optimal value of r or the increase in r. If environments were constant and full of
infinitely abundant resources, the populations of individual species would realize
r max ' But environments are heterogeneous to varying degrees and in different
ways, creating an evolutionary challenge for species to maximize r as much as
possible given prevailing environmental conditions. The response of selection for
a given species will be a function of the generation time, generally expressed as
T = !,L(x)m(x)(x)
!,L(x)m(X)

where T is the mean generation time in a particular population. The rate of

increase of the popUlation will therefore be
log,Ro
r=-T-

where Ro is the net reproductive rate of the sum of reproductive expectations (Ro
= !, Lx mx)' These formulations are based upon the assumption of a stable age
distribution in a population, a condition which virtually never exists. Populations
generally fluctuate to some degree in terms of density, so that the distribution of
age classes is not stable from generation to generation. Reproductive expecta-
tions, the product of each Lx and mx value for females, are changing in different
age classes, so that fitness is determined by an additive effect summed over all
reproductively active age classes. Periods of colonizing episodes in species favor
age classes with high fecundity in early adult age classes so that the production of
offspring is maximized. (87) Populations experiencing K selection will be charac-
terized by fecundity being distributed over older age classes (see also Fig. 6.4).
One type of environmental heterogeneity affecting populations is the transi-
tion from favorable to unfavorable conditions, a situation that promotes coloniz-
 246 CHAPTER 6

ing episodes in space: the species migrates to a new environment. Periodic

temporal fluctuations of this sort may occur in very seasonal lowland tropical
regions where species migrate into higher elevation areas where conditions have
generally remained benign (in terms of moisture and responses of resources to
moisture gradients). The studies of Dingle(14()-142) have indicated that Oncopeltus
milkweed bugs exhibit survivorship curves that correspond to different environ-
mental character states such as light and temperature regimes (Fig. 6.12).
Temperature change may alter the survivorship curve even when day-night
cycles remain 12 : 12 hours: the survivorship curve for 27°e indicates considera-
ble mortality in early adult age classes, whereas 23°e conditions promote longev-
ity and continued fecundity (Fig. 6.12). The implication is that hot tropical
conditions, such as those of the dry season in some lowland areas, may promote a
form of r selection in which bugs are programmed to reproduce early in adult
life. Such a population structure may be an effective adaptation to bugs locating
and colonizing patches of suitable habitat along an elevational gradient or hori-
zontally into moist refugia as the dry season advances. More temporally favor-
able lowland tropical regions, such as those of tropical rain forest, would tend to
select for extended reproduction to match a survivorship curve similar to that of
Oncopeltus for the 23°e and 12: 12 cycle (essentially tropical). Relative to

1.00.::----....,.--.,..--......- -
....\\
.75 \
\
\

,,
\
\

~ .50
\,
.=" \
o \
2:0.25 \
~ ""'"
~-"'"":~-.....,~~-',~--"*~-~ FIGURE 6.12. Survivorship curves for
o:J ~ W the milkweed bug, Oncopeltus, under the
0%1.00 same light regimes but different temperature
i=
a:: regimes, in the laboratory. Female sur-
~ 0.75 vivorship is indicated by solid lines, males
~ by dashed lines. x is a unit of age and Lx is
0.50 the same as Lx in the text. [Modified and
redrawn from' 'Life history and popUlation
consequences of density, photoperiod, and
0.25 temperature in a migrant insect, the milk-
weed bug, Oncopeltus, by H. Dingle,
O~_~...._ _~_ _......_ _~_......... American Naturalist 102:149-163 (1968)

o W by permission of the University of Chicago

AGE IN WEEKS (x) Press.]
POPULATION RESPONSES TO THE ENVIRONMENT 247

TABLE 6.3
Rates of Increase in Laboratory Cultures of O. fasciatus under Different
Environmental Conditions a

Temperature Initial Increase Doubling time

Photoperiod CC) density per day (r) (days)

16L-8D 27 10 pairs/box 0.0861 8.42

16L-8D 27 20 pairs/box 0.0810 8.90
16L-8D 23 10 pairs/box 0.0736 9.75
12L-12D 27 20 pairs/box 0.0593 12.03
16L-8D 23 20 pairs/box 0.0499 14.24
12L-12D 23 20 pairs/box 0.0369 19.13

aFrom Dingle.(14O)

Temperate Zone conditions, however, tropical regions tend to select for lower
population growth rates (Table 6.3). In tropical conditions, it is the age to first
reproduction that is extended (Table 6.4), thus lowering r. In some Oncopeltus
populations it is the joint effect of earlier age of first reproduction and higher
levels of fecundity that promote Temperate Zone populations to be larger. (458)
The timing of migration to favorable habitats in insects such as Oncopeltus will
be prior to reproduction, and these individuals suppress feeding and reproduction
for the sake of maximizing energy for flight. (141.142) These individuals possess
high reproductive values (V x), since the expectation for breeding is high once the
colonization or migration is completed. The reproductive value is the expected
contribution of an individual of specified age to future growth of the popula-

TABLE 6.4
Development Time from Birth to Adult and Age at First Reproduction
for O. fasciatus in Laboratory Culture a

Temperature Development time (days) from Age (in days) at first

Photoperiod ce) birth to first adult reproduction

l6L-8D 27 27 38 (l0 pairs/box)

47 (20 pairs/box)
16L-8D 23 36 46 (10 pairs/box)
63 (20 pairs/box)
l2L-12D 27 27 61 b (20 pairs/box)
12L-12D 23 36 95 (20 pairs/box)

aFroID Dingle.(14O)
'One female began egg laying at 48 days. All others began at 61 to 65 days.
248 CHAPTER 6

tion.(459) Reproductive value is defined relative to the value of this statistic at

birth (v o ) for the species:

Integration is from x to X, so that only future births are considered.

The studies of Dingle, Landahl, and Root primarily concerned laboratory
populations of milkweed bugs, and there is some difficulty in assuming that
similar survivorship and fecundity curves occur in natural populations. That the
tropical form begins reproduction later in adult life and has a lower mx curve in
the laboratory may not hold in the field, where survivorship will also be a
function of biotic control agents affecting adults, a factor ruled out in the design
of laboratory studies. Nonetheless, such laboratory studies do measure relative
differences between forms or species under controlled conditions of abiotic fac-
tors and food supplies, and these studies provide estimates of the physiological
capacities of forms or species to build up populations. One might attempt to infer
that lx and mx curves of the kind found in the tropical form by Landahl and Root
result in gradual, low recruitment of adult bugs in the field and that predation
rates on adults are low. Field studies are needed to confirm such inferences.
Colonizing species should maximize reproductive value in order to
maximize the likelihood for success in a new environment. In the lowland
tropics, I envision the following situation promoting a genetic structure of an
insect population that includes an ability to undergo occasional or periodic col-
onizations of the kind discussed above for milkweed bugs.
1. Annual severe dry seasons, leaving behind a mosiac of relatively small
moist refugia-type habitats that some insects can exploit, by horizontal movement
into them, until the rainy season. In addition to this horizontal movement, some
vertical movement is expected in which adult insects move into moist forests at
higher elevations during the dry season and recolonize the lowlands the following
rainy season. This phenomenon is best described as an ecological contraction-
and-expansion cycle that selects for some colonization in species unable to main-
tain the population structures established in the rainy season during the severe dry
season.
2. Periodic floodings of some stream or river-edge flood-plain forest
habitats. Rather than the environmental stress being water-stress effects on in-
sects or their resources, as is the case for dry-season effects, the stress here is the
alteration of plant communities so that associated insect communities are also
affected adversely, to the extent that individuals may make long- or short-
distance movements to colonize unaffected areas. The success of such events will
be relatively less frequent than for effects of seasonality, since the former catas-
trophes are probably irregular events, and thereby making it difficult for selection
to adapt the species to them. Success will largely be defined by the intensity of
POPULATION RESPONSES TO THE ENVIRONMENT 249

the perturbation and the amount of phenotypic flexibility or genetic variation in

the population of individual species.
3. Pronounced perturbations in the structure and composition of habitats
resulting from natural and human-induced activities. Again, predictability of
such events is low and local extinctions may be high if species are unable to
colonize new habitats.
In all three categories of tropical environmental heterogeneity, success in
colonization will also be determined by the interaction of a colonizing species
with species already established in the communities associated with the environ-
ments being colonized. If lowland tropical habitats on the mainlands tend to be
saturated with species,(l5) competition may preclude successful establishment of
new species. But if harvestable productivity is sufficiently high, resources might
be sufficient for new species to be added to some communities under these
conditions. Tropical mountain tops may be relatively easier to colonize if the
arriving species can adjust to the stressful effects of the physical environment.
The above considerations deserve comment with respect to trophic organiza-
tion in tropical terrestrial communities. Some river-edge tropical insects in
Amazonian forests are efficient mobilizers of detritus from the ground litter, (460)
and such forms are expected to adapt well to new areas where litter is present.
Other generalists, such as endoparasitic flies and wasps and some predators on
insects, are also expected to do well, assuming they can migrate successfully.
Perhaps the greatest opportunities for local extinctions during such periods are to
be found in the phytophagous insects, especially in communities where species
are specialists on a few kinds of host plants.
Ecological escape from an unfavorable environment by an insect species
may not necessarily be an "all or none" strategy in which a major portion of the
population leaves and the rest of the membership goes extinct. Particularly in the
lowland tropics, various kinds of disturbances may alter only portions of a
habitat, and some portion of the insect communities found there may survive
without leaving. The effect of disturbance may often be patchy, and if so,
selection may be for a mixed strategy within a species: genetic polymorphisms
may permit some morphs to survive in the altered habitat while other morphs are
prone to dispersal to new habitats. Populations in some unpredictable environ-
ments are predicted to be polymorphic for parameters underlying population
growth.(460 Polymorphic insect populations may have a higher average fitness
because they endow a mixed strategy in terms of how temporal environmental
heterogeneity is responded to within a population. As discussed above, various
kinds of temporally fluctuating habitats or environments can exist in the tropics,
with the major differences among these being the regUlarity and intensity of
fluctuation upon insect communities. The two morphs predicted under such
conditions are morphs adapted for early reproduction (r selection) and morphs
adapted for late reproduction (K selection). Each kind of morph may increase or
250 CHAPTER 6

decrease in the population according to the size of the population at anyone

time.(462)
The lowland tropical dry season of dry forest regions in particular represents
a major, regularly occurring environmental fluctuation to which resident species
may respond by a mixed strategy. Some butterfly species in such regions have
wet- and dry-season morphs.(317) Within the dry season, a period of environmen-
tal stress on many insects and other organisms, there may be a mixed strategy
within a species, in that one morph adapts to the dry conditions while another
migrates to less stressful habitats. Different habitats provide cues favoring the
timing of reproduction within a species at a certain time and with a certain pattern
of pulsation through time. Peak reproductive activity should occur at times when
the chances of survival of juveniles are highest or when predators are most
readily satiated.(363) But the timing of reproduction, and the expression of spe-
cific kinds of Ix and inx trajectories within populations, is expected to be fluid
for most environments in the sense that every species requires some degree of
genetic variation and phenotypic flexibility in terms of the potential for popula-
tion growth. Even in the lowland tropics, where physical conditions are generally
held to be benign relative to very high altitude tropical regions, there will be
some degree of temporal uncertainty in most habitats. This uncertainty or
heterogeneity arises from occasional to regular fluctuations in the environment,
and shifting distributions of species resulting from colonizations, range expan-
sions, and shifts in harvestable productivity. These effects may be somewhat
patchy in space, since some communities may be more sensitive to such distur-
bances than others. It is therefore the interaction of temporal variation of the
environment, and the resulting spatial patchiness of its effect, that selects for
some degree of adaptive flexibility in species. Insects are expected to exhibit
response patterns to environmental heterogeneity because of their relatively short
generation times, and these organisms therefore provide suitable empirical
models for investigations on the dynamics of communities in the tropics.
Temporal heterogeneity is also evident in the tropics in terms of how sea-
sons affect the patterns of plant growth and reproduction in different regions. The
reSUlting patterns, in tum, determine resource distributions for animals includ-
ing insects. Levels of harvestable productivity in tropical communities change
with seasons, and some plant structures (i.e., vegetative biomass) are replaced
with others (i.e., reproductive biomass: flowers and fruits) during a season in
some regions. But regional effects also come into play: the tropics are variable
in terms of spatial heterogeneity of the environment.
Spatial heterogeneity in tropical regions can be classified as follows:
1. Among regions, differences exist in terms of climate, topography, and
the kinds of organisms capable of being residents.
2. Within a region there may exist several distinct or similar habitats, each
with a more or less distinct flora and fauna, perhaps with some degree of restric-
tion in gene flow among the habitats.
POPULATION RESPONSES TO THE ENVIRONMENT 251

3. Within a habitat the distributions of organisms will be patchy to some

degree; each species may have a characteristic kind of patchy distribution, and the
sum of these distributions within a trophic level defines the pattern of community
structure associated with that habitat. Patchiness of this sort is believed to vary
among some trophic levels: plant species distributions and their associated guilds
of phytophagous insects are expected to be patchy in species-rich habitats, while
generalist predators, parasites, and scavengers may be less patchy since much of
their foraging and feeding is opportunistic. Omnivores may present a mixture of
degrees of patchiness in distribution because periods of specialized feeding on
certain plants may be intermingled with opportunistic and generalized feeding on
other insects. A good example may be some tropical orthopterans.
The spatial heterogeneity of vegetation depends upon many factors, includ-
ing seed dispersal mechanisms, optimal seed germination and seedling growth
conditions, and microclimate.(463.464) The bulk of the insect biomass found in a
habitat depends in some way upon the vegetation found there, either directly
(feeding on leaves, stems, roots, flowers, seeds, and fruits) or indirectly (feeding
on fallen and drying or dead plant parts and products). The spatial distribution of
a plant species within a habitat is molded by the interaction of both physical
factors and biotic interactions among individuals of the species and among
species. (463) Such factors also underlie the mechanisms of habitat selection in
plants. In fact, an optimal habitat is sometimes referred to as an optimal
patch,(463) although patchiness can also refer to the distribution of resources
within a habitat. Wiens(463) distinguishes between true habitat selection and
habitat correlation. True habitat selection occurs when individuals of a species
have a choice between two or more habitats and choose the optimal one. Habitat
correlation refers to an enforced distribution among habitats resulting from some
environmental factor affecting the species. The response to habitat is determined
for a species largely by the patch structure of the region. (465) Habitat suitability is
also a function of population densities of species: as density increases, habitat
suitability declines(463) (Fig. 6.13). Under low population density for a species,
habitat selection will be most effective as a mechanism to maximize fitness
within the population. But comparisons in habitat quality for a species should be
made only when popUlations among different habitats are examined at their
equilibrium densities.(464) Niche breadth of a species within a habitat is often
density dependent. (49)
These considerations may also apply to insects because these organisms are
closely associated with vegetation as a resource. In many tropical habitats, espe-
cially in the lowlands, true habitat selection may occur since plant species do not
reach very high densities. This is not universally true as habitats exists where
densities of a few plant species are very high. But in habitats with high species
richness of plants, the observed distributions of species may reflect true habitat
selection, with minimal or no clouding effects of density (Fig. 6.13). The same
reasoning may follow for phytophagous insects. Densities of individual insect
252 CHAPTER 6

FIGURE 6.13. The Fretwell-Lucas

model of habitat selection and distri-
bution. Habitat 1 has the highest
quality. When total population size is
low (A), all members of the popula-
tion should reside in it. But as density
increases following colonization, the
quality of this habitat declines to a
point where fitness prospects be-
tween habitats 1 and 2 become equal,
resulting in the individual selecting
both (B). As density increases fur-
ther, other habitats (C) will also be
colonized. [Reproduced, with per-
mission, from 1. A. Wiens, Ann.
Rev. Ecol. Syst. 7:81-120 (1976).
Copyright 1976 by Annual Reviews,
POPULATION DENSITY IN HABITAT Inc.]

species may not attain high levels in some habitats, and the distribution of insect
species under such conditions will reflect optimal use of resources.
The actual distribution of individuals over the resources within a habitat or
among habitats will depend also upon the "grain-response pattern". (463) Levins(3)
distinguished between fine- and coarse-grained environments in terms of the
manner in which organisms encounter different environments or different re-
source patches. An environment is coarse grained if the organism spends all of its
time in one patch, whereas an environment is fine grained if the organism passes
through many patches.(3,381) Individuals of a species adapted to exploiting re-
sources in a fine-grained manner are predicted to spread more or less uniformly
over resource patches as popUlation density increases, whereas an initially
clumped coarse-grained specialist at low population density may spread out over
other resources as population density increases and suitable resources become
available (Fig. 6.14). Thus an insect species specializing on one kind of legume
herb in an advanced secondary habitat in the lowland wet tropics may be clumped
on patches of that resource when density is low. But as density increases, and if
the insect is genetically or physiologically flexible enough to exploit allied
legumes in the same area, the population may spread out over the other patch
types in the habitat. The key factor is determining the degree to which insect
popUlations in fact increase to densities in the tropics to bring about such pre-
dicted changes in distribution. Increased niche breadth may not always be a
reflection of increased popUlation density, since some insects may be adapted to
feeding on a broad range of plant species regardless of density considerations.
An interesting problem to examine is the transition of an insect species from
an association with a natural habitat to outbreak conditions associated with ag-
 POPULATION RESPONSES TO THE ENVIRONMENT 253

Fine-grained Coarse- grained

LOW

. ..
,..
u
z
<l
0
z
::J
- •
... .0 _0

~. -
•
6- ..
. _.J.
~
..
., e.. -
~.

...
lD

. .". ... .- . ....... .....

<l

. . . . . .. . . -. ..
~
-
............... .
• It-.
. --=-
~
... - • !.!"~t
.~

HIGH .! ~ • • .. . • ~;;
- -

clumped

z
o
iii
a:
1&.1
Q.
II) " coarse- grained
o
uniform L-_-_-_-_-_-_-_-_-_-_-_-_-_-_-_-_-_-~_-_-_-_ fine- grained

law - - - - - - - _
.. high
DENSITY

FIGURE 6.14. A model of habitat selection in fine- and coarse-grained environments. The fine-
grained population is expected to retain a uniform dispersion pattern over a broad range of population
densities, whereas the coarse-grained population initially may have clumped distribution (due to
colonization of preferred patches) but later distribution that switches to a uniform pattern as density
increases (occupancy of suboptimal patches are imposed upon the expanding population). [Repro-
duced, with permission, from 1. A. Wiens, Ann. Rev. Ecol. Syst. 7:81-120 (1976). Copyright 1976
by Annual Reviews, Inc.]

roecosystems: niche breadth may actually decrease as resources increase (super

patch) and population density also increases. Wiens(463) summarizes a variety of
investigations suggesting that predators forage so as to maximize patch prof-
itability rather than necessarily selecting individual prey types. This is somewhat
a restatement of the predicted generalist feeding habitats of predators and para-
sites in the lowland tropics. Predators lock into a patch where several different
kinds of suitable prey may occur, and remain there until it is profitable to
discover a new patch with similar properties. Habitats with many patches of
suitable prey will promote considerable dispersal within predators and contact
with many patches of prey.(466) The number of different patch types to be ex-
ploited by a predator or parasite species will be a function of interpatch travel
254 CHAPTER 6

time and the increase in foraging time per prey capture resulting from adding
another patch to the foraging route. (467)
Many tropical habitats will support a large biomass of insects in various size
classes, and these organisms in tum will be prey for predators and parasites.
Distances between patches of suitable prey are expected to be relatively low, and
predators and parasites are expected to move among many patches and patch
types during foraging. Such an effect results in a generalist feeding strategy,
since predators and parasites in such habitats may exploit many different kinds of
prey or hosts without causing the extinction of individual patches.
Wiens(463) has discussed the possible role of a species exploiting many
resources or patch types as a means of decreasing the likelihood of extinction when
only a few resources are exploited. Population stability, it is argued, may be
increased if the species spreads the population's membership over many patch
types, thereby ensuring that the species survives if some patches prove unsuita-
ble. Each patch type presents a unique sphere of selection pressures, and each
patch type may respond differently to local disturbances in the environment.
Thus survival of the population increases as more patch types are added to the
total resources occupied by the population. (463) The ability to exploit a greater
number of resources will be limited to some degree by the character states of the
parameters underlying population growth. Changes in population density may
result in the degree to which subpopulations become differentiated in response to
exploiting different patch typesYl2)
If a species in the tropics has a wide geographical distribution and gene flow
is restricted between many of the populations, there may be considerable evolu-
tionary divergence in response to this broad-scale spatial heterogeneity (re-
gional). Relaxation of gene flow may cloud such effects, however. The expan-
sions and contractions of forests in the Pleistocene in South America have
resulted in mixed distributions of completely and partially differentiated species of
butterflies and other organisms.(468,469) Differences among popUlations, when
considering each population in terms of a "populational phenotype, "(112) result
from selection and drift, processes dependent upon population size and density.
The geographical complexity of tropical regions(16) provides the opportunity for
considerable divergence in a species in response to regional spatial
heterogeneity. Thus the phenotype of the very widespread butterfly M. peleides
in Central and South America varies greatly over large regions (A. M. Young,
unpublished data).
What relationship, if any, such variation has to differences in the environ-
ments from region to region is not known. Phenotypic variation in geographically
widespread species like M. peleides may be related to differentiation of popula-
tions in response to regional environmental effects. The degree to which different
popUlations of the butterfly are continuous among regions is not known. (I use
the term region to refer to large land-mass areas such as northern Central
POPULATION RESPONSES TO THE ENVIRONMENT 255

America, southern Central America, northeastern South America, etc. Within a

region there may be considerable spatial heterogeneity through mountains and
different climatic conditions over large areas, or zones. Thus Costa Rica is a
region within southern Central America in which there is considerable geograph-
ical complexity and elaboration of subregions or zones.) M. peleides exhibits at
least three distinct phenotypic patterns in Costa Rica, corresponding roughly to
major climatic zones of the country. A morph found in some parts of the central
mountains in Costa Rica has deep blue, almost purple wings; an Atlantic
watershed morph has bright blue wings with a black band, and a western coastal
form has almost completely brown wings (A. M. Young and P. DeVries, unpub-
lished data). The boundaries of these forms have not been studied in Costa Rica
or elsewhere in Central America. How many distinct populations there are of the
species within each subregion is not known. Yet an understanding of the ecology
of this insect in Central America necessitates determining distributional patterns
and pinpointing areas of restricted or absent gene flow. Similar analyses should
be performed for other groups of tropical insects as a means of elucidating the
relationship between regional spatial heterogeneity and changes in phenotypic
means for distinct populations. Special attention should be given species with
broad distributions across elevational gradients in the tropics, since environments
along these gradients, in terms of habitat type and number of habitat types, are
expected to change in response to transitions from biologically accommodated
communities in the lowlands to relatively physically controlled communities(104)
in the high altitudes.
Both spatial and temporal patterns of environmental heterogeneity provide
selection favoring the elaboration and maintenance of genetic variation and
phenotypic flexibility within populations of insects. The amount of genetic varia-
tion, for example, is expected to be greatest for insect species requiring the
qualitatively greater shifts in resource utilization patterns in response to environ-
mental fluctuations in space or time. Some tropical specialist species may be
coarse grained in distribution in response to patchy distributions of resource
patches. Such species will experience different selection pressures in different
resource patches if the patches are small; effects of drift may also influence the
direction of adaptation in small patches.(463) Populations may be characterized by
a "popUlational mean" for certain characters, but we need to consider the
character states unique to subpopulations or demes as well, as emphasized in the
recent studies of McCauley and Wade(l12) with laboratory populations of
Tribolium. Genetic polymorphism is expected under such conditions, with dif-
ferent morphs associated with different patch types. For fine-grained species the
environment is encountered as an average of conditions by each individual. Such
conditions favor high phenotypic flexibility within each individual rather than a
mixture of specialized morphs.(463)
Young(275) examined such effects for the relationship between patch struc-
256 CHAPTER 6

ture of host plant resources and the population structure of tropical butterflies. Of
considerable interest is the possibility that fine-grained phytophagous insects
possess a diversity of enzyme forms to degrade plant toxins associated with
different plant species. Different levels of fine-grained responses in plant-feeding
tropical insects may exist: (1) the insect is polyphagous and requires a diverse set
of enzymes to handle diverse sets of plant secondary substances; (2) the insect
possesses a broad monophagy in which it feeds on several closely related plants
or plants with similar secondary substance profiles, requiring a limited number of
enzymes or perhaps one all-purpose enzyme. The first situation might be geneti-
cally controlled by a heterozygous system of alleles, while the latter requires
fewer alleles or homozygosity. But the alternative explanation, namely, that
coarse-grained adaptations imply homozygosity and fine-grained adaptations re-
quire heterozygosity, has also been considered. (463) If considerable diversifica-
tion has occurred among the patches or patch types exploited by a coarse-grained
insect species, then genetic polymorphism may be the optimal strategy of adapta-
tion. The key factors appear to be the degree to which the resource types are
different and the capacity for certain loci to handle each resource type.
It is clear, however, that phytophagous insect species in particular, en-
counter considerable spatial heterogeneity within tropical habitats, as a function
to the plant species richness of these habitats. The overall number of plant
species found in the habitat and the spatial distribution of individuals of each
plant species in that habitat are major factors promoting the evolution of feeding
strategies in tropical insects. If an insect species is locked in as a specialist on a
certain plant species, and if that plant species is too patchy within one habitat but
not in another, the insect will select the latter habitat, based upon a premise of
degree of relative profitability. A complete understanding of the observed pat-
terns of insect species found in a particular tropical habitat will involve the
construction of testable hypotheses related to the following issues:

1. The structure and dynamics of the plant community found in the

habitat.
2. The patch structure of the vegetation within the habitat, in terms of the
patchiness of individual plant species.
3. The degree to which both plants and insects are regulated in abundance
by fluctuations in physical conditions of the local environment.
4. The niche breadth of each insect species and changes in niche breadth
with respect to changes in population density of the insect species, and
similar changes in potentially or actually competing species.
5. Changes in the patch structure of the vegetation over time and possible
results of shifting the degrt;e of food limitations for insects.
6. The frequency of local colonizations and extinctions of individual
species, for both insects and for plants.
POPULATION RESPONSES TO THE ENVIRONMENT 257

7. Assays of the potential feeding capacities of individual insect species

on alternative resources in the habitat.
8. The proximity of other similar habitats as candidates for colonizing
episodes of insect species from the habitat studied.
9. The dispersal abilities of insect species.
10. The degree to which insect species are regulated by biotic control
agents, and an examination of the natural history of these control
agents.

Although much of what is advocated here is strictly natural history in approach,

included is the potential for some interesting simple experiments.
The above discussion strongly implies the need to look at the natural history
of individual insect species in determining why a species is associated with a
particular habitat or community. Mentioned, however, was the shaping of the
distribution of a species in a habitat by interactions among closely related species.
Tropical lowland faunas in particular are full of taxonomically closely related spe-
cies of insects, as well as taxonomically unrelated species that may be ecologically
similar. Many workers have emphasized that many closely related insect species
(i.e., many congeneric species and many allied genera within a tribe or subfamily)
co-occur in some regions of the tropics through divergence in habitat preferences
or modes of exploiting common resources without conflict. The degree to which
each species is a specialist in the habitat, if it co-occurs with other species in the
habitat, is also pertinent.
Three species of African cotton stainer bugs Dysdercus illustrate different
patterns of reproduction and migration that correlate with their differing patterns
of exploiting resources. All three species have the same basic life history pattern:
a short period of intense flight soon after the molt to the adult stage is followed by
feeding, wing muscle degeneration, and maturation of ovaries.(470) Furthermore,
when all three species are deprived of food, wing muscles do not break down and
eggs do not mature: flight continues until food is encountered. The ecological
similarities among these three species of Dysdercus stop here, as each species
has evolutionarily modified this basic life-history pattern to meet its own ecolog-
ical requirements. Dysdercus Jasciatus is the most opportunistic species in that it
feeds on abundant fruit supplied at certain times of the year, reproduces early in
adult life, and has the highest fecundity of the three species. It is a colonizing
species characterized by large adult populations at certain times of the year. At
the opposite end of the ecological spectrum is D. superstitiosus. which feeds on a
greater variety of fruits than D. Jasciatus. reproduces much later than D. Jas-
ciatus. and has a much lower fecundity (1<0). It is a strongly K-selected species
that thrives in a more certain environment, being genetically and physiologically
flexible enough to feed on many different resources; it is a fine-grained species,
while D. Jasciatus is a coarse-grained species with respect to food supplies. The
258 CHAPTER 6

third species, D. nigrofasciatus, represents an intermediate or compromise

strategy between the other two species in that the age of first reproduction and
total fecundity are intermediate between those of the other species, and diet
breadth is also intermediate. The result is that each species possesses a distinct
population structure and is integrated into communities to varying degrees. The
differences among the species may contribute to their ability to co-occur in the
same region of Africa but not necessarily imply that selection favoring such
ecological divergence came from interspecific competition or competition with
other ecologically similar species.
We need to be able to historically distinguish between true habitat selection
and habitat correlation(463) in attempting to understand the present-day life his-
tories and distribution of these insects. Whenever possible, it is also advantage-
ous to consider the previous distributions of the species as well as learn whatever
we can of the past habitats occupied by the species. Present-day selection on a
species in a particular region may have little to do with the kind of selection
previously operative on the population. Tropical communities are characterized
by the great numbers of closely related taxa of insects and other organisms. The
observed distributions may have resulted from past biogeographic events, past
selection pressures and habitats, and the forces of contemporary selection.
The incidence in tropical communities of the occurrence of aggregations of
conspicuously colored insects on plants is higher than in comparable Temperate
Zone communities. Many aposematically colored tropical insects are specialist
feeders on plants rich in toxic secondary substances,(343) and aggregative be-
havior may reinforce distasteful properties to potential visually hunting predators
such as lizards and frogs. The conspicuous aggregations of some phytophagous
hemipterans on different food plants(183.355) in the tropics may also represent
coarse-grained exploitation of a patchy resource. Aggregating behavior may
evolve to keep the membership of a subpopulation together in a resource patch
while at the same time providing for effective defense against predators. I am not
advocating that all patchily distributed insect species on plants have effective
means of deterring predators of the kind discussed here. If the insect is associated
with a food plant type rich in toxic secondary substances, and if the food plant
has a patchy distribution, the juvenile stages in particular of the insect may
develop aggregative behavior to deter predators. Perhaps such behavior was
made possible by the insect's being a coarse-grained specialist on that plant: the
coarse-grained response kept the juvenile stages together on the same plant or in
the same small plant patch and provided a preadaptation to the evolution of true
aggregative behavior of the kind noted for hemipterans such as P. blandus on
some Cucurbitaceae.(355) Adults tend to disperse, facilitated perhaps by a shift to
feeding on seeds and fruits.
Cluster oviposition might have also evolved as a response to a coarse-
POPULATION RESPONSES TO THE ENVIRONMENT 259

grained resource for a specialist insect: cluster oviposition is often associated

with highly coordinated aggregational behavior in phytophagous insects. In the
past evolutionary history of the insect species, the original host plants might
have been very patchy in distribution and the coarse-grained strategy prevailed.
Such conditions provided the opportunity for increasing population or subpopula-
tion densities of juveniles on the host plant patch or individual plant, and this was
followed by the development of adjustments in the phenotypes so as to maximize
survival under crowded conditions, namely, through the evolution of cluster
oviposition and aggregated juveniles (and in some cases adults). If the plant
species involved was very patchy and the subpopulation was associated with a
large patch, some forms of aggregational behavior could be carried over into the
adult stage, such as communal roosting in some tropical butterflies. A key factor
is the unpalatability of insect to predators and the reinforcement of unpalatability
by aggregates on the food plant. Such an explanation for cluster oviposition in
butterflies, for example, is very different from others. (265) Even though today the
insect exhibiting such traits might be exploiting fine-grained resources,the origi-
nal pattern of exploitation, perhaps no longer present, would have been coarse
grained under this hypothesis. Some aspects of this hypothesis are testable in the
tropics today.

6.6. COMPETITION AS A SELECTIVE FORCE IN POPULATION

RESPONSES

Because tropical floras and faunas tend to be markedly diverse and rich in
terms of taxa, it is tempting to ascribe such diversity, at least in part, to the
effects of interspecific competition. Although some forms of character displace-
ment result from competition among species, character displacement or di-
vergence also results from an overabundance of resources rather than from re-
strictions of resources. (471) If species populations are kept well below the carrying
capacities of the environment, and if a suitable resource becomes more abundant,
one or some species may diverge to exploit the abundant resource, even though
the species may not have been competing. Hespenheide's reasoning(471) is well
taken for the tropics, where there can be shifts in harvestable productivity of
habitats, increasing or decreasing resources for species. If harvestable produc-
tivity is increased, a species previously kept well below its carrying capacity
may respond by shifting to exploit the abundant resources. This shift may involve
character displacement even though the species was not in competition with other
species. Individual selection predicts that populations will increase in size in
response to individuals increasing their own fitness through increased reproduc-
tion. One way to increase reproduction is to mobilize more energy from an
260 CHAPTER 6

abundant food supply. When two or more ecologically similar species in the
same habitat are kept below the carrying capacity of the environment, from
predation or parasitism, competition cannot be taking place since resources, even
if limited in supply, are not fully utilized by the species. When species exploit
patchy resources and are fine-grained, sUbpopulation densities associated with
individual resource patches remain low, so that competition with other species is
unlikely. If two or more species are coarse-grained specialists on the same
resource, competition may occur if the populations are not kept down by preda-
tion, parasitism, or disease.
Insects associated with terrestrial habitats in tropical lowlands probably
exhibit various forms of limitation on their populations. Both coarse- and fine-
grained species may be limited in abundance by biotic mortality factors such as
parasites, in which case populations are below K. Food resources tend not to
limit in this case. Competition is unlikely. The fine-grained species probably
experience fewer episodes of resource limitation if biotic controls are not operat-
ive, because their populations are distributed over several resource types.
Coarse-grained species stand a better chance of entering into competitive
interactions with other species, especially if all participants are specialists on the
same resources. Resource limitation may result for such species if biotic control
is low or absent. Phytophagous insects may be phased in and out of such situa-
tions if they are specialists on young or old plant parts. Vegetative growth
patterns of plants may not be continuous throughout the year, especially if a dry
season occurs in the region, so fresh meristem may be a limiting resource for
some insects specializing on both a particular plant species and the new growth
of individuals. Similar patterns may occur for insects exploiting flowers and
fruits, structures which do not necessarily occur evenly throughout the year.
Thus insect populations may fluctuate considerably if resources are fluctuating in
abundance and if the species in question tend to be specialists. If many plant
species respond to the environment in the same direction, even generalist species
may be similarly affected. Detailed natural history studies of each plant species
and its associated insects are needed over a several-year period to determine such
patterns of abundance, and it is necessary to determine the full spectrum of
resources utilized by each of the insects in the guild.
Individual insect species may be temporarily phased in and out of competi-
tive interactions with other insect species, even those feeding on other structures
of the food plant,(93) as a function of fluctuations in the physical environment
affecting vegetative and reproductive cycles in plants. Communities associated
with high-altitude areas of the tropics may exhibit seasonal-type fluctuations in
the abundance of some resources (such as new meristem for pierid but-
terflies(472», but since the overall diversity of insects within a community tends to
be relatively low, competitive interactions are far less likely. Ecological di-
vergence and speciation in high-altitude insects may be relatively infrequent once
POPULATION RESPONSES TO THE ENVIRONMENT 261

these species colonize such regions, because the environment tends to be harsh
even though perhaps predictable in terms of fluctuations. Harvestable produc-
tivity is kept low by the harshness of the physical environment. Unusual
topographic local effects may create localized and patchy habitats of a particular
kind, and these "islands" may serve as sources of speciation in some insects at
high altitudes; other than these, the environment is rather homogeneous and
harsh. Interspecific competition or resource limitation effects on insect popula-
tions are less likely at these altitudes, even though populations may be controlled
by harsh climatic conditions resulting in some density-independent mortality and
suppression of breeding activity at certain times of the year. Clearly, long-term
studies at high altitude regions of the tropics are needed to examine these ideas,
as the data base is lacking.
Although there are cases of competition among parasitic ichneumonids for
the same host,(473) the predicted generalist feeding strategies of endoparasitic
insects in the lowland tropics as discussed in this book probably reduce the
intensity of competition among species for any single host species. Clearly many
instances of multiparasitism are likely in tropical insects, but parasitoids may be
largely fine grained with low investment per host patch. The same may be the
case with insects and other arthropods predatory on other insects. However,
some predators, such as ambush bugs requiring a certain kind of flower to sit in,
may experience resource limitation if flowers are not always available in the
habitat. In the case of sit-and-wait predators such as ambush bugs, perching sites
such as certain species of flowers may be a limiting factor in their populations,
with annual periods of abundance of these insects corresponding to periods of
flowering. If there is a temporal succession of different but suitable species of
flowers for these bugs throughout the year, populations may stabilize. Highly
mobile predators such as asilid flies and various solitary or subsocial hymenopte-
rans may show less responses. Insects which visit certain flowers for nectar or
pollen may be capable of facultative shifts to other flowers as species composi-
tion of blooms changes throughout the year. Insect species incapable of such
facultative switches within and among adjacent habitats may exhibit large fluctu-
ations in population sizes at different times of the year. Fluctuations in the
physical environment may also limit populations of pollinators in some in-
stances. Although cacao (Theobroma cacao) trees flower throughout the year
with some fluctuations, regions in Africa with a strong dry season adversely
affect populations of the pollinating midges (Ceratopogonidae), presumably as
the result of dessication effects on these small insects. Pollinator activity de-
clines, and fruit set is even lower than usual. Small-bodied pollinating insects in
seasonal lowland tropical regions may exhibit considerable fluctuations in abun-
dance between dry and rainy seasons, and fruit set in preferred plant species may
also reflect these events.
One of the problems in attempting to explain the distribution of closely
262 CHAPTER 6

related species in tropical communities is the lack of information of the history of

species in previous environments. Closely related species in the same region may
exhibit various forms of allopatry or microallopatry, but it cannot be concluded
that such distributions result from competition. Biogeographical events and op-
portunistic responses of species to new and abundant resources(471) may result in
shifts in distribution having little or nothing to do with interspecific competition.
But Janzen(93) has called attention to a much more subtle kind of competition that
may be prevalent among species in tropical communities. Janzen argued that
unrelated insects utilizing different portions of the same plants within a popula-
tion of a plant species may be in competition with one another. Insect species A
may be attacking the leaves of plant individual A while insect species B is
attacking the roots of individual B of that plant species, and since each insect
species takes energy away from the plant species, competition is taking place
among the species. The combined effects of all such species attacking the plant
population, from the viewpoint of anyone species in the interaction, is increasing
resource limitation because each species will have less energy available through
successive generations. The studies of Rockwood(474) demonstrated that increas-
ing herbivore attacks on a plant species lowers the fitness of the plant through
reduced production of fruits and seeds. Janzen's concept represents a "super-
herbivore effect" arising from many different herbivores removing energy
(biomass) from the plant, which in tum reduces its fitness. As a result it becomes
less attractive a resource for each predatory insect species.
Andrews(475) presented some interesting data on the arthropod communities
associated with cacao plantations in Costa Rica and Dominica that illustrate
another aspect of community structure. She found a greater range of arthropod
species and numbers per species in Costa Rica for the same number of sweeps
taken in the two localities (Figs. 6.15 and 6.16). The Costa Rican (La Lola)
plantations seem to support a larger community of arthropods than the Dominica
plantations, with ants and diperans forming the bulk of species in both localities
(Table 6.5). Yet the Costa Rican samples had greater percentages of other insect
orders such as beetles and lepidopterans (Table 6.5), suggesting a greater degree
of complexity in community structure assuming that these species are residents of
cacao plantations. One is tempted to suggest that the likelihood of competition is
greater in the Costa Rican cacao plantations, especially for a group such as ants.
Yet the data may reflect differences in habitat complexity within cacao planta-
tions having little to do with competition. Mainland tropical communities are
expected to have a greater number of insect species than island communities,
even though trophic structure may be different. (415) The great diversity of poten-
tial nesting and feeding sites for ants in cacao plantations may be the consortium
of environmental features promoting ant diversity. (476)
Carroll and Janzen(l96) discuss the relationship between ant diversity and
availability of nesting sites in the tropics. Ecological diversification may have
POPULATION RESPONSES TO THE ENVIRONMENT 263

,
is.
CII
1000
CII
~
V'\ 500
0
~
-00
a. 100
0
.£
4 50
'0
'"
4i I
-'l I
E I
:> 10 I
Z I
I
I
5
-.., r---,
I
1-1 r ...
I I
I I
1 I I I
I I I I
I I
I I
I I
I I I I
1
4 8 12 16 20 24 28

Arthropod length (mm)

FIGURE 6.15. Size-frequency distributions for arthropods collected in the understory foliage in cacao
plantations in Dominica (solid histogram) and Finca Experimental La Lola, Costa Rica (dashed histo-
gram), for all forms shorterthan 28 mm. [From R. M. Andrews, Breviora, No. 454 (1979). Copyright
1979 by Harvard University Press.]

nothing to do with competition in such instances, but rather be the result of

different species encountering different opportunities for fitting into the prevail-
ing environment. In the case of cacao plantations, resident insect species bring to
these agroecosystems the host of ecological traits evolved in natural habitats
before the plantations were established. Depending upon the type of shade cover,
number of tree species providing shade, and other factors, cacao plantations vary
considerably in complexity and as environmental mosaics for insects.
Competition may in some instances playa prominent role in determining the
diversity of a particular group of insects, as illustrated by Johnson and Hubbell's
study of stingless bees.(433) Niche shifts involving insect species making tran-
sitions to secondary habitats from forest habitats in the tropics may be another
example. Young and Moffett(221,222) suggested as a working hypothesis that the
ithomiid butterfly M. isthmia may have moved onto tough and hairy solanums in
264 CHAPTER 6

10

a
4>
5

4>
~
0
0 ,,-,,
.. ,, ,-, ,-, ,,-' :,
0
.:::::

'- , ,, ,.- ...... -. ,,

,, ,-
,- ,
-0
0 5

.£
C-
0
J , _1-
'-'
~
j:
1
-,
a>
~
>0- 05
Ci

01
4 B 12 16 20 24 28

Arthropod length (mm)

FIGURE 6.16. The distribution of dry weights for different size classes of arthropods censused
from cacao plantation foliage in Dominica and Costa Rica (solid and dashed histograms, respectively).
[From R. M. Andrews, Breviora, No. 454 (1979). Copyright 1979 by Harvard University Press.]

pastures in response to competitIOn among ithomiids for thin, papery-leaved

solanums in forest habitats. Most ithomiids exploit such solanums, and cases
exist where more than one species exploits the same plant species. (290,302,304,310)
Furthermore, woody forest trees and shrubs may be less attractive to insects as
host plants than some related species in secondary habitats. (477) Forest solanums
are small trees or shrubs, while pasture solanums tend to be spine-covered
herbaceous plants. Mechanitis might be one of the few ithomiids capable of
making the transition from a less suitable resource base to a more suitable one
(assuming the results of Scribf'r and Feeny<477) for Temperate Zone plants) even
though the defenses of the pasture solanums may be more difficult to overcome
initially. If there was ecological convergence by ithomiids upon the more suitable
of forest solanums in response to many others being unsuitable,(477) competition
may have resulted among some or all of these species. Forest ithomiids tend to
lay few eggs, and larval densities per species are generally low (A. M. Young,
unpublished data), suggesting an adjustment in population growth parameters to
reduce competitive interactions. These species may be K strategists in the sense
of allocating, by necessity, considerable energy to overcoming the nonnutritive
properties of the host plants, thereby reducing reproduction. Such effects lower
the carrying capacity for each species and reduce the likelihood of competition.
POPULATION RESPONSES TO THE ENVIRONMENT 265

TABLE 6.5
The Composition of Arthropod Collections Made by Sweeping
Understory Vegetation on the Dominica and Costa Rica Cacao Sites. a •b

Dominica Costa Rica

Cacao Forest Cacao Forest

foliage foliage foliage foliage

Lepidoptera larvae 0.2 0.0 1.9 0.3

Lepidoptera adults 0.3 2.7 5.8 0.7
Orthoptera 1.9 0.5 2.3 4.1
Blattidae 0.1 0.0 0.0 1.1
Hymenoptera: Formicidae 40.8 52.5 10.3 21.4
Hymenoptera: other 10.7 3.1 18.5 12.1
Homoptera 5.5 5.8 10.6 6.4
Hemiptera 1.5 0.0 1.8 0.7
Coleoptera adults 1.2 2.3 13.3 11.2
Diptera adults 32.3 20.6 29.3 21.7
Holometabolous larvae 0.1 0.0 0.1 0.1
(except Lepidoptera)
Araneida 4.5 11.8 5.5 6.3
Isopoda 0.0 0.0 0.1 2.4
Isoptera 0.3 0.0 0.0 9.6
Miscellaneous 0.5 0.7 0.5 1.9
Sweeps 700 450 250 350
Total arthropods 3552 3180 1952 1219
Number of 5 mm/lOOO sweeps 243 44 972 526

UProm Andrews.(475)
'Values are percentages of total numbers in each collection and do not include collembola and mites.

In such a system, in which one or a few resource types in the forest understory
are more or less saturated with butterfly species, there will be opportunistic
selection for some species to migrate to new habitats as they appear and where
suitable host plants occur. The creation of pastures in the lowland tropics of
Central and South America may have provided the opportunity for a few
ithomiids such as various species of Mechanitis to invade these habitats and
overcome an ecological barrier. Such transitions may represent population re-
sponses to competition in the original habitat, although all of this is speculation.
Indirect evidence is provided, however, by the observations on the natural-
history and host plant exploitation patterns of ithomiids today. (290.298.30L302.304.310)
The transition might have been done by a pre-Mechanitis form, and the
genus may have subsequently undergone its major adaptive radiation in tropical
pastures.(478) Other butterflies and tropical cicadas may also be experiencing
niche shifts, some of which may result from competition. (71.138) But such sugges-
266 CHAPTER 6

tions require confinnation with long-tenn field studies and analysis of the extent
to which resources of individual species are limiting in different adjacent
habitats. It is necessary to differentiate between noncompetitive responses to new
resources or changes in the abundance and accessibility of original resources and
to actual competitive interactions when examining shifts in insect species to new
resources within a habitat or to new habitats. Regardless of origin, such shifts
may involve drastic changes in population structure and dynamics. But a shift
arising from a competitive syndrome for a species may be more challenging since
the species was probably strongly K-selected in the original habitat. Species with
populations characterized by a mixture of r- and K-selected genotypes may have
an easier time making such transitions, and such species would be primarily
those responding to new resources without coming from a sphere of strongly
competitive influences.

6.7. SPATIAL ENVIRONMENTAL HETEROGENEITY AND DEMES IN

THE TROPICS

The lowland tropical wet or dry forests consist of mosaics of different

habitats, each possessing some dynamic pattern of plant species and insects
associated with them. Each habitat has a characteristic distribution of plant taxa,
and those with many species with patchy distributions can be tenned species rich.
Species-rich tropical habitats may include both primary and secondary growth.
Janzen(29) has pointed out the patchy distributions of tree species in tropical
forests, showing that the observed patchiness could be explained by specialized
fonns of seed and seedling predation by insects and other animals, coupled with
the effects of long-distance pollinators promoting outcrossing.
In habitats where plant species are abundant and patchily distributed, some
insect species may have their populations subdivided into demes. A deme of a
phytophagous insect species may be associated with one or few resources patches
(e.g., host plants), with the demes being separated by the patch structure of the
habitat. An insect species may be associated with resources in several adjacent
habitats, in which case the deme structure is distributed across those habitats.
Wright(448) pointed out that the subdivision of a popUlation into small demes is an
effective means to bring about rapid evolutionary divergence if gene flow among
the demes is at least partially restricted. In the lowland tropics in particular,
whole guilds of insects may be associated with the patchily distributed individu-
als or clumps of a host plant, and under such conditions one might speak of a
fonn of community evolution: mUltispecies superdemes may diverge considera-
bly in response to localized selection pressures affecting all members of a guild.
Such effects may be more pronounced in the lowland tropics, where many plant
species may have patchy distributions over one or more habitats in the region.
POPULATION RESPONSES TO THE ENVIRONMENT 267

There is some evidence of differentiation among demes for some Temperate

Zone species of flies(391) in which species are specialists on certain resources. In
some tropical habitats and regions we expect to encounter many specialist insect
forms, particularly in the case of phytophagous insects: the rich diversity of
floras in some regions promotes ecological diversification in phytophagous in-
sects associated with them, even though in some cases a breeding population of a
species becomes partially subdivided as a consequence of host plant individuals
or clumps being highly dispersed. Such conditions may reduce some gene flow
within the population. Ecologically enforced partial isolation of this kind for
some insects may result in various levels of drift effects depending upon deme
sizes, at the same time buffering against environmental perturbations.(448)
The extinction of demes can be high if environmental perburbations are
intense and frequent. The spatial distribution of partially isolated demes for a
species promotes and maintains some forms of genetic polymorphisms as a result
of disruptive selection. Thus if we discovered an insect species with a distribu-
tion on a plant population associated with several habitats in a lowland tropical
region, we would expect to find some degree of isolation among subgroups, as is
known for some species of Heliconius butterflies(479) and probably also occurring
in other specialized insect groups. The constancy in membership in some com-
munal roosts of these butterflies may provide some indirect support for the deme
concept as applied to tropical insects. (190,479) Yet sometimes whole roosts of these
butterflies are disturbed by environmental factors to the point that they disappear
from an area. (455) Demes in phytophagous insects and in other specialized insect
species in the tropics may represent an enforced diversification in response to the
patchy spatial distribution of critical resources necessary to maintain a breeding
population. But such demes may be vulnerable to frequent extinctions in habitats
where unpredictable environmental perturbations occur. Yet there is also evi-
dence that some demes of tropical butterflies are very stable over long
periods. (451)
In strongly seasonal tropical environments, demes may expand and contract
periodically and according to season. Thus we might envision that a population
of the common nymphalid butterfly Anartiafatima in the northwestern lowlands
of Costa Rica is distributed rather evenly over large pastures and young
secondary-growth areas during the rainy season. These butterflies exploit a va-
riety of herbaceous Acanthaceae as larval host plants, (322) and these resources
tend to be abundantly distributed over large areas of secondary growth and
pastures in many regions of Costa Rica. Population density of adult butterflies may
be quite high in an area. But during the 6-month dry season of this region, food
plant patches disappear or shrink in size owing to cessation of vegetative growth
and initiation of leaf drop. Such a drastic change in the host plant population
results in a contraction of the adult population into many small pockets associated
with edges of moist forest patches and similar areas. Late into the dry season the
268 CHAPTER 6

butterflies enter into reproductive diapause. There might be some breeding in

these subpopulations or demes early in the dry season, and there might be some
gene flow among them. But late into this season the demes cease breeding and
gene flow is cut off. In the following rainy season the populations again expand,
but the genetic structure of the population may have been changed by "sampling
error" arising from deme formation in the dry season. Demes of different size
may be subject to different kinds of mortality during the long dry season, enhanc-
ing shifts in gene frequencies by the beginning of the rainy season. Similar
effects may occur in other insects during severe tropical dry seasons in the New
and Old World tropics.
Some species of Lepidoptera in the Temperate Zone with broad geograph-
ical ranges often exhibit popUlations with varying amounts of genetic variation as
measured by allozyme patterns.(480) At the same time, there may also be some
evidence of substructuring of individual popUlations as reflected by a degree of
intrapopulational variation in allozymes. (480) Depending upon local population
sizes, the observed patterns of allozyme variation result from differential effects
of natural selection and genetic drift. In the moth Yponomeuta studied by Men-
ken, (480) the gregarious habits of the larvae are a factor in promoting the substruc-
turing of breeding populations. Gregarious larvae or nymphs are found in many
species of Temperate Zone and tropical insects, and together with the fractionat-
ing effects of spatially patchy plant species may result in considerable substruc-
turing of breeding populations (subdivision of demes). On the average, many
herbaceous plant groups have very patchy distributions in some tropical habitats,
as do tree species,l29) and therefore a greater incidence of substructuring of
popUlations might occur in such environments. Alternatively, some insect
species may exhibit considerable phenotypic flexibility of a nongenetic form in
exploiting different microenvironments along a gradient of resource patchi-
ness,(481) and in such cases drift may be the major component governing shifts in
genetic composition within subpopulations.
The dry season (see Chapter 7) is a periodic form of environmental perturba-
tion affecting the abundance and distribution of plants in some habitats. Such
patterns of temporal heterogeneity create changes in the spatial heterogeneity of
some resources for insects, which in tum may result in changes in the genetic
structure of insect popUlations. Such popUlations may be subject to various forms
of directional selection and random genetic changes (drift) that shift in direction
and intensity between seasons according to variables such as initial population
size, genetic variation in the population, dispersal, and the extent to which
resources such as food plants are affected in abundances (popUlation sizes, den-
sities) and distribution (patchiness). In the Temperate Zones some insects ex-
hibit polymorphisms that correspond to the exploitation of different kinds of
resources over sizable geographical regions. Halkka(482) showed that different
colored morphs of a spittle bug Philaenus are found in different kinds of host
POPULATION RESPONSES TO THE ENVIRONMENT 269

plant patches, and that gene flow is somewhat restricted among these demes. The
morph found in each deme is adapted to feed on a particular combination of host
plants (' 'minimeadows"). Dissimilar degrees of dispersal ability between adults
and nymphs preserve the polymorphism in response to ecological dissimilarities
among widely scattered host plant patches, but speciation is prevented by some
exchange of adults. In the tropics we might expect to encounter many examples
of such effects since many plant species have patchy spatial distributions in some
habitats. Studies of groups such as the Homoptera, Hemiptera, and Orthoptera
may reveal some interesting patterns of deme structure and quasi-isolation among
demes and how such patterns vary between seasonal and relatively nonseasonal
regions. There is probably a greater occurrence of effects similar to those de-
scribed by Halkka in the Temperate Zone for many lowland tropical areas and
involving a greater number of species of insects associated with plants.
Whenever individuals in the population of a given species enter into a period
of colonization of new habitats or regions, the species often encounters a change
in the spatial heterogeneity and the way the species responds to such a shift in the
environment is also different. It is important to bear in mind that colonization is a
relative condition of one population compared with another of the same or
closely'allied species. If a population of a particular species is highly subdivided
into several distinct groups or demes, the dispersal among these units is low or
absent, one may speak of colonization episodes associated with particular demes
rather than with the entire population in the habitat or region (collection of
habitats). For a given insect species, we can envision two general types of
ecological situations in the tropics that promote colonizing episodes.
First, man-induced changes in habitats may result in a mixture of extinctions
and colonizations among the insect species affected by the disturbance. If popula-
tions are small and fragile, if genetic variation for dispersal-prone genulypes
is low, and if feeding flexibility is low, we would expect the species to experi-
ence a localized extinction at a rate directly proportional to the size and age
distribution of the breeding population. If thtse ecological traits are of the oppo-
site character states, colonization is more probable. But the chances of success in
colonization are low. (87) Second, recurring localized catastrophes brought about
by natural conditions may lead to colonization. In the tropics, these may be
classified as landslides, floodings, volcanic eruptions, and hurricanes. These
agents function in a density-independent fashion to alter large portions of vegeta-
tion cover in some regions. Schwerdtfeger,(45) for example, analyzes the prop-
erties of hurricanes affecting northern Central America and the Caribbean ar-
chipelagos. Localized landslides may alter the vegetation cover in a particular
habitat, making the area potentially available for colonization by other or-
ganisms. For example, in the central highlands of Costa Rica, frequent small
mud slides provide bare patches of earth that are colonized by seeds of
Machaerium seemani, a leguminous vine, within a few months (A. M. Young,
 270 CHAPTER 6

BIOTIC
POTENTIAL

ill Y ~TlC
I-~~------------------------~
PERIODICITY

FIGURE 6.17. Four general theories of the

regulation of population size. Broken line in
periodicity section: cycles of abundance re-
sulting from population's being exploited by
POPULATION SIZE a predator or parasite.

unpublished data). These young vines in tum are colonized by the butterfly M.
peleides (Fig. 4.2), a species adapted to a variety of secondary habitats.(311,316) In
more stable secondary habitats in the region, the butterfly oviposits primarily
on mature leaves of Mucuna urens and several other adult-size leguminous
woody vines and trees. (16 ) The species appears to opportunistically exploit
young seedlings of Machaerium as they appear in patches following landslides
and the construction of new road cuts. Such colonization is somewhat different
from full-scale colonizing episodes since the species is probably only occasion-
ally trying to exploit new habitats and resources, while the major bulk of the
population continues to thrive in more stable conditions. Yet we may expect that
some insect species are in fact experiencing colonizing episodes in response to
changes in the availability of optimal resources in existing habitats. As habitat
quality declines for a particular species, the likelihood of colonization increases,
i.e., migration to another habitat. Colonizations shift the species richness of insects
in some tropical habitats in response to changes in the distribution and abundance
of resources. Comparisons of species experiencing colonizing episodes with allied
species not experiencing colonizing episodes under the same or similar condi-
tions are needed to elucidate the ecological and genetic traits fostering coloniza-
tion. (483)
POPULATION RESPONSES TO THE ENVIRONMENT 271

It is the interplay of intrinsic and extrinsic properties of insect populations

that results in the observed pattern of abundance within a breeding population of
a species and the temporal and spatial changes in the population. In this chapter
an attempt was made to elucidate some of the mechanisms that regulate insect
populations, even though such mechanisms will vary considerably among
species, habitat, and geographical region. The array of genotypes in a population
during a single generation endows the species with a "biotic potential" or the
ability to increase population size. The biotic potential usually results in a level-
ing off of abundance (Fig. 6.17) as the result of density-related factors curtailing
the biotic potential. Regionally, various kinds of climatic factors promote the
increase or decrease of the population, and the interaction of the species with its
resources and predators or parasites may generate cycles of abundance over time
(Fig. 6.17). When the population size is largely a function of prevailing climatic
factors and their fluctuations, the population will exhibit periods of both excep-
tionally low and exceptionally high abundances at different times (Fig. 6.17).
The analysis of the census histories of tropical insect populations should take into
account the varied sources of regulatory factors affecting anyone population of a
species.
 CHAPTER 7

EFFECTS OF SEASONALITY
ON INSECT POPULATIONS
IN THE TROPICS

7.1. TEMPERATE ZONE AND TROPICAL SEASONALITY

The entomological and natural history literature is replete with examples of the
effects of Temperate Zone cycles of annual seasonality on insect populations. In
the Temperate Zone regions there is a succession of four major seasons at most
latitudes, although the duration and intensity of each season vary considerably.
The growing season or northern summer alternates with the winter season, a
period of little or no growth in plant communities and associated animal com-
munities. The transitional seasons of spring and fall are periods of buildup and
decline, respectively, of plant and insect populations at northern latitudes. As
discussed recently by Lawton,c42S) there is considerable interest in how the
dynamics of Temperate Zone plant communities influence the diversity of
phytophagous insects throughout the growing season.
Seasonality in the temperate zones influences plant species in at least two
ways significant to insect populations associated with them: (1) herbaceous plant
species with few or no effective chemical defense systems build up large quan-
tities of new biomass each year, a major portion of which is harvestable by
insects; and (2) forest trees have most of their biomass already in place each
growing season but show seasonal changes in the production of chemical de-
fenses such as tannins.(258) In regions of the world where abiotic factors fluctuate
considerably and unpredictably, phenotypic differences among individuals
within a population, as related to feeding, disease resistance, and other factors,
become of paramount importance in establishing and maintaining populations.
For different reasons, a spectrum of phenotypic variation within populations of
tropical insects provides the flexibility for responding to shifts in biotic factors
affecting average fitness. Lawton,(425) citing other studies, presents data showing
that herbaceous species exhibit a gradual buildup of phytophagous insect popula-
tions throughout the growing season (data from Price(484) and others) and that
insect species on forest trees exhibit a rapid increase in diversity at first, followed

273
274 CHAPTER 7

by a period of decline and some subsequent recovery (data from Feeny<258l).

These effects are stylized in Fig. 7.1. A greater number of insect species are
found on trees and woody vines than on herbs in both the temperate zones and the
tropics, because trees and woody vines possess a greater biomass per species and
greater structural complexity. Thus trees and woody vines have more "mi-
crohabitat space" for insects than herbs. This difference between woody plants
and herbs for building up insect communities is most pronounced in seasonal
environments since populations of individual insect species have a relatively
shorter annual time period (compared with the nonseasonal and benign tropics) to
build up and maintain breeding populations prior to the arrival of the unfavorable
season.
As shown in Fig. 7.1, some interesting similarities in the seasonal distribu-

TEMPERATE TEMPERATE
(Forest trees) (Herbs)

f/)
w
U
W
SUMMER WINTER SUMMER WINTER
Q. TROPICAL NON-SEASONAL TROPICAL NON· SEASONAL
f/) (Forest trees & vines) (Herbs)
I-
U
W
f/)
3!;
u..
0
a:
w
III
:::E
::I
Z TROPICAL SEASONAL TROPICAL SEASONAL
(Forest trees & vines) (Herbs)

RAINY DRY RAINY DRY

SEASONS SEASONS

FIGURE 7.1. Hypothetical model for the effects of seasonality, or lack of it, on the species richness
of herbivorous insects associated with forest trees and herbs.
EFFECTS OF SEASONALITY ON INSECT POPULATIONS 275

tion of insect species are predicted for both Temperate Zone and strongly sea-
sonal lowland tropical regions: the tropical dry season overall is a period of
reduced abundance of insect species, whereas the northern winter generally
results in a complete cessation of insect activity on vegetation at most latitudes.
In the tropics a generally higher harvestable productivity, is expected particularly
in herbeceous secondary growth plant species, resulting in a greater accumula-
tion of insect species as the rainy season advances (Fig. 7.1). This effect may
even be greater in the nonseasonal tropics for herbaceous plants, since vegetative
growth continues to some degree throughout most of the year (Fig. 7.1). Insect
faunas of woody vines and forest trees may be relatively less rich in both the
seasonal and nonseasonal tropics, if annual vegetative growth is less relative to
herbaceous plants in the same region and if trees and vines elaborate large
concentrations of defensive compounds in a manner similar to that found in
Temperate Zone oakS.(258)
The predicted depression of overall insect species abundance during the
tropical dry season in both the seasonal and nonseasonal tropics (in the latter there
is usually a short and erratic veranillo) does not imply that all species decline:
there is some replacement of species in that some become more active than others
in the dry season, but in most cases the net result is a decline. Excluded from this
reasoning, of course, is the concentration of many species into moist forest
refugia during the dry season in strongly seasonal tropical regions. Janzen and
Schoener(l34) reported that many small-bodied insects become less abundant in
the dry season in lowland Guanacaste Province, Costa Rica. Large-bodied bee
species, however, become more active during the dry season of this region since
many plant species are flowering at this time.(116,118) The depression in insect
species during the tropical dry season at some localities is probably heavily
skewed toward those with small body size distributions. Comparing the seasonal
and nonseasonal tropical regions, fluctuations in species abundance may be more
noticeable in the strongly seasonal regions since the popUlations of individual
insect species are believed to be synchronized with the rainy season (Figs. 7.1
and 7.2). Such oscillations are expected to be less noticeable in the nonseasonal
tropics since many overlapping generations may occur throughout the year.
Several overlapping generations are expected for the seasonal tropics during the
growing season, and because defensive systems against herbivores are observed
to be well developed in both trees (and woody vines) and herbs, popUlations sizes
are expected to be similar for both kinds of plant growth forms. The nonseasonal
tropics provides a greater opportunity for predators and parasites of herbivores to
regulate their host (prey) populations, thereby keeping the populations of these
insects well below the carrying capacity of the habitat(s) for the average herbi-
vore species.
It must be kept in mind that this is an attempt at a general synthesis, and that
the nonseasonal tropics will contain many insect species with marked cycles of
276 CHAPTER 7

IEMPERATE TEMPERATE
(Forest trees) (Herbs)

II)
w
irlII.
II)

....
~~F-----S-U-M-M-ER------rl~W-IN-TE-R~~----S-U-M-ME-R----~~W~I-NT-E-R~
a:: TROPICAL NON-SEASONAL I TROPICAL NON-SEASONAL
~ (Forest trees A vines) : (Herbs)

W I
N I
iii I
I
Z I
Q I
!c :
5~--------------~~ ______ 1

i~--------------~----~--------------~------~
TROPICAL SEASONAL TROPICAL SEASONAL
Q (Forest trees & vinas) (Herbs)

~
::::i
c(
a::
w
Z
~

RAINY DRY RAINY DRY

SEASONS SEASONS
FIGURE 7_2_ Hypothetical model for the effect of plant growth form and seasonality on the popula-
tion census histories for plant-feeding insects in the tropics and Temperate Zone. On a per species
basis, population size is probably greater during the growing season in seasonal environments, and in
theTemperate Zone a species may be multivoltine (solid lines) or univoltine (dashed line) depending
upon the developmental time and absolute length of the growing season.

species abundance and population size fluctuations. Critical factors determining

such cycles include the mean generation time of the population for a species the
degree to which its food plants are affected by slight or strong seasonal effects
throughout the year, and the degree to which the predators and parasitoids ofthe
insect species are also affected by seasonal changes in climate.
We need to examine the popUlation census histories of individual insect
species over several years at a locality to determine the extent that populations
fluctuate and how fluctuations correlate with changes in the local environment.
Such data are generally lacking for most tropical insects. Tropical populations in
general are believed, in most species, to have many overlapping generations,
EFFECTS OF SEASONALITY ON INSECT POPULATIONS 277

although Temperate Zone species may be univoltine or multivoltine (Fig. 7.2).

Tropical insect populations, especially those associated with herbaceous vegeta-
tion, are predicted to build up more rapidly than comparable Temperate Zone
popUlations because the first rains of the rainy season result in a sudden and
massive flush of new growth in many plant species(116) and the rate of increase in
new harvestable biomass from plants is generally higher. Temperate Zone plant
communities are marked by a more gradual buildup in harvestable biomass,
particularly in herbaceous species.
In the strongly seasonal tropics, the number of active insect species may
surge at the beginning of the rainy season, followed by a gradual decline and
quasi-equilibrium near the end of the rainy season (Fig. 7.1) resulting from a
reduction in harvestable biomass, maturation of surviving foliage, and the bring-
ing into synchrony of biotic control agents on phytophagous insect popUlations.
In tropical regions insect communities on vegetation are predicted to come into
some form of equilibrium with resources as food quality diminishes in the latter
half of the rainy season, and the pattern will be similar to that presented by
Lawton(425) (his Figure 7.10) for Temperate Zone communities.
Not all insect orders will respond to seasonality in the same manner, since
feeding habits differ greatly. J anzed 72 ) suggested that plant-sucking bugs exhibit
fewer gradients in species composition among habitats in the tropics since they
probably come into contact for fewer defensive compounds than chewing insects
such as Lepidoptera or Coleoptera. As host plants dry up in open fields in the dry
season, homopterans may move into moist forest pockets, undergo reductions in
population size, and feed on plant species very different from their original hosts.
Homoptera are generally less abundant in forest understories in the tropics since
there is less biomass of soft meristems for feeding.(72) As with Temperate Zone
faunas, orders such as Lepidoptera and Coleoptera(425) are expected to be the
most abundant groups in both forest trees(413) and herbaceous plant communities
in the tropics.
The greatest number of discontinuities in species within these orders are
predicted in moving from one kind of habitat to another in the same region,
and by virtue of their feeding habits Coleoptera and Lepidoptera are specialists
on a narrow range of plant species. Thus the spatial rearrangement of the
lepidopterous and coleopterous (and others such as Orthoptera) subsets of the
local phytophagous insect communities in the tropics will be most drastic in
strongly seasonal regions since both trees and herbaceous species may respond
considerably to seasonality Y 16) Both temperate and tropical communities of
bracken-exploiting insects are overwhelmingly dominated by chewing insects,
notably Lepidoptera and Coleoptera. (425) The influence of the tropical dry season
is probably greatest on chewing insects if these insects exhibit host plant selectiv-
ity on the basis of chemical and other defense properties. In temperate forests,
disturbed forest patches where the canopy is thinned have higher rates of removal
278 CHAPTER 7

of fleshy fruits and appear to be bimodally synchronized with the end points of
the growing season, periods when damage to fruits by insects and other inverte-
brates is minimal.(485) Light gap patches in mildly seasonal or nonseasonal tropi-
cal forests tend to have a high number of bird species as well as insects, and
because of these conditions, both fruit damage and fruit removal are expected to
be high throughout most of the year, especially since a succession of fleshy fruit
types might be available over the yearY°l)
The above considerations suggest that the tropical dry season, where and
when severe and long in duration, disrupts insect populations for many species
by reducing the population and reducing the overall species richness. If the
number of species remained about the same during such a dry season, and only
population sizes of individual species declined, species diversity would of course
increase, as calculated from the well-known formulation
log H = !. Pi logpi
where H is the index of diversity, and Pi is the proportion of individuals in the ith
species.(3) Yet biologically a species may be replaced with others, and the ob-
served net changes in community structure with regard to insects would be due to
factors such as physiological constraints of small body sizes and constraints
imposed by specialized feeding habits and requirements.
Examining other levels in the trophic organization of ecosystems indicates
that changes in the species composition and population sizes of various
phytophagous insects will influence the distribution and abundance of predators
and parasitoids on these forms. Again, there might be a replacement of species
from season to season in some cases: parasitoids of large-bodied insects may
become more active in the dry season if they can remain in a moist enough
microhabitat for sufficient periods to allow breeding. Then again, parasitoids and
predators may shift to moist forest refugia as a means of "tracking" their hosts.
The average amount of time a tachinid fly can spend searching for a succulent
lepidopteran larva in an overgrown pasture in a strongly seasonal tropical region
is probably far less in the dry season that it is during the rainy season. The same
is true for braconid wasps and other forms.
Communities of scavenging insects such as ants and some Coleoptera may
also shift if the abundance of dead or dying insects and other suitable animals is
less during the dry season in some exposed habitats. Spatial escapes of these
kinds for many insects in different trophic levels of tropical ecosystems may
allow sustained breeding activity in suitable (even if suboptimal) moist forest
refugia during the dry season. Reproductive diapause is another form of escape in
which an insect popUlation escapes in time and waits out the dry season and until
resources are once again abundant. The incidence of reproductive diapause is
predicted to be generally high in specialized plant-feeding insects with chewing
mouth parts and in ones which plant resources are greatly curtailed as the dry
EFFECTS OF SEASONALITY ON INSECT POPULATIONS 279

season advances in a strongly seasonal tropical region. The physiological

mechanisms and environmental cues bringing into phase such diapause responses
to a time of environmental stress will be different in the tropics than in the
temperate zones. Tropical dry seasons are not accompanies by noticeable
changes in photoperiod length on a daily basis,(45) but perhaps subtle changes in
relative humidity and temperature and large changes in moisture somehow trig-
ger diapause and the behavioral syndromes that accompany it in tropical insects.
The enzyme systems involved in physiological responses to the severe tropical
dry season must be qualitatively different from those regulating diapause re-
sponses in Temperate Zone insects in which other environmental cues are in-
volved.
When a species has a broad geographical distribution encompassing both
tropical and extratropical regions, it is often necessary to determine the
mechanisms underlying differential adaptation to the different kinds of environ-
ments present. A major distinction is the necessity for an overwintering diapause
mechanism in temperate popUlations of such an insect species, while individuals
in popUlations occupying highly seasonal tropical regions require a different kind
of mechanism to pass the dry season. From recent studies of Central American
and Temperate Zone locust popUlations ,(486) when such divergent phenotypes are
required in the temperate vs. tropical populations, strong selection is predicted to
be against hybrids along geographical borders between the two regions, since
such forms presumably would not be adapted to either kind of environment. It is
not a question of cIinal gradation from one kind of environment into another, but
rather an abrupt change. Selection, under such conditions, favors the expression
of effective isolating mechanisms between the populations as a means of reduc-
ing the frequency of possible hybrid forms.
Janzen(116) has stressed the synchronization of flowering with the dry season
in tree species of the tropical dry forest zones of Central America (see also
Section 7.3). The timing of peak flowering in many tree species with the dry
season requires that pollinators be active at this time of the year. In fact, one
could argue effectively the other way around as well: it is optimal for pollinating
animals to be active at this time of the year, and tree species therefore syn-
chronize peak flowering with it. Regardless of cause-and-effect direction, selec-
tion has resulted in some insects such as large-bodied bees, butterflies, and moths
being active as adults during the dry season and in pollination occurring as a
result of this activity. The energy allocation tradeoff between vegetative growth
and differentiation of flowers in some tree species during the long dry season is a
response pattern imposing a sort of grain on the environment for various groups
of insects: phytophagous insects may drop out of the community temporarily or
make compensatory shifts to other communities (as discussed above), while
pollinating insects of dry-season-flowering tree species respond to a grain pattern
(flower resources) replacing other grain resource patterns (meristems, young
280 CHAPTER 7

leaves, etc.). Yet for the average insect species in a seasonal tropical environ-
ment, some form of selection operative is expected at least to some extent upon
active stages of the life cycle throughout most or all of the year, although the
direction and intensity of such selection are changing with time, especially over
different successive seasons. This is a very different effect than for most Tem-
perate Zone insect species, in which activity virtually ceases during winter.
Thus the lowland tropics as a whole provides a relatively greater opportunity
through time for species of insects to respond to fluctuations in the environment,
both biotic and abiotic. The response to selection is faster than in Temperate
Zone communities under such conditions. Most tropical insects probably have
mean generation times of less than 30 days, as evidenced for many species of
tropical butterflies, with notable exceptions such as cicadas, which probably
have generation times of 3-5 years in the tropics. Thus most insects can pass
through about 6 to 11 generations per year compared with about 1 to 4 genera-
tions per year at northern latitudes. Of course, reduction in breeding during the
dry season represents the rate-of-response-reducing effects at the low end of the
generation time range for tropical insects. Nonetheless, phenotypic changes re-
sulting from selection in seasonal tropical environments endow such species with
a means of coping with at least one form of environmental perturbation. Such
insect species may be better adapted to responding successfully to broad climatic
changes in a region and perhaps for invading high-altitude harsh environments in
the tropical regions of the world and to tropical deserts as they appear and
expand.
This introductory discussion of seasonality is an extension of Chapter 6, in
that seasonality in the tropics represents one form of environmental fluctuation
affecting the distribution and abundance of insect species in both space and time.
Below we examine some illustrative patterns.

7.2. SOME REGIONAL PATTERNS OF SEASONALITY IN THE

TROPICS

The regulation of the size of insect populations is sometimes mediated by

fluctuations in abiotic factors in the environment: periods of favorable weather
have promoted increases in insect populations, and periods of unfavorable condi-
tions have resulted in decline. Density-independent regulation of insect popula-
tions through severe and intense fluctuations in abiotic factors such as rainfall
and temperature enforce fluctuations in the census histories of insect populations.
In general, Temperate Zone insect populations exhibit considerable fluctuations
in size from year to year over several successive years, with outbreaks occurring
in years of favorable growing conditions (Fig. 7.2). In Temperate Zone ag-
roecosystems, weather has been shown to locally influence the ability of some
EFFECTS OF SEASONALITY ON INSECT POPULATIONS 281

endoparasitic insects to regulate the populations of targeted pest insect

species ,(487) and such effects vary considerably among different localities in a
region. In general, biotic control agents are predicted to be less successful in
regulating pest populations at higher latitudes, principally because of the variable
effects of weather and other physical factors in the environment.(104) In these
cases, climatic factors exerted positive or negative influences on population
growth over several generations of the insect during a single year, resulting in
years of outbreak and years of low abundance. In the tropics, however, lowland
regions tend to have fairly stable temperature and rainfall conditions from year to
yeat45 ) even though rainfall, for example, may fluctuate considerably from day
to day (Fig. 7.3). In the lowland tropics insect populations in general may be
subject to less pronounced patterns of temporal variation in temperature than in
the temperate zones. There may be considerable diurnal fluctuations in air tem-
perature in the tropics (Fig. 7.4), but month-to-month evenness in the lowlands is
usually high. (45) Seasonal shifts in temperature are less likely, with the major
shift being in the rainfall regimesY16) Therefore, insect populations in lowland
tropical regions are less likely to be caught in sudden, unpredictable shifts in
temperature of the kind typical for the temperate zones. In some insect popula-
tions in the temperate zones, shifts in temperature interact with fluctuations in

580

540

500

460

420

380

340

300

260

220

180

140

10 1 5
August
SUCCESSIVE DAYS

FIGURE 7.3. An example of daily variation in rainfall during the rainy season within the pre-
montane tropical rain forest zone of northeastern Costa Rica. Day-to-day variations in rainfall may
influence the activity of insects such as butterflies and cicadas, which often depend upon sunny
conditions for reproduction.
282 CHAPTER 7

a b
r
30

.......
.~' ,

0' 2 ~
S2
u
..
' "-
E
0
.§.
e
.
28

.a~ /'1
,.-, •.9""-;"'"
",
~
I:
<II
Co
/ .... :~: 'iii
ex:
E ....... c. ..'

~ "

......
~ /-=
,,'
.....
26

I•
\ ........ __ ••• , •••• -5.
,

Jan
"'" ..•.
Dec
0
Jan
I~ r Dec

SUCCESSIVE MONTHS

FIGURE 7.4. (a) Monthly temperature at Puerto Ayacucho (e) mean calculated from 24-hourly
values for 1958-64), and at Manaus, Brazil (0) (b) Monthly rainfall at San Fernando (solid) (total:
1491 mm) and at Manaus, Brazil (hatched) (total: 2095 mm) for 1931-60. On the ordinate for the
rainfall graph, mm/100 means a factor of 100 to adjust the scale shown (0-3). [Modified and redrawn
from W. Schwerdtfeger (ed.), Climates aJCentral alld SaurhAmerica. Elsevier, Amsterdam (1976).)

population density to determine patterns of reproduction.(487) In some Temperate

Zone butterfly species the first killing frost date is unpredictable and results in
considerable density-independent mortality of immature stages. Seasonal im-
pacts of temperature changes of these kinds are less likely to occur in the low land
tropics but may affect high-altitude insects. Some tropical lowland regions have
very stable monthly temperature regimes (Fig. 7.5) but fluctuating rainfall pat-
terns reflecting seasonality (Fig. 7.3). The result is a seasonal pattern of flower-
ing periodicity in some Brazilian forests (Fig. 7.6), where many tree species
flower during the drier months. Different regions within the wet tropics may
exhibit very different peak periods of rainfall on an annual basis, reflecting the
localized effects of wind patterns and topography. (16.45)
Over relatively short or long distances different climatic and vegetation
regions in the tropics may have similar total annual rainfall, but the monthly
distribution is different: lowland tropical wet forest regions (rain forests) tend to
have uniform rainfall each month but with one or more short dry seasons or
veranillos, while lowland tropical dry forest regions have an extended period of
little or no rainfall, the dry season or verano, and high elevational regions have
dry seasons of varying intensities but generally far cooler temperatures. In Costa
Rica, the length and intensity of the dry season differ considerably between
EFFECTS OF SEASONALITY ON INSECT POPULATIONS 283

-r--r--
30 Locality and Altitude

/ ~ /""
~
I-- /
Barronquilla, Sea level

25

- Ocalla, 1100 m

- -- Pamplona, 2100 m

10 f"" - I--
- COQUO, 2600 m

5
Jan. Feb. Mar. Apr. May June JlAy Aug. Sep. Oct. Nov. Dec.

FIGURE 7.5. Along altitudinal gradients in the American topics, the greatest monthly variation in
temperature occurs at or near sea level and least on mountain tops, as exemplified here for Colombia.
[From G. W. Miskimen, Biotropica 4:85-92 (1972).]

tropical lowlands and mid-elevation areas (Fig. 7.7). Depending upon the region
and locality, the annual dry season in seasonal regions may have predictable or
unpredictable beginning and termination dates, a property of considerable impact
on insect populations where generation times are generally less than 30 days. In
some Temperate Zone insects such as crickets, the details of how the life his-
tories of individual species have been shaped by selection for exploiting specific
microhabitats have been analyzed,(489) and given the overall greater taxonomic
diversity within orders in the tropics, such patterns oflife history synchronization
with seasons should be even more apparent in strongly seasonal regions. Being
relatively large-bodied, crickets and other acoustical orthopterans probably ex-
hibit considerable behavioral and ecological specializations, giving rise to both
rainy and dry season forms.
Schwerdtfeger45 ) provides an extensive and exhaustive treatment on the
origins and patterns of climatic patterns in Central and South America. The
reader is urged to consult this work for detailed analyses of climate in the New
284 CHAPTER 7

a
400

300 30
f u
E o
---0 ....
200 __ 0---
__ 0 _ _ _ _ _ 0 - - - - _0 ___ - - 0 ___
....... _ ...
20
...0 - - - _ _ 0 ___ - - 0 - - ____ 0-

100 10

NOV DEC JAN FEB MAR APR MAY JUN JUL AUG SEP OCT

b
50 .I'
," '
<II
>, 40
, ,,
, J

III
"0 ,
,,
S
"-
<II
30
'.
J I
i
~
v 20
~
"-
<II

10

NOV DEC JAN FEB MAR APR MAY JUN JUL AUG SEP OCT

FIGURE 7.6. Annual distribution of mean monthly temperature (dashed line and circles) and rainfall
at Reserva Nova Lombardia, Santa Tereza, Espirito Santo, Brazil (a) and flowering phenology for the
same period (b) for a representative sample of flora. [From J. F. Jackson, Biotropica 10:38-42
(1978).]

World tropics. An important general contribution from Schwerdtfeger's book,

however, is an appreciation for the considerable horizontal and vertical variation
in temperature and rainfall patterns in the tropics, a form of broad-scale temporal
and spatial heterogeneity. This heterogeneity is often preserved from year to year
in that the lowland tropics in particular tend to have fairly predictable annual
rainfall and temperature regimes as do some high-elevation regions. In some
portions of South America the seasonal distribution of rainfall, for example, is
remarkably similar over large areas such as Surinam. (45) Larger areas such as
Venezuela exhibit greater heterogeneity in monthly rainfall patterns.(45) Tempera-
ture fluctuations in tropical lowlands are on the order of a few degrees throughout
the year (e. g., Fig. 7.5). Tropical regions, however, are characterized by large
diurnal fluctuations in air temperature (e.g., Fig. 7.3), a feature of the physical
environment that promotes separation of activity patterns of groups of organisms
EFFECTS OF SEASONALITY ON INSECT POPULATIONS 285

NORTHERN DRY PACIFIC REGION

600 (LOWLANDS)

500
400 DRY DRY
SEASON SEASON
300
200

100

600 CENTRAL VALLEY REGION

(j) (SAN JOSE)
0: 500
UJ 400 DRY DRY
I- SEASON SEASON
UJ
300
:::2:
::::i 200
....J
~ 100

Z 600 CENTRAL MOUNTAINS REGION

0 500
~
I-
400

0::: 300
(3 200
UJ
0: 100
c..

600

500
400
300

200
100

F M A M A SON D

MONTHS

FIGURE 7.7. Variation in the length and intensity of the annual dry season at some representative
elevations in Costa Rica. Ten to thirteen years of data are generalized for these regions.

over the day-night cycle. Tropical nights in the lowlands remain warm enough to
permit insect activity at night throughout the year, a feature less pronounced in
the northern latitudes. The diurnal patterns of insect activity in the lowland
tropics may shift in different seasons for aerial forms due to differential effects of
air dryness on body size,(134,152) which in turn may alter the diets of some
insectivorous birds (Table 7.1).
Temperature regimes in the tropics over the 24-hr daily period thus allow for
the evolution of different peak periods of activity in insects, while temporal and
spatial variation in rainfall regimes in the tropics promotes the divergence of
whole assemblages of plants and insects and whole periods of activity or inactiv-
ity (rainy and dry seasons). High-altitude regions in the tropics may be more
286 CHAPTER 7

TABLE 7.1
Average Abundance of Major Prey Taxa in Stomachs of Chactura spinicauda,
C. brachyura, and Stelgidopteryx ruficollis and among Flying Insects at
Different Seasons a

Prey of Chacturd' Prey of Stelgidopteryx

Insects Insects
flyingC Individuals Species flyingC Indi vid uals Species

Dry season (Costa Rica)

Diptera 69.9 6.5 ± 2.4d 5.5 ± 1.5 65.5 1.5 ± 0.6 1.2 ± 0.4
Hymenoptera 9.2 69.0 ± 13.7 36.6 ± 3.5 13.5 26.2 ± 1.1 14.7 ± 1.7
Coleoptera 10.7 39.8 ± 9.7 28.3 ± 4.3 12.6 8.7 ± 0.8 8.2 ± 0.9
Total 100 155.3 ± 22.4e 80.5 ± 8.2 100 38.0 ± 0.8 25.7 ± 2.6
Wet season (Costa Rica)
Diptera 75.1 5.6 ± 2.4 4.3 ± 1.1
Hymenoptera 12.7 98.6 ± 18.3 33.1 ± 2.8
Coleoptera 10.2 37.6 ± 19.3' 15.6 ± 1.2
Total 100 151.1 ± 22.9' 59.5 ± 3.6
Wet season (Panama)
Diptera 69.6 10.8 ± 3.3 9.3 ± 2.6
Hymenoptera 19.6 207.0 ± 91.0 25.5 ± 3.8
Coleoptera 6.5 27.8 ± 7.0 19.5 ± 4.3
Total 100 253.5 ± 92.9 60.3 ± 8.7

a From Hespenheide.{19l)
be. spinicauda (Costa Rica) or C. brachyura (Panama); numbers of prey based on six stomachs examined for
each season and site for Chactura. four for Stelgidopteryx.
cProportions of insects in samples at site where birds were collected; see text.
d Mean ± standard error.
elncludes termites of a single species and 27.6 ± 8.7 individuals.
fIf nitidulid beetles of a single species found in one stomach are excluded. the number of Coleoptera individuals
per stomach is 18.3 ± 1.6 and the total number of individuals per stomach is 134.8 ± 19.7.

uniformly stable in terms of monthly temperature regimes than adj'acent low-

lands. (45) Miskimen(490) presented climatic data showing that altitude is a critical
factor in determining the amount and effectiveness of rainfall in the tropics.
Rainy seasons usually begin and end at about the same dates along an elevational
gradient from lowlands to mountain tops, but evapotranspiration varies.(490) In
terms of physiological effects on plants, the effective rainy season is greatest and
most uniform at high altitudes along such a transect, and if so, can support a lush
growth of vegetation in some areas.
Several groups of insects in Neotropical environments have rather distinct
bimodal daily activity periods, often coinciding with dawn and dusk. Several
genera and species of cicadas exhibit highly structured chorusing bouts at dawn
EFFECTS OF SEASONALITY ON INSECT POPULATIONS 287

and dusk, possibly to maximize acoustical communication in still air,(150) al-

though some of these forms also chorus during other hours of the day. (215) The
interplay of physiological constraints related to dessication; exertion of energy
associated for foraging, mating, and predator avoidance; and maximization of
activity at times when predators are least active, may represent the major selec-
tive responses by such insects that result in bimodal, dawn-dusk activity periods.
Although some Amazonian simulid flies exhibit clear-cut bimodal biting periods
each day, others in the same region do not. (491) Different species of simulids are
affected differently by changes in air temperature, light intensity, and humidity,
and most species probably have a basic circadian rhythm for biting activity that is
modified to varying degrees by exogenous factors. Different species, or different
populations of the same species in different regions or habitats, may exhibit
differences in response patterns to sets of exogenous factors. (491)
Very high altitudes in the tropics, such as those of the Andes in South
America, are cool and have distinct wet and dry seasons at some elevations.
Biologically, the effects on insects of seasonality at these altitudes and under
cool-temperature regimes may be less than at comparable lowland areas in the
tropics. General discontinuities in temperature regimes on tropical mountain tops
create ecological barriers for insects(46) much more intense than in mountains of
the temperate zones. This is not to say that insects do not respond to seasonality
at high elevations in the tropics: they do. But the cool-temperature regimes of
these regions lower overall species richness of insects through adverse effects on
development and resource heterogeneity. The overall pool of insect species re-
sponding to seasonality is less at high elevations (mountain tops) in the tropics,
so seasonality is less of a selective factor in shaping the partitioning of insect
communities by seasons under these conditions. Long-term studies of high-
altitude insects in tropical regions are needed to examine these ideas, and more
data are needed on what the forests in disturbed mountain regions such as the
Andes were like in the past. Tropical mountain tops are somewhat analogous in
effect to the temperate zones in that both promote lower equilibrium number of
species in communities and tend to have individual species with larger popula-
tions that are physically controlled.o o4 )
Tropical dry seasons influence insects and other organisms in the following
ways.
I. Steadily increasing dryness, from a cessation of rainfall, reduces the food
supplies normally available to some species residing in a particular community.
Tropical communities with low or herbaceous canopies, i.e., exposed, will ex-
perience the effects of the dry season more rapidly than plant species forming the
understory of forest. Although the forest canopy may be similar to the open
pasture or field in this regard, the spectrum of species present is different as are
perhaps the physiological tolerances of the plant species.
288 CHAPTER 7

2. The dry season may reduce the intensity of competition among ecologi-
cally similar species in the habitat if some of the individuals of species are more
vulnerable to dessication than others.
3. Likewise there may be a relaxation of biotic control agents such as
predators and endoparasites in open habitats as the dry season progresses, since
these organisms are generally much smaller than their hosts. The result might be
that former host insect populations experience a surge of growth in the dry season
if dessication can be avoided and food supplies remain abundant.
4. For many orthopterans, lepidopterans, and other groups, the incidence of
egg and juvenile mortality resulting from pathogeneic fungi and other microor-
ganisms may decrease markedly in the dry season as a result of these control
agents becoming dessicated, especially in open habitats. Such an effect would
contribute to increasing popUlation growth in these insects during the dry season,
everything else being equal (which it is not).
5. Small and soft-bodied insects (adults or immature stages) may decline in
the dry season in exposed habitats due to water loss constraints on individuals.
6. A portion or all of the membership of a population of either a small-
bodied insect species or an insect species whose food supply has dried up in
exposed habitats will exhibit lateral or vertical shifts (mini-migrations) to moist
forest refugia in lowland dry forest regions where forest cover is discontinuous
and scattered. Such movements result in considerable density-independent
mortality and increase the concentrations of insect species occupying forest re-
fugia for the dry season. At the beginning of the rainy season such contracted
populations may again expand as food supplies and other resources (nesting sites,
etc.) are added to the surrounding communities. But the intervening processes
may form a sort of deme selection in which drift will change the genetic makeup
of popUlation by the end of the dry season. Thus dry seasons and enforced
formation of isolated to quasi-isolated demes make up an evolutionary filter
determined largely by drift effects (especially if the species enters into a repro-
ductive diapause in the dry period).
7. In some insect species, "dry-season morphs" may increase in frequency
in the popUlation while the frequency of "rainy-season morphs" declines as the
dry season progresses. Dry-season morphs may be in reproductive diapause,
have different thermoregulatory habits and physiology from wet season morphs,
be active at different times of the day, and perhaps utilize different sets of
resources from rainy season morphs. The distribution of such morphs is regulated
by a form of frequency-dependent selection in a polymorphic population ex-
periencing periodic fluctuations in the environment.
S. The insect faunas of dry forest lowlands adjacent to mountains may
represent fluctuating subsets of the mountain insect faunas that expand into the
lowlands at the beginning of each rainy season, and a portion of which retreats
EFFECTS OF SEASONALITY ON INSECT POPULATIONS 289

back into the mountains in the following dry season (and perhaps several genera-
tions later). Such shifts in lowland insect faunas may be greater than for insect
faunas associated with lowland rain forests adjacent to other mountain slopes in
the tropics.
9. In some groups of insects, there might be evidence of considerable
seasonal selection among ecologically similar species in which some species are
active in the adult stage during the rainy season and other species are similarly
active in the dry season in the same habitats. A good example is provided by
annual peak emergence periods of some cicadas in northwestern Costa Rica. (139)
Seasonality may differ greatly between dry forest and rain forest lowlands in
the tropics. First, the dry season of tropical rain forest regions tends to be less
pronounced than that of dry forest regions, increasing the uncertainty of the
environment for insects and other organisms to synchronize life cycles and life
histories. The beginnings and ends of dry seasons in the wet tropics are less
predictable events, as is the total period of dryness. How insect species interpret
the dry season in terms of maximizing fitness will thus differ between the two
regions. For example, the frequency of courtship and oviposition may actually
increase during the dry season in wet tropical regions because many butterfly
species are more active as adults in sunny weather (although others are active in
cloudy and misty weather) and the incidence of successive sunny days is greater
in the dry season. Many butterfly species at 200 m elevation on the Caribbean
watershed of northeastern Costa Rica exhibit increased oviposition during the
short dry season between January and March (A. M. Young, unpublished data).
The stress of the dry season on physiological functions related to water stress in
insects and their resources (i.e., plants and plant products) in tropical rain forest
regions is probably nil since humidity remains in many pockets of vegetation and
generally more forest cover acts as a moisture source for insects (although rain
forests are disappearing rapidly in all tropical regions). Thus in rain forests many
insects may actually synchronize reproductive activity with the dry season in an
opportunistic fashion, while the opposite response may be true for dry forest
regions.
The butterfly fauna of Brazil exhibits considerable variation in both num-
bers of forms active throughout the year in a given region and the estimated
numbers of individuals per species.(492) Cycles of seasonality in the tropics,
particularly at low and intermediate elevations, tend to reduce popUlations of
butterflies and to a lesser extent the number of species active in the part of the wet
season. (492) Clench(493) found that the greatest abundances of species of butterflies
on Andros Island, Bahamas, occurred during the early phase of the rainy season
(June) when rainfall was highest. Clench noted that some butterflies, particularly
pierids, exhibited a seasonal change in phenotypic appearance and also that adult
preferences for species of flowers changed with seasons. Subtropical and tropical
290 CHAPTER 7

changes in climate will alter the distribution and quality spectrum available for
foraging insects such as butterflies, and switches in patterns of flower visitations
are expected.
The large Neotropical cicada F. mannifera again provides some interesting
insights into responses within populations in regions with different climatic con-
ditions. This robust insect (Fig. 3.15) is distributed in forests in Central America
generally below 800 m elevation and is found in both wet and dry forest regions.
In northeastern Costa Rica adults exhibit a distinct diurnal singing pattern in
which males chorus mostly within a 15- to 20-min period at dusk, and to a lesser
degree at dawnY50) But in northwestern Costa Rica, adults confined to small
patches of forest sing intermittently throughout the day (A. M. Young, unpub-
lished observation). Analysis of songs and external morphology of adult cicadas
from both regions currently under way (A. M. Young and T. E. Moore) suggest
in a preliminary fashion that the species is experiencing very different environ-
ments. The data suggest that chorusing behavior in this cicada, a major compo-
nent of fitness in cicadas in general, varies considerably between seasonal and
nonseasonal regions in the tropics, a phenomenon perhaps largely regulated by
the differential effects of temperature and other regimes of the physical environ-
ment. The observed differences in daily chorusing patterns may be due to evolu-
tionary divergence within the species or else represent differing response patterns
in a flexible phenotype. The system warrants further study. Adult F. mannifera
from the dry forest region have light-green prothoracic markings and are smaller
(Fig. 3.15), while those from wet forest have brown prothoracic markings and
are larger, suggesting a possible pleiotropic response of gene pools to different
forms of selection in the regions studied. There is some merit to long-term field
studies of single species in the tropics. The studies of Baldry and Molyneux(494)
of the tsetse fly Glossina medicorum in the Northern Guinea Savannah region of
Africa during the dry season indicate that popUlations of this insect exhibit
definite dawn-dusk cycles of flight activity in certain kinds of forest habitats.
Such cycles, in small flying insects, may be adaptive in permitting maximal
activity at those times of the day when water stress conditions are minimal.
Depending upon the relative amount of exposure to the sun, different
habitats in the tropics will contain organisms that respond to different degrees to
seasonality. Plants and insects of the primary forest understory may exhibit less
seasonality in peak adult activity periods and synchronization of popUlations with
one season or another than plants and insects in adjacent young secondary
habitats and primary forest canopy. Data are lacking on precisely how informa-
tion about seasonality is translated into a cue for a change in popUlation dynamics
in insects in the tropics. The species composition and population densities of
various species change in foliage-inhabitating insect communities between rainy
and dry seasons.(72,!34) Large-bodied insects such as cicadas exhibit fairly dis-
EFFECTS OF SEASONALITY ON INSECT POPULATIONS 291

tinct allochronic adult emergence periods in tropical regions with a dry sea-
son .<70,71, 135, 136, 138,139)
The flowering periods of many tropical trees are also allochronic to vary-
ing degrees,(117) and large-bodied pollinating insects are active in the dry sea-
son.(116) It is of interest to speculate that changes in the physiology of indi-
vidual plants within a species population in response to a shift in season may
function as a cue to many insects exploiting these plants. The nymphal stages
of cicadas in the tropics probably last 3-5 years, as suggested by collections
of nymphs: three to five size classes of nymphs occur together in the ground
for some species studied (A. M. Young, unpublished data). Although sub-
terraneous nymphs presumably feed on relatively nutrient-poor fluids of the
root systems of various plant species in a habitat,(366) some cicada species may
exhibit a preference for the root crowns of certain trees.(71) Dense accumulations
of nymphal skins from the last molt occur around the base of adult-size trees
in many tropical forests.(137) If such patterns reflect actual preferences for
the root systems of those trees, is there some physiological change within the
xylem fluids,(495) and do cicada nymphs, preparing for the final molt, use these
changes as a time cue for emergence? The correlations of emergences with
seasons in the tropics are there,<70,71,135,136,138,139) although the mechanisms are
not known. Other candidate cues such as changes in soil moisture content and
hydrostatic conditions of soil come to mind as cues, but there is probably less of a
gradient in soil moisture content established vertically in primary tropical rain
forests than in other, more exposed habitats. Yet cicadas exhibit very seasonal
emergence patterns in primary forests.(70,7I)
In the lowland and mid-elevation regions of the tropics in particular, the
observation that many species of insects occur in such regions suggests a possible
relationship with the distribution of sizes for individuals within populations of
individual species. In other words, when species richness is high, the level of
richness is in part a function of the size distributions of these species, a distribu-
tion undoubtedly correlated with the degree of within and between heterogeneity.
As the result of about 1000 sweep samples in Costa Rica and Massachusetts,
Schoener and Janzen(416) have studied the size distributions of insects in the
tropics and Northern Temperate Zone. They examined how the number of
species and size distributions of insects varied between these two regions. Al-
though the range of size distribution for insects is greater in the tropics, there is a
skew toward larger insects as well in this region. (416) Yet it may be difficult to
generalize that insects are bigger in the tropics, because the pattern may be
patchy in terms of habitats and regions sampled within the tropics. We would
expect a greater range of habitats in the tropics, and within habitats a greater
range in resources for insects, but such ranges in resources may vary considera-
bly in different habitats and regions. Within- and between-habitat heterogeneity
292 CHAPTER 7

in the tropics is great, and such a pattern of resources promotes a greater number
of species and greater range in body sizes among species.
Tropical dry seasons influence the rates of leaf drop and of drop of dying or
dead plant products, fruit and seeds, and dead insects into forest streams and
rivers.(460,496,497) The result is that nutrient patterns available to aquatic insects and
other invertebrates may vary at different times of the year and in different kinds of
waters varying in nutrients.(496,497) Shifts in seasons in central-Amazonian inunda-
tion forests influence the distribution of forest litter invertebrates, causing migrations
or seasonal dispersal.(497) Periods of flooding result in migrations by some inver-
tebrates to other sites. The rainy season not only increases flooding in tropical
forest flood plain regions but may also lead to an irregular but continuous pattern
of litter production due to falling debris from the forest, as studied by Puig(498) in
French Guyana. The result of course is a pattern of pulsing changes in the
distribution and abundance of organic materials for ground cover, cryptozoic, or
subterraneous insects and other arthropods in such habitats.
Both seasonal and long-term changes in forest litter composition will influ-
ence the insect communities exploiting this resource. During rainy periods litter
production may increase,(498) and this may be a time of optimal popUlation
growth for litter-inhabiting insects, including ant species. A greater incidence of
fruit drop and seed drop taking placeool) at certain times of the year than at others
may signal an increase in the abundance of ground-dwelling and seed-eating
Orthoptera in tropical forests. Thus an understanding of tropical seasonality can
be approached from the standpoint of influences on the ecology of individual
species in a habitat or from a consideration of how communities respond to
potential shifts in resources. Several workers (43,101,116.118,499) have measured the
phenological patterns of flowering and fruit production in some tropical forests
with a consideration of synchronization with seasons and with activity of pol-
linators. The annual pulsing of flowering and fruit production represents a pre-
dictable fluctuation in the resources available to insects and other animals(SOO) in
tropical forests. Forests with a strong dry season exhibit distinct phenological
differences among tree species for flowering, (17) whereas the patterns, while still
present, are less abundant in tree species in comparatively nonseasonal forest.
Brown and BensoU<204) reported distinct annual fluctuations in the abun-
dance of various species of ithomiine butterflies in Brazil and provided evidence
that the maintenace of a multiple mimetic polymorphism in a heliconiine but-
terfly, H eliconius numata, was due to environmental heterogeneity. Potentially
mimetic species of ithomiines were far more abundant, and populations fluc-
tuated greatly each year, conditions making H. numata a sort of "pseudo-
Batesian" mimic of these butterflies.(204) Ithomiine butterflies in tropical rain
forest in Costa Rica display a conspicuous increase in abundance during the first
half of the rainy season (Fig. 7.8), a phenomenon perhaps related to increases in
new host plant biomass for oviposition and larval development. Even though this
EFFECTS OF SEASONALITY ON INSECT POPULATIONS 293

2.5

>-
<t
Q
2
tt::
W
Il.

...
Q

tt::
::> 1.5
l-
ll.
<t
o
III
W
...J
I.L.
0::
W
l-
I-
~05

February Marcn April May June

FIGURE 7.S. Abundance of butterflies near Rincon in southern Costa Rica, measured or number of
captures per day. The various lines and symbols depict changes in abundance of various species. [From
K. S. Brown, Jr., and W. W. Benson, Biotropica 6:205-228 (1974).]

region (Osa Peninsula) is relatively nonseasonal, there is a short dry season.

Other groups of butterflies may exhibit increased reproductive activity in the dry
season within tropical rain forest zones, depending upon the condition of adult
and larval resources: the availability of meristem tissues suitable for oviposition
and larval development on host plants during the dry season, and an absence of
high adult mortality resulting from desiccation, increased rates of predation, etc.,
function together to maintain stable or slightly depressed breeding popUlations
and in some cases actually elevate abundance.
In northeastern Costa Rica the number adults of many butterflies markedly
declines along forest edges near the end of the rainy season (October-December)
each year, as emergences of most cicadas decline. (70) This observation presents a
curious situation in terms of some tropical insects escaping from too much
wetness rather than too much dryness in the dry season. If being active (courting,
294 CHAPTER 7

mating) in very rainy weather is not adaptive, adult emergence can be suppressed
during the wettest months of the rainy season. The reduction in emergence may
result from a synchronization of life histories around this period or from annual
irregular responses to wetness during this period.
Alternatively, it would be interesting to determine whether in some insects
in the tropics a rainy season diapause-like mechanism exists that is locked into a
cue of excessive moisture and that actually arrests development in early stages of
life cycles. As with reproductive diapause mechanisms in insects of strongly
seasonal tropical regions such as lowland dry forest, the mechanisms of control
would be different from those of Temperate Zone insects: moisture or lack
thereof, in critical amounts, and perhaps some physiological cues from plants,
may govern such diapause-like systems, if they exist, in tropical insects. Some
field studies of tropical butterflies (e.g., that of Parides by Cook et ai.(444) in
Trinidad) indicate that periods of very heavy rains may result in increased
mortality of adult butterflies. The heavy rains associated with the wet season in
Ethiopia are a major cause of mortality in gelada baboon populations, with
annual variations in birth rate varying inversely with the severity of rainfall. (50])
Localized shifts of individuals among several subpopulations in the region pro-
vided a buffering mechanism against major fluctuations in overall annual popula-
tion density in the baboon population. In tropical lowlands, the heaviest rains
often come near the end of the wet season. This may also be a period when the
vegetation consists primarily of mature leaves and in which population densities
of folivores and other herbivores may be high. Such conditions may favor a
reduction in reproductive output in some insects since food supplies are reduced
and saturated with herbivores. Conditions may also be unsuitable for courtship
and oviposition. Such effects may be most prevalent in lowland tropical rain
forests.
One of the major difficulties concerning our current knowledge of how
tropical seasonality cycles affect insect populations is a lack of data over many
successive years. Explaining the extent to which year-to-year changes in the
intensity of dry seasons in particular regions result in changes in abundance and
distribution of insect species in certain kinds of communities and habitats re-
quires such long-term studies. How such correlative patterns change with altitude
in the tropics should also be studied, since harsh environments may suppress the
effects of seasonality on fluctuations in population size.
The monarch butterfly, Danaus plexippus (Danaidae), illustrates an example
of a tropical insect that each year attempts a major recolonization of northern
latitudes, a migration prompted in part by the expansive geographical distribution
of more than 100 species of the larval host plants, milkweeds (Asclepiadaceae),
into the Northern Temperate Zone.(981 Tuskes and Brower(981 suggest that clus-
tering behavior in overwintering congregations of this butterfly results in random
mating during the spring and before the northern migration begins. Such mating
EFFECTS OF SEASONALITY ON INSECT POPULATIONS 295

ensures a generalized monarch genome that permits ecological flexibility in

recolonizing northern latitudes, a condition made essential by the physiological
divergence among milkweed species in terms of the chemical properties of these
host plants underlying defense against herbivores. Such flexibility in a species
optimizes the chance for successful colonization of a relatively less predictable
environment than that from whence the founders cameY04) The ecological fac-
tors in the evolutionary history of tropical populations of the monarch that re-
sulted in the migratory syndrome being built into the genome of this species
remain obscure. On the other hand, the geographical divergence and spread of
the larval host plants into northern latitudes would result in a form of selection on
milkweed herbivores to exploit the expanded resource base undet individual
selection of an initially directional form. An intriguing possibility, although
exceedingly difficult to confirm, is that monarchs originated in those regions of
tropical America that were climatically unstable at the time the species evolved.
Alerstam and Enckell<s02) suggest that migration in many bird species resulted
from species originally occupying tropical regions of high environmental insta-
bility, such as dry forests and savannas found today in North and East Africa.
Such regions, at least for birds, are unpredictable in terms of providing sufficient
food and other necessary resources during the breeding season. Analogizing
tentatively to the monarch, the question is asked: did the monarch originate in the
tropical dry forest lowlands of the Neotropics and evolve the migratory habit in
response to a decline in larval host plants during the long and severe dry seasons
of these regions? Is what we see today in some dry forest butterflies such as
Opler's Eurema daira, which migrate vertically into nearby higher elevations
during the dry season to reach wetter habitats and survive the dry season, an
intermediate step in the evolutionary filter the monarch passed through early in
its evolutionary history?
From studies on annual emergence curves for various species of saturniid
moths in northern latitudes, Waldbauerls03 ) suggested that the observed
polymodal emergence curves of these species are caused by unpredictable en-
vironmental conditions: a pattern of peak emergence variation within a single
generation in these univoltine species permits survival if early spring conditions
are favorable one year, late spring conditions favorable the following year, etc.
The adaptation is expected to tum up at the marginal areas occupied by insect
species at northern latitudes.<s03) Waldbauer analogizes the situation to the model
of patch structure developed by Levins and MacArthurl481l in the sense that the
environment each year is coarse grained in terms of optimal periods for adult
emergence: there is an early emergence patch and a late emergence patch, and
fitness is maximized in the popUlation by the species exploiting both patches.
Could a similar mechanism exist in tropical insects? Overall tropical envi-
ronments are probably less temporally unpredictable than comparable Temperate
Zone environments, even for regions with strong seasonality such as lowland dry
296 CHAPTER 7

forests. Within a single generation there is probably an absence of selection for

polymodal emergences in most tropical insects, since the environment is essen-
tially fine grained with respect to optimal emergence periods. Such predictions
apply to both seasonal and nonseasonal species since the shift between seasons in
the tropics is probably very predictable each year for most localities. The exis-
tence of many overlapping generations throughout the year in many insect
species in the tropics functions as a form of ecological buffer against any unpre-
dictable changes in the beginning or termination of a season. Overall, polymodal
emergences of the kind discussed by Waldbaud 503 ) are probably less numerous
in tropical insect faunas. A further buffer in the tropics against environmental
uncertainty through time is the edible condition that most kinds of resources
remain in to some degree throughout the year, particularly with regard to living
and dead plant tissues. Pollinator populations may be more susceptible to the
fluctuations in the boundaries of season from year to year or over a run of several
years. ~

For insect specjes occupying relatively restricted habitats, population den-

sities may undergo annual cycles of abundance in direct relation to the degree of
seasonal influence on habitats. Vepsalainen(504) found a wing-length polymorph-
ism (dimorphism) in Gerris species occupying small ponds subject to periods of
increased population densities of these insects. Because the populations are food
limited, selection for long-winged morphs occurs when densities are high as a
means of colonizing new ponds. Populations cannot exceed the carrying capacity
of the habitats, and therefore density is relaxed by colonizing episodes. Short-
winged morphs are adapted to low-population-density conditions and allow the
species to build up populations. Density-dependent selection favors the mainte-
nance of morphs in Gerris populations in regions which experience environmen-
tal catastrophes resulting insects vacating many ponds. Such events are followed
by colonizations, and when the ponds are filled, selection occurs for short-
winged forms until density becomes critical relative to food supplies. At this time
long-winged morphs become frequent. In this system disruptive selection favor-
ing both morphs in a population periodically shifts to directional selection for
short-winged morphs if the population is isolated. (504) If some degree of genetic
variation is preserved following directional selection, the long-winged forms can
again increase in frequency if favored by selection. Vepsalainen calls this phe-
nomenon a morphism cycle and analogizes it to taxon cycles.(9Il Such considera-
tions may have a more general application for the analysis of insect populations
in the tropics.
In strongly seasonal regions, a form of density-dependent selection may
occur for different morphs adapted to different seasons. Some butterflies of
lowland dry forest in Central America, for example, may have dry- and rainy-
season morphs. P. A. Opler (personal communication) has suggested such an
adaptation in the pierid E. daira in lowland Guanacaste Province, Costa Rica: a
EFFECTS OF SEASONALITY ON INSECT POPULATIONS 297

rainy-season morph is relatively mobile and breeds in various kinds of habitats

throughout the rainy season. During the early part of the long dry season, popula-
tions of the butterfly consist of both morphs, with a gradual increase in frequency
of the dry-season morph. This morph is capable of withstanding increasingly dry
conditions and enters into a reproductive diapause for the dry season. The rainy-
season morph migrates both horizontally and vertically into moist forest habitats,
where it continues to be active. The increase in frequency of the dry-season
morph is a response to decreasing availability of resources needed by the rainy-
season morpho The reverse form of selection is predicted for the rainy season.
The adaptation represents a facultative response by a'butterfly species to a very
seasonal tropical environment. The information available on this system is pre-
liminary (P. A. Opler, personal communication) and warrants further study.
Similar density-dependent selection may occur for other insects in very seasonal
tropical regions. It is expected to tum up in species whose popUlations contain
sufficient genetic variation and phenotypic flexibility to permit the existence of
different morphs, with different sets of ecological traits, and for which a drastic
change (usually a decrease) in the abundance of resources occurs in the dry
season.
Insect species occupying tropical habitats where resources are very patchy
in space and where environmental catastrophes may result in considerable
density-independent mortality of individuals in breeding populations may pro-
vide another general case of such density-dependent selection in the tropics.
Species of insects forming small colonies associated with highly dispersed
patches of resources (host plants, fruits, certain kinds of litter, etc.) in regions
where periodic floodings occur are candidates for such ecological adaptations. If
the patch size is large enough to support a colony or small breeding population
and if biotic control is low or density independent, popUlations or colonies may
exceed the carrying capacity of the environment unless colonizing morphs in-
crease in frequency under such conditions. The implication is that insect com-
munities associated with certain kinds of resources and habitats may experience
periodic cycles of morphisms if selection pressures can reverse and residual
genetic variation is present within popUlations of each component species for
adaptive response. Insect populations may undergo periodic and cyclic genetic
tracking of the environment in response to fluctuations in abiotic and biotic
factors. Biotic control agents such as parasitoids may experience cyclic changes
in abundance, producing fluctuations in abundance in host populations. If popu-
lation density is high and genetic variation is present, selection may favor
morphisms adapted for colonization and integration into the community, respec-
tively. Mixed life history strategies of this kind permit ecological flexibility and
the timing of generations with changes in the environment. (505)
The studies of Masaki(506) on Temperate Zone and subtropical cricket
species reveal some interesting information on which life stages are apt to enter a
298 CHAPTER 7

diapause in the tropics. Diapausing nymphs are found in some subtropical crick-
ets, whereas Temperate Zone species tend to have diapausing eggs.(506) Subtrop-
ical forms do not have an egg diapause. The evolution of an egg diapailse may
represent a greater evolutionary commitment associated with a complete disrup-
tion of the annual growing season, whereas nymphal diapause is a less intense
response to the less intense seasonality of subtropical regions. Photoperiodic
regulation grows less intense as one moves toward the equator, and this gradient
may provide the regulatory cue for the growing lack of egg diapause as one
moves toward the tropics. In regions of the tropics where a pronounced annual
dry season occurs, reproductive diapause of adult insects may be the most com-
mon form of responding to the interruption of the growing season. Reproductive
diapause would represent the next s~ep from nymphal diapause in the scheme
developed by Masaki(506) and extended here for the tropics. Reproductively
diapausing adult insects in the tropics may have a greater resiliency for resuming
breeding activity once the rainy season begins. In Temperate Zone regions, the
severity of winter temperatures precludes activity in adults or larval forms in
most interrestrial insects and therefore limits the expression of such an adaptation
in temperate regions.
Mooney et ai. (507) studied the pattern of defoliation of leaves of the Temper-
ate Zone desert shrub Dipiacus aurantiacus by the larvae of the checkers pot
butterfly, Euphydryas chalcedona, in California. The region has a severe sum-
mer drought in which these plants lose all but their terminal leaves. These
remaining leaves have high concentrations of a leaf surface resin which impedes
the development of the larvae, and the larvae enter into a diapause to "wait out"
the summer until the plants flush out again with new leaves suitable for consump-
tion. The timing of diapause in the butterfly's popUlation dynamics coincides
with the period of least suitable food availability. In seasonal tropical regions,
similar hebivore-plant interactions may occur, e.g., many plants lose their leaves
and cease new growth, leaving only older leaves. The evolution of a diapause
mechanism to escape not only a severely reduced food supply during the tropical
dry season, but also perhaps in many cases unsuitable remaining plant tissues,
would allow the species to wait out the dry season without extinction. Alterna-
tively, the species could migrate to a more suitable environment where suitable
food plants are found during this period. Such a strategy would be expected for
species which exploit a food plant that undergoes severe reduction of edible parts
during the dry season and cannot undergo a reproductive diapause. A host of
other factors, such as flight behavior, adult resource availability, and distance to
more favorable habitats will be some of the phenotypic constraints determining
whether the diapause or migration strategy evolves.
In some groups of insects such as carabid beetles, extensive habitat surveys
in the temperate zones indicate that ecologically similar, taxonomically closely
related genera and species often occur together in the same habitats,(SOS) suggest-
EFFECTS OF SEASONALITY ON INSECT POPULATIONS 299

ing that interspecific competition may not be as prevalent for some groups as
generally believed. Similar surveys in the tropics may also indicate that the high
incidence of closely related and ecologically similar species of insects in some
habitats is accounted for in terms of other kinds of ecological interactions than
competition. In Temperate Zone regions, fluctuating patterns of physical fea-
tures of the environment may impose a density-independent mortality on insects.
Coupled with a limited amount of biotic mortality, such density-independent
mortality regulates the abundance of individual insect species. In the lowland
tropics, however, the causation may be different: intense predation and
parasitism act in a density-dependent fashion to keep insect popUlations below
the carrying capacity of the environment, precluding the expression of interspe-
cific competition arising from the limitation of resources. This mechanism, of
course, may be only one of several operative in tropical insect communities as
well as to some degree in Temperate Zone communities as well. There may also
be resource limitation of some species, especially in highly seasonal regions
where temporal fluctuations in food supplies may be high throughout the year.
Various behavioral mechanisms may ensure that a population of a particular
insect species does not become too high on a particular resource. As described
above, the ability to form colonies of breeding individuals within a population
results in a fractionation of the popUlation over many resource patches. In the
case of relatively few resource patches or an insect with considerable dispersal
ability among patches, selective oviposition on unoccupied patches where larvae
or nymphs are absent, and the avoidance of oviposition on occupied patches,
distributes the population's membership equitably over resource patches. Such
behavior has been reported for Heliconius butterfiies(283) and chrysomelid bee-
tles. (509)

7.3. SOME SEASONAL DISTRIBUTIONS OF TROPICAL INSECTS

Wolda(SIO-SI2) summarized several years of light-trap data on the seasonal

abundance of some insect taxa (Homoptera, Orthoptera) in different tropical
forest habitats in Panama. His surveys of insects were conducted in areas with
distinct wet and dry seasons. W olda found that many insect species exhibit
marked changes in abundance correlated with seasonal patterns of rainfall. Many
species of insects increased in abundance in the wet season, presumably as the
result of new foliage being present. (510) Wolda's data point to the effect of
increasing the harvestable productivity in terms of new foliage on the abundance
of folivores in tropical forests. Less seasonal tropical rain forests may have a
steady, even distribution of some insect species throughout most of the
year.(70,71,513) The taxonomic and ecological complexities of insect communities
in the tropics(47) suggest that different groups may respond differently to the
300 CHAPTER 7

effects of seasonality or lack thereof. For example, some folivores such as

lepidopteran larvae and chrysomelid beetle larvae may exhibit rather marked
fluctuations in abundance between seasons because the availability of new leaves
also changes. Taxa that are specialized feeders on young or new leaf tissues may
exhibit greater fluctuations than those feeding on mature leaves. Phloem and
xylem feeders may exhibit fewer responses although be affected in other
ways.(72)
Because of such inherent ecological complexities in tropical insect com-
munities, most studies of seasonal changes in abundance of insects in the tropics
have dealt with relatively few taxa.(SI4-SI6) Seasonal synchronization in the
phenological patterns of flowers and fruit(JOJ.lIS) may dramatically affect the
seasonal abundance patterns of pollen-feeding and nectarivorous insects and alter
patterns of niche breadth with respect to feeding habits. Clearly we need to
examine the population dynamics of insect species in seasonal tropical environ-
ments, since some recent studies with vertebrates under such conditions(S!7)
indicate that individual folivore species change their foraging habits in different
seasons. Seasonal changes in the susceptibility of vegetative parts of various
plant species to attack by insects(474) may result in fluctuations in population size
for various insect species in seasonal tropical habitats. Young(315) discovered
marked declines in the early stages of the butterflies Anaea (Zaretis) itys and
Anaea (Memphis) morvus (Nymphalidae) during the dry season in northwestern
Costa Rica and attributed the decline in populations to a reduction in larval host
plants (Lauraceae, Flacourtiaceae) (Fig. 7.9). A reduction in both mature and
young leaves during the dry season(J 16) reduces the food supply for the folivorous
larvae (Fig. 7.10) of these butterflies, and a small pool of adults presumably
passes the dry season in reproductive diapause, primarily in riparian forest
habitats. GIS)
The breeding populations of other butterfly species may also decline stead-
ily as the dry season advances in such a region,(J90,299,3IS) resulting in relatively
small and fractionated adult populations left to colonize once the wet season
commences. An abundance of flowers in the tropical dry season provides a food
supply for the adults of many butterfly species.(J!6) Seasonal changes in the
species composition and abundance of Drosophila as seen in Brazil reflects
cycles of expansion and contraction of resources needed by these insects. (74) Not
all groups of insects exhibit such population reductions in the dry season. Jan-
zen<72l reported that some insects increased in abundance during the dry season in
secondary vegetation. Not only are large-bodied insects such as many Orthoptera
capable physiologically of remaining active during the dry season, but their
populations may increase since some kinds of preferred foods (seeds, pollen,
etc.) may be more abundant this time of year. Population increases in bees and
wasps may also occur during the dry season in response to the abundance of
flowers. It is the interaction of body size, age distribution in the population,
EFFECTS OF SEASONALITY ON INSECT POPULATIONS 301

PLAYAS DEL COCO

(CAESARIA SSP. HOST PLANTS)

800

FIGURE 7.9. Seasonal phenology of new

meristems for host plants of Anaea butterflies
within the tropical dry forest zone of north-
western Costa Rica.

kinds of food supplies, changes in food supplies in response to seasonality, and

dispersal abilities of individuals in an insect population that determines the re-
sponse strategy of an insect species to the pronounced dry season in some
lowland regions of the tropics. Information is needed on the natural history of
individual species in particular habitats to determine such response patterns.
Tropical insect communities have traditionally been thought of as diverse in
tropical rain forest regions,(SlS) but such diversity may also be high in tropical dry
forest, where the alternation of wet and dry seasons, with intervening short
periods of dryness within the wet season,(1l6) promotes a dynamic and shifting
pattern of community structure in response to seasonality. In tropical rain forest,
insect and other invertebrate species may have relatively smaller popUlations that
exhibit fewer fluctuations in size or density than similar species in seasonal
regions.(519)
In one of the few studies of its kind in the tropics, Brasell et al. (520) analyzed
the heterogeneity in mineral content in litter between rain forests and adjacent
agricultural habitats in tropical Australia. Brasell et al. found significant dif-
ferences in the concentrations of various minerals and nitrogen between the two
habitats even though annuallitterfall was similar in both. In an Araucaria planta-
tion, they found a seasonal shift in the litter concentration of calcium during the
fruiting period of the tree. Such data suggest that the composition of litter, rather
than its abundance, may vary considerably between forest and agricultural sites
in the tropics and that plant phenologies of fruiting may influence the chemical
composition of litter. Such effects in tum may influence the cryptozoic and
302 CHAPTER 7

FIGURE 7.10. Caterpillars of the nymphalid butterfly Anaea (Zaretis) itys in lowland Guanacaste
Province, Costa Rica on the food plant (Caesaria spp., Flacourtiaceae). Older instars (A and B) rest
on leaves while younger ones (C and D) perch on the bared midrib of an eaten leaf, tying leaf
fragments and fecal material to the midrib to enhance crypsis.
EFFECTS OF SEASONALITY ON INSECT POPULATIONS 303

terrestrial insect faunas associated with such habitats. Depending upon nutritive
needs for individual species and the temporal patterns of breeding, populations of
these organisms may shift in response to fluctuations in food supplies. Massive
annual floodings of rain forest flood plains may also result in significant changes
in the nutrient profiles of litter and soil and in the distribution of arthropods found
there.(496,497) Janzen(332) pointed out that nutrient-poor soils in the tropics have
fewer plant species.
Although the forest flo~r litter in the temperate zones is generally charac-
terized by various stages of decay, the decomposition of leaf litter in tropical
forests, particularly in humid or wet regions or in the wet season in seasonal
regions, is relatively fast, with at least one full decay cycle within a year. (521)
Less than 10% of the total fallen litter is carried over into a second year, (520 and
in some stages of secondary succession, rates of decomposition may even be
higher (1. Ewel, personal communication). The leaf litter of a mature tropical
forest will be very heterogeneous owing to high local species composition of the
flora, and turnover rates of-litter from different species are therefore expected to
vary considerably owing to different chemical and physical properties of the dead
plant tissues. Such a mosaic of litter may promote the maintenance of complex
communities of litter organisms, including insects such as collembolans and ants.
Because turnover rates are generally high, such communities depend upon a
continuous fall of new litter in order to survive. Thus the degree of residentiality
of a litter-inhabiting ant colony in the tropics will be determined largely by the
turnover rate in the litter as related to rates of replenishment and growth rate of
the colony. In tum, the activity of the colony may hasten the rate of turnover of
the litter making up the essential microenvironment for the ant species. The
communities of dipterans that utilize most litter as breeding substrates, such as
many Neotropical Ceratopogonidae,(522) will also be influenced by the same prop-
erties of leaf litter. Yet our current knowledge of the structure and dynamics of
tropical forest litter communities, in relation to (l) the degree of physical and
chemical heterogeneity of the local litter patch, (2) rates of turnover for various
species of trees contributing to the litter, and (3) feedback effects of the litter
organisms on the availability of litter and rates of nutrient release from it, is
meager. The diameters of fallen branches on the floors of tropical forests also
determine the rate of wood decomposition by some termites, and fallen branches
form another major but rapidly turned-over component of tropical litter.
R. E. Buskirk and W. H. Buskirk(523) analyzed fluctuations in abundance
and species composition of foliage-inhabiting arthropods in the understory of a
lower-montane rain forest in Costa Rica for 19 successive months and found that
both peak abundance and greatest species diversity occurred in the late dry
season and early wet season. They also found that arthropods with the largest
body sizes were present in greatest numbers at about the same time. They explain
the observed patterns by suggesting that the within- and between-season succes-
304 CHAPTER 7

...)

"...z
...)

:;c
a:

~i!'
015 30
,..:1:
0 20
-..,
za
;),,-
10

~ 2'
:I:~ 24
"''''
:1: ....

.....
-><a:"
:II>.
22

...
:I:
I-
20
--'
JUL JAN JUL JAN
1970 1971 1'71 1~7?

FIGURE 7.11. Monthly rainfall . temperature, and wind patterns at the Monteverde Cloud Forest
Preserve in Costa Rica. Shown are monthly mean values (solid lines) with one S.D. (shaded area) for
1956-69, and monthly means for 1970-72 (dashed line). [From R. E. Buskirk and W. H. Buskirk,
Am. Midi. Nat. 95:288-298 (1976). Copyright 1976 by the American Midland Naturalist and Univer-
sity of Notre Dame Press.]

sion of arthropod species is due to a succession of whole subfaunae, each as-

sociated with a specific set of environmental conditions relating to the plant
community and climatic factors. The region experiences a distinct dry season
accompanied by increase in winds but no consistent change in temperatures (Fig.
7.11). The observed increase in abundance of arthropod abundance and body size
at the end of the dry season and early into the wet season (Fig. 7.12) from the
suitability of results large-bodied insects for the dry season and the increase in
plant productivity with the first rains. The data of the Buskirks illustrate the
synchronization of whole arthropod communities with seasonality in the tropics.
The lower-montane rain forest environment experiences considerable seasonality
in rainfall, which in tum results in changes in levels of harvestable productivity
EFFECTS OF SEASONALITY ON INSECT POPULATIONS 305

Vl
..J
<

t
=>
0
:>
15

f
~
L<-
a

t
a:
LLJ
10
::I 2
=>
z

t t t t t t t
4

2
t
JUL SEP NOV APR JUN AUG OCT JAN
1970 1970 1970 1971 1971 1971 1971 1972

FIGURE 7.12. Changes in the abundance (above) and body size (below) of arthropods as seen in
200-sweep samples (means one S.D.) in the Monteverde Cloud Forest Preserve, Costa Rica. [From
R. E. Buskirk and W. H. Buskirk, Am. Mid!. Nat. 95:288-298 (1976). Copyright 1976 by the
American Midland Naturalist and the University of Notre Dame Press.)

in the plant communities. This in tum changes the community of herbivorous

insects associated with the vegetation. Predatory arthropods were found to lag
behind the phytophagous groups in cycling through the seasons, a pattern ex-
pected for most communities.
Malaise trap sampling of the insects at the interface of grassland and forest
habitats in the Nairobi National Park over a 5-year period revealed that peak
abundance periods of various taxa of phytophagous insects and their associated
parasitic Hymenoptera occurred during periods of heaviest rains in this seasonal
region of Africa, although many other groups of nonphytophagous insects exhib-
ited fewer distinct seasonal fluctuations in population sizes and diversity. (524.525)
The seasonal periods of heaviest rainfall were characterized by an increase in the
 306 CHAPTER 7

abundance of individual phytophagous species as well as in the number of

species.(S2S) Many of the nonphytophagous species sampled appeared to be uni-
voltine, with only one peak period of abundance each year.
The time required for an insect population to exhibit a response to a change
in seasons will depend in part upon the duration of the life cycle, generation time,
and whether the dry season was passed in a diapause condition. Thus an insect
experiencing a reduction in population size during the dry season may not show a
recovery response immediately during the early wet season: there may be a lag,
the duration of which is directly proportional to the mean generation time and the
degree to which the population's membership is synchronized reproductively
with the dry season. Diapausing forms may take even longer to exhibit the
recovery response.
Bates(526) analyzed patterns of seasonal abundance of mosquitoes at
Villavicencio, Colombia, over a 2-year period. Bates found that adult popula-
tions of most species peaked in abundance in the early wet season (Fig. 7.13).
These species were active nocturnally and therefore must susceptible to desicca-
tion during the dry season. But a few diurnally active species inhabiting the forest
habitat had rather uniform distribution throughout the year. These species are

20] HAEMAGOGUS CAPTURES

1~]
SCALE - MOSQUITOS PER MAN-HOUR

11
CJ
10 a
u
9 ;:
8 <

7
t
6
240
:2: 5
:2: 200
z 160 4
.J
.J 120 3
«
u.. 2
80
z 40 1
~
a: 0 0
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25
WEEK
FIGURE 7.13. Weekly phenological patterns of rainfall, water levels in experimental containers
(breeding sites), and abundances of mosquitoes (Haemagogus capricornii) in Colombia (1943).
[From M. Bates, 1. Anim. Ecol. 14: 17-25 (1945). Copyright 1945 by the British Ecological Society.J
EFFECTS OF SEASONALITY ON INSECT POPULATIONS 307

apparently more adaptive to somewhat dry conditions associated with daytime

activity in the forest. Bates suggested that it was the uniform and predictable
annual dry season that regulated the abundance cycles of most species of
mosquitoes at this site. But because different groups of mosquitoes have quite
different ecological requirements in the tropics,(526) not all groups respond to
seasonality in the same manner. In addition to desiccation effects on adults, even
in a region with a mild dry season such as Villavicencio, species which breed in
small standing pools of water may experience considerable reduction in breeding
sites during the dry season. On the other hand, species that breed in running
waters may exhibit less of a depression in abundance if streams do not dry up
completely.
Other insects associated with small pools of standing water may exhibit
population declines if suitable breeding sites are decreased in abundance during
the dry season. Thus even though the annual dry season at a premontane tropical
rain forest site in Costa Rica is short and erratic in intensity, there is a marked
reduction in the abundance of adults of the giant damselfly Mecistogaster
coerulatus in the understory of primary forest during the dry season (A. M.
Young, unpublished data). These insects breed in tank bromeliads and small
pools of rainwater collecting in depressions on the trunks of fallen trees. (527) Such
resource patches may be less abundant in the dry season. The distribution of tank
bromeliads in tropical forests is regulated by many factors, of which availability
of light is an important one,l528) and many bromeliads are therefore found in or
near the forest canopy, a microenvironment subject to increased exposure during
the dry season.
Insect species that act as vectors for various diseases of man may exhibit
considerable variation in abundance and life history as a result of seasonal effects
on the local environment. Dalmat(529.530) studied the distributional patterns of
Simulium black flies in Guatemala and found marked seasonal fluctuations in the
abundance of adults and larvae. Low densities of larvae occur at two different
periods: in the late wet season when heavy rains dislodge larvae from their
substrates in small streams as the result of flooding, and during the dry season
when small streams dry up. Fly species in larger, stable streams tend to exhibit
less fluctuations, since these habitats do not change markedly with seasons.
Dalmat suggests that popUlations of black flies, the vector of onchocerciasis in
Central America, are regulated by seasonality for species and popUlations as-
sociated with small streams.
In some tropical regions, broad-scale changes in habitats may result in
considerable change in the composition and trophic organization of communities.
Sometimes these changes are caused by the impact of seasonal rainfall regimes
on tropical ecosystems. For example, IrmlerC460,496,497) has examined the changes
in the macroinvertebrate bottom fauna in inundation forests in the central Ama-
zon region. The inundation forest represents a unique environment in terms of
308 CHAPTER 7

both physical and biotic features. (53!) The canopy of the forest is a source of
nutrient materials for the bottom fauna of aquatic organisms, including macroin-
vertebrates, through fall of leaves, fruits, and twigs. Three different water types
occur in the region, and each has inundation forests associated with it: white-
water type, mixed-water type, and black-water type. Each individual forest type
has a unique fauna. (496,497) On an annual basis, the inundation forest passes
through various phases linked to the prevailing seasonal rainfall pattern: (1) the
inundation phase, which is the transition from dry to wet conditions in the forest;
(2) the immersion phase (flooded condition); and (3) the dry phase, in which the
water level recedes due to lower levels of water in the rivers from the dry season.
Irmld496 ,497) found that various macroinvertebrates exhibit different pat-
terns of passing through the annual cycle of water abundance fluctuation. Some
insects, such as cicindelid beetles, exhibit highly synchronized seasonal life
history patterns related to the alternation of wet and dry phases of inundation
forests in central Amazonia. Irmler found that the cicindelid Pentacomia egregia
completes larval development in the dry season, and the adults occupy the forest
canopy during the wet season when the water level is high in the inundation
forest. Irmld496 ,497) found that inundation forest associated with black water
have a greater diversity of terrestrial soil invertebrates (mostly decomposers) than
white-water inundation forest. Black-water rivers are believed to be produced
from nutrient-poor white sand soils as the result of the vegetation growing on
these soils being exceptionally rich in secondary substances. (32 ) Such rivers are
rich in humic acids such as tannins and other phenols while poor in nutrients. (32)
The result is that inundation forests on black water tend to have low productivity
and biomass because of poor nutrient content. (332) White sand soils tend to have
low diversity stands of trees, while nutrient-rich soils have higher diversity of
plant species. Janzen suggests that plant species growing in low-diversity areas
require more effective chemical defenses against herbivores. Irmler's discovery
of a lower biomass of macroinvertebrates in white-water inundation forest agrees
with Janzen's predictions in terms of the litter and soil nutrient components being
less than in other kinds of inundation forests. Lower litter diversity and biomass
cannot support as large a community of animals as nutrient-rich litters in tropical
forests. Furthermore, secondary compounds passing through white-water soils
may repel many organisms.
Irmler<496,497) noted that inundation forests in central Amazonia support a
high diversity of carabid and staphylinid beetles (about 319 species studied) and
that different ecological classes of these insects are based on how they respond to
the seasonal cycle of the region. During the inundation phase some beetles
migrate to higher areas as the rivers flood, only to be eliminated as inundation
continues. Many others do not migrate and synchronize their activity with the dry
phase. Within the dry phase distinct phenological groups appear at different
times as the water level fluctuates. Some species active in the dry phase switch
EFFECTS OF SEASONALITY ON INSECT POPULATIONS 309

habits from trees to soil and the reverse in the wet season. (496,497) Some species
live in the soil throughout this period, while others remain in the canopy. The dry
phase represents a period during which a formerly inundated region can be
colonized by beetles from elsewhere and also during which ecological release of
some arboreal species can exploit resources on the ground or in the soil. Some
species synchronize their life cycles with the dry phase if they are permanent
residents of the inundation forest. (496,497) Carabids and staphylinids may exhibit a
mixture of generalized and specialized ecological traits in terms of utilizing a
highly fluctuating environment such as the ground and vegetation of river-edge
flood plains periodically inundated. It is interesting to speculate if the invasion of
arboreal habitats by these insects and others may have arisen in response to
selection pressures from inundation cycles in central Amazonia. Irmler'496,497)
suggests that insect communities associated with inundation forests exhibit con-
siderable change in composition and abundance of component species and that
they are therefore somewhat unstable. Such conditions may provide selection for
occupying more stable habitats such as forest canopy (relative to the inundated
forest floor), and generalist species occupying fluctuating and seasonal tropical
environments such as inundation forests may eventually specialize to the forest
canopy, where resources may be more predictable through time.
Not all species are expected to display such evolutionary response patterns,
since alternative strategies for adapting to the seasonal cycle are available along
with the genetic and phenotypic flexibility to exploit new habitats. Nutrient-rich
inundated tropical forests are expected to support a greater number of insect
species overall than white-water, low-diversity stands of vegetation, and pre-
sumably more selection favors transition from ground-dwelling to arboreal exis-
tence in the former type of forest: the greater diversity of the forest results in
increased selection for niche specialization if resources are periodically depleted
during cycles of inundation-emersion-dryness. Irmler and Furch(426) found that
some litter-feeding organisms such as cockroaches in inundation forests are faced
with food supplies that vary considerably in accessibility and nutritional content
at different times in the annual inundation cycle. This is yet another variable to
consider as a form of selection pressure: the quality of resources fluctuates with
the cycle and may periodically reduce fitness to the point where it becomes
optimal to switch to arboreality or other habits.
The same or closely related species occupying different areas within a
tropical region may exhibit different adaptations to prevailing conditions of sea-
sonality. In some instances individuals will migrate to areas far from the breeding
sites, the pattern determined largely by prevailing winds. In tropical Africa some
species of tabanid flies enter into a dry-season diapause, while others migrate to
other areas. (532) Adult tabanids generally exhibit very narrow temperature and air
moisture preferenda, ecological traits that determine the level of adult activity at
anyone time. Bowden(532) examined the phenologies of several species of
310 CHAPTER 7

tabanids in Africa and found that some species occur in areas where adult
emergences are regulated by successive bouts of wet and dry weather. These
species, Tabanus guineensis and T. taenioia, pass the dry season in reproductive
diapause and wait for the wet season to resume breeding. (532) The flies construct
mud cylinders to pass the dry season.(533) Adult emergence occurs after a heavy
rain, and the timing of emergences in these species is regulated by the moisture
statUs of the soil where the flies pupate. Bowden found that adult emergence is
triggered by heavy rainfall about 4 or 5 weeks before the emergence occurs.
The time span is needed to bring soil moisture to the appropriate level to trigger
emergences.
Tabanids in tropical Africa may require periods of alternate wetting and
drying for successful pupation and emergence.(532l Reproductive diapause in
tabanids is broken by the heavy forest rains and eggs mature followed by oviposi-
tion in moist soils. Shifting bouts of wetness and dryness in the dry season
represent an unpredictable environmental fluctuation with an oscillatory period
less than the developmental time of these insects. Thus the reproductive diapause
ensures that breeding takes place when the wet season is underway and soil
moisture gradually increases. Mating and fertilization take place during bouts of
wetness when adults emerge. Greathead found that periods of wetting trigger the
emergence in only a portion of a given brood of bombyliid flies that parasitize
acridid eggs in Africa, so that the emergence is staggered over a long period. The
emergence is synchronized with the oviposition of Acrididae. In tabanids,
emergence is delayed long enough to permit soil moisture to reach the point at
which it would be suitable for development of the eggs and larval survival
requirements.(532) The diapausing flies exhibit a delayed response to the early
rains, which is very adaptive in that the delayed response by vegetation to early
rains also delays the appearance of large wild herbivores bitten by the adult flies.
Bowden points out that as the wet season continues soil moisture increases,
resulting in an increase in feeding by the larvae and an eventual large hatch of
adult flies beyond the first rains. The response delay in adult emergence di-
minishes as the wet season progresses because soil moisture stabilizes and the
cue is always there for larvae and pupae to complete development. The overall
result is that adult flies continually emerge throughout the wet season, and the
number of overlapping generations increases at the same time. Synchronization
in this system is therefore largely restricted to the dry season as a means of timing
egg and larval development with the first rains of the new wet season and timing
emergences of adult flies with the return of large grazing animals that tabanids
bite for blood meals.
Other tabanids exhibit migratory behavior in which migrations follow pre-
vailing winds that eventually give rise to rainfall at some locality. (532) The system
is similar to that of migratory locusts in Africa responding to the dry season and
its effect on food supplies. (532) Many Acrididae require periods of wetness for
EFFECTS OF SEASONALITY ON INSECT POPULATIONS 311

oviposition, with dry conditions in the tropics promoting dispersal rather than
oviposition. (534) Yet prolonged wet conditions tend to increase mortality of both
eggs and nymphs,(534) so optimal conditions tend to be periods of alternating
wetness and dryness. Locust populations occupying such areas are capable of
building up to very high densities before migrations begin. Density-dependent
emigration is a response to increasing population densities and dwindling food
supplies, and individuals under such conditions prepare for emigration through
physiological changes promoting dispersal and nullifying reproduction. The
density-dependent regulation of migratory locust populations in the tropics fits
the familiar model developed to show the interaction of increasing densities and
decreasing food supplies that regulates population size. Predators, parasites, and
pathogens on locusts do not operate in a density-dependent fashion, so popula-
tion regulation is intrapopulational and related to changes in food supply as well.
Migrations in locusts result when the dry season reduces the effective vegetation
cover that can be harvested by juveniles, and as density increases, migratory
behavior ensues.(534) African tabanids migrate but not necessarily as a density-
dependent response to the dry-season effects on food supplies. The available data
suggest that adult tabanids follow moisture gradients in migration, as they are
sensitive to moisture levels. (532)
One of the interesting facets concerning fluctuations in the populations of
insect vectors of diseases in tropical regions is the degree to which populations
are regulated by other organisms. For example, levels of pathogenic attack by
various microorganisms in larval black fly populations in Guatemala are gener-
ally low and highly varied among different black fly species.(39) Since many of
these fly species are vectors of onchocerciasis, there is interest in whether these
organisms can regulate black fly populations. It appears, however, that natural
levels of infestation are low. The larvae of some Lampyridae are predators of
snails that are schistosome vectors in Liberia,(535) although few data are available
on the extent to which these insects regulate the snail populations. Lampyrid
larval populations under natural conditions may have very patchy distributions,
thus precluding a capacity for long-term effective control in a habitat, but data
are needed to confirm such a prediction. There are efforts to establish populations
of selected nematode taxa as control agents for Anopheles mosquitoes in EI
Salvador. (536) Many other environmental factors, such as host distribution and
feeding behavior, affect the degree to which a population of mosquitoes infests a
particular area,(457.537l and effective control programs must examine as many of
such factors as possible and how they change with seasonal fluctuations in the
environment. Approaches emphasizing genetic control of targeted mosquito
populations in the tropics(538) should examine the effects of seasonal cycles in
rainfall upon the ability of selected genotypes (i.e., those with male-linked
translocation heterozygotes) to survive at different times of the year or at least
during periods when breeding is maximized. In an interesting study, Whitlaw
312 CHAPTER 7

and Chaniotis(539) found that the distribution of triatomine bugs as vectors of

Chagas' disease in Panama is determined by the distribution of Scheelea palms.
Seasonality enters into the picture in the way fluctuations in rainfall influence the
interactions of vector or control insect species with their resources, including
host abundances and water levels in species with aquatic immature stages.
Although directed movement or migration into more favorable habitats is
one known strategy of escape in the tropical dry season in some insects,(540,541l
forms incapable of such movement may enter into a period of arrested develop-
ment or diapause. Seasonal changes in ambient temperature are diapause induc-
tion cues in some tropical insects,(524,525,542) and rainfall changes may induce
breeding activity in some forms. (543,544) Diapause is terminated in the mosquito
A. gambiae by increase in rainfall. (545) Induction of diapause in the red locust,
Nomadacris septempfasciata, is caused by changes in photoperiod near the
equator,(546) although such a mechanism is probably less prevalent in most tropi-
cal insects in seasonal regions. Physiological changes in host plants provide the
cue for larval diapause in some Lepidoptera(547) as well as for the termination of
adult diapause in the desert locust, Schistocera gregaria. (548) Denlinget525 ) dis-
cusses the synchronization of life history patterns of insects with seasons in the
tropics. Seasonal response systems in tropical insects must be viewed in terms of
(1) effects of different seasons on food supplies, (2) phenotypic constraints
associated with long-range mobility or lack thereof, (3) availability of nearby
riparian forest refugia for colonization during the dry season within a dry forest
zone, and (4) phenotypic constraints associated with the feasibility of diapause in
different stages of the life cycle.
Owen and Chantet515 ) present evidence that seasonal changes in the abun-
dance of Ichneumonidae in Kampala, Uganda, correspond to a similar phenol-
ogy in the availability of lepidopteran larvae which are hosts. Peak abundances of
both ichneumonids and lepidopteran larvae occur early into the wet season, and
lags a bit behind the flushing out of the vegetation of the region. The region has
two seasonal peaks in rainfall each year, but the insects exhibit a marked increase
in abundance only at the peak in rainfall following the dry season, a period of
reduced insect populations. Owen and Chanter suggest that insects synchronize
their emergences with the average wettest months rather than having maximal
emergences in the absolute wettest months. They point out that such abundance
correlations with rainfall phenology at this locality are taken with caution since
data is lacking on the interactions of the two groups of insects. More information
is needed before the abundance of lepidopteran larvae can be concluded to
determine the abundance of ichneumonids. Owen and Chanter point out an
interesting distinction of their Malaise trap data from similar studies in the
temperate zones: in the tropics the abundance of some individual species in-
creases rather than the total number of species at a time of peak abundance, while
EFFECTS OF SEASONALITY ON INSECT POPULATIONS 313

in Temperate Zone communities the number of species present increases.

Owen(514) also reported the same phenomenon in tropical Sphingidae.
Young(7\) noted that some species of cicadas in premontane tropical rain
forest stagger emergences among bouts of wetness and dryness but that a species
may have a peak period of emergence in either the dry or wet season. Young
suggested that a minimal and critical number of wet or dry days were needed to
trigger emergence and that a season of peak emergence was really a composite of
smaller emergence periods of this kind so that the emergence accumulates as the
season progresses. Although samples are small, one might argue from such
observations that in the tropics overall species richness remains about the same
throughout the year but that diversity goes up and down as a few species domi-
nate in the community in response to peak emergences being synchronized with a
particular season. On a per-unit-of-contemporary-time basis, progressive and
seasonal changes in tropical environments may not be as drastic as in Temperate
Zone environments. In the latter a greater turnover in resources may be available
to certain groups of insects as the seasons change, whereas in the tropics the
effect, on the average, is more quantitative than qualitative. This is not to say,
however, that seasonal phenologies in resources do not occur in the strongly
seasonal tropics. There is ample evidence that they exist.°16.118) The issue at hand
is that different groups of insects and other organisms may exhibit very different
types of phenologies in response to tropical seasonality and that the crucial
evidence resides in the effect of seasonality on resources and the type of
physiological control mechanisms underlying the timing of emergences or lack
thereof.
Young(!38) found definitive evidence of seasonal adult emergences in vari-
ous species of cicadas in the central highlands of Costa Rica. Along tree-lined
stream edges in otherwise heavily disturbed habitats in this region, one large-
bodied cicada, Q. gigas, experiences peak emergence each year during the dry
season, and there is a good correlation in the abundance of nymphal skins in
successive years (Fig. 7 .14a). A similar pattern was found for a medium-sized
cicada, Z. smaragdula, at this site (Fig. 7.14b), but peak emergence is during the
long wet season. Emergences of other cicadas at other localities also vary consid-
erably at different times of the year (Figs. 7.15 and 7.16). In lowland tropical dry
forests of Costa Rica and elsewhere, cicada species vary considerably in both the
size and breadth of the emergence each year and the time of year for peak adult
emergence (Fig. 7.17). A major component of fitness in cicadas is the choruSing
by males for mating,05OJ and highly synchronized seasonal emergence patterns in
these insects may reduce possible crossing of courtship signals while optimizing
transmission of songs for species active at a particular time of the year. (150,215)
The various genera and species of cicadas studied in Costa Rica by
Young<70,7\,\35-i39) exhibit spatial distributional patterns in which the nymphal skins
314 CHAPTER 7

650

600

550

500
•
!Il
Iii
« 450
u •
....I •
« 400
J:
11. •
:::E
>
z 350
~~
«~
~ 300
u
u..
0
!Il
250 •
cc:
W
IJI
:::E 200
::J
z •
150

QUESADA GIGAS

o 350

FIGURE 7.14. The relationship between abundances of cicada nymphal casts (skins) in the same
study plots at one locality in the central highlands of Costa Rica in different years for two species of
cicadas. [From A. M. Young, Am. Midi. Nat. 103: 155-166 (1980). Copyright 1980 by the American
Midland Naturalist and University of Notre Dame Press.]

of two or more species accumulate beneath large forest trees, particularly Legumi-
nosae, but different species exhibit peak adult emergence at different times of the
year. As discussed earlier, the mechanisms of synchronized seasonal emergences
in tropical cicadas have not been studied. The possible adaptive significance of
allochronic emergences among species in the same habitats, in addition to op-
timizing courtship, may also minimize competition for oviposition sites. The
females of various tropical cicadas tend to prefer similar or the same oviposition
sites in the forest understory.(70,71l In lowland and premontane tropical rain
 180

160

III 140
....
(l)
•
CIl
Iii 120
~
u
<t 100
::c:
!l.
::i!! •
~ 80
~
c
(3 60
U
u.
o 40 ••
CIl
a: • ZAMMARA SMARAGDULA
w
~ 20 ••
:::l
Z

o 20 40 60 80 100 120 140 160 180 200

NUMBERS OF CICADA NYMPHAL CASTS 1972

FIGURE 7.14. (Contillued)

150 ....
140 I-
rJ)
Z 130 I-
;;;:
C/) 120
..J
« 110
I
a. 100
::;
>-
Z
90

« 80
0
«
()
70

(3 80
l.L.
50
0

Jll~~
rr: 40
(/)
::J
W (/)
zw

,. ll~l
CD
::; 30
()
=>

.,
20 0
Z z
10

0
i
" ci i ci
<Ii J:
1< w
~ w
~ ci ~ <Ii <Ii ci J: ai
w a: z Z w w
!l :; 0
() ::J !!j
:; ,::J
W
:; ,::J
::J ::J ::J
U. a: a. ::; u. u. a: u.

"'" "'"
::; "'" "'" "'" "'" "'"
::;
1973 1974 1976 1977 1978 o 1979 019800

MONTHS CENSUSED OVER SEVERAL YEARS

FIGURE 7.15. Discontinuous census histories for a Proarna cicada at Puntarenas. Puntarenas
Province, Costa Rica, over several years. [From A. M. Young, Milwaukee Pubi. Mus. Contrib. Bioi.
Geoi. No. 40 (1980). Copyright 1980 by Milwaukee Public Museum.]
316 CHAPTER 7

DICEROPROCTA SP.
1.2

1.1

...J
1.0
« U)
52 UJ
tuI G 0.9
UJ
o
... 0.
U)
0.8
0. «
>- 0
I «
o 52 0.7
Z 0
« a:
U) 0
UJ u. 0.6
!;: U)
:;; UJ
- > 0.5
tii a:
UJ a
~ UJ
- 0
0.4
U)Z
Z UJ
~ ~ 0.3
oZ- :;;UJUJ
!;: ...J 0.2 F.AMOENA
...J « ~
::J ::J
0. Z 0.1
o0. « Z

0.0
J F M M o N D

SUCCESSIVE MONTHS

FIGURE 7.16. Predicted annual emergence curves for adults of several species of cicadas (Homop-
tera: Cicadidae) in the lowland tropical dry forest zone of northwestern Costa Rica. The positions of
the peaks of these curves are based on actual observed densities, but the end points in several
instances are hypothetical projections. The dashed-line curves for two species represent data from
other localities and projected for this region. [From A. M. Young, Milwaukee Publ. Mus. Contrib.
Bioi. Geol. No. 40 (1980). Copyright 1980 by the Milwaukee Public Museum.]

forests, alternating bouts of wetness and dryness during the wet season are
sometimes accompanied by emergences of a small number of dry-season
cicadas. (7J) Such observations suggest that the cue for emergence is related to the
amount of wetness in the soil or to some related indirect effect of soil moisture
status.
At Khartoum, several species of dragonflies display fluctuations in adult
abundance related to the annual cycle of seasons. (549) The activity of the intertrop-
ical convergence zone (ITCZ) determines the local patterns of rainfall and dry-
ness in this region, and the dragonflies synchronize adult emergences with hot
but still moist conditions following the period of heavy rains.(549) Happold stud-
ied some 13 species and found that most appeared after the rains but also disap-
peared as conditions became cool and dry. Apparently periods of heavy rain fill
up breeding sites near Khartoum, and nymphal development is accelerated. In
the vicinity of Khartoum the periodic flooding of the Nile destroys dragonfly
nymphs so that adult popUlations appearing in the area are migrants from other
 EFFECTS OF SEASONALITY ON INSECT POPULATIONS 317

700

600 /\~
Cf)
II ',
f- 500 I
«
Cf)
I
u I

,
...J
« I
J: 400 I
0- I
~ I I

, ,,
>-
z I I
I I

,,
LL
0 300
a: I
UJ I

,
co I
~
::J I I
z 200 I
I I

100
I
I
I
I
,,
I

'~I
II

0
I
I , (I,
/

July I July Aug

1972 1973 1974

CENSUS DATES IN SUCCESSIVE YEARS

FIGURE 7.17. The wet-season adult emergence pattern for the cicada Z. smaragdula (Homoptera:
Cicadidae) in the central highlands of Costa Rica, based on discontinuous samples of nymphal casts
(skins) in successive years. The solid line connects actual data points for each year, and the dashed
lines show projected or hypothetical beginning and termination points for each emergence. [From A.
M. Young, Am. Midi. Nat. 103:155-166 (1980). Copyright by the American Midland Naturalist and
the University of Notre Dame Press.]

areas. Conditions are generally either too dry for breeding in small bodies of
water or too flooded so that the Khartoum odonate fauna comes from more
favorable areas upstream.(SSOl The annual northward movement of the ITCZ
brings in some species from more southerly locations. Happold's data illustrate
the considerable instability in the insect fauna of a desert tropical locality in
which insects require an aquatic environment for development. While the Khar-
toum area itself is relatively hostile for colonization by dragonflies as a result of
harsh fluctuating conditions over the year, further instability in the adult popula-
tions turning up there is a result of shifting seasonal patterns of rainfall else-
where. Temperate Zone dragonfly species exhibit highly synchronized life his-
318 CHAPTER 7

tories associated with seasonal changes in aquatic environments and duration of

the annual growing season.(551-555) Many Temperate Zone species of Odonata at
the higher latitudes are either summer or spring species in terms of the adult flight
season. But at more southern latitudes within the Temperate Zone, the distinction
tends to be obliterated because overwintering nymphal populations contain sev-
eral age classes. (554) In the nonseasonal lowland tropics similar patterns of over-
lapping generations, even more pronounced, may stagger emergences of adult
dragonflies throughout most of the year. The species believed to exhibit such
patterns would be those breeding in relatively permanent bodies of water such as
swamps and forest streams. Tropical dragonflies breeding in less stable bodies of
water in seasonal regions may exhibit rather disjunct annual emergence pat-
terns.(556)
In the Temperate Zone of southern Australia, annual emergences of the
pasture scarab beetle, Aphodius tasmaniae, a univoltine species, vary considera-
bly, depending upon the extent of dryness at different times of the year.(557)
Outbreaks of adult beetles occur when unusually dry winters and wet summers
occur at about 4-year cycles. Wet summer conditions promote increased hatcha-
bility of eggs, and dry winters prevent young larvae from drowning.(557) Dry
autumns result in drought conditions that kill off first-instar larvae, and wet
winters promote drowning and pathogenic fungal attack of larvae. The fluctua-
tions from year to year in seasonality and incidence of attack by pathogenic fungi
regulate the sizes of adult hatches. In tropical areas, emergences of many insects,
including scarab beetles, are expected to be staggered throughout the year, ex-
cept for species with long life cycles and very specialized ecological require-
ments. The tropical fauna is expected to contain a high number of multivoltine
species so that populations are somewhat buffered against unexpected changes in
climate. Halffter and Matthews(558) state that population cycles in tropical dung
beetles are synchronized with rainfall regimes: new adults generally appear at the
beginning of the wet season and do not breed until the end of the rainy season but
before the dry season. The dry season, in strongly seasonal regions, is passed in
the egg stage, and the adults die off. (558) But in regions without a distinct and long
dry season, population structure is usually very different and adults are equally
abundant throughout the year. (558) As tropical dry seasons tend to be more pre-
dictable in terms of duration and intensity, dung scarabs and other insects can
adjust their life histories to seasons of stress without appreciable losses due to
mortality. But in Temperate Zone systems such as Maelzer studied for pasture
scarabs, annual fluctuations in rainfall result in either considerable mortality to
some stage in the life cycle or increased survivorship in others. Even in univol-
tine tropical forms such mortality may be relatively lower than in Temperate
Zone forms, since the annual cycle of seasonality tends to be more precise,
particularly in the lowlands. (116) Thus the tropical dry season may regulate the
annual cycle of abundance and popUlation structure for many insect species in
EFFECTS OF SEASONALITY ON INSECT POPULATIONS 319

highly seasonal regions, but it does so in a relatively more uniform or even

manner from year to year.
Seasonal tropical regions do contain some insect species capable of display-
ing outbreak conditions.(51l) The advent of the dry season may alter community
structure in such a way that one or a few herbivorous insect species become
abundant. If the dry season alters the abundance and foraging habits of insects
and other organisms that feed on these herbivores, and if the herbivores exploit
plant species that remain evergreen to some degree in the dry season, outbreaks
may occur since biotic control on populations is diminished while food supply is
still present (even though perhaps somewhat reduced in total biomass). Kenyi<360)
noted outbreaks of the lepidopteran Belenois aurota on Capparis tomentosa
thickets during the first rains following the dry season in the Ruwenzori National
Park in Uganda. The first rains represent a period of increased new growth of
plants, increasing the food base for herbivorous insects, and the adult butterfly
population passed the dry season in substantial numbers. (360) Kenyi attributes the
observed severe defoliation by the butterfly larvae to the prior partial or complete
defoliation of thickets by elephants during the dry season, an event which by
increasing desiccation lowered the abundance of various predatory arthropods
dwelling beneath or in leafed thickets. Thus the series of biotic control factors
normally present in thickets by the time of the first rains were diminished and
herbivore populations increased, causing defoliation. Other changes in the as-
semblage of animals found in thickets in the dry season were also noted, such as
lower frequency of nesting by various bird species. In this system, the annual
variation in the grazing habits of elephants and other factors regulate the degree
to which particular thickets are defoliated during the dry season. Heavily de-
foliated thickets are ideal candidates for outbreaks of selected herbivorous insects
at the beginning of the wet season since (1) biotic control agents are no longer
present in defoliated bushes and (2) these bushes exhibit pronounced flushing out
of new foliage at the time of the first rains.
Some relatively thin-cuticle cryptozoic arthropods such as millipedes may
exhibit facultative changes in cuticular physiology in response to the alternation
of seasons in tropical regions. Such mechanisms permit these species to continue
to be active in both seasons without appreciable losses in fitness. Some genera of
millipedes such as Cingalobolus and Aulacobolus in India exhibit changes in the
composition of the cuticle in different seasons. (559) The epicuticle consists of
lipoproteins that function during the dry season to conserve water in millipedes.
Molts during the early wet season result in the deposition of cuticle without the
highly impervious epicuticle, since transpiration rates are lower in the wet season
and water loss is not a problem. Such synchronized cycles of cuticular organiza-
tion may also be present in many other arthropods, including insects in highly
seasonal tropical regions. The mechanism would be most adaptive in relatively
cryptozoic forms that feed on detritus and litter that is generally present through-
320 CHAPTER 7

out the year even though in varying degrees of dryness. In other words, a
relationship may exist between general feeding habits and ability to remain active
in a particular habitat in the dry season for arthropods, including some insects. In
species that can remain active, physiological changes such as periodic switches
in cuticle permeability may permit activity to continue under very different
conditions of dryness, so long as food is present.
Hoffman(S60) has demonstrated that the butterfly Colias eurytheme in the
Temperate Zone exhibits sex-related changes in wing pigmentation in response to
photoperiod and temperature. In this species the males are orange and the
females are white. The orange color results from the deposition of nitrogen-
containing pigments in the wing scales. Male butterflies exhibit sensitive changes
in color in response to changes in photoperiod but not to changes in ambient
temperatures.(S60) Low temperatures, on the other hand, promote some deposition
of color pteridines in the white females. Hoffman suggests that the effects pro-
mote the appearance of light-colored butterflies when the weather is hot and dry
(summer), an adaptation to thermoregulatory tolerances. Hoffman(56J) concludes
that in very predictable seasonal environments, physiological adjustments to
changes in the environment function to eliminate losses in fitness that could arise
from genetic tracking. It is a restatement of the phenotypic flexibility response
pattern that permits organisms to make short-term adjustments to different en-
vironmental conditions. (99)
In highly seasonal tropical lowlands, the majority of adult butterflies seen in
open habitats during the dry season tend to be light-colored, while the forms
(various satyrids and nymphalids) that aggregate in shaded strips of forest along
streams tend to be darker (A. M. Young, personal observations). Dark-colored
species such as Anartia fatima that remain active in open habitats often exhibit
daytime aggregating behavior under leaves in shady places. (562) During the dry
season in El Salvador the adults of Metamorpha stelenes tend to be lighter in
color than during the wet season.(3J7) The condition is caused by changes in the
breadth of the light-green portions of the wings and to the darkness of the green
color. In less seasonal lowland tropical areas, the color mix of butterflies in open
secondary habitats tends to be greater during the short dry season (A. M. Young,
personal observations). The general implication from such observations in the
tropics is that butterflies may experience a high incidence of thermoregulatory
stress during the dry season, a condition somewhat analogous to Hoffman's
summertime Colias popUlations in Kansas. Of considerable interest, however, is
the control mechanism. With relatively few generations per year and unpredicta-
ble boundaries to seasons in the temperate zones overall, the optimal strategy, as
predicted by Hoffman(56J) may be the physiological response for changes in
pigmentation. In the tropics, however, where there may be more generations per
year and seasons tend to have predictable boundaries, the genetic tracking
strategy may be optimal. Comparisons of temperate and tropical populations of
EFFECTS OF SEASONALITY ON INSECT POPULATIONS 321

Colias and other butterflies of seasonal zones may elucidate different or the same
mechanisms underlying the regulation of thermal stress in the dry season or
summer. Emmel and Leck(563) reported that butterfly species occupying clearings
in Panama showed a marked decrease in abundance in these habitats, whereas the
forest fauna increased during the same period (Fig. 7.18). They interpret these
results in terms of physiological stress from increased dryness in the clearing
during the dry season and the forest becoming an increasingly important moist
refigium as the dry season advances. The data illustrate the dynamic state of
tropical butterfly faunas in seasonal regions.
Owen et ai. (516) describe the seasonal abundance of some butterflies in
Africa in garden habitats. Further preliminary evidence from the nymphalid
genus Anaea from the northwestern tropical dry forest zone of Costa Rica indi-
cates that annual fluctuations in adult abundance are a function of phenological
patterns in the growth of larval host plants. (315) In a very deciduous forest site two
species of understory trees, Caesaria spp. (Flacourtiaceae), cease vegetative
growth of new meristems during the long dry season, during which time eggs and
larvae of A. (Zaretis) itys are absent. At a semi deciduous forest site (Barranca

55

50

:::
5 45
.i
~

:> 40
z

Clearing

OCT NOV DEC JAN FfB

,
MAlt

•
••
15
•••
••
FIGURE 7.18. Relationship between tropical season-
••
••

...
ality and the abundance of butterfly species in a forest
and forest clearing in Panama. [From T. C. Emmel
10
•••
and C. F. Leck, J. Res. Lepid. 8:133-152 (1970). Forest .4
Copyright 1970 by the Lepidoptera Research Foun- ~
••
dation.] 5
• ••
322 CHAPTER 7

site) the understory tree N ectandra sp. (Lauraceae), the larval host plant of A.
(Memphis) morvus, also ceases growth of new meristems, lagging somewhat
behind the Caesaria host plants at the more deciduous site (Fig. 7.9); in both
instances adult butterfly populations dwindle as the dry season progresses and
presumably adults enter into a reproductive diapause.(315) Populations of M.
peleides in such forest habitats also dwindle as the dry season advances( 9O ) (A.
M. Young, unpublished data), presumably since the growth of new meristems
slows down and ceases. Many tropical butterflies selectively oviposit on new
meristems, and the young larvae feed on' them.

7.4. MOLDING OF ECOLOGICAL TRAITS IN TROPICAL INSECTS BY

THE DRY SEASON

Tropical terrestrial regions are conceptual patchworks of heterogeneous

climatic regimes and vegetation cover. The continuous interplay of general
weather patterns and localized unusual or unique weather effects generates this
patchwork in space and time. Thus although one readily recognizes the lowland
tropical dry forest zones of Central America and elsewhere, upon closer examina-
tion one begins to distinguish considerable heterogeneity over short distances:
topographic effects interact with wind patterns to produce rain shadows and other
patchy environments. Such effects may result in marked discontinuities in the
vegetation cover of the region, and superimposed upon this is the degree of both
natural and man-induced disturbances of habitats that add more heterogeneity. In
a manner similar to that of Lewontin and White/ 54 ) who conceptualized an
adaptive landscape for Maraba grasshoppers in which peaks and valleys formed
a highly textured contour surface with peaks of different fitness levels, one may
cautiously envision a tropical region as being composed of different environ-
ments, each assigning a characteristic level of fitness for individuals of a particu-
lar insect species.
Thus although the large cicada F. mannifera occupies different habitats at
different actual elevations within the tropical dry forest zone of Costa Rica, the
popUlation size and density and the average fitness of individuals within different
populations may vary considerably in different habitats. And the degree to which
a population is exposed to seasonality will also vary. Local effects on the general
seasonal phenology of rainfall of the region may shorten or lengthen the adult
emergence period in cicadas and other insects, possibly resulting in different
populations being somewhat out of synchrony with one another in terms of peak
courtship periods. The effect is especially pronounced if patches of habitat and
insects are widely scattered.
Differences in local vegetation in different patches may result in different
sets of insects and other animals found there, altering the sphere of selection
EFFECTS OF SEASONALITY ON INSECT POPULATIONS 323

pressures in different patches. Host tree associations of cicada nymphs, pred-

ators, or pathogens on nymphs and adults(136,564) and interactions with other
cicadas may also vary considerably. Young(139) noted that different sites within
the dry forest zone of northwestern Costa Rica vary considerably in the numbers
of cicada species found. For some groups of highly mobile insects such as
butterflies and moths, such a patch structure may result in considerable turnovers
in insects associated with different habitats in different seasons. The migrations
of various butterflies vertically to higher elevations in the dry season and a
presumed return to the lowlands in the following wet season suggest that but-
terfly faunas associated with various habitats at different elevations may be far
from static. The response mechanisms to survive the dry season shape the ecolog-
ical traits of such species, as well as those that remain in a particular habitat
during wet and dry seasons. In this section we will examine how insects and
other organisms respond to tropical seasonality, with an emphasis on synchroni-
zation of individual life histories and changes in whole communities. In some
species of African mosquitoes with very seasonal adult emergence periods at a
particular locality, emerging adult populations are often highly congregated, with
mating taking place in less than 4 days-the maximal life-span of individuals. (200)
The timing of flowering in tropical plants must be related to the optimal time
of year for pollination. Distinct seasonal patterns of flowering in the
tropics OI6,118) are correlated with changes in rainfall. Temperature and humidity
may also regulate flowering in some tropical plants. (565) The observation that
many tropical plants are triggered into flowering by short-day cycles(566) suggests
that plants are sensitive to small shifts in photoperiod, conceivably of the kind
associated with day-to-day variations in cloud cover during the wet season.
Janzeno 16) was one of the first workers to conceptualize the adaptive significance
of plants and animals synchronizing their life cycles with the tropical dry season.
His argument was based on observations of the synchronized t10wering of many
tree species with the dry season in the lowland tropical dry forest zone of Costa
Rica. As the dry season progresses, trees lose most of their leaves and vegetative
growth is reduced. These trees burst into t1owering, and the flowers tend to be
large and conspicuous. Pollinating insects such as bees and wasps can locate the
t10wers more readily during the dry season when foliage density is reduced by
deciduousness. It follows that pollinating insects also synchronize their life his-
tories with the dry season in order to maximize utilization of resources (nectar
and pollen). Some of these insects track the phenologies of various tree species in
the sense that different tree species are visited at different seasons. Thus there is a
definite structure to communities in seasonal tropical environments as well as
relatively nonseasonal ones.(I17) The pollination systems of tree species in both
wet and dry tropical forests are very diverse, involving a variety of animals such
as insects, birds, and bats. oI7 ) Many overstory trees of tropical forests produce
large quantities of relatively small flowers, each with a small food reward.o 17>
324 CHAPTER 7

These trees are visited and pollinated by bees, moths, butterflies, wasps, and
beetlesOl7 ) and exhibit distinct seasonal flowering patterns. Understory trees and
shrubs tend not to have the mass-flowering strategy; bloom times are considera-
bly longer than in most overstory tree SO 17) and do not necessarily adhere to a
seasonal pattern.
In lowland tropical dry forests, many tree species flower in the middle of the
long dry season and are pollinated by medium- to large-sized bees. OJ7) Fran-
kie(117) also points out that other pollination systems predominate in this zone:
moths and small bees opportunistically pollinating various tree species in both
wet and dry seasons. In some tropical communities occupying habitats with
distinct seasons, closely related bee species show allochronic activity periods,
with one species active in the wet season and the other in the dry season.(S67) But
such differences in abundance of bees between seasons was not due to shifts in
the availability of pollen but rather to differential effects of rainfall on activities
of the bee species. One species tended to collect a greater amount of pollen from
a few plant families, while the other species showed a more even distribution of
pollen-collecting for many plant families (Fig. 7. 19). Frankie also discusses the
organization of pollination systems within seasons and in the transitional periods
between seasons. The greatest diversity of pollination systems in the dry forest
zone occur at the middle of the dry seasonY 17) Frankie points out that the life

70

60

50
(/J

'"
Ci
E 40
0
(/J

'0
~ 30
'"
.0
E
::J
Z
20

10

1 5 10 1
Number of families per sample
FIGURE 7.19. Pollen-resource differentiation be-
Exomalopsis Exomalopsis tween closely related bee species at the same locality
globosa similis in Jamaica. [From A. Raw, BiotropicQ 8:270-277
(300 samples) (120 samples) (1976).J
EFFECTS OF SEASONALITY ON INSECT POPULATIONS 325

cycles of pollinating moths such as Sphingidae may be synchronized in order for

adults to be most active at the early phase of the wet season when many moth-
pollinated plant species are in bloom.
Other species may be active in the middle dry season as the result of a
buildup of larval populations thriving on various shrubs and trees in riparian
forest communities of the region: many of these plants put out flushes of new
growth in the dry season in these moist habitats, providing a continuous food
base for some insects. Many moths involved in tropical pollination systems tend
to be large and can therefore withstand some amount of dryness. Larval popula-
tions may peak twice a year in tropical dry forest, at the beginning of the wet
season and at the beginning of the dry season (citation of D. M. Windsor in
Frankie(JJ7)). These peaks are maintained by the different vegetative growth
cycles of various trees and shrubs in selected habitats of the region; the early
wet-season peak is presumably the result of the plant communities overall re-
sponding by massive leafing out at this time, while the early dry-season peak is
presumably the result of many insects synchronizing the last wet-season genera-
tion with the end of that season (see Fig. 7.20). For cultivars such as T. cacao
and other endemics of tropical rain forest, flowering is maximized during the wet
season, and a build up of insect pollinator populations follows (Fig. 7.20). Such
plant species, when cultivated in seasonal environments as shown in Fig. 7.20,
do not maintain high levels of flowering during the dry season, and pollinator
populations also decline at this time. This effect is expected to be most pro-
nounced for pollinating insects of small body size, where the rate for a species to
drop out of the pollinating guild is directly proportional to the length and severity
of the dry season (expressed by the number of days or weeks of absolute dry-
ness), and the magnitude of the deleterious effects of dryness on body size and
physiology. Large-bodied insects may be less drastically affected and, in some
regions with long dry seasons, such pollinators may increase in abundance in the
dry season. In nonseasonal tropical regions or habitats, both flowering and pol-
linator abundance may be more or less evenly distributed over the year.
The dry-season synchronization of activity is a response to the impending
contraction of food supplies for larval popUlations that occurs in the dry season. One
might suggest that the early dry-season flush is smaller than the early wet-season
flush of insects, since more new foliage is available in the latter and predators,
parasites, and some pathogens may lag in response to the appearance of
phytophagous forms that are hosts. But there is also evidence that some species
of vespid wasps begin rearing brood in the dry season,<568) and these wasps are
predators on a variety of soft-bodied insects used to feed larvae. Such informa-
tion is indirect evidence that prey insect populations are available during the dry
season, and the greatest densities are found in moist habitats.(J34)
Frankie et al. (18) found that some solitary bees forage in great abundance at
mass-flowering overstory trees. Some bees visit many different individuals of a
326 CHAPTER 7

SEASONAL
! I
I
I
I
DRY WET DRY : WET
I

I
i
UJ I
a::
ez
I
I
I
I

""::::;
..J I
I
I
I
I
I
~
,,",,,
I
I
I
C I
Z
:
I
00(
\,
' .... _-,...-..._,-"
I

UJ / I
I
a:: / \1
w ,, ~ I
I ,
,

~..J , J.--.... I
... --------- I '

II.
NON-SEASONAL
II.
0
W
U
Z
g
Z
:;)
III
00(

I
I I

;....'
;'--',__. . --------------1...,
1','1
I I
~4fII"""-""''''

-------...... I I ... _--..... _", I

I
Feb Mar AQrMaY .....n.lu1 Aug$eQt Oct New Dec Jan Feb Mar Apr MEoy

SUCCESSIVE MONTHS

FIGURE 7.20. Hypothetical model for the changes in abundance of flowers and associated pol-
linator populations in seasonal and nonseasonal tropical environments. Flowering patterns are shown
by the solid lines and pollinator abundance by the broken lines.

certain tree species, providing a mechanism for outcrossing.(1l6.11S) During the

dry season medium- to large-sized solitary bees often forage at the top of the
forest canopy in tropical dry forest, (569) which has the greatest concentration of
flowers and lower densities of other bees such as Trigona species. (117) Tropical
bees exhibit diverse means of collecting pollen and exhibit specialized to
generalized foraging patterns.(S70) Schlising(S70) analyzed the insect visitors to
flower of Ipomoea vines during the dry season in the tropical dry forest zone of
Costa Rica. These vines exhibit mass flowering during the dry season at this site.
He found that although the composition of visitors varied at three different
habitats, there was evidence of considerable structure in the timing of foraging
for insects visiting the flowers throughout the day. Bees belonging to four dif-
ferent families exhibited distinct nonoverlapping foraging periods throughout the
EFFECTS OF SEASONALITY ON INSECT POPULATIONS 327

day. Species visiting flowers early in the day generally had longer visits, and as
air temperature increased, visits diminished in length, presumably as a result of
increased thermoregulatory problems. These vines and another studied, Aniseia
(also Convolvulaceae), are pollinated by these bees. Thus pollinating as-
semblages of bees associated with certain plant species during the dry season
exhibit considerable temporal organization on a daily basis. Such behavior is
undoubtedly linked to physiological constraints associated with desiccation in
bees, differences in nutrient requirements (qualitatively and quantitatively)
among bee species (nectar composition may change over a daily cycle,c572.573)
and the reduction of possible competitive conflicts over a patchy and seasonally
abundant resource.
Data on the succession of flowering seasons among 10 plant species in the
Monteverde cloud forest in Costa Rica that are pollinated by 14 species of
hummingbirds suggest that these plants compete for pollinators and that the
allochronic flowering periods lower competition for pollinators.(574) The simul-
taneous flowering of several plant species in the same area may distract hum-
mingbirds from making an optimal number of visits to each species and thus
lower fitness in these plant popUlations. The succession of Sphingidae in the
relatively nonseasonal tropical rain forest near Puerto Viejo in Costa Rica(575)
suggests that flowering periods in some sphingid-pollinated plants may also be
staggered throughout the year for similar reasons, although more data are re-
quired to confirm such a prediction. Seasonal succession of sphingids also occurs
in Amazonian rain forest. (576) The staggering of flowering times in many plants
species is likely to turn up in less seasonal tropical forests where both plant and
pollinator populations may have even distributions in abundance throughout most
of the year. Yet the marked suppression in cicada emergences near the end of the
wet season in lowland tropical rain forest°J) suggests that this period (1-2
months) may suppress activities in other insects as well. More seasonal areas are
more apt to exhibit the bimodal peak abundances in insects (Frankie(117»). Again,
a dearth of data from long-term studies prevents us from conforming or rejecting
such ideas.
As with pollination, different regions in the tropics contain seed predation
systems that mayor may not be synchronized with seasons. D. H. Janzen has
explored the ecological traits of seed predation systems involving insects and
vertebrates in both wet and dry forests of Central America.(29.362.363) In seasonal
tropical regions as well as some nonseasonal regions, different tree species will
exhibit distinct fruiting periods each year. For example, in the lowland tropical
dry forest zone, peak fruiting may occur in the early to middle phase of the wet
season for tree species with mass flowering in the dry season. Smythe(1OJ) has
summarized the major coevolved features of seeds and seed predators.
Smythe(lOJ) found on Barro Colorado Island in Panama that tree species with
large-seeded fruits tended to have very seasonal fruiting periods while those with
328 CHAPTER 7

small-seeded fruits tended to have rather nonseasonal periods. Competition for

dispersal is suggested to be greatest among tree species with distinct seasonal
patterns of fruiting. (IOJ) In lowland tropical forests where different tree species
exhibit distinct fruiting periods each year,(101,517) organisms which use these
fruits as resources may adjust their life histories to them, particularly if they are
specialists or very limited generalists (feeding on a few kinds of fruit).
Morton(577l discussed the synchronization of avian breeding cycles in frugivorous
species in tropical forests with.fruiting periods. A wide variety of insect groups
exploit fallen, rotting fruits as food supplies and breeding sites (butterflies,
beetles, flies, etc.). As it is known that various volatile products of fermentation
attract many different groups of Lepidoptera, (578) it is not surprising that a variety
of sweet-to-foul-smelling rotting fruits and seeds on the floors of tropical forests
attract a variety of insects, and in some cases, the resources emitting these odors
are the major food sources for insects. Many groups of Neotropical butterflies
regularly exploit such resources.(J86) Depending upon the degree to which the
insect is a specialist on a certain kind of fruit and the phenotypic constraints
arising from life cycle and tolerances to subtle fluctuations in temperature, air
humidity, etc., the population structure of such an insect species may be very
viscous or fluid. Young075 ,186) noted considerable localized movement patterns
in adult butterflies feeding on rotting fruits replenished experimentally in a low-
land tropical rain forest. Patchy spatial distributions in small insects such as
drosophilid flies may arise as a function of locations of rotting fleshy fruits and
changes in such distributions throughout the year. In other words, the distribution
of fruit-feeding insects on the forest floor is very dynamic and changing as fruit
distributions change. Changes in fallen fruit distributions will be a function of the
seasonality of fruiting in tree and shrub species in the forest. In lowland tropical
rain forests, a predicted organization of such insect communities is one charac-
terized by relatively low densities of insects associated with an accumulation of
fruit beneath an individual adult tree (in the fruit shadow) and the utilization of
many different resource patches (different types of fruit) per unit area. If an
insect species is a specialist under such conditions, the distribution is very patchy
and each patch may become a distinct subpopulation with its own breeding
structure.
Sterculia apetala trees are found in lowland tropical dry forest in Central
America and develop large crops of pods that mature in the dry season.(365) The
seeds fall to the ground and are fed upon by Dysdercus cotton stainer bugs,
which may kill virtually all seeds in forest habitats. (365) Individual trees in pas-
tures produce seed crops that stand a greater chance of surviving from bugs but
die from desiccation. Various small mammals such as squirrels and monkeys are
the major dispersal agents of the seeds, as they carry off pods from trees and drop
some of the seeds elsewhere, a behavior resulting from the exposed use of these
animals to hairs from the inner walls of the seed pods. (365) The cotton stainer bugs
EFFECTS OF SEASONALITY ON INSECT POPULATIONS 329

opportunistically feed on the abundant seed crop of Sterculia in the dry season
but presumably switch to other foods at other times. In southern Ghana, for
example, Dysdercus bugs exploit several different tree species throughout the
year, moving to different species as they come into fruiting and returning to
Hibiscus, which flowers and fruits throughout the year when the other species
cease fruiting.(470) Hibiscus in the Savanna zone acts as a reservoir for cotton
stainer populations to build up and migrate from when suitable host trees are
fruiting. In the forest zone, Sida trees make up the predictable reservoir in terms
of flowers and fruits for these bugs. (470)
A major component of the adaptive strategy of coping with the tropical dry
season in the apparently migratory bug Dysdercus bimaculatus is the exploitation
of those host seeds with the highest amount of energy. (579) Nutrient availability
places a constraint on the evolution of optimal body size in relation to selection
favoring either migratory behavior or diapause to pass the dry season.(579) Mi-
gratory behavior, even if relatively localized, permits an insect to exploit all
available energy-rich food resources and becomes the primary strategy, since
diapause involves a more drastic reorganization of the phenotype.
These examples of pollination systems and seed predation systems illustrate
how insects synchronize population cycles and feeding habits with a seasonal
succession of resources in the tropics. Many insect species exploiting plant parts
or products as food resources are adapted to exhibit facultative switches in
response to a changing spectrum of available resources if the average fitness of
individuals within breeding populations is to be preserved through periods of
environmental fluctuation, even if predictable.
Rockwood(22s.226) analyzed the flexible foraging behavior of two species of
Atta leaf-cutting ants in the lowland tropical dry forest region of Costa Rica.
During the wet season, foraging was primarily diurnal, while during the dry
season it was nocturnal. Rockwood suggests that these changes in foraging
behavior might have been related to effects of seasonal changes in rainfall on
vegetation cover and other features of the habitats. Various factors also influence
the selection of leaves for cutting by Atta, young leaves generally being pre-
ferred,(580l and shifts in foraging patterns may be related to availability of pre-
ferred leaves in a habitat. Rockwood found that the greatest quantities of leaves
and flowers were harvested at the beginning of the dry and wet seasons, periods
during which new growth and flowers were abundant in the habitat. Mature
leaves are only minimally harvested in both seasons.(22S) Rockwood's data pro-
vide another example of how seasonality affects the availability of optimal food
resources for specialist insect species: the supply of such resources (new plant
tissues and flowers in this case) varies in different seasons, and the insects
display facultative changes in foraging habits to compensate for the changes. The
growth of Atta colonies depends upon the selectivity of food resources brought to
the nest for fungus gardens by the ants. (581) These data point to the existence of a
330 CHAPTER 7

phenomenon now generally accepted by biologists: the tropics are not uniform in
space and time in terms of the availability of resources required by individual
species fragmented into breeding popUlations, and whole assemblages of or-
ganisms are organized in specific ways that may change both through cycles of
contemporary time such as seasonality and through evolutionary time.
Temperate Zone assemblages of insects such as Drosophila species occupy-
ing the same region or habitat may contain species that exhibit different peak
activity periods corresponding to different seasons.(582) Some species may peak
in annual abundance during the spring and summer, while others peak in late
summer and autumn, and some species may exhibit different preferences for
habitats.(582,583) In these species there is considerable dietary overlap with niche
separation manifested in spatial and temporal separation of populations. Pipkin et
al. (584) found mixtures of specialized and generalized species of flower-feeding
Drosophila species at several tropical sites. The flower-feeding habit is syn-
chronized with the seasonal flowering periods of the preferred plant species, and
some flies enter into a diapause in the dry season when these flowers are not
available. (584) Although such species exhibit specific morphological and be-
havioral adaptations for flower feeding, on a per species basis there appears to be
considerable flexibility in feeding so that some species can be polyphagous while
others, more restricted, are monophagous. Monophagous species are ideal can-
didates for enering into reproductive diapause during seasons when their pre-
ferred host plants are not blooming in sufficient numbers to maintain breeding
popUlations.
A similar system may be present for phytophagous insects in seasonal
tropical regions: if a portion of the host plant spectrum utilized by a particular
polyphagous insect species in a seasonal region remains lush to some degree
during the dry season, the population may continue to be active, barring effects
of other ecological constraints, but perhaps reduced in size proportional to the
loss of resources. Such a situation may be expected for some species that utilize
host plants with new growth in riparian forest habitats. If the host plant distribu-
tion of a monophagous species is such that a major portion of the population goes
deciduous in the dry season, the optimal response strategy would be to evolve
facultative diapause or migrate to a higher elevation where conditions are moist
and the host plant population is still evergreen. Thus we might expect a greater
incidence of reproductive diapause (or pupal diapause) or migratory ability in
monophagous insect species in very seasonal lowland tropical regions than in
polyphagous species, assuming that plant distributions include habitats where
some degree of evergreenness is preserved throughout the dry season.
Monophagous-feeding Drosophila in the tropics exploit predictable and
abundant flower species(584) sufficient to maintain breeding populations of these
species. Polyphagy seems to appear when anyone potential resource type is
neither predictable nor abundant. For some insect groups, one might expect that
EFFECTS OF SEASONALITY ON INSECT POPULATIONS 331

polyphagous species are more prevalent in strongly seasonal regions where some
resources of these particular groups are only seasonally abundant and predictable
in distribution. Ecological specialization in some fungus-feeding Drosophila
species is generally limited to specie's exploiting mushrooms available in dif-
ferent seasons and habitats throughout the year, even though competitive interac-
tions may reduce niche size in such species. As preferred resources fluctuate in
quality and abundance temporally, there will be a greater tendency for insect
species exploiting such resources to become ecologically flexible and possess a
broad niche with respect to those resources. Any single tropical forest ecosystem
will contain species of plants with varying intensities of physiological response to
seasonality, even in regions where the dry season is relatively short. Thus tropi-
cal forest ecosystems represent highly fluid mosaics of resources through con-
temporary time, requiring or promoting both specialist and generalist forms
within different communities, depending upon the degree of fluctuation in prop-
erties of the resources affecting fitness in insect populations. Such mosaics may
be more oscillatory in very seasonal tropical regions.
The sweep-sample studies of Janzen and Schoener in the tropical dry forest
zone of Costa Rica<l34) show that insect species are more abundant in moist
habitats than in dry habitats during the dry season and that insect communities in
close proximity may have disjunct distributions of species depending upon the
position of the community in a moisture gradient. As the dry season becomes
more intense, such discontinuities are expected to increase.(134) Yet in all com-
munities examined, Janzen and Schoener found a greater number of insect
species during the dry season than in comparable Temperate Zone communities
sampled in the growing season. Thus even a very dry tropical habitat during the
dry season tends to support more insect species than Temperate Zone com-
munities. Information on the natural history of component species in such tropi-
cal communities would reveal why species richness is maintained during a stress-
ful season.
Size of suitable habitat, and abundance and distribution of essential re-
sources within the habitat, are two major determinants of the success or failure of
an insect species invading the habitat in response to selection or changes in
environment elsewhere. If the habitat is too small or the resource base within it
too small or of low quality during periods of peak reproduction (in a seasonal
environment), the invading species will go extinct within a short period. Duf-
fey(585) found that a habitat in Britain was too small to maintain a breeding
population of the large copper butterfly, Lycaena dispar. If the resource base or
habitat is too small, the population cannot become large enough to (a) buffer
against mortality of various sources and (b) allow for sufficient breeding activity
to produce the next generation. In the lowland and mid-elevation tropics, one can
conceptualize tropical forest habitats as being dynamic mosaics of natural and
artificial (human-induced) light gaps interspersed with forest, and an ever-
332 CHAPTER 7

changing size distribution will be associated with such habitats. The likelihood
for colonizing species to survive in them will be determined in part by the size
and quality (in terms of resources) of each habitat at anyone point in time.
Turnover of insect faunas associated with such heterogeneous environments is
expected to be high.
Insect species associated with plants in wetter tropical regions may remain
active in the short, often erratic dry season characteristic of many tropical rain
forests. Perhaps some of the best evidence for such sustained activity is that from
the analysis of population size fluctuations in the butterfly H. ethilla in
Trinidad.(45I) There were remarkably stable popUlation sizes throughout most of
the year, with only a slight depression in the few dry-season data available (Fig.
7.12). Ehrlich and Gilbed45I) ascribe the observed high stability of the butterfly
population to a combination of factors, including high adult longevity, high
reproductive output, predictability of adult and larval resources, and an absence
of severe fluctuations in the environment, i.e., rainfall. The relatively mild dry
season of this region has runs of wet days, during which many insects can remain
active. The observed continued activity of many butterflies in forest understories
throughout the year in tropical rain forests 086 ,563) results from (1) the generally
weak effect of the short dry season on the flora and (2) the preservation of wet
conditions in the understory, an environment generally shielded from the direct
effects of the dry season. Forest-edge popUlations of Parides butterflies in tropi-
cal rain forest zones are generally active throughout the year (A. M. Young,
unpublished observation; see also reference 586). Cook et al. (444) found sustained
adult activity of H. charitonius in a mid-elevation tropical rain forest zone
throughout the year, even though the study period included the short dry season.
The extent to which the annual dry season is intense determines in large part the
degree to which arthropod communities associated with rotting carcasses are
cycled through the ecosystem and the organisms expediting the recycling.(191l
The tropical cash crop T. cacao is an understory tree species with small,
cauliflorous flowers generally pollinated primarily by tiny midges of the dipteran
family Ceratopogonidae.(406,587) Because of the general lack of dense populations
of these midges in many cacao plantations in both the New and Old World wet
tropics, pollinator abundance is believed by some workers to be a limiting factor
in fruit set. (522,588) An interesting feature of cacao phenotypes is that they occur in
both seasonal and nonseasonal tropical climate regimes, although a distinct wet
season is needed for normal levels of flower production. PopUlation sizes of
midges in some cacao plantations may be determined in part by the abundance of
suitable breeding and development sites on the ground cover beneath mature
trees.(58&-590l In nonseasonal cacao-growing regions throughout the year, midge
populations may exhibit small oscillations and fruit set, although low, tends to be
the same (A. M. Young, unpublished observations). There is some evidence,
however, that in seasonal regions of Africa where cacao is grown, massive
EFFECTS OF SEASONALITY ON INSECT POPULATIONS 333

flowering occurs early into the wet season, and fruit set at this time is very low.
According to Myers(591l the native habitat of cacao is the understory of Guiana
rain forest, and because of this origin, the species does best in cultivation where
daily moisture fluctuations are minimal, a property requiring the use of shade
cover in plantations.(592,593) In the Gold Coast, the dry season is a limiting factor
for adequate growth of young cacao trees.(594) Because the early stages of these
particular groups of ceratopogonid midges require very moist conditions for
successful development and because adult midges are subject to desiccation very
easily, pollinator populations may lag behind flowering periods following a dry
season. Fruit set in subsequent flushes of flowers later in the wet season may be
higher since midge populations have the opportunity to recover from the dry
season. In seasonal environments, cacao production provides an example of how
pollinating insect populations may be severely curtailed in abundance by the end
of the dry season and how the ensuing asynchrony between flower abundance
and pollinator abundance may result in very low fruit set. The variance in fruit set
from tree to tree for a given strain of cacao in a single plantation is expected to be
greater at the beginning of the wet season in a very seasonal region than in
subsequent periods of fruit set in the wet season. Pollination systems involving
larger-body insects may not exhibit such behavior, since these insects may sur-
vive the dry season without appreciable losses in popUlation size or density.
Keep in mind, however, that for cacao and other trees of tropical forests,
production of mature fruit is influenced by factors other than pollinator
abundance. It is interesting to note that cacao originally evolved as an understory
tree in Amazonian rain forests,(595) where there was probably a well-established
equilibrium between pollinator abundance and abundance of flowers in small
clumps of the tree. The environment was essentially nonseasonal, and pollinator
popUlations were probably stable throughout each year. The' 'invasion" of cacao
as a cultivar in more seasonal tropical regions has been only partly successful
from a commercial viewpoint, since seasonality of rainfall imposes annual
periods of environmental stress on pollinator populations under these conditions.
 CHAPTER 8

DYNAMICS OF
ORGANIZATION OF INSECT
COMMUNITIES IN TROPICAL
ECOSYSTEMS

8.1. NOTES ON TROPICAL ECOSYSTEMS

The relatively few studies of tropical ecosystems available have emphasized the
pathways of nutrient cycling in some tropical forests(S96) or descriptive and ex-
perimental studies of biotic organization of communities.(29.65,43S) Okwakol(597)
demonstrated in an African termite Cubitermes a relatively high rate of removal
of organic materials from soil, suggesting that in habitats where termites are
abundant, cyclic depletions of the nutrient content of soil are frequent, thus
imposing some constraints upon other litter organisms. To what extent, for
example, termites compete with collembolans for certain nutrients under such
conditions may indicate food as a limiting factor in collembolan populations. It is
clear, regardless of the interactions, that the litter mosaic of tropical forests is
very dynamic in terms of turnover of accessible nutrients and breeding sub-
strates. Historically, much of tropical biology has focused upon expedition-type
collecting of specimens for museums and universities, development of cultivars
and agronomic systems, and medically related entomology. The entomological
literature, for example is full of detailed descriptive and experimental studies on
insects as vectors of various tropical diseases and the effects of changing weather
patterns on predicting outbreaks. Much of the earlier work in the tropics has
deployed a "targeted" approach: in a particular study, an emphasis was placed
on few taxa or single species. Surprisingly, a need remains for detailed natural-
history studies of tropical organisms to feed the conceptual and theoretical
modeling of how communities and whole ecosystems are put together.
In this book, I define an ecosystem as the sum of abiotic and biotic compo-
nents of a distinct environment or habitat such as a forest. Within that environ-
ment, there are one or more communities of organisms. Each community is
associated with a particular trophic level in the ecosystem, and each occupies a

335
336 CHAPTER 8

distinct microenvironment or microhabitat. The community of decomposer or-

ganisms lives in the forest floor litter microhabitat. A community is composed of
a distinctive but dynamic set of species each maintaining a breeding population.
Sometimes communities contain transient species that do not maintain breeding
populations in them. It is necessary to be able to recognize transient and resident
species in communities. Depending upon the kinds of resources exploited by the
species in a particular community and the temporal and spatial distributions of
these resources (i.e., their patchiness), some resource patches may accumulate
over evolutionary time a variety of species exploiting different parts or products
of the resource. A group of closely related and ecologically similar species, a
guild, exploits a specific portion of a resource, although different sets of species
may occur together on the same resource. Thus seedlings of P. macroloba in
lowland and premontane rain forests on the Caribbean watershed of southern
Central America may have several distinctive sets of species feeding on them.
Both the number of sets and their species composition may change over ecotonal
and geographic gradients within the distribution of the tree species. Thus the
seedlings support one or more guilds of phytophagous insects, and the commu-
nity is the collection of such guilds in a particular area of habitat. Usually the
species within a community are taxonomically and ecologically similar, and they
usually exploit the same structures or substances from the resource. Different
subfamilies are represented in the rain forest understory butterflies exploiting
juices liberated from rotting fruits molding on the forest floor. (86) Although
within this community there is a guild of satyrid butterflies, many other satyrids
do not exploit the same resource. Guilds are almost always composed of closely
related species.
In the tropics, ecosystems are generally more complex because the vegeta-
tion cover is generally more diverse in lowland areas, there is greater structural
complexity to habitats, and consumer and decomposer trophic levels are more
complex in terms of numbers of species. The vertical stratification of a lowland
tropical rain forest,r394) for example, imposes a vertical gradient in light which
differentially attracts different kinds of insects and other animals. The rain forest
is the ecosystem, and different insect communities are found at different levels
because plant and debris resources change vertically. The convergence of dif-
ferent groups of butterflies, for example, at different strata may determine the
kinds of mimicry complexes found at different levels of tropical rain forest. (74 )
Horizontally, tropical regions display a great deal of heterogeneity in the organi-
zation and biotic complexity of plant communities in different successional
stages. Each successional stage may be considered as a separate ecosystem, with
a unique set of communities of plants and animals, with a different array of
structural complexity.
Connell(6S,43S) has provided evidence that tropical rain forest ecosystems
may not be as stable as generally believed: a lot of turnover in species results
DYNAMICS OF ORGANIZATION OF INSECT COMMUNITIES 337

from a variety of environmental perturbations. Tropical forest ecosystems, pri-

mary or secondary successional stages, are dynamic in that there are continual
localized changes within the species composition and structure of communities
and guilds within them. Because such complex ecosystems contain interactions
among species that are specialized or generalized, an understanding of tropical
communities, especially in terms of plant-insect interactions, is achieved by
examining some examples of how various kinds of insects interact with resources
in tropical habitats. In one sense the stability of an ecosystem is determined by
the degree of stability and nutrient profile of soils. If soil is nutrient poor but
supports a stand of trees with well-developed chemical defenses that leach out
into the soil from plant parts, the ecosystem may be stable though low in diver-
sity. <3(2) Nutrient-rich soils may promote a greater diversity of plant species, each
with its own characteristic "defensive compound shadow," and it is this sort of
soil profile heterogeneity that endows some degree of instability in tropical forest
ecosystems. Instability also arises from natural perturbations such as tree falls
and floodings. Yet at the same time that various factors promote instability in
tropical forests, many others promote stability and the resulting accumulation of
species-rich ecosystems. Some of these factors are explored in the sections to
follow.

8.2. PATTERNS OF ORGANIZATION IN TROPICAL FORESTS: SOME

COMMENTS

Pires et al. (325) measured the abundance of tree species in an Amazonian

rain forest and emphasized the high number of tree species found per unit area
and the very different tree species distributions that may exist within different
regions within Amazonia. Although previous workers maintained that much of
the great diversity of tree species in tropical forests was the result of the interplay
of stable and benign climate with ecological interactions such as competition,(63)
Janzen(29) provided evidence suggesting that specialized interactions between
seed and seedling predators (herbivores) and tree species generated that patchi-
ness characteristic of many tree species in tropical forests. Janzen proposed that
each tree species has a host-specific set of herbivorous insects that kill large
amounts of seeds and seedlings beneath the parent trees, while the few survivors
are those carried off by various vertebrates. The net result is a lowering of
population density for each tree species for both adult and preadult age classes,
and an increase in the distance between parents and offspring. Such a system
results in a greater number of tree species found over relatively small areas of
tropical forest. Janzen proposes that the system is most prevalent in lowland
tropical rain forests, and becomes less likely as one moves into dry forest and
into the Temperate Zones. Severity of physical factors such as temperature and
338 CHAPTER 8

rainfall fluctuations tend to lower the opportunities for ecological specialization

by herbivorous insects and seeds and seedlings through evolutionary time. Jan-
zen's model also predicts that specialized coevolved relationships between her-
bivorous insects and seeds and seedlings will keep population densities of indi-
vidual tree species below abundance levels which otherwise result in competition
among species. An important implication of the model is the historical com-
ponent: such specialized interactions have evolved over long periods, yet strong
stabilizing selection should not be operative to preserve these traits.
The integrity of the tropical rain forest flora, so to speak, depends upon the
maintenance of sufficient genetic and phenotypic flexibility to cope with changes
in the coevolved association. The relationships are not static, since coevolved
associations involve the expression of chemical defense barriers by plants to
herbivores, and herbivores must change physiologically or biochemically to over-
come such changes if the species is to continue exploiting that resource. Such
coevolved associations are expected to be most frequent for species of insects
that lack alternative strategies of exploiting other resources without going extinct
locally or suffering appreciable losses in fitness. The greater the frequency of
coevolved associations, which are more likely to evolve under continually favor-
able physical conditions of the environment(!Q4) as long as they are nonoverlap-
ping, the greater the evolutionary accumulation of species in a habitat or region.
The data are impressive for many groups of tropical butterflies.(6)
Going back to the formulation of MacArthur for adding additional species to
tropical communities (Chapter 2), herbivore specialization of a coevolved form
provides space in the ecosystem for other plant species that are poor compe-
titors. (29) In tum, these new species in the ecosystem can provide new resources
for other insect species and develop their own coevolved associations. Janzen
points out that tropical herbivores are apt to be effective in overcoming the
chemical defenses of plants and that the more effective such interactions become,
the less likely it is that plant species will survive in a habitat. Thus extinction
rates of plant species are predicted to be higher in the tropics than in Temperate
Zone regions, and tropical rain forest communities are therefore very dynamic
over time. Connell(6S) shows that the great species richness of tropical rain forests
is a nonequilibrium condition of the sort discussed by Janzen(29): species abun-
dances are well below an equilibrium that would result if each species were to
realize the carrying capacity of the habitat. Perturbations will result in a lower
diversity equilibrium state if coevolved associations between plants and their
herbivores are altered, thus making it possible for some species of trees to
become dominants. In the evolution of lowland tropical rain forests, the over-
story tree species make up a dominant feature of the ecosystem, and these trees
were pioneers relative to the rest of the forest (understories). Federov(S98) remarks
that the spatial distributions of overstory trees are great, while those of under-
DYNAMICS OF ORGANIZATION OF INSECT COMMUNITIES 339

story trees tend to be lower. Understory trees are specialists physiologically for
shaded conditions and are more recent members of the community.
At present, we have too few data on the extent to which many tropical forest
trees have host-specific seed and seedling predators, and comparing these associ-
ations between overstory and understory tree species in a habitat would be
especially revealing. Early stages of succession in the wet tropics are charac-
terized by large, monospecific clumps of trees, a result of the absence of host-
specific seed predators in these habitats. Either such species never evolved such
associations with insects, or else the associations were lost as the result of
changes in the vegetation cover of a region. Monospecific clumps of trees may be
stable over long periods of time and thereby provide predictable resources for
some phytophagous insects. But one would expect the frequency of outbreaks of
a few insect species associated with mono specific clumps of trees along the
borders of primary forest to be relatively rare. The combination of benign physi-
cal environmental conditions and the high predictability of the food base (leaves,
etc.) for these phytophagous species results in predictable resources for some
predators and parasites of these insects, and effective biotic control may result.
The control system may be less stable in highly seasonal tropical regions, and
some outbreaks are expected under such conditions. Depending upon the con-
straints on life history traits shaping r, an insect species feeding on a large food
base may not exploit the food base and thereby provide space for other species.
In this manner monospecific stands of plants in lowland tropical rain forests may
accumulate phytophagous species, either as a community associated with that
plant species or as separate guilds on the plant.
Studies show that insect outbreaks are related to the aftermath of typhoons.
A typhoon may drastically alter the resource base for an insect species, as
indicated in an analysis of outbreaks in the spruce bark beetle I. typographus in
Hokkaido. (599) One of the limiting factors operating on populations of this insect
is the availability of fallen trees of the host species as the insect tunnels beneath
the bark of these trees. Equilibrium maximal densities of the beetle popUlation
are estimated by counting entrance holes of burrows, and the average is 60-70
burrows per square meter of bark on fallen trees. (599) Yet after intense storms and
typhoons, the density of fallen trees increases tremendously and the density of
entrance holes of the beetle exceeds 350 per square meter.(599) Thus outbreaks
result from density-independent factors affecting the resource base of the insect,
that is, by increasing the resource base and number of insects responding in a
density-dependent fashion to an increased food supply. The data illustrate how
large-scale abiotic factors can disrupt the equilibrium population densities of
insects in complex ecosystems. A recent analysis of light-gap dynamics in tropi-
cal forests(337) suggests that tree falls occur frequently and alter the vegetation of
such ecosystems. Depending upon the size of the area affected and the tree
340 CHAPTER 8

species in it, and the life history properties of insect species in it, and the life
history properties of insect species in the area, outbreaks of tropical insects may
occur as the result of such environmental perturbations.
The ebb and flow of coevolved associations between insects and plants in
tropical forests is mitigated by the genetic structure of both insect and plant
populations. Outcrossing is expected to be frequent in tropical forest trees as a
means of providing a mechanism for maintaining genetic variation to cope with
changing selection pressures. (40,499) The evolution of nectar reward properties is
shaped to a large degree by the presumed adaptive significance of outcrossing in
tropical forests.(216.217) Pollinating insects such as bees exhibit a diversity of
methods for collecting pollen,(600) and their foraging strategies are molded by
energy requirements and the spatial distribution of nectar and pollen sources. (40)
In some groups of bees there is considerable ecological specialization for exploit-
ing specific groups of plants, a classic example being orchid bees. (601l The degree
to which a bee species is specialized for visiting certain kinds of flowers and
other resources, the spatial distribution and abundance of these resources, effects
of seasonality on resources, and the nesting requirements (including how nest
sites are determined by proximity of resources(21l), will determine the population
structure of such species. Different species within a single genus or subfamily are
often very specialized in the tropics, suggesting that evolutionary divergence in
foraging habits evolved as a means of lowering competition. If so, then bees tend
to be at equilibrium population sizes in the tropics.
Along with hummingbirds and some flies, some tropical bees are oppor-
tunistic feeders on other products such as the excreta of coccoids.(602) Such
products may be supplementary sources of nitrogen needed for dietary optimiza-
tion. Some tropical seed-eating forest floor grasshoppers are opportunistic pred-
ators on insects at lights (A. M. Young, unpublished observations). Different
genera and species of sand wasps in South America exhibit different prey prefer-
ences and nesting habits. (603) Such differences may reflect ecological diversifica-
tion in response to an abundant and diversified supply of prey insects and mic-
rohabitat differences. Limiting distributions of resources, for example, optimal
nesting sites in the case of Neotropical sphecid wasps such as Trypoxylon, may
necessitate the evolution of male guarding behavior if strong competition exists
for nest sites or mates.(604) Strong intermale competition in these wasps at nesting
sites suggests that females develop a mating strategy maximizing mate choice.
Optimal mate choice systems may be more prevalent in tropical insects than
originally suspected, since the intensity of selection pressure from multiple
sources, including predation and parasitism, is expected to be high. This condi-
tion favors the maximizing of opportunity to promote the maintenance of the best
suited alleles within breeding populations. Limiting resources such as food
supplies, nesting sites, and mates are expected to play major roles in shaping the
mating strategies of tropical insects. Although clearly many different kinds of
DYNAMICS OF ORGANIZATION OF INSECT COMMUNITIES 341

insects in the tropics are specialists, there may also be considerable phenotypic
flexibility in such species, expressed as opportunistic feeding on other resources.
Thus although biological organization in tropical ecosystems is very evident,
many species in associated communities will exhibit varying amounts of be-
havioral and physiological flexibility. (99) Such capacities may be significant in
permitting species to switch feeding habits in response to changes in community
or guild structure. There can be facultative switches associated with predictable
changes in resource abundances (seasonality, floodings), or obligatory switches
associated with addition of new species to the community and possible competi-
tion. We need to gather more data on whether insect species in different com-
munities are at equilibrium population sizes or densities.

8.3. NOTES ON THE DYNAMICS OF TROPICAL INSECT

COMMUNITIES

The majority of insects found in tropical habitats exploit plants and plant
products as food resources. An observation of a particular insect species feeding
on a particular plant species is to be interpreted in terms of (1) the degree to
which the relationship is the original coevolved association or one developed
from the original one in the past ,(436) (2) the degree to which the insect is a
specialist on this plant species, and (3) the spectrum in fitness within the insect
population on each plant species if the insect is polyphagous or exhibits broad
monophagy.(27S) It might also be advantageous to know something about the parts
or substances of the plant used as the resource and the extent to which the resource
is shared with other species. Such measurements should form a basis for con-
tinued research on the natural history of tropical insects.
At the conceptual level, phytophagous insects regulate the biomass and
diversity of plants in an area.(60S) Severe defoliation of plants by insects, espe-
cially in areas where species numbers are high and there are seasonal periods of
peak abundance, results in losses in fitness in plant populations ,(474) which in tum
should result in counterselection to reduce defoliation in future generations.
Some amount of defoliation contributes to the maintenance of low-density popu-
lations, along with host-specific seed predation, and when plant species are
regulated in this way, diversity within communities increases over evolutionary
time. (606) Patch structure for a given plant species will be generated by the effects
of dispersal agents and biotic mortality agents affecting the survivorship of adults
and juveniles, and also by various density-independent factors such as flooding
or fires. In lowland tropical dry forests, fire is a major factor contributing to
patch structure of vegetation. Coevolved interactions between plants and her-
bivorous insects can reduce the likelihood of competition among plant species as
a source of patch structure: host-specific predation by insects on seeds and
342 CHAPTER 8

seedlings reduces the likelihood for some tree species to become dominants in
tropical forests ,(29,65) and insects grazing on leaves and other vegetative structures
may promote diversity by reducing population sizes of adults. In specialized
habitats with several ecologically similar invertebrate species, reduction in
niche size of some resident species can result from competition with other
species.(607,60S) The evidence for such niche reduction would be nonmicrosympat-
ric distributions in which each species occupies a different microenviron-
ment.(60S)
In some Temperate Zone communities, plant diversity and structure corre-
late positively with diversity of Homoptera(414); similar patterns are expected in
tropical communities. But for some insect taxa, host nonspecificity in the tropics
is increased(4IS) in response to individual resource types being too patchy in
spatial distribution.
Depending upon the parameters of population growth and dispersal abilities
within populations of individual resident species in tropical communities, there
may be specialization or lack of it. Long-distance pollination of some trees in
tropical forests by euglossine bees(40) makes it possible for these insects to
specialize on these resources, whereas insects with more restricted movement
patterns may become generalists on several alternative resources distributed as
small patches in a relatively small area of habitat (where small is defined relative
to the dispersal prowess of the individuals). Specialization can evolve if patch
size is very large relative to dispersal ability and popUlation growth rates.
Karr<609) demonstrated with comparisons of temperate and tropical avian faunas
of forest and field habitats that the higher diversity of birds in the tropics is due to
several different factors such as reduced energy requirements for breeding,
greater vertical stratification of faunas, and increased abundance of insectivores.
Further heterogeneity in avian distributions in the tropics results from con-
vergence of many species to exploit fruits that are distributed in a patchy fashion
within forests. (609) Bird species specializing in insects as food are less abundant in
the tropics.(43) Insectivorous bird species specializing in larger insects are more
abundant in the lowland tropics than in the temperate zones or at higher eleva-
tions in tropical zones.(4J) But in seasonal tropical lowlands many bird species
feed on both fruit and insects(6!O) because seasonal fluctuations in insect abun-
dance reduce insects as a food source at certain times of the year. Thus selection
pressure from avian predators may be greatest on some insect popUlations at the
beginning of the wet season in strongly seasonal regions of the tropics, but the
degree to which popUlation increases at this time satiate such predators is not
known. In less seasonal regions, predation pressure from birds may be more
evenly distributed throughout the year.
Sweep samples of foliage insects in various tropical habitats reveal that
groups such as Homoptera tend to be very diverse on herbaceous vegetation.(72)
Working in the Temperate Zone, McClure(2sJ) concluded that leaf hoppers exhibit
DYNAMICS OF ORGANIZATION OF INSECT COMMUNITIES 343

considerable niche diversification by utilizing different portions of the same

plant. Microhabitat differentiation is well documented in Homoptera in salt
marsh grasses in the Northern Temperate Zone.(73) If the species within a com-
munity are in equilibrium, resource partitioning results in coexistence. (281)
Frequency-dependent competitive ability within a community may result in
coexistence of species if high densities attained by a species reduce its fitness
more than that of a competing species.(61l) Although in the nonseasonal tropics
potentially competing species have little or no opportunity to shift abundance
levels allochronically, so that an unstable equilibrium of coexistence is main-
tained,(612) such effects may occur in seasonal areas. Markedly allochronic
emergence patterns of cicadas in the tropiCS(71,138,139) mayor may not be indica-
tive of such interactions compensating for losses in fitness from competition. It is
necessary to determine what, if any, resources utilized jointly are in sufficiently
limited abundance to invoke competition among microsympatric species.(7J) Al-
though the adult popUlations of several species of cicadas with markedly seasonal
emergence cycles can be very dense,(39) there is little evidence to date of the
operation of density-responsive predators or pathogens, even though a path-
ogenic fungus may kill off large numbers of P. biolleyi,(J36) and some species
of birds in lowland and premontane wet forest successfully stalk singing males
of the large and loud cicada Z. smaragdind 7 1> (And A. M. Young, unpub-
lished observations). Such mortality appears to be patchy. Similarly, army ant
swarm raids may take heavy tolls on the immature stages of paper wasps,(193) but
such effects must be very patchy and thereby not regulating prey populations in a
density-dependent fashion. Other factors, such as the degree to which a habitat
has been modified by human disturbance, can influence the likelihood for out-
breaks of herbivorous insects associated with a particular plant species surviving
in the remnant habitat. (613)
Regions in the tropics that have repeated and localized catastrophes such as
floods or fires may provide sufficient density-independent mortality of individu-
als in populations of two or more competing species to result in another form of
unstable equilibrium of coexistence that fluctuates considerably each year. Such
a form of nonequilibrium coexistence was originally conceptualized by Hutchin-
son. (612) If it exists in the tropics, it is probably very patchy in effect and perhaps
not repeated over successive years. It seems probable that examples of coexis-
tence among several insect species on the same host plant or other resource in the
tropics are explained by microhabitat diversification on the resource and the
maintenance of population densities at low levels due to biotic control agents.
However, few data confirm such a prediction. Rolled-leaf hispine beetles exhibit
both monophagous and polyphagous feeding habits on various Zingiberales.(614)
Polyphagous species are those utilizing several different hosts in a region where
no one host is abundant. The number of rolled-leaf hispines correlates positively
with the number of common host plant species (Fig. 8.1). Those species of
344 CHAPTER 8

<Il HEUCONIACEAE
w 8

~
<Il
7
6 MARANTACEAE
w
z
a::
<Il
5
I 4
LL
«
w
3
-.J
2 COSTACEAE
0
W • ZINGIBERACEAE
FIGURE 8.1. The abundance of rolled-leaf
-.J
-.J
0 • CANNACEAE his pine beetle species as a function of the
~ frequency of common species in their host
o 234 5 678 plant families, at Finca la Selva, in north-
eastern Costa Rica. [From D. R. Strong, Jr.,
COMMON HOST SPECIES Biotropica 9:156-169 (1977).]

Zingiberales that are widely distributed grographically have a greater number of

hispine species feeding on them because of considerable geographic replacement
of species. (614) But the number of hispines coexisting in a habitat is a function of
the abundance of the host plant, sizes of host plants, and diversity of host plant
families.(614) When many common host plants are in an area, there will be many
species of these beetles, presumably since the rolled-leaf microhabitat is a limited
resource for these specialist insects. On the other hand, Slater(615) found no such
correlation between the commonness of a host plant and its utilization by various
species of blissine bugs in the tropics. The host plant associations of blissines
appear to reflect the evolutionary history of the host plant groups in that the
insects diversified for feeding on three major groups of monocotyledons early in
their evolution. (423.424,615)
A significant result of studies such as that of Strong(614) is that mechanisms
of species coexistence in tropical habitats are determined largely by within-
habitat ecological associations and are not related to the geographical distribution
of species. Host plant range may provide the evolutionary opportunity for the
geographical replacement of species, but these species will in many instances
have similar ecological roles as their counterparts in other portions of the range.
The localized coevolution between an insect and its host or resource will repre-
sent a biochemical and physiological commitment on the part of both partners in
the interaction, although some degree of genetic or phenotypic flexibility remains
in the gene pools. Janzen(363) and Wilson and JanzenC 6(6 ) found considerable
evidence that seed-eating bruchid beetles are host specialists in the dry forest
lowlands of Costa Rica, each species usually having only one host (range of one
to six host tree species). In some cases, two or three species of bruchids were on
DYNAMICS OF ORGANIZATION OF INSECT COMMUNITIES 345

a host seed. Because of this almost 1: 1 correspondence, it appears that the

number of host seed tree species determines the number of bruchid beetle species.
Gilbert and Smiley<329J summarize data on the host plant associations of
ithomiid butterflies to provide additional evidence on the frequency of coevolved
associations between insects and plants in the tropics. The results of Young<88,
270,290,298,301.302,310) provide additional evidence of such a pattern. Drummond,(617)
working in Ecuador, reported 53 ithomiines co-occurring with 44 potential host
plants. Of 27 species whose life cycles were determined, 81 % were limited to a
single host plant, while others used two to five larval host plants.(617) Drummond
did not find a correlation between the size of larval host plant populations and the
number of ithomiine species found on them. Gilbert and Smiley(329) believe that
such patterns of resource partitioning do not arise from competitive interactions
among the larvae of different ithomiine species on the same host plants but rather
as the result of the coevolved associations of each ithomiine species with a partic-
ular host plant species within the Solanaceae. The degree to which the larval host
plants studied were in fact different in terms of secondary defensive compounds
underlying host plant specialization was not determined. The Solanaceae as a
group are well known for high concentrations of alkaloids as defensive sub-
stances ,(27JJ a general trait underlying the coevolved association of the Ithomiinae
with the group.(6) Some primitive ithomiine genera are coevolved with Apocy-
anaceae(270,273) even though they co-occur with several solanaceous-feeding
ithomiines.
Gilbert and Smiley<329l discuss the generally low abundance of immature
stages of ithomiines on larval host plants and show that there is a positive
correlation between the abundances of individual ithomiine species and the
abundances of their host plant species. The butterflies do not appear to be
food-limited in the larval stage.(617) Gilbert and Smiley explain the pattern by
assessing the effective density of larvae per unit of leaf area as the primary
variable influencing the foraging behavior of predatory ants and wasps:
ithomiines using an abundant host plant species are expected to attain greater
levels of abundance before a functional response is elicited in the predators. (329)
Gilbert and Smiley argue that rare ithomiines may have similar abundance levels
of larvae when assessed in terms of leaf areas occupied, but because the host
plants are rare, these species are limited before host plants are defoliated or
inaccessible to ovipositing females. The result is that such ithomiines never build
up on such a patchy host plant and therefore populations remain at low densities
unless the abundance of the host plant increases in the habitat. Communities of
ithomiine butterflies should also be examined with regard to the kind of habitat
they occupy and the degree to which certain potential host plant species increase
or decrease in abundance as a result of perturbations in the habitat. In general,
however, the' 'indirect host limitation" idea of Gilbert and Smiley(329) shows that
346 CHAPTER 8

a highly specialized ithomiine exploiting a rare host plant species will have
population densities considerably lower than those of another specialized species
exploiting an abundant host plant in the same habitat. Gilbert and Smiley038l also
report that ithomiine species richness increases in mid-elevation forest habitats
since these areas have a greater abundance of naturally disturbed sites invaded by
various Solanaceae.
As pointed out by Gilbert and Smiley, such findings agree with those of
Janzen,(29l who found greater densities of beetles and moths at mid-elevations.
Janzen concluded that plants at mid-elevation sites in the tropics can increase net
productivity of harvestable biomass because cool nights reduce respiratory los-
ses. Each host plant species therefore presents a larger food base that makes
space for additional specialist species to coexist. This idea is an extension of
MacArthur's theory as to how new specialist species can be added to tropical
communities through increases in productivity.(J6l But for ithomiines, Gilbert
and Smiley<329l provide evidence supporting the view that increased diversity of
host plants, rather than increased biomass, is the primary factor in increasing
species richness of ithomiines. The natural history data on various ithomiines
from various elevational sites in Costa Rica collected by Young(88,270,290,298.301,
302,310) support the view that both geographically and locally ithomiines appear to
be larval host plant specialists. The observation by Young 88 ,89l of severe defolia-
tion of S. rugosum shrubs by the ithomiine H. euclea in premontane tropical rain
forest provides further support for the prediction that specialist ithomiines ex-
ploiting an abundant host plant will have larger adult popUlations. In this particu-
lar case no functional response was observed by predatory insects such as wasps
or ants, and larval densities reached outbreak conditions and many individual
host plants were heavily defoliated. (88) During periods of wet weather, clusters of
eggs were killed by an unidentified pathogen but no predation on eggs or larvae
was observed. Young suggested that recent changes in the forest edge habitat at
this site resulted in rapid expansion of the host plant population, providing a
greatly increased food base for Hypothyris.
The observed absence of other ithomiines on this abundant host plant was
attributed to the plant having unusually tough leaves, a trait overcome by the
communal feeding habits of the gregarious larvae. A similar system occurs in the
specialized association of the ithomiine M. isthmia with certain solanums found
in pastures and along roadsides: these plants have thick, spine-covered stems and
leaves that form a formidable physical barrier to many insects with the exception
of a few aposematically colored Orthoptera and some chrysomelid beetles. Cluster
oviposition and gregariousness in the larvae are two conspicuous behavioral traits
of Mechanitis and Hypothyris generally not found in other ithomiines, and it was
suggested 88 ,221,222l that they are adaptations to breaking down tough leaf tissues by
group feeding. Larvae may grow faster in large groups (Fig. 8.2). These host
plants establish ecological barriers(6) to colonization by other ithomiines with single
DYNAMICS OF ORGANIZATION OF INSECT COMMUNITIES 347

17
16
- - - EGG MASS no.14
15 - - - - - EGG MASS no.18
14

~ 13
~ 12
Z
;;; 11
!7l10
..J
9
~
0::
<I:
..J 8 ,/'/
FIGURE 8.2. Rate of development for the 7
Neotropical butterfly M. isthmia (Ithomiidae) ~
6 3rd INSTAR
as seen for two egg masses (clusters) in the
wild. Although this sample is very small, of 5
interest is the faster rate of larval development 4
for the clustered group oflarvae from No. 14,
the larger egg mass. [From A. M. Young and 3~~~~--~~~~~--~~ __
1 2 3 4 5 6 7 8 9 10 11 12
M. W. Moffett, Dtsch. Enlomol. Z. 26:21-38 t SUCCESSIVE DAYS
(1979).] (Day of hatching)

oviposition and solitary larvae as well as other insects. In the Mechanitis system,
however, considerable mortality of eggs and young larvae results from predation
by orthopterans and wasps.(220,221l The complete absence of such predation in
Hypothyris suggests that this plant species possesses some ecological traits that
deter searching behavior by such predators. Aggregative behavior in the early
instars of Lepidoptera larvae is believed to be related to the physiological vul-
nerability of young larvae to food supplies and weather conditions, while such
behavior in older instars is related primarily to evasion of biotic mortality agents
such as predators.(61S) In some aggregating insects, groups break up in response
to the presence of a predator. (619) Not all tropical plants and their associated
herbivores are subject to the same kinds of effects as described by Gilbert and
Smiley 029J in terms of functional responses by predators as a means of effective
biotic control of herbivore populations. Nothing is known about the chemical
defense system of S. rugosum and whether leaf toughness is the only defensive
trait of this plant. The plant may produce other signals that deter predators and
parasites, so that wind-blown pathogenic microorganisms are the only biotic
mortality agents operative on Hypothyris populations at this locality. Further
studies are needed to compare the intensity of predation and parasitism in
selected herbivorous insects associated with a variety of plant species in the
tropics. If predators and parasites are generalists, the availability of more acces-
sible food supplies at anyone time may shift biotic control from other herbivore
populations, possibly resulting in outbreaks. The effect of pathogenic microor-
348 CHAPTER 8

ganisms is probably largely density-independent and determined largely by day-

to-day patterns of rainfall.
Another example summarized by Gilbert and Smiley(329) illustrates the diffi-
culty in making broad generalizations about the strategies and patterns of re-
source utilization in tropical insects. Gilbert's analysis of local patterns of larval
host plant exploitation by 14 species of the satyrine genus Euptychia in Trinidad
revealed a high percentage of polyphagous species present in an area does not
reflect the number of Euptychia species. (329) In fact one site studied had 10
species of this genus and only one larval host plant species, and another four sites
had twice as many butterfly species as host plant species. (329) But the distribution
of specialist species of the genus (monophagous species) tracks that of the
selected host plant species. The pattern breaks down for polyphagous species as
explained above. Gilbert and Smiley, citing some unpublished studies of their
colleague Michael Singer, suggest that unusual features of host plant clumps
(grasses and sedges), coupled with unusual oviposition behavior of Euptychia
species, determine the patterns of several butterfly species on relatively few host
plant species. Gilbert and Smiley also hypothesize that a lack of "chemical
distinctiveness" among various grass species determined the patterns of host
plant utilization by the butterflies. They analogize the Euptychia patterns to the
organization of stem-boring guilds studied in the Temperate Zone by
Rathcke.(620) At several sites in lowland and premontane rain forest in Costa Rica
the butterflies M. peleides and M. granadensis exploit at least one larval host
plant together, namely, the forest edge woody vine M. seemanii, even though
peleides exploits many other leguminous species in the same region and habitats
(A. M. Young, personal observation). In Costa Rica, several species of Morpho
butterflies partition the tropical rain forest environment in terms of habitat type,
larval host plants, and, to some degree, the height adults t1y above the ground
(Table 8.1). All of these species have closely related allies in South America, and
in most instances one might expect similar patterns of environmental partition-
ing. Yet the reasons for such patterns remain obscure, especially since adult
popUlation densities of these butterflies are generally quite IOW. 075 •176 ,190)
At least three species of troidine swallowtail butterflies overlap considerably
in using the larval host plant Aristolochia constricta in tropical rain forests in
Costa Rica (A. M. Young, personal observation). Given the interesting findings
of Gilbert with other buttert1y-plant systems in the tropics, it would be interest-
ing to explore the organization of Morpho and troidine swallowtail butterfly
communities, with an emphasis on the relationship between the number of poten-
tial host plant species in an area and the number of allied buttert1y species. Many
species of Morpho are in Central and South America, and host plant records
suggest considerable ecological diversification within the genus.(176) Critical var-
iables, as emerging from the studies of Gilbert,(38) include patch structure of host
plants, habitat species densities of the butterflies, habitat species densities of the
DYNAMICS OF ORGANIZATION OF INSECT COMMUNITIES 349

TABLE 8.1
The Species and General Habits of Morpho Butterflies Found in
Premontane-to-Lowland Tropical Wet Forest Region a of Northeastern
Costa Rica b

Major adult Flight Relative Larval food-

Species habitat(s) height range abundance plant taxa

M. peleides Various secondary; 1-3 m High Several

large forest light gaps Leguminosae
M. granadensis Advanced 1-3 m Low Probably few
secondary; large forest Leguminosae
light gaps
M. amathonte Primary forest 1-6 m Intermediate Probably few
Leguminosae
M. theseus Primary forest 6-30 m Loa Probably few
Menispermaceae
M. cypris Primary forest 10-30 m Low Probably few
Leguminosae

"Defined here as 0-200 m above sea level.

bFrom Young.(331a).

host plants, the degree of distinctness among host plant patches of a particular
plant species in terms of features attractive or repulsive for oviposition in one or
more butterfly species, and the degree of chemical and physical distinctiveness
among patches of different host plant species co-occurring in the same habitats.
Benson(282) has shown that there is considerable resource partitioning in
heliconian butterflies for patterns of exploiting their passifloraceous larval host
plants. In these butterflies, populations are generally limited by the abundance
and distribution of the larval host plants, and there is considerable competition
among heliconian species for oviposition sites.(282) Because of such effects,
ecological shifts have resulted from selection pressures of interspecific competi-
tion as a means of reducing losses in fitness where species overlap in distribution.
Benson shows that such resource partitioning among heliconians at several
localities has taken place with respect to (l) plant species utilized, (2) habitat of
preferred host plant, and (3) specific part of plant used for oviposition and
development of early larval instars. The organization of heliconian communities
or guilds is similar in regions where faunistic complexity is high or low. (282) In
tropical rain forest regions considerable larval host plant specialization is evident
among sympatric heliconians, although at one subtropical site in Brazil where the
number of heliconian species is low, the species tend to be generalists exhibiting
considerable overlap in larval host plants. Yet each species exhibits different
habitat preferences, making it a patch generalist, thus minimizing competition
for individual host plantsY83) This subtropical site is strongly seasonal, which
350 CHAPTER 8

results in periodic reductions in the harvestable productivity of each larval host

plant, and each heliconian species tracks the fluctuations in host plant abun-
dances through life history adjustments. The seasonal fluctuations in host plant
cycles impose the generalist strategy upon these heliconian species as a means of
maintaining breeding populations through periods when one or more host plant
species may be scarce.(283) But competition is prevented under such conditions by
distinct patterns of habitat separation among the species, reducing resource parti-
tioning in terms of host plant species.
A fine equilibrium appears to be established at each locality between the
number of host plant species, host plant population sizes, habitat separation
among host plant species, and the number of heliconian species found there, to
the point that Benson et al. (283) conclude that heliconian communities are at
saturation levels. Competition is believed to be the mechanism excluding other
species from being added to the communities.(283) Gilbed 338 ) suggests that the
monophagous habit of many specialized herbivores as the result of coevolution,
implies that popUlations of the specialist butterfly will increase as the host plant
popUlation increases. Another case cited by Gilbert, an alternative pattern, is the
herbivore species generally adapted to feed on many members of a single plant
family: when a host plant becomes abundant in a habitat, it is likely to be
colonized by several herbivore species. Gilbert points out that most heliconiine
butterflies and Passiflora-feeding flea beetles fall between these two extreme
conditions in terms of the relationship between host plant abundance and insect
abundance. Thus an insect may be a coevolved specialist for feeding on several
different host plant species within a certain family yet feed on only one or two as
the result of ecological factors.029.338) To some extent it is true that increases in
the abundance of a larval host plant increase the number of insect species feeding
on it if these species are generally pre adapted to exploit several members of the
host plant family. Gilbert and Smiley(29) show a significant correlation between
numbers of herbivore species on individual species of Passiflora as a function of
the local abundance of each plant species. Higher harvestable productivity of
Passiflora results in a greater number of herbivorous insect species on it, even if
many members of such a guild are monophagous.(3291 Gilbert and Smiley0291
have examined other ecological factors affecting the numerical relationship be-
tween host plant diversity, host plant abundance, and herbivore species numbers,
and the reader is referred to this work for a detailed analysis.
In developing his arguments with regard to heliconiine butterflies exploiting
the Passifloraceae, Gilbert stresses the role of biogeographic events in determin-
ing the broad and restricted geographical distributions of various heliconiine
species. Turner 6 2J) reported that heliconiines experienced considerable adaptive
radiation while isolated in the glacial forest refugia areas of South America.
During such times the group underwent considerable habitat separation as evi-
denced in their present-day distributions in some cases. Cyclic refuge isolation
DYNAMICS OF ORGANIZATION OF INSECT COMMUNITIES 351

resulted in some species spreading out over much of the range of potential
habitats, while others were more restricted. (621) Brown et ai. (622) suggest that
localities in South America adjacent or close to Quaternary refugia have a high
number of species to colonize new habitats. Some sites extremely rich in terms of
numbers of heliconiine species are within the overlap zones for the expanded
faunas of at least four surrounding forest refugia. (468,478.62J-626) Forest plants
experienced considerable diversification within Quaternary refugia, and many of
these host plant species (e.g., in the family Passifloraceae) could have resulted in
the rich diversity of heliconiiine larval host plants found in the overlap zones
today. (622) These biogeographic heterogeneous effects make it necessary to select
sites where refugia influences are similar in order to examine the effects of local
ecological factors regulating herbivore communities associated with particular
groups of plants. Gilbert(338) examines the role of such ecological factors, includ-
ing resource variety, habitat structure, climatic predictability, and local topog-
raphy in determining the local species abundances of heliconiine butterflies in
relation to how such factors have influenced the local Passiflora communities
and adult resources of the butterflies. Thus although strong coevolution between
groups of insects and their resources is a common phenomenon in the tropics, it
is also necessary to examine the role of a variety of ecological factors in shaping
the organization of local herbivore communities or guilds and, whenever possi-
ble, to recognize the influence of biogeographical events.
In Chapter 5 the concept of habitat selection in tropical insects was dis-
cussed, yet clearly some of the considerations here on the maintenance of insect
communities with a set number of species also relate to strategies of habitat
selection pointed out by Benson's studies of heliconian butterflies and their larval
host plant associations.(282) The degree to which a resource is in limited supply
will clearly mold the strategy of habitat selection for an individual herbivorous
insect species that is a specialist on that plant. The intensity of habitat selection
will also be a function of the quality of individual resource patches, and anyone
habitat is expected to exhibit a spectrum of quality determined by the condition of
the resource on a patch to patch basis. Anyone herbivore species exploiting the
resource should maximize the choice of high quality resource patches while
minimizing the colonization of poor-quality resource patches. If several ecologi-
cally similar species are converging upon the same resource, there will be either
(1) intense competition within a habitat between these species for the high-
quality patches, perhaps resulting through evolutionary time in resource parti-
tioning along different resource parts or by divergence onto different host plant
species, or (2) segregation of the herbivore species into different habitats where
the same host plant species is found, thereby reducing effects of competition.
Furthermore, Janzen(93J has suggested that even herbivores exploiting different
parts of a single plant, or different individuals of the same host plant species in a
particular population of that plant, are, in effect, in competition.
352 CHAPTER 8

In an interesting study of the food-limited and specialist Temperate Zone

aphid Pemphigus, Whitham(627) found considerable evidence of variation in re-
source patch (size range of Populus leaves) resulting in exceedingly high levels
of colonization for optimal patches and very low levels of colonization for
poor-quality patches. Whitham(627,628) found that fertilizer-treated individuals of
one host plant species increased the breeding success of the aphids, indicating
that nutrients were a limiting factor affecting these herbivores. Overall resource
quality can be improved by fertilizer treatment for this host plant species. Yet for
a second species of Populus, the breeding success of this aphid species could not
be significantly increased by adding fertilizer. (627) This second species of
Populus was the preferred host of the aphid, and there was evidence of consider-
able intraspecific competition for optimal resource patches. In another paper,
Whitham(628) conceptualizes the Populus host plant population as a heterogene-
ous environment or habitat in terms of the great variation in quality of individual
leaves for colonization by Pemphigus. As habitat quality is increased, so does
the density of competitors, and there is a clear pattern of adjustment in densities
in different habitats of varying quality so that average fitness is about the same in
all habitats.(628) In tropical species in similar systems of habitat quality variation,
it would be interesting to determine how popUlation densities of a single her-
bivorous insect species are regulated in terms of the effects of interspecific
competition on optimal resource patches. The greater density of closely related
and ecologically similar species in many tropical habitats, and the higher inci-
dence of ecological specialization, suggests that the maintenance of an adequate
level of average fitness within a population is a function of competition between
those herbivore species that have populations at or approaching the carrying
capacity of the habitat. Such effects may be less significant in species where
biotic control agents such as predators and parasitoids keep populations well
below such levels.
A growing body of evidence shows that mate recognition in some butterflies
is the result of ultraviolet reflection cues rather than color cues in the visual
spectrum. (629) If generally true, visual coloration in butterflies, especially in
tropical forms, may be more of an adaptation to thermoregulation and sheltering
behavior and for active or passive escape from visually hunting vertebrate pred-
ators. Neotropical Morpho may exhibit pursuit invitation behavior toward such
predators, depending upon the degree of blue reflection from the wings and flight
behavior.(207) In EI Salvador and presumably elsewhere, anoles successfully
knock down ithomiid butterflies with their tails and devour them (A. Muyshondt,
personal communication) and wing color patterns in these butterflies may be
more effecti ve against other kinds of predators. Some butterflies undoubtedly use
visual colors in courtship as well. Turner<630) has given a detailed discussion of
the evolution of butterfly mimicry and other forms of defense against predators,
with an emphasis on tropical faunas. Many rain forest understory butterflies in
DYNAMICS OF ORGANIZATION OF INSECT COMMUNITIES 353

the great families Brassolidae, Satyridae, and Morphidae, have conspicuous

eyespot markings on the ventral sides of the wings. These butterflies typically
fold their wings when feeding or resting, and the colors are generally cryptic to
the human observer, making the insects blend into the forest floor or shadows of
understory vegetation. Many of these butterflies exhibit a peak of feeding near
dusk in lowland and premontane tropical rain forest,(88) and several individuals
of a species may accumulate daily at rotting, fermenting fruits on the forest floor.
The early stages are cryptic and those of some species are associated with mono-
cotyledons (Fig. 8.3). Home range movements may be involved in such feeding
behavior, with individuals returning over successive days to feed at the same re-
source.l!7S,186,190) Young(J88) suggested that eyespot markings are enhanced in
their appearance to humans, and perhaps to predators other than butterfly col-
lectors, when streaks of sunlight near dusk hit the wings of feeding butterflies.
Stradling 631l suggested that such eyespot markings in some butterflies together
with other wing markings resemble the heads of amphibians and reptiles, two
possible groups of predators on these insects. If interpreted as sucy by predators,
these animals may avoid feeding on resting butterflies. Young 188) noted a decrease
in the size and incidence in eyespot marking for butterflies active in the higher
strata of tropical rain forests and suggested that selection would not maintain such
eyes pot phenotypes if "eyespot-sensitive" predators were less abundant in such
microhabitats.
In the context of Pimentel,<632) the degree to which an herbivorous insect
species affects the fitness of a plant species, as a result of specialized predation,
depends upon the degree to which both plant and insect enter into cyclic patterns
of frequency-dependent selection for traits underlying the coevolved relation-
ship. Tropical forest ecosystems may consist of many such coevolved relation-
ships, each one within the many communities and guilds present constantly
changing in response to reciprocal selection from each partner. Thus a given
plant species may possess a certain spectrum of monoterpene compounds that
function as chemical defense compounds, and there is a range in variation for
effectiveness against specific herbivores. The habitat consists of many plant
species, each endowed with a spectrum of defensive compounds genotypes and
phenotypes that may qualitatively differ among species. (256) As selection pres-
sures from herbivorous insects shift with respect to a particular tree species (and
such shifts may be frequent given the high species densities of herbivorous
insects in most tropical terrestrial habitats below 2000 m), directional selection
may segregate a relatively rare toxic genotype of the plant in response to a change
in the coevolved relationship. Subsequently, strong stabilizing selection (Chapter
2) maintains and perfects the new toxic phenotype, which then increases in
abundance in the plant population. But the more abundant the formerly rare toxic
phenotype becomes in the population, the rate of increase being a function of the
life history pattern of the plant species in that habitat and the intensity of preda-
354 CHAPTER 8

FIGURE 8.3. Life stages of the brassolid butterfly Opsiphanes tamarindi associated with Heliconia
and Musa in Costa Rica. (A) Egg, (B) third- and (C) fifth- instar caterpillars, (D) pupa, (E) newly-
eclosed adult on pupal shell. [From A. M. Young and A. Muyshondt, Stud. Neotrop. Fauna 10: \9-
56 (1975).]
DYNAMICS OF ORGANIZATION OF INSECT COMMUNITIES 355

tion on it by one or more herbivore species, the greater the contact with herbivor-
ous species and the more rapid the counter (evolutionary) response by the insects.
In this manner, there is a frequency-dependent overcoming of the new defensive
barrier by the insects, and if biotic control agents are inoperative in this system
(i.e., predators and parasitoids of the herbivores) or if interspecific competition is
nil, the insects will become pest species and outbreaks may occur. The new toxic
phenotype is now ineffective in deterring attack by the herbivores. Thus an
optimal strategy for a tree or other growth form of plants in tropical forests in
particular, where plant predation rates can be at least potentially very intense, is
to segregate a range of toxic phenotypes at each generation, with some propor-
tion of these toxic phenotypes remaining rare within the population.
The strong coevolution in Central America between some Cucurbitaceae
and the luperine chrysomelid beetles that feed on them is mediated by the use of
bitter substances, cucurbitacins (oxygenated tetracyclic triterpenes), in the plant
tissues as feeding stimulants by these herbivores, even though these substances
probably originally evolved to deter such predation or plant parasitism.(633) Such
studies also point out that although some coevolved feeding associations between
plants and insects have evolved in the tropics, the chemical or physiological
mechanisms underlying the interaction have been retained through the evolution-
ary radiation of both the host plant group and the herbivore group into other kinds
of environments, including those promoting the domestication of the plants as
food sources.
The observation in nonseasonal lowland tropical regions that outbreaks of
herbivorous insects on plants in natural ecosystems are relatively rare suggests
that most plant species in such environments maintain breeding populations with
a high degree of genetic variability with regard to defensive compound chemistry
and other traits relating to defense against herbivores. Such properties, in con-
junction with disjunct spatial distributions, fluctuations in physiological cycles
(season or nonseasonal) regulating vegetative and reproductive growth, and the
probable limitation of herbivore populations by predators and parasitoids, lower
the likelihood of outbreaks in natural ecosystems. Communities and guilds of
plant-feeding insects, as emphasized in the studies of Gilbert,(338) Strong,(634) and
Janzen,(47) are well organized in space and time, and such properties reflect both
the highly coevolved interactions between plants and herbivores and the interac-
tions of both plants and their herbivorous insects with other components of
tropical forest ecosystems (i.e., interactions between plant species, interactions
between insect species including competition and predation, other features of
food chain associations).
Some surveys of avian species in lowland tropical forests indicate a high
proportion of insectivorous species relative to that found in Temperate Zone
forests.(609) Some studies indicate that predation on herbivorous insects by birds
in forest ecosystems reduces the population densities of these insects mark-
356 CHAPTER 8

edly. (635) In tropical forests such predation by birds(6!O) may take on even greater
significance in terms of regulating insect populations: many insects have rela-
tively low population densities to begin with, because plant species tend to have
patchy distributions and many insects are specialists on a few plant species,(47)
and under such conditions of low population densities, avian predation may act
as an even greater selective agent and source of biotic regulation of popula-
tions. (635) Thus it is necessary to examine, whenever possible, the food chain
associations for herbivorous insects in determining the degree to which indi-
vidual species are resource limited or regulated by systems of predation.
Chapter 9 considers how systems of agriCUlture in tropical regions have
altered such coevolved associations and the resulting high structuring of insect
and plant communities in the tropics. Detailed studies on the ecological traits
associated with coevolved interactions between plants and their specialized
seed-eating insects in both natural and disturbed tropical habitats are given in
Janzen.(29,362,363) The reader is encouraged to examine these papers for a com-
prehensive understanding of the dynamics of such coevolved systems in the tropics
and their implications for the organization of tropical forest ecosystems.
Some portion of community structure in tropical forests is determined by
selection pressures promoting the association of certain kinds of insects with
plants. Chapter 4 included a brief discussion of the highly coevolved mutualisms
between plants and ants(33,373,376) that emphasized that ants replace chemical
substances as the major line of defense against herbivores in plants. The studies
of Rehr et al. (262) showed that chemical defense in acacias turns up when the
mutualist ants, Pseudomyrmex species, are absent. The analysis of the functional
role of extrafloral nectaries of Bixa for attracting ants indicated that herbivory is
less on plants with actively patrolling ants. The greater incidence of such
mutualisms in the tropic'> indicates that (1) the benign physical environment has
been conducive to the evolution of such coevolved relationships, and (2) in plant
lineages where chemical defenses have not evolved, possibly as the result of
evolutionary history or other forms of contemporary selection tending to erode
genetic variation that would permit such a system to evolve, intense herbivory
has resulted in the evolution of mutualisms with ants. In some cases, the relation-
ship has evolved to the point where the activity of the resident ant species results
in an increased nutritional supply for the plant. <3(6)
In an interesting study of ant-plant mutualisms in the Temperate Zone,
Inouye and Taylot 636 ) found that the effectiveness of ants as herbivore deterrants
may not always be as effective as chemical defense. They studied the relation-
ship, along an elevational gradient in Colorado, of various ant species with the
aspen sunflower, Helianthella quinquenervis, a plant with conspicuous extra-
floral nectaries that produce large amounts of sugar and amino-acid-rich nectar.
At high elevations the ants are effective in lowering the amount of herbivory by
various insects on seeds prior to dispersal and are effective in preventing oviposi-
DYNAMICS OF ORGANIZATION OF INSECT COMMUNITIES 357

tion by various flies. But at lower elevations, where such seed predators and ants
are more dense, ant protection was less effective in reducing predispersal seed
predation. Other plant species in the same habitat appeared to be better protected
by chemical defenses than by ants.(636). Compared with some tropical systems
such as Acacia x Pseudomyrmex or Cecropia X Azteca, (33,374,375) in which the
plants provide a nesting site for the mutualist ants, Inouye and Taylor's Colorado
system lacked ants actually living in the plants. (636) Perhaps under such condi-
tions, the relationship is not as well evolved to the point that selection has
resulted in ants actually living in the plants, But such a relationship may not
evolve if reductions in fitness are not appreciable (unlikely, given Inouye and
Taylor's data) and if other genetic and phenotypic constraints within the plant
popUlations preclude the expression of the necessary physiological and
morphological changes in phenotype needed to accommodate the ants. If the
average plant species within certain habitats in the lowland and mid-elevations of
the New and Old World tropics experiences greater selection pressures from
herbivory (insect and vertebrate) than plant species in comparable Temperate
Zone environments, either there will be a greater incidence of highly coevolved
associations of insects with plants with many other insects screened out by the
chemical defense systems of each plant species,(6) or ant mutualisms will evolve
in lineages where chemical and physical defenses are not effective against herbi-
vores, reducing fitness, and there will also be a greater incidence of the evolution
of extrafloral nectaries.
The evolution of extrafloral nectaries, as seen in many tropical plants, may
represent intermediate stages in the evolution of plant-ant mutualisms if such
plant species are genetically flexible enough to evolve the traits necessary for
such mutualisms and if the selection pressures favoring such an interaction are
sustained through enough generations. If, on the other hand, the plant species is
incapable of evolving such a mutualism (Janzen(33) has stressed the extreme
phenotypic commitment by both partners that are involved in the evolution of
such mutualisms.) due to genetic and phenotypic constraints affecting the overall
average phenotype within the popUlation, the expression of extrafloral nectaries
may be the end point in herbivore defense strategy for that species (although the
system is not static and subtle changes in the phenotype related to effectiveness
of extrafloral nectaries will continue). The likelihood of an ant-plant mutualism
evolving also depends upon the available pool of ant species, the degree to which
each species is already ecologically specialized, and energy requirements related
to maintenance of colonies. The incidence of plant -ant mutualisms in the tropics
may be higher than what is known today.
The integration of the tropical forest ecosystem is highly dependent upon the
breeding systems of the various plant species found there. Primarily through the
efforts in recent years of workers such as D. H. Janzen, H. G. Baker, K. S.
Bawa, and G. W. Frankie, there is a growing body of information on the
358 CHAPTER 8

breeding systems of tropical trees and shrubs in different kinds of tropical forest
environments.(118,218) The subject, although fascinating and extremely relevant to
matters of community structure in the tropics, is too vast to summarize here, as
the breeding systems and associated pollinators are varied and diverse, so that
only a minimal treatment is given.
Van Der Pijl(334) has emphasized the overall diversity of pollination systems
seen in the tropics and has attempted to examine general evolutionary patterns
within the angiosperms in the context of pollinators and dispersal agents. The
striking diversity in floral morphology and physiology seen in tropical floras
indicates that some sets of pollinating insects may be specialized ecologically to
exploit unusual substances secreted by flowers. Some groups of tropical bees, for
example, may exploit noncarbohydrate or nonprotein substances such as fats or
lipids secreted by certain plant taxa. Although such insects may visit a variety of
other plant species for pollen and nectar (e.g., Neotropical solitary bees), the
feeding niche or nest provision niche component may include these kinds of
flowers. The degree to which the exploitation of such unusual resources is carried
out by a particular bee species will be a function of many factors, including the
ecological role of the substance in the biology of the insect, accessibility of the
resource, and competition among species for the resource. Different groups of
tropical pollinating insects may have spatial popUlation structures permitting the
exploitation of such resources opportunistically or as a regular component of the
diet. In the tropics, there is a greater degree of specificity by pollinators and
animals acting as dispersal agents of seeds for certain plant groups, although in
other cases the relationships are more generalized. Floral structures are generally
designed to attract certain kinds of pollinators while excluding others in many
specialized plants, and exhibit considerable nonspecificity in other groups. Floral
structure, seed dispersal, life table, and resulting patchiness in distribution of the
adult population of a plant species together constitute the primary determinants of
the optimal breeding system for that species. A large percentage of plant species
in the tropics are pollinated by a variety of insects. Insects and other animals are
attracted to flowers by nutritional rewards associated with retrieval of nectar and
pollen, and such rewards vary considerably in quality and quantity among dif-
ferent plants and within the same plant species at different times of the day,
etc.(216.217) Some insects, e.g., certain tropical Lepidoptera, also obtain sub-
stances from vegetative parts and seeds of various plants077-179) and do not enter
into the pollination association. The insects may use these substances as precur-
sors in the synthesis of courtship-related substances in one sex or the other,(l77)
and such systems represent a different form of coevolution between insects and
plants.
The relationship of pollinators to a particular plant species must in part be
viewed in terms of the role of flowers as attractants, sources of rewards (energy),
and ecological screens for permitting only certain kinds of insects and other
DYNAMICS OF ORGANIZATION OF INSECT COMMUNITIES 359

animals to visit them. The role of a particular kind of flower as a food source for
some insects will be a function of its phenotypic traits, including the
morphological and physiological. In turn pollinating insects evolving into
specialists for a particular flower will possess specific morphological, be-
havioral, and ecological traits that optimize retrieval of required nutrients from
the flowers. Such specializations are diverse in the tropics, where both insects
and floras are diverse, locally and over geographical regions.
A specialist insect species on a particular flower species or cluster of flower
species will exhibit regular foraging patterns for exploiting that resource(s), since
individuals will receive virtually a full complement of a nectar or pollen diet from
that resource( s). Yet there will also be a spectrum in the amount of specialization
closely related insect species will possess for exploiting a particular set of flow-
ers, and some species will be generalists in some geographical zones, drop out
completely in others, or become specialists in still others. Que'stions of pollinator
interaction with tropical plants must be viewed at different levels of taxonomic
organization: one can ask about the regional proportions of plant species polli-
nated by bees, butterflies, sphingid moths, or dipterans (and others), perhaps
over an altitudinal gradient in the tropics and expect to find distinct patterns for
different zones. Within a given zone, one will expect to find a wide range of
specialization by insects to the flora of that zone: a set of bee species in the humid
lowlands may be associated with a different set of plants from the same or similar
set of bees in the mid- or high-elevation zones, and the proportions of plants
pollinated by these bees may also change over such a gradient. The number of
sphingid-moth-pollinated plant species may be greater at mid-elevation sites than
in adjacent lowland sites, with constraints being set by the composition of the
floras at such sites, both in terms of floral structures represented and in terms of
the availability of larval host plants. Within a given zone, each pollinating insect
species may have primary and secondary food sources on daily foraging routes,
and the composition and ratio of such resources may also change among different
zones. In the lowland tropical rain forest zone, the average Starchypheta bush
may have a greater number of butterfly species associated with its blooms(637,638l
than a similar bush of the same species at a mid-elevation site. Such differences
will have great impact on the popUlation structure and foraging behavior of the
individual insect species in both zones.
Studies in tropical pollination biology therefore stress the community of
pollinating organisms associated with a particular plant species, the population
structure of that species and flowering phenology, and the degree to which that
particular plant species fulfills the food requirements of each pollinating insect
species. We also need to examine the impact of joint vertebrate and insect
pollinating systems associated with individual plant species. Such considerations
illustrate that one explanation for the great diversity of insects associated with
flowers in the tropics is the regional diversification of floras and associated
360 CHAPTER 8

pollinating insects. The other explanation is more ecological than regional: diver-
sity of such insects is also determined by the composition of communities within
a particular region and the degree to which component species are specialists or
generalists in terms of exploiting food sources supplied by flowers.
The majority of pollination systems in the tropics involve birds, bats, but-
terflies, moths, flies, beetles, bees, and wasps.(216.217) Various aspects of pollina-
tion systems were already discussed briefly on page 340, and also in terms of
effects of tropical seasonality (Chapter 7). Baker et al. (218) believe that plant
species in the early stages of tropical succession are self-compatible while those
of primary forest are predominantly self-incompatible. Mass flowering and self-
incompatibility seemed to be linked as indicated Frankie et al. in their study of
bee pollination of the papilionaceous legume tree Andira inel'mis in lowland dry
forest of Costa Rica.(499) The authors found that two daily peak periods of nectar
flow occur in this tree and that pollen is released during the first period. About 70
species of bees were collected from the flowers, and most of the solitary bee
species exhibit limited movement among con specific trees. Low intertree move-
ment accounted for the observed patterns of fruit set. (499) Mass flowering pro-
vides a large resource base in some tree species of tropical forests for many bee
species, and competition for these resources (nectar and pollen) is minimized due
to satiation effects relative to patch size of the resource. Apparently, however,
there is sufficient movement by foraging bees among conspecific individual trees
to maintain a certain level of fruit set. Self-incompatibility appears to require
mass flowering in order to attract a sufficient number of bees to ensure some
low frequency of exchange with con specifics in the same habitat. Self-incom-
patibility functions in tropical forest trees as a means of outcrossing 638 a,639)
presumed necessary for providing adequate genetic variation to allow the tree to
respond to shifting selection pressures associated with herbivory and other
ecological factors. (40) Mass flowering may also function to minimize potential
competition for pollinators by providing a sufficiently large resource base for
many species. The high local diversity of various Leguminosae and other groups
of forest trees in tropical forests, and the similarities in floral structure and peak
flowering periods among these trees, (I 17) can result in considerable interspecies
competition among trees for a set of pollinators. Pollinator constancy is one
means of reducing competition and allowing each tree species to have a comple-
ment of pollinators(640,641) when the potentially competing tree species co-occur
in the same habitat. Mass flowering in several closely related tree species at the
same time in the same habitat presents many large resource patches of regular
size to pollinators, a condition that reduces competition for pollinators. Both
conditions may be within colony segregation of resource patches by a bee species
or different colonies of the same species may forage at different resource patches,
resulting in increased resource availability for other bee species foraging in the
same habitat. Although it is necessary to examine nest locations of individual bee
DYNAMICS OF ORGANIZATION OF INSECT COMMUNITIES 361

species, foraging distances, and how workers within each nest divide up resource
patches, ifthey do at all, the high diversity of bee species observed by Frankie et
al(499) suggests such mechanisms to be operative.

Closely related plant species in tropical forests may exhibit considerable

divergence in floral characters as a means of partitioning the local pool of
potential pollinating insects. Schemske(642) found that two sympatric species of
the shrub Lantana (Verbenaceae) in tropical rain forest possess markedly dif-
ferent floral characters, including inflorescence size and corolla length. The
result is that different sets of lepidopteran pollinators visit the flowers of the two
species and competition for pollinators is minimized between these species. (642)
A recent survey of the amino acid composition of nectars from floral and extra-
floral nectaries of many temperate and tropical plants revealed that (1) amino acid
complements vary considerably between species for both kinds of nectars, (2)
extrafloral nectars were always different from floral nectars of the same plant
species, and (3) certain kinds of amino acids are invariable and consistently
found in extrafloral nectaries, a condition suggested to be related to nutritional
requirements of protective guard insects such as ants and wasps.(218) Such pat-
terns of specificity at the biochemical level function to partition pollinator com-
munities in environments where species densities of plants are high.
Other forms of behavioral and physiological changes in flowers, such as the
color change syndrome studied by Barrows(643) in Pachyptera hymenaea (Bigon-
oniaceae) in Mexico, in which flowers changed color even after nectar and pollen
ceased being produced, build in further specificity in pollination systems in such
habitats. Barrows suggested that patches with mixed flowers attract greater num-
bers of pollinating insects than patches of dark flowers (ones with nectar and
pollen) alone. The level of synchronous blooming in patches results in a more
effective signal for pollinators, thereby increasing the likelihood of pollination in
young inflorescences with pollen and nectar still present.
In some tropical forest ecosystems, the primary pollinators of many trees,
shrubs, and vines are the stingless bees (Trigona spp.). A. Wille (personal
communication) of the University of Costa Rica has been studying the abundance
and specificities of these bees for various plant species in tropical rain forests and
suggests that these insects are the major pollinators of many plants in tropical
forests. Stingless bees exhibit flexible foraging habits and preferences for in-
florescences, and make seasonal shifts, habitat shifts, in foraging easily. Such
responses enable these insects to exploit a wide variety of plant species both in
primary and in secondary successional stages in tropical forests.
In terms of the overall organization of pollinator communities in the lowland
and mid-elevation regions of the tropics, the ranked order of abundance and
diversity is believed to be as follows: stingless bees (principal pollinators), vari-
ous large-bodied bees, lepidopterans, other insects such as flies and beetles, and
various vertebrate groups such as bats, mice, and hummingbirds. This ranked
362 CHAPTER 8

order probably follows a reverse gradient in specificity for certain groups of

plants, although far more data are needed on the specificities of pollinating
insects such as small flies, beetles, and wasps, groups often assumed but not
confirmed in most cases to be generalist pollinators. The composition of local
pollinator communities, especially in the case of insects with their complex life
cycles, will be limited by the availability of other resources such as nesting sites
in the case of bees, or larval host plants in the case of the Lepidoptera. An
analysis of pollination systems within a particular region or forest ecosystem
must consider these other resources as affecting pollinator abundances and popu-
lation cycles. Abiotic factors such as winds, air moisture content, and tempera-
tures will also affect the composition of pollinator communities, especially along
altitudinal gradients or within localized rain shadows on mountain slopes. Such
effects will influence the overall organization of pollinator communities.
H. G. Baker and I. Bakd217 ) have discussed the role of nectar and pollen
composition in molding the specificities of pollinating animals to flowering
plants. Hummingbird-pollinated flowers tend to have nectars rich in sugars,
whereas wasp-pollinated flowers are generally low in sugars. Large bees tend to
be associated with sucrose-rich nectars and small bees with hexose-rich nectars;
hummingbirds and butterflies also exhibit preferences for sucrose-rich nectars.
Amino acid concentrations tend to be high in lepidopteran-pollinated flowers in
tropical forests. Pollen composition also influences pollinator choice. H. G.
Baker and I. Bakd216 ,217) recognize two major groups of pollen: those with
starch and low lipid content, and those with no starch and high lipid content.
Bees tend to prefer starchless pollen, while hummingbirds prefer starchy pollen.
Small pollen grain size and low starch content are characteristic of bee or fly
pollens, and starchy pollens tend to be larger. Starchy pollens can be harvested
by some species of Heliconius butterflies as a source of amino acids for egg
maturation(3[) and perhaps also to increase energy reserves, from the breakdown
of starch components of collected pollen, for sustained flight associated with
oviposition behavior.
Ceratopogonid midges in the Canadian region exhibit considerable speci-
ficity for feeding on the pollen of certain plant species.(245) The Ceratopogonidae
(Diptera) are known for their diverse feeding habits, which include pollen feed-
ing in some lineages. Pollen feeding may be a link in why these midges are the
principal pollinators of tropical trees such as T. cacao and Hevea brasilien-
SiS.(644) Many dipterans require environmental sources of amino acids for proper

egg development, and this may also be the case for Euprojoannisia and other
genera associated with cacao flowers in the New and Old World tropics.(S22.s87-S89)
Successful pollination of cacao requires the deposition of composite aggregates
of pollen grains (pollen balls) of the correct variety on the style of an open flower
(Fig. 8.4), so pollination in this system depends upon the frequency of deposition
of aggregates, the size distribution of pollen aggregates, and the relative effec-
 c
<
Z
l>
s:
~
o."
o
:xl
C)
l>
Z
N
~
o
z
o."
Z
t/)
m
~
(")
o
s:
s:c
z
::::j
;;;
t/)

FIGURE 8.4. (A) Buds and open flower of T. cacao (Sterculiaceae); (B) buds, flowers, young and mature fruit of T. cacao showing ei
cauliflorous habit.
364 CHAPTER 8

tiveness of insect pollinators in depositing pollen aggregates. (645) Some studies in

Africa indicate that a variety of insects, in addition to ceratopogonid midges, are
involved in the pollination of cacao. (646) A review of the systematics of For-
cipomyia (Euprojoannasia) midges involved in the pollination of cacao is given
by Bystrak and Wirth.(406) Life cycle studies of these midges (Fig. 8.5) have been
conducted in both the field and the laboratory, (587Mr-652) and analysis of natural
breeding sites has also been undertaken. (590.649.651.653.654) Considerable interest
focuses upon the environmental factors affecting the abundance of pollinating
midges in cacao plantations.(522.588-590,647,648.651.655) The success of cacao as a
major cash crop of the tropics depends upon environmental factors such as
pollinator availability and disease and upon intrinsic genetic and physiological
mechanisms determining fruit set. (565.644.656-658) The structure of the cacao planta-
tion habitat, and the kinds of insects and other arthropods found in it, influences
the healthiness of cacao trees.(659-661) Mutualisms involving ants, including those
associated with cacao, are often successful, although a major focus concerns
ant-plant associations.(662.663) From systematic studies of the genus Theobroma
(Cuatrecasas),(595) it is evident that the group has been successful in undergoing
evolutionary differentiation in tropical forests of Central and South America.
Bawa(638U) has examined the breeding or sexual systems of tree species in
lowland tropical forests and has noted the high incidence of dioecious species
associated with these ecosystems. Trees bearing small flowers pollinated by a
variety of small bees are usually dioecious. Dioecism is an adaptation in plants
that permits outcrossing. The separation of sexes on different plants ensures
some level of gene flow among plants. Hemaphroditism in conjunction with
self-incompatibility also promotes outcrossing. Self-incompatible flowers pos-
sess structural features to minimize pollen interference, i.e., to prevent pollen
from the anthers of a flower ending up on the stigma of that same flower. If this
were not so, incorrect pollen grains would compete with space on the stigma for
correct pollen grains arriving on insect pollinators from other flowers. Accumula-
tion of the incorrect pollen grains on the flower lowers the likelihood of success-
ful zygote formation. Self-pollination generally includes the situation in which
incorrect pollen results in partial growth of the pollen tube and thus represents a
different level of incompatibility from pollen interference. Sexual or breeding
systems in tropical plants must be viewed in terms of constraints imposed by the
particular pollination systems involved for that species, and the optimal partition-
ing of resources (energy) to female and male functions (regardless of
monoecism, dioecism, hermaphroditism, etc.), so that female and male compo-
nents of fitness are optimized each generation. As Janzen(42) has indicated, a tree
full of hermaphroditic flowers may set few fruit but have a high fitness in terms
of performing the male function, i.e., be a pollen donor, and selection would
favor a breeding system, given the type of pollination system present, that
optimizes both male and female components of fitness within the breeding popu-
DYNAMICS OF ORGANIZATION OF INSECT COMMUNITIES 365

FIGURE 8.5. Schematized life cycle of Forcipomyia midges (Diptera: Ceratopogonidae). the pol-
linators of T. cacao in both the New and Old World wet tropics. Midges are about 2 mm long and
undergo life cycle in moist, rotting organic debris. (Illustration courtesy of Dr. Willis W. Wirth,
USDA and Smithsonian Institution.)

lation of the tree species under consideration. Level of fruit set, therefore, is not
the only measurable component of fitness for the population. In T. cacao, for
example, self-incompatibility exists because there is often partial growth of the
pollen tube when incorrect pollen lands on the stigma of the same flower. (656)
366 CHAPTER 8

Estimates of successful pollination in such trees must thus take into account the
frequency of partial pollen tube growth on individual trees within a specified time
period.
Wild species of Theobroma, including cacao, are self-incompatible,<595)
while many artificially developed cultivars of cacao are self-compatible.(656)
Self-compatible cacao usually has higher levels of fruit set than self-incompatible
cacao, presumably because pollinators visiting self-compatible varieties are able
to pollinate more flowers. Pollinator abundance may not be a limiting factor in
fruit set in self-compatible cacao. In its original wild state, cacao evolved a
certain maximal physiological level for flowering and fruit set. Ecological fruit
set is probably lower than the physiologically programmed level of fruit set for
most trees in self-incompatible stands of cacao due to effects of pollinator abun-
dance and diversity. Wild cacao is known to occur in small aggregate clumps in
the understory of tropical rain forest, (595,644) and the pollination system must have
been adapted to such a distribution. Purseglove(644) suggests that the question of
apparently low pollination of cacao in plantations should be approached with
viewing the plantation as a composite aggregate of many small clumps of cacao,
a pattern similar to the spatial arrangement of cacao in the wild.
Richards (94 ) mentions that many trees and shrubs of wet forest understories
bear small inconspicuous flowers, many cauliflorous and pollinated by a variety
of small insects other than bees, wasps, moths, and butterflies. The origins and
adaptive significance of cauliflory is still obscure.(394) Such pollination systems,
as in cacao, are probably highly buffered against shifts in seasons by the forest
overstory. Yet when forests are cleared away or destroyed by other means, such
systems are vulnerable in the direction of lowering pollination and fruit set,
thereby increasing the likelihood for local extinctions. Cacao probably represents
an exception that has been developed into a cultivar for human consumption, but
still the matter of low pollination in exposed environments remains.
An interesting question concerns the relationship between physiologically
programmed levels of fruit set and levels determined by ecological factors such
as pollination and disease.
K. S. Bawa (personal communication) has opened the possibility that cacao
and other tropical trees may exhibit a form of "functional andromonoecism" in
which some trees produce a greater abundance of physiologically and
morphologically bisexual flowers but which function as a source of pollen only.
The strategy of the tree under such conditions is to maximize the output of male
gametes. Such trees are characterized by many flowers but relatively few fruits.
The tree is physiologically set for relatively low fruit set but produces an abun-
dance of flowers, which function as a source of genetic material for pollination of
other trees in self-incompatible systems. Functionally andromonoecious plants
may exhibit either large, showy flowers or small, inconspicuous flowers (K. S.
Bawa, personal communication). Those with large, showy flowers are most
DYNAMICS OF ORGANIZATION OF INSECT COMMUNITIES 367

likely pollinated by bees; the others, by relatively small insects such as midges.
The absence of morphological andromonoecism in tropical plants such as cacao
may be due to selection favoring the maintenance of functional pistils: in
functional andromonoecism any flower has the capacity to be pollinated, while in
morphological andromonoecism only a small fraction have this capacity, possi-
bly greatly disrupting pollinator activity; investing in extra pistils, so to speak,
increases the option for selection of zygotes in future generations. Such a system
is adaptable to changing selection pressures affecting other portions of the
phenotype, such as interactions with herbivores. In hermaphroditic flowers such
as cacao, there is physiologically an equal distribution of reproductive effort
between male and female portions of the flower before fertilization, but if there is
functional andromonoecism, there is greater portion of flowers with the male
function because the tree only "needs" a number of fruit determined by the
number of functionally hermaphroditic flowers (i.e., those producing fruit).
Under such considerations, pollinator abundance would be less a limiting factor
in fruit set than if functional andromonoecism did not exist.
Alternatively, an absence of functional andromonoecism in self-
incompatible cacao suggests pollinator limitation of fruit set since levels of fruit
set are very low. Again, optimal fruit set will be determined by the physiological
state of the tree, but ecological factors such as pollinator deficiencies in the
habitat may lower fruit set even further than what normally occurs when pol-
linator popUlations are in some form of equilibrium with abundance of flowers in
a patch of trees. In such a system, increases in pollinator populations may result
in increased fruit set if pollinator popUlations are synchronized through time with
flowering periods. The observed lack of fruit set in the majority of flowers per
tree is the result of insufficient pollination rather than of functional andromonoe-
cism. Such distinctions are especially meaningful when examining fruit set pat-
terns in cultivated tropical trees such as cacao or at in trees in disturbed habitats.
In natural habitats one might expect some form of quasi-equilibrium between
pollinators and flowers, particularly in species that are self-incompatible and
require certain kinds of pollinators.
The maintenance of fruit set in self-incompatible plants such as cacao may
depends in part upon the movement patterns of pollinating midges relative to the
abundance of flowers and degree of spacing among flowering trees. For exam-
ple, in a small clump of wild cacao trees in flower, the average pollinating midge
may encounter a greater number of flowers on different trees within the clump,
which results in relatively greater fruit set than for a similar midge in a cacao
plantation. In the plantation habitat, midges may be distracted by a far greater
number of flowers by surrounding trees, which may result in a lower fruit set
than in wild clumps. Thus although satiation points may be the same in both
habitats, midges in the first situation visit a greater number of flowers per tree
than in the plantation and total fruit set is higher. The size, density of individual
368 CHAPTER 8

plants, and synchrony of flowering within resource clumps of different sizes may
have profound effects on the movement patterns of pollinating insects and the
percentages of flowers pollinated per tree. Intertree distances may be on the
average shorter in clumps of wild cacao than in planted cacao in plantations. If
so, pollinating midge efficiency may be greater in wild clumps than in planta-
tions since less distance is needed to reach different trees. Many tropical pollinat-
ing insects may exhibit density-dependent foraging if their preferred resources
are of relatively few kinds per unit area of habitat, thereby precluding diet
switches in response to rarity of anyone preferred type. But it is also predicted
that such pollinators would switch to other resources as the preferred one be-
comes too patchy. (] 5.640) As patch size decreases, it is less desirable to be a patch
specialist. (664) Pollinators should exhibit considerable patch selectivity when
there is a broad spectrum in patch quality and size. ([66) Charnov et ai. (665) show
that a pollinator or patch specialist will abandon that patch when its quality falls
below the average quality for all patches of that resource in the habitat or region.
Givnish(666) has reviewed the various hypotheses regarding the evolution
and adaptive significance of dioecism and monoecism in plants. He presents
information relating to when dioecism is adaptive to plants. One of his predic-
tions is that dioecism should appear in a tropical tree species when flowers or
floral rewards are large relative to the costs of pollen and seed production and in
which flowers are adapted to specialized pollen vectors such as some bees and
moths. The preponderance of trees in tropical forests indicates that dioecism
should be common, but not from the standpoint of promoting outcrossing.(666)
The reader is encouraged to read the review of concepts in Givnish.(666) Givnish
points out that the study of HubbeH<28) shows that dioecious species in tropical
forests are no more clumped in their spatial distribution than hermaphroditic
species, suggesting that the clumping of individual trees is not necessarily corre-
lated with breeding systems. Givnish(666) has also emphasized that the selection
pressures resulting from the seed dispersal and seed predation systems influence
breeding systems. The diversity in growth form of tropical vegetation indicates
that tropical ecosystems consist of plant species with diverse breeding systems
and that one must examine the breeding system of a particular group in terms of
many ecological factors. A cornerstone of such an effort is an examination of the
various factors affecting fitness within a breeding population: herbivory, seed
dispersal, pollination, and ecotypic effects. The studies of Bawa(638a) indicate a
diversity of breeding systems in tropical trees alone.
The "fruit investment strategy" is expected to vary considerably among
different plant species. It is that portion of available resources allocated to the
development of fruit as opposed to the development of flowers. When this
physiological potential is high in a given plant species, pollinator abundance is
more likely to be a limiting factor in production since the plant is capable
physiologically of producing a relatively high number of fruit. Frequency-
DYNAMICS OF ORGANIZATION OF INSECT COMMUNITIES 369

dependent selection is associated with the allocation of resources to fruit produc-

tion. Fruit set is limited by the amount of resources required to support the
development of fruit, pollinators, and mortality of flowers and fruit. In the case
of simultaneous hemaphrodites such as cacao, excess production of flowers
increases the number of pollinator visits and the amount of pollen dissemination.
In such a system, pollinating insects limit the distribution of pollen in the breed-
ing population. The flowering strategy, including the abundance of flowers rela-
tive to fruit set, is believed to be a compromise between fruit production and
pollen donation. Although it is traditionally maintained that dipterans and other
nonhymenopterous and nonlepidopterous insects are inefficient pollinators, we
need to examine the pollination systems of tropical trees and shrubs pollinated by
these insects in light of current theory and interest in tropical pollination systems
in general.(638.639) Mesler et al., (667) for example, working in coastal redwood
forests of Northern California, found high fruit set in two plant species and
attributed it to high abundance of pollinating fungus gnats (Sciaridae and
M ycetophilidae).
Janzen(42) notes that flower-to-mature-fruit ratios of from 100: I to 500: I
are common in large tropical trees in natural habitats and indicates that such trees
are physiologically programmed to observed levels of fruit set. Pollination ap-
pears to be a challenging process for tropical trees in that a variety of factors
influence fruit set. (42) For example, Janzen suggests that a large pool of stigmas is
necessary in order that the correct pollen, particularly in self-incompatible sys-
tems, ends up in the proper amount on at least a minimum number of stigmas
through activity of pollinators. It is also suggested that a large flower crop is
necessary to satiate pollinators that may be of limited numbers owing to competi-
tion with other plants.(42) Janzen suggests further that each tree species may
possess a strategy of functioning as pollen donors rather than capturing pollen
and developing zygotes, including hermaphroditic species. Thus individuals of a
certain tree species may have in fact reproduced more because they provided
pollen for other individuals of that species, even thought fruit set is low or
absent. Viewed in this light, low pollination may not be due to limited pollinator
supplies in natural habitats, but this may not necessarily be the case in crop
species such as cacao. The cultivation process of tropical tree species such as
cacao might have altered the average ratio of flowers to mature fruit, with an
even greater skew toward flowers because the agricultural habitats of such
species may not be optimal for pollinator populations. The arguments developed
by Janzen would be most applicable, therefore, to cacao in its wild state, and
with some offset in agricultural habitats due to insufficient abundance or spec-
trum of resources required by pollinating midge populations. Individuals that are
pollen donors and little else may occupy habitats suboptimal for seed bearing.(42)
Janzen also suggests that for species in which there is seed predation, greatest
fitness is achieved when large seed crops are produced so that such predators are
370 CHAPTER 8

satiated and a portion of seeds survive. If there are periods when large seed crops
cannot be produced, the optimal strategy of such a tree may be to be a pollen
donor.(42)
Pollinating insects in tropical ecosystems may be specialized for exploita-
tion of one or few resource types if the quality of these resources is high and if
their distribution and abundance are predictable.(411.4S!) Pollinators actively seek
out and select high-quality resources and not necessarily on the basis of their
abundance. Specialized pollinators are believed to be abundant in the
tropics,034,66S) and seasonal tropical regions may support a great wealth of pol-
linating species if plants partition the environment into different peak flowering
periods,(116) thereby permitting a larger pool of species to co-occur with allo-
chronic activity periods. Seasonal tropical regions may necessitate timed and
synchronized switches to different flowering species by a pollinator guild as the
resource spectrum changes and nectar or pollen rewards fluctuate in abundance
accordingly. Any single plant species may offer a spectrum of resource patches
to a set of pollinators, with considerable variation in patch quality. Depending
upon the amount of competition, if any, for high-quality patches among member
species in a pollinator guild, the exploitation of low-quality patches may result
from such interactions and if persistent may result in evolutionary or facultative
switches to other resources. Competition among pollinating s'pecies associated
with a particular plant species or group of plant species may result in one or more
of such species experiencing a temporary contraction in home range movements
as related to exploiting a certain number of resource patches in limited supply,
although overall diet preferences may not be altered and may expand in response
to competition.(34!) But if competition persists, and if the species in competition
possess sufficient genetic or phenotypic flexibility, selection may favor the evo-
lutionary switch to an alternative and abundant resource to alleviate losses in
fitness due to competition on the original resource base. (669)
Because many tropical plants have distinct flowering periods and are a
limited resource, closely related or ecologically similar pollinating species are
expected to partition available resources. (433,642,670) If there is considerable varia-
tion in patch size for nectar or pollen rewards, pollinators will exploit the most
abundant patches since nutritional intake is faster than for associations with rare
patches. Some tropical butterflies as well as Temperate Zone species exhibit
considerable constancy in daily exploitation of flower patches, and such home
range patterns are known in the tropics for some large bees.(40) Trapline home
range movements are known for tropical bees and butterflies,(31,40.33sl and such
pollinators may exploit several different species along the daily trapline route. In
some nonseasonal tropical forests, there may be a gradual temporal pattern of
flowering in some plant species, making the spatial distribution of such resources
constant throughout most of the year. Successional patterns of preferred re-
sources may impose shifts in foraging behavior among pollinating insects, thus
DYNAMICS OF ORGANIZATION OF INSECT COMMUNITIES 371

changing through time the organization of pollinator guilds and communities in

response to temporal patterns of flowering of resources.
Tropical forests undoubtedly consist of diverse mixtures of patch structures
among the many plant species found there, each one having a specific kind of
breeding structure. Some species may have disjunct breeding populations con-
sisting of widely scattered individuals, i.e., patchy distribution, while others may
be characterized by large patches of several individuals. Patch structure can also
be assessed in terms of flowering strategies: some tropical trees exhibit mass
flowering in which there is synchronized blooming of a large number of flowers
per tree, while others stagger flowering throughout the year. Mass flowering
makes available large patches of resources for pollinators, while staggered, low-
level flowering provides for small patches at anyone point in time. Mass flower-
ing invokes a density-dependent response in foraging by many tropical bees,
while small patches promote the trapline-type foraging. (40,499) Levin(64Ol predicts
that in the absence of strong selection distinct patterns of local differentiation in
tropical plants with patchy distributions will be low. Outcrossing promoted by
long-distance pollinators in such species prevents local differentiation. It is
stressed that many tropical tree species of low adult densities have very skewed
sex ratios, (638a,639) which reduces effective densities of breeding individuals even
more. (640l Density-dependent foraging in tropical pollinating insects is expected
to occur in plant species with large patch sizes. Levin<640) discusses the evolu-
tionary implications of patchiness in plant populations as related to foraging
behavior of pollinators.
A growing body of evidence indicates that various kinds of floral parasitism
involving insects are prevalent in tropical forest ecosystems. (67],672) Flowers,
with their nectar and pollen rewards being predictable resources, are targets for a
variety of insects and other animals, many of which are pollinators. (216) Ants form
a conspicuous component of the rain forest fauna,(673) and some of these species
routinely visit flowers. Nectar or pollen robbers take away rewards for pol-
linators and thus reduce the likelihood of pollination in some or all individuals
within the patch or breeding population. Thus intense and routine nectar or pollen
robbing reduces the average fitness and alters the breeding dynamics of the target
plant species. Barrows(671) distinguishes three kinds of floral robbers of Lantana
camara (Verbenaceae) in Costa Rica: nectar foraging, nectar-foraging perforat-
ing, and pollen foraging. A guild of floral robber insect species may therefore
accumulate on various plant species, depending upon factors that include (1)
selection pressure arising from limited resources for potential floral robbers, (2)
abundance of preferred hosts for plundering, (3) the pollinator guild or commu-
nity associated with the plant species, and (4) the floral structure and reward
physiology, including diurnal-nocturnal rhythms of production. Additional fac-
tors include the predictability and abundance of the resource in space and time.
Nectar robbery decreases cross-pollination and seed set in Lantana. (671) In this
372 CHAPTER 8

system the stingless bee Trigona Julviventris is a nectar-foraging perforator of

yellow flowers of this species.(67J) Perforating behavior by nectar robbers dam-
ages the flowers and removes the nectar reward for pollinators. Nectar robbing of
four-o' clocks, Mirabilis jalapa, by worker paper wasps (Fig. 8.6) can result in
considerable performation damage to corolla tubes,(674) a factor that may reduce
breeding success in the plant popUlation.
Nectar and pollen robbing in tropical plant groups may be much more
prevalent than indicated by the amount of information available to date. McDade
and Kinsman(675) found that pollen robbing by Trigona bees on hummingbird-
pollinated Aphelandra (Acanthaceae) in Costa Rica did not reduce pollination.
Flowers pierced by various insects and other bird species produced less nectar
after perforation in addition to removing nectar in Aphelandra. (675) When much
damage is frequent within a popUlation, pollinators avoid visits to damaged
flowers or increase foraging time on damaged flowers since the nectar reward per
flower is less.(675) Such floral parasitism may result in reduced fruit and seed set
in affected popUlations. McDade and Kinsman make the interesting suggestion
that floral parasitism may result from human disturbance affecting plant popula-
tions and that early successional communities in the tropics may have a greater
incidence of such relationships.
If the natural distributions of resources of potentially floral parasites are
altered as the result of such disturbance, there may be convergence to exploit
available high-energy and nutrient-rich resources of plant species characteristic
of such communities. Early successional communities in the tropics are charac-
terized by large patches of principally herbaceous species, and such distributions
provide readily available resources in large supply for floral parasites. Trigona
bees are expected to tum up frequently as floral parasites, given their opportunis-
tic and flexible foraging habits. (21) Introduced plant species in tropical regions
may be ideal targets for floral parasites(67L672) either affecting or not affecting
pollination.
When optimal resources are patchy in space, the popUlation structure of a
species exploiting these resources will adjust to the patchiness(676.677l or new,
more locally abundant resources may be used. In Costa Rica, the introduced
ornamental vine Thunbergia grandiflora (Acanthaceae), originally from India,
does not set seed but is visited by Trigona bees and hummingbirds, including
nectar and pollen thieves, as discussed by Young.(677a) At some sites currently
being studied, at least two species of hummingbirds routinely visit the large,
showy lavender flowers, and several species of Trigona exhibit two kinds of
thieving: (1) bees enter the corolla and bite through the tubular anthers and
collect the released pollen grains, and (2) bees land at the base of the calyx,
perforate the corolla tube, and collect nectar. In studies conducted to date, more
than 50% of the open flowers within a patch of the vine are so damaged at any
one time when flowering is high. The striking absence of fruit set in this intro-
 c
-<
z
»
s:
oen
o."
o
:u
C)
»z
N
~
6
z
o."
Z
en
m
~
o
"s:
s:c
z::j
iii
en

FIGURE B.6. Nectar-robbing behavior by a social paper wasp, Epipona querini (Vespidae: Polistinae: Polybiini), on
the inflorescences of four-o 'clocks, Mirabilis jalapa (Nyctaginaceae), in Costa Rica. (A) Wasp lands and seeks base of
w
corolla tube; (B) and (C) wasp bites through calyx and corolla, leaving large hole after taking nectar. C:l
374 CHAPTER 8

duced vine in Central America is possibly due to a lack of effective pollinators.

In the New World tropics, hummingbirds and Trigona bees have converged to
exploit the nectar and pollen rewards, but the vine reproduces only by vegetative
growth. Within a patch, flowers are present throughout the year, making a
predictable resource for hummingbirds and insects, as discussed by Young.(677a)
These preliminary studies of the lack of effective pollination of the introduced
ornamental T. grandiflora in Costa Rica, and the associated rampant floral
parasitism, suggest that introduced plant species in tropical regions may be
subject to rapid exploitation by endemic insect species. If resources for some of
these insect species are limited, there might be short-term facultative switches to
novel resources or long-term evolutionary switches if the novel resource remains
and becomes abundant. Likewise, cases may exist where introduced pollinating
insects displace native pollinating species if the former have a competitive advan-
tage over resident species. For example, there is some evidence that the Af-
ricanized honey bee is displacing resident pollinating stingless bees in Neotropi-
cal regions. (678) These studies point to the high instability of interactions between
some pollinators and their resources in the tropics.
Tropical forest ecosystems are highly structured assemblages of organisms,
with characteristic patterns of vegetation in distinct successional stages. An
understanding of the guilds and communities of insects associated with different
plant species, and the trophic organization of insect communities, is obtained by
examining the diversity and patch structure of plant species in different succes-
sional stages. In this section, we have touched on a few illustrative aspects of
how ecological properties of plant species determine the structure of insect com-
munities and guilds. Detailed studies of the natural history of individual insect
species are needed to determine the role of abiotic factors (temperature, rainfall,
light intensity, etc.) and biotic factors (competition, predation, mutualism, re-
source limitation, etc.) in order to explain the population structure and dynamics
within different communities.
The considerations discussed by Janzen(42) of fruit and seed crop patterns of
variations on individual large trees in tropical forests indicate that tropical or-
ganisms may possess inter-individual patterns of variation within species that are
adaptive to specific forms of selection pressures. Spatial variation in the distribu-
tion of fruits or seeds in the canopy of a tree or on the forest floor may influence
the manner in which such resources are exploited by insects and other animals.
Lowland Neotropical rain forests exhibit continuous production of fruits and
seeds throughout most of the year/lOt) even though there may be a succession of
fruiting species of large trees over the year. Yet within the fruit or seed crop of an
individual tree of a particular species, a spectrum of variation may exist in
weight, position, and other traits affecting attractiveness as a resource for insect
species. In forests of the Old World tropics, some tree species are monocultures
DYNAMICS OF ORGANIZATION OF INSECT COMMUNITIES 375

due to growing on poor-quality sandy soils, (332) and such large trees exhibit
mast fruiting. These trees are believed to possess well-developed systems of
chemical defenses against herbivores, at the same time show pulsing of fruit-
ing and generally low productivity due to poor nutrient profiles of the soil. Such
monoculture forests have a low biomass of insects. <3(2)
A high species richness of insects, both herbivores and other trophic groups,
is found in mixed-species tropical rain forests with high harvestable pro-
ductivity.(47,412) Pollinating insects are more diverse in Neotropical mixed-species
rain forests, thus providing the raw materials for the recycling of plant resources
in the form of seeds and fruits and their breakdown products.(42) Specialized
coevolved interactions between plants and insects provide the basic ecological
fabric of Neotropical rain forests, with plant diversity made potentially possible
by the high nutrient profiles of volcanic soils. Janzen(57.679) points out that when
mixed-species tropical rain forests are altered, the coevolved interactions may be
dissolved although some of the participants may persist outside of the context of
such interactions. The annual succession of vegetative growth and sexual repro-
duction of plant species in tropical rain forests provides a diverse resource spec-
trum for insects and other animals. The species richness of herbivorous and
pollinating insects, as well as the predators and parasitoids exploiting them, is
determined by the plant species richness of the habitat. The review of various
complex ecological interactions between plants and insects by Janzen(4!) illus-
trates the interdependence of these organisms in determining the overall structure
and composition of tropical forest ecosystems.

8.4. SOME ECOLOGICAL FACTORS GENERATING PATCHY

DISTRIBUTIONS OF INSECTS IN TROPICAL HABITATS

The membership of an insect population will have a spatial distribution in a

habitat determined to a large extent by the distribution of resources. Population
growth is a function of the size of the existing population and the effects of
intrinsic and external factors. (680.681) If the resources are widely scattered as small
patches, and if the breeding population of the insect species being investigated
possesses life history traits(92) so that reproductive effort is distributed over more
than one patch at any moment in time, the insect population is fractionated over
some or all patches of the resources. Thus we have defined one parameter
affecting patchiness, namely, the spatial availability of those resources influenc-
ing fitness.
From previous considerations in this book, we also know that a variety of
other ecological factors affect patchiness of an insect popUlation.
I. When there is a quality distribution associated with a spectrum of neces-
376 CHAPTER 8

sary resources. Selection in most situations will favor the exploitation of optimal
resource patches even though other patches of the same resource exist in the
habitat.
2. When there is habitat selection, in which one or more contingent habitats
differ considerably in resource quality and abundance. Selection should in most
instances favor the exploitation of high-quality habitats by an insect species, with
or without gene flow between habitats.
3. Optimal quality or quantity of resource patch or optimal habitat selection
reflects the physiological capacities of the species, but other ecological factors
may alter the choice of potentially optimal resource or habitat patches. Intraspe-
cific and interspecific competition may result in a density-related dispersal of
individuals over a spectrum of resource or habitat patches, some of which may be
suboptimal in terms of providing resources that maximize average fitness.
4. Coactions (Chapter 2) such as predation and parasitism may vary consid-
erably among different resource or habitat patches, but an optimal patch from the
standpoint of food or resource quality may not necessarily be an optimal patch in
terms of low predator or parasite density. Selection in most instances should
favor choice of patches where predation or parasitism is low on prereproductive
or young reproductive age classes, assuming that the loss in fitness from such
mortality is less than loss of fitness from selecting resources or habitat of subop-
timal quality in another dimension of the niche.
S. Some habitats or resource patches may be better for the insect in terms of
microclimatic or other abiotic factors affecting fitness. Selection should favor
patterns of site specificity and degree of residentiality in part due to variations in
microclimate and abiotic conditions in general.
For all of these classes of ecological factors influencing the degree of
patchiness in an insect population, the extent to which these factors fluctuate
through time must also be considered, especially in terms of tropical seasonal
regions and degree of regional and habitat stability (see the discussions of light-
gap dynamics in tropical forests in Hartshorn(337) and Denslow(682) for some
recent thinking and evidence indicating considerable instability in lowland tropi-
cal rain forests). The degree to which individual host plant species, for example,
will be patchy resources for phytophagous insects and targets for the predators
and parasitoids of these insects depends in large part upon the successional stage
of the community.(683,684)
Milld685 ,686) states that competition between species results in either (1) an
increase in niche size and concomitant elimination of some ecologically similar
species in the same habitat or (2) a reduction in niche size with increased ecologi-
cal diversification, so that overall diversity increases in the community and the
incidence of specialization is greater. Populations at or near equilibrium levels
with the available resources in the habitat may enter into competitive interactions
DYNAMICS OF ORGANIZATION OF INSECT COMMUNITIES 377

with other species. But if there are considerable temporal fluctuations in the
availability of resources, the intensity of such competitive interactions will vary
considerably within such a guild or community. (463,464) Vertical distributions of
the populations of closely related species of scorpion flies in a Temperate Zone
habitat exhibit patterns in which more aggressive species occupy the higher
quality strata in terms of feeding sites.o 84 ) In the Temperate Zone, scorpion flies
maximize foraging above the ground level, a behavior exhibited by many tropical
ants that are scavengers (Thornhill; (184) Thornhill's citation of C. R. Carroll,
personal communication), and ThornhilP84) suggests that the essentially extra-
tropical distribution of scorpion flies may be due to competition with arboreally
foraging ants in the tropics. If so, the implication is one of complete displace-
ment of different insect groups as a result of competition for food supplies in
arboreal microhabitats in the tropics.
Baldry and Molyneux(494) have shown that the dipteran Glossina medicorum
in West Africa exhibits considerable preference in terms of habitats where
trypanosome infections are high. Although the species is widely distributed,
different populations vary considerably in the extent of infection at anyone time
of the year and the differential effects are correlated with the type of forest
habitat.(494) In some species of Simulium flies that function as vectors of on-
chocerciasis, the distribution and abundance of the aquatic larval stages are
determined largely by the availability of host arthropods such as crabs and
mayfly nymphs, to which the larvae attach themselves.(687) An analysis of the
habitat requirements of the host arthropods is often needed to determine the
potential for outbreaks of the flies in a particular region of Tanzania.(687) Such
studies and others concerning correlations between resource quality, patch size,
and equilibrium popUlation sizes for a species(677) show that population growth
and size are directly related to the quality of resource patches. Dethier(279) has
considered the problem of host plant popUlations limiting population size in the
Lepidoptera in a general way: reduced food supplies result in mass dispersal of
larvae when food supply is very low, which in tum results in considerable
mortality of larvae associated with dispersal. Yet other studies of Lepidoptera in
particular suggest that larvae disperse from individual host plants as a result of
diminished quality of the resources provided,(688) even though these species may
have a strong coevol ved association with a certain host plant species. (689) Some
tropical butterflies exhibit specific forms of home range behavior determined in
part by the distribution of larval and adult food supplies. (479) The carrying capac-
ity of the habitat may be very different for closely related species, and density-
related effects may influence population size in different ways. (81,82) In some
parts of Uganda, the dipteran Simulium neavi and two other species, vectors of
onchoceriasis, have larval populations requiring completion of the life cycle in
crabs, but considerable mortality is associated with crab-locating behavior. (687)
378 CHAPTER 8

The results indicate that shifting distributional patterns of the crabs, even in
confined conditions, make for an unpredictable "resource" for the fly larvae,
contributing considerably to regulation of larval popUlations.
One kind of patchy environment characteristic of lowland and mid-elevation
tropical rain forests is the tank bromeliad, so named because its structure results
in the accumulation of rainwater. The water is held long enough to allow the
buildup of invertebrate communities. There are many species of such
bromeliads, and each species may have a characteristic vertical distribution in the
forest habitat. Because tank bromeliads harbor communities of microorganisms
and potential disease-vector dipterans, there is a medical interest in the species
breeding in these patchy microhabitats. The tank bromeliad, with its mini-aquatic
environment, has been analogized in the literature as a microecosystem com-
posed of invertebrate herbivores that feed on the plant's tissues, algae, and other
plants growing from it; decomposers exploiting accumulated organic debris such
as twigs and fallen leaves; and carnivores that feed on the organisms in these
other trophic levels. The system is very open, primarily as the result of verte-
brates (birds, lizards, frogs, snakes) feeding on the invertebrates in tank
bromeliads. The studies of Laessle(407) on the physical and chemical properties of
the water in tank bromeliads indicates considerable heterogeneity among
bromeliads in a region or habitat, suggesting heterogeneous selection pressures
or nongenetic physiological response characteristics for organisms inhabiting
these microhabitats. Laessle mentions that a variety of insect larvae, such as
Diptera and Odonata, thrive in such environments along with many other ar-
thropods and other invertebrates. Calvert(690) studied the association of nymphs
of giant Neotropical damselflies (Megaloprepus and Mecistogaster) with tank
bromeliads in Costa Rica. In some instances, what appear to be rather simple
interactions between arthropods in the tropics, tum out to be more complex when
studied further. Thus although the large, familiar orb-spinning spider Nephila
clavipes constructs elaborate, sticky webs to attract a variety of prey, the spider
itself also attracts a variety of parasitic and predatory or scavenging flies that form
a small assemblage of arthropods interacting in specific ways.(40S) Further com-
plexity is added by the exploitation of prey lodged in these webs by the giant
damsel fly Megaloprepus coerulatus(S27J
Vandermeer et al. (690 studied the dynamics of protozoan communities of
tank bromeliads in Costa Rica, showing evidence of competitive exclusion of
Paramecium. Because of the large biomass provided by these epiphytes in tropi-
cal rain forests, and because of the association of many different organisms with
them, many of which are probably highly specialized to such microhabitats, the
distribution and abundance of tank bromeliads in tropical rain forests may ac-
count for an appreciable portion of the animal diversity found in the tropics. For
example, an analysis of 11 large tank bromeliads sampled once in the Costa
Rican highlands (Bajo La Hondura) revealed a total of 65 species of inverte-
 TABLE 8.2
The Distribution and Abundance (Number of Individuals) of Aquatic and Terrestrial Arthropods a
among 11 Tank Bromeliads at Bajo la Hondura Pasture, Costa Rica, February 1975 b

Aquatic Terrestrial
arthropods arthropods Totals
Height
above Diameter'" Number of Number of Number of Number of Diversity
Bromeliad ground (m) (em) species individuals species indi viduals Species Individuals (H')

I 2.0 40 0 0 5 35 5 34 0.428
2 2.8 99 3 31 2 2 5 33 0.607
3 4.0 78 7 23 0 0 7 23 0.576
4 2.0 18 3 3 4 4 7 7 0.845
5 4.0 25 3 3 5 5 8 8 0.905
6 1.6 40 0 0 4 118 4 118 0.149
7 1.6 40 4 23 6 15 10 38 0.813
8 1.8 72 1 10 6 33 7 43 0.467
9 1.5 51 2 7 2 12 4 19 0.939
10 2.0 30 0 0 2 223 2 223 0.144
11 2.4 80 8 5 94 6 102 0.377

"The following taxa were present and censused (A, aquatic; T, terrestrial): insects: Diptera (A), Neuroptera (A), Coleoptera (T), Orthoptera (T), Homoptera (T); other arthropods:
Isopoda (T), Arachnida (T), Annellida (T), Diplopoda (T), Chilopoda (T), Collembola (T).
·Unpublished data (A. M. Young, instructor) collected by students in the Lawrence University (Appleton, Wisconsin) winter term "Tropical Ecology" program.
'Diameter as measured from tips of leafaxils.
380 CHAPTER 8

TABLE 8.3
Changes in the Biomass of Insects Attracted to Experimental Baits a in the

Hymenoptera
Diptera (excl. ants)

Time Total Total d Total Total

interval SC Ne S N

8:30-9:30 A.M. 20 116 3 15

9:30-10:30 A.M. 19 129 3 7
10:30-11:30 A.M. 17 91 10 42
11:30 A.M.-12:30 P.M 18 117 8 64
12:30-1:30 P.M. 12 69 7 80
1:30-2:30 P.M. 17 51 7 105
2:30-3:30 P.M. 19 84 7 185
3:30-4:30 P.M. 18 66 6 156
4:30-5:30 P.M. 21 77 4 158
5:30-9:00 P.M. 3 0 0
X± S.D. X± S.D. X ± S.D. X ± S.D.
8:30-9:30 A.M. 1.81 ± 1.40 10.54 ± 11.23 0.27 ± 0.46 1.36 ± 2.83
9:30-10:30 A.M. 1.72 ± 1.55 11.72 ± 12.67 0.36 ± 0.50 0.63 ± 0.92
10:30-11:30 A.M. 1.54 ± 1.63 8.27 ± 10.37 0.90 ± 1.04 3.81 ± 7.44
11:30 A.M.-12:30 P.M. 1.63 ± 1.36 10.63 ± 12.55 0.72 ± 0.78 5.81 ± 9.42
12:30-1:30 P.M. 1.09 ± 1.30 6.27 ± 8.58 0.63 ± 0.80 7.27 ± 14.96
1:30-2:30 P.M. 1.54 ± 1.12 4.63 ± 4.67 0.63 ± 0.67 9.54 ± 19.27
2:30-3:30 P.M. 1.72 ± 1.27 7.63 ± 10.23 0.63 ± 0.67 16.81 ± 45.58
3:30-4:30 P.M. 1.63 ± 1.74 6.00 ± 6.51 0.54 ± 0.68 14.81 ± 24.75
4:30-5:30 P.M. 1.90 ± 1.92 7.00 ± 8.02 0.36 ± 0.67 14.36 ± 32.59
5:30-9:00 P.M. 0.09 ± 0.30 0.27 ± 0.90 0.00 ± 0.00 0.00 ± 0.00

a Eleven types of bait. including 8 odoriferous ones (meat, banana, pineapple, orange, papaya, tomato, jelly,
eugenol) and controls (salt water, mud) were used to examine species richness and abundance of insects in the
forest understory and their distribution throughout the day.
b The locality is Finca la Tirimbina near La Virgen (220 m elevation), Heredia Province, and the habitat is under-
story (bait height range 1-2 m) in mixed primary and advanced secondary tropical rain forest.

brates and 649 individuals, of which more than 50% of the latter were isopods
(Table 8.2). The communities were found to consist of recognizably terrestrial
and aquatic insects and other invertebrates (Table 8.2). Although the sample size
is small, the data indicate considerable variation from bromeliad to bromeliad in
terms of species diversity in the communities, when using the diversity measure:
H' = (C/N) N 10glON - ni log ni
There are no apparent correlations between diversity and size of bromeliad or
height above the ground. All 11 bromeliads were sampled in a single pasture
dotted with various trees. Those bromeliads with very low diversity measures
 DYNAMICS OF ORGANIZATION OF INSECT COMMUNITIES 381

Understory of Premontane Tropical Rain Forest in Northeastern Costa Rica b

Ants Coleoptera
Overall total
Total Total Total Total
S N S N S N

3 70 9 44 35 245
II 331 8 24 42 491
13 1,090 9 25 49 1,248
16 1,816 14 30 56 2,027
13 1,270 11 32 43 1,451
8 441 17 41 49 638
5 771 20 55 51 1,095
12 1,059 17 33 53 1,314
8 411 16 51 49 697
3 504 16 74 20 581
X ± S.D. X ± S.D. X ± S.D. X ± S.D. X ± S.D. X ± S.D.
0.27 ± 0.46 6.36 ± 19.47 0.81 ± 0.98 4.00 ± 4.57 3.18 ± 2.27 22.27 ± 25.79
1.00 ± 1.00 30.09 ± 87.70 0.72 ± 0.78 2.18 ± 2.67 3.81 ± 2.89 44.63 ± 88.52
1.18 ± 1.32 99.09 ± 305.66 0.81 ± 0.75 2.27 ± 2.61 4.45 ± 3.58 113.45 ± 305.51
1.45 ± 1.43 165.09 ± 509.55 1.27 ± 1.42 2.72 ± 3.60 5.09 ± 3.30 184.27 ± 513.62
1.18 ± 1.07 115.45 ± 359.84 1.00 ± 1.09 2.90 ± 3.36 3.90 ± 3.36 131.90 ± 369.87
0.72 ± 1.19 40.09 ± 29.33 1.54 ± 2.20 3.72 ± 5.58 4.45 ± 3.35 146.30 ± 58.00
0.45 ± 0.68 70.09 ± 225.53 1.81 ± 2.99 5.00 ± 7.61 4.63 ± 4.03 99.54 ± 268.36
1.09 ± 1.04 96.27 ± 303.12 1.54 ± 2.16 3.00 ± 4.42 4.81 ±4.19 119.45 ± 323.48
0.72 ± 0.64 37.36 ± 113.68 1.45 ± 1.80 3.72 ± 4.79 4.45 ± 4.10 62.45 ± 142.99
0.27 ± 0.46 45.81 ± 150.63 1.45 ± 1.69 6.72 ± 8.13 1.81 ± 1.83 52.81 ± 148.55

'Number of species.
d Totals, N values, are best considered as "bits" rather than as individuals because some or all of the actual indi-
viduals present during a particular time interval may occur in subsequent time intervals, thereby biasing the
actual number of individuals present, as no marking was done.
e Number of individuals.

(such as Nos. 6, 10, and 11) were characterized by a very high abundance of
either isopods (for Nos. 6 and 10) or dipteran larvae (No. 11) (Table 8.2).
Bromeliads such as No.5 with a high diversity were characterized by a low
number of individuals, high number of species, and in some cases, approxi-
mately equal representation of both aquatic and terrestrial forms. The composi-
tion of individual communities was determined, of course, to a large degree by
the degree of wetness or dryness and the amount of litter in each bromeliad.
In an ongoing study with about 14 months of data obtained to date, "artifi-
cial" plastic-cup bromeliads established in two cacao plantations in Costa Rica
reveal some interesting distributional patterns of insects and other invertebrates.
382 CHAPTER 8

The cups were anchored to branches of cacao trees, one per tree and to 20 trees in
each plantation, perforated in the bottom for drainage but initially filled with a
mat of washed leaf litter. The two plantations are very different in terms of
degree of shade cover over cacao trees and the heterogeneity of shade cover trees:
one plantation (La Tigra) has a mixed canopy of various wild tree species, and
the other (El Uno) has a rather uniform cover of Hevea rubber trees. Eleven of
the 20 artificial bromeliads censused in La Tigra after 14 months had thriving ant
colonies, a total of four species present, but only one species in a cup. The
colonies had larvae and pupae. The remaining cups had low densities of roaches,
isopods, and a few other invertebrates. In the El Uno plantation, 10 of the cups
had ant colonies, but species different from those in La Tigra, and the colonies
were generally smaller than those of La Tigra. A few isopods were found in the
other cups along with one or two species of spiders. One cup, clogged and filled
with rainwater, had a belastomid bug and many mosquito larvae. Such data,
while preliminary, indicate that experimentally introduced artificial arboreal
habitats are rapidly colonized by ants and a few other invertebrates. Privat(649J
found that ceratopogonid and cecidomyiid midge larvae breed in natural
bromeliads in Costa Rica, but none of these were found in the artificial
bromeliads in this study. Colonization of natural or artificial bromeliads in the
tropics is a function of the suitability of individual bromeliads for certain or-
ganisms and the degree to which a community builds up in each one.
Short-term temporal variation in abundances of insects at resources in the
understory of tropical rain forest can be studied by placing a series of "experi-
mental" and "control" food resources (baits) in the habitat, replicating these,
and observing the numbers of species and individuals per species on the baits
over a diurnal cycle. Sometimes the data generated by such a study provide an
estimate of temporal evenness in the species composition and abundances for the
community of insects and other arthropods exploiting, or attracted to, these food
types. Because the system is experimental and combines both odoriferous, nutri-
tive baits and nonodoriferous, nonnutritive ones, the degree to which the ob-
served patterns of insect distribution approximate natural distributions over the
naturally distributed spectrum of resources remains unknown. Yet the data can
reveal some interesting comparative differences among various taxa086J and
thereby tell us something about the relative differences or similarities between
species associated with the resource spectrum studied.
When a series of experimental and control food resources was placed in the
understory of a mixed primary and secondary forest within the premontane
tropical wet forest region of northeastern Costa Rica during the short and erratic
dry season, the predominant groups of insects found on the baits were Diptera,
Hymenoptera, and Coleoptera. Each of these orders was represented by several
subfamilies, genera, and species, and the number of species and individuals
(abundances) remained markedly stable from 8:30 A.M. to 5:30 P.M. for all
DYNAMICS OF ORGANIZATION OF INSECT COMMUNITIES 383

orders, lumping the data from all baits (Table 8.3). A marked decline was found
for most groups, save for some ants, after dark (Table 8.3). Virtually all groups
peaked in species abundance and numbers of individuals between 10:30 A.M. and
1:30 P.M., and ants made up the greatest biomass of insects present. The nine
experimental baits yielded more than 80% of all insects censused, with the
remainder on controls (Table 8.4). While the distribution of flies was remarkably
even over the experimental baits, the Hymenoptera were bimodally skewed,
with two groups, predatory wasps and bees, being associated with the meat
and eugenol baits, respectively (Table 8.4). Eugenol, while not a food source,
tests for the effect of a strong odor, mimicking the odor components of some
flowers, while the other experimental baits contain both odoriferous and nutri-
tive substances.
Ants were heavily associated with papaya in terms of species richness, but
the greater abundances were found on meat (Table 8.4). Coleoptera, mostly
Staphylinidae, were greatest in species richness on fruit baits such as pineapple,
orange, and tomato, with greatest abundances being on pineapple and tomato

TABLE 8.4
The Distribution and Abundance of Various Groups of Insects on Baits
for the 11 :30 A.M. to 12:30 P.M. Time Interval for the Premontane Tropical
Rain Forest Data from Costa Rica

Hymenoptera
Diptera (excluding ants) Ants Coleoptera All insects

%of % of % of % of % of % of % of %of % of % of
all S" all ]Vb all S all N all S all N all S all N all S all N
Baits (18)< (117)d (8) (64) (16) (1816) (14) (30) (56) (2027)

Meat 16.6 5.9 25.0 34.3 18.8 93.6 7.1 3.3 16.1 85.3
Banana 11.1 7.6 0.0 0.0 6.3 0.4 7.1 10.0 7.1 0.9
Pineapple 16.6 29.6 12.5 40.6 0.0 0.0 21.4 26.7 12.5 3.4
Orange 16.6 4.2 0.0 0.0 6.3 0.1 28.6 3.3 14.3 0.8
Papaya 16.0 17.8 12.5 10.9 31.3 4.2 7.1 10.0 17.9 5.2
Tomato 11.1 25.4 12.5 1.6 0.0 0.0 21.4 20.0 10.7 1.8
Jelly 11.1 8.4 12.5 1.6 12.5 0.5 0.0 0.0 5.4 0.5
Eugenol 0.0 0.0 25.0 10.9 6.3 0.1 7.1 3.3 10.7 0.9
Salt water 0.0 0.0 0.0 0.0 6.3 0.7 0.0 0.0 1.8 0.6
Water 0.0 0.0 0.0 0.0 6.3 0.1 0.0 0.0 1.8 0.0
Mud 0.0 0.0 0.0 0.0 6.3 0.3 0.0 0.0 1.8 0.3

as, Number of species.

·N, Number of individuals.
"The numbers in parentheses refer to the number of species found on baits for each insect order.
dThe numbers in parentheses refer to the number of individuals for all species found on the baits, for each insect
order.
384 CHAPTER 8

100
GHANA' HABITATS

,-
60 90
_ Coastal scrub
~

i
10
.0
and grassland
~, nigricons 70 _ Guinea sovonno-
~ woodland
40 60
_ High for.st

I
L
'0
0 30

I
40

30

~
20
20
10
10
fiJ

. B
0
A , , , , c
co ~ c
:: ~ i. , iz ~ co

.
~

Ii : ~
~
IIj ~
~
~ i g
~
~ 0 i5 c ~ ;;

-.......
100

90

80 iiI_uta.eked
§ ~
I~
70 ~
ni"ricans
i~
60

.fI!
L

'0
0 0

il
40 ~fI!

~
z
u 30

. . . .. ..
20

~ ='" _2
0

i
0
g
C co
0
0
2

.
z
z
~
~
~ ~
~
~
;;
~
~ ~

FIGURE 8.7. Prey selection patterns by African driver ants, Anomma. (A) Insects make up the
major kind of arthropods taken as prey in all habitats and by different species of these ants. (B)
Resource partitioning for insect prey by two species of African driver ants, (C) and by three species.
(D) Different species of these ants forage in different habitats, resulting in the observed pattern of
prey resource partitioning. [From W. H. Gotwald, Jr., Ann. Entomol. Soc. Am. 67:877-886 (1974).
Copyright 1974 by the Entomological Society of America.]

(Table 8.4). As expected, control baits were virtually empty in many instances
save for a low occurrence of ants (Table 8.4).
The data indicate that some insects, e.g., ants, are patchy in distribution in
that both species and numbers of individuals peak on certain food resources and
not on others. And although there is a diurnal replacement of some species in all
of the orders studied, the data show that overall species richness and biomass of
this "synthetic community" associated with the baits are stable over the period
of 1 day. In some instances, such as the interaction of Coleoptera with tomato
bait, species richness is inversely related to abundance, but in several other
associations of insects with baits, there appears to be a positive relationship
between these two components of community structure (Table 8.4): as species
accumulate at baits, so do their numbers or population sizes. On a per hour basis,
DYNAMICS OF ORGANIZATION OF INSECT COMMUNITIES 385

the range in number of species, for all groups of insects studied, was from 20
(9:00 A.M.) to 56 (12:30 P.M.), while for numbers of individuals for all species
and orders was from 245 (9:00 A.M.) to 2027 (12:30 P.M.). Of the maximum
number of individuals seen at one time (2027), about 90% were ants, and after
9:00 A.M., ants made up more than 50% of all insects on baits. Since 60-90% of
all ants were found on the meat bait, it appears that these forms were predatory.
These preliminary data point to the need for further long-term and replicated
studies of this approach.
In some predatory insects such as African driver ants, while there is often a
preference for a general category of prey among all species, different species in
different habitats will exhibit distinct prey preferences within the generalized
category, as exemplified by the study of Gotwald. (430) The explanation for such
differences is that the abundance of different kinds of prey will vary in different
habitats (Fig. 8.7). It is the interaction of habitat type with prey types and
predators that mold the distinctive features of divergence in resource exploitation
in these insects.
 CHAPTER 9

INSECT SPECIES IN
AGRICULTURAL HABITATS
IN THE TROPICS

From the summaries in the previous chapters concerning the well-established

patterns of species richness of insects in lowland and mid-elevation tropical
forests, it is clear that various kinds of crop lands adjacent to mixed tree species
forests in the tropics will be invaded to varying degrees by insect taxa in space
and time. It is evident that mixed species tropical forests of the New World
represent complex and dynamic assemblages of organisms and that a major
evolutionary and ecological interface is the sets of co evolved interactions be-
tween insects and their host plants. These interactions impose an unstable
equilibrium in the distribution and abundance of species in tropical forest ecosys-
tems. For some insect species, there may well be the case in which breeding
populations in natural ecosystems are kept well below the carrying capacity of
the habitat by carnivorous predators and parasitoids. Janzen<72J has amply shown
that such organisms tend to be most diverse in the very same habitats where both
herbivorous insects and plants are also diverse, namely, lowland and mid-
elevation tropical rain forests where the harvestable productivity from plants is
presumed, but not demonstrated, to be high. Yet for other species such controls
may not be as effective and insect populations under such conditions may ap-
proach an equilibrium states with available resources. It is not hard to imagine
that insect species in productive tropical ecosystems exhibit both forms of popu-
lation structure in space and time.
The floral and faunistic components of tropical forest ecosystems, espe-
cially the insects for the faunistic part, are very old and have had considerable
evolutionary opportunity to develop coevolved interactions. Janzed 41 ,47,57) has
stressed the dynamic state of tropical insect-plant interactions and the adaptive
role of variations in resources from the viewpoints of both insect and plant for a
specific kind of interaction. We know that the occurrence of natural successional
stages in plant communities in the tropics is a gradual affair, with specific sets of
ecological traits associated with each stage.

387
388 CHAPTER 9

9.1. PLANT -INSECT INTERACTIONS AND TROPICAL

AGRICULTURAL HABITATS: SOME COMMENTS

The occurrence of successional stages implies that natural tropical forest

ecosystems undergo change, mostly the result of natural perturbations such as
floods, mud slides, hurricanes, and large- and small-scale tree falls. Undoubtedly
such events alter the structure and composition of insect communities associated
with the vegetation affected. Compared with adjacent forests, tropical fields left
alone tend to build up distributions of plant species with large patch structure and
relatively low diversity, and such habitats may have a high harvestable pro-
ductivity. The result is that herbivorous insects under such conditions may build
up large popUlations on dense host plant patches, especially if carnivorous pred-
ators and parasitoids cannot regulate herbivore populations under these condi-
tions. (72) Such herbivore species are well adapted for intraspecific competition,
and the patch structure of plant species might be such that interspecific competi-
tion is minimized by abundant food supplies. The result is an evolutionary
accumulation of many herbivore species in secondary succession habitats. Be-
cause insects can evolve rapidly to meet changes in the environment, and because
lowland and mid-elevation tropical insect faunas can be very rich,(72.412) these
organisms become integral components of agricultural habitats in tropical re-
gions.
Studies of tropical forest ecosystems show that the bulk of nutrients are kept
in the flora and fauna, not in the soils. Thus when tropical forests are cleared
away by humans for agricultural endeavors, sustained crop yields for food-type
crops generally last only a few years. Furthermore, such crop species are vulner-
able to intense attack by herbivorous insects. Monocultures of food crops in the
lowland and mid-elevation tropical regions have a different spatial arrangement
of the vegetation from natural forests: individuals of the same plant species are
packed closely together in monocultures. Janzen(57) explains that a plant species
has three lines of defense against attack by herbivorous insects: (1) periods of
unfavorable weather reducing the insect populations or resulting in a period of
inactivity each year, (2) a complex of carnivorous predators and parasitoids that
regulate the abundance of the herbivorous forms, and (3) the scattering of indi-
vidual host plants of a given species in the habitat, making it difficult for her-
bivorous insects to locate them. The combination of host plant spatial
heterogeneity and carnivorous predator-parasitoid complexes is probably suffi-
cient to regulate herbivorous insects on many plant species in the Temperate Zone
since the intervening winters decimate insect populations and the annual growing
season is short. (57) Furthermore, unpredictable patterns of fluctuations in weather
at the boundaries of Temperate Zone growing seasons provide for density-
independent regulation of insects. Yet in the lowlands and mid-elevations of the
tropics weather fluctuations are generally predictable and many regions are non-
INSECT SPECIES IN AGRICULTURAL HABITATS 389

seasonal, conditions favoring a steady buildup of insect populations over a long

annual growing season.
Regions with a strong dry season are somewhat analogous to Temperate
Zone regions in that some insect populations decline during this period, and
Janzen(S7,692) points out that farming successes are highest in the lowland seasonal
tropics. The absence of annual density-independent control of insect populations
by weather in the nonseasonal or mildly seasonal tropics necessitates host plant
populations having a patchy spatial structure in which individuals or small
clumps of a species are scattered throughout the habitat in order to counter any
attack by herbivorous insects.(4LS7) With the notable exceptions of tropical cash
crops such as coffee, cacao, spices, and rubber, which to a large degree have
well developed chemical defenses against herbivores, dense plantings of food
crops in the nonseasonal or mildly seasonal tropics provide the opportunity for
herbivorous insects to build up large populations. Food crops are characterized
by palatable foliage or fruit, and palatability in the eyes of humans may also be
palatability in the eyes of herbivorous insects. (42) In such environments herbi-
vores are believed to be essentially free from population regulation by carnivor-
ous predators and parasitoids, since these insects are not adapted to feeding on
high densities of anyone kind of insect species.(57)
The question of whether species richness of carnivorous predators and
parasitoids of herbivorous insects is high in most tropical terrestrial habitats is an
open one at this time. Janzen et al.(412) have shown that parasitic Hymenoptera are
more abundant (numbers of individuals) and more diverse (numbers of species)
at a mid-elevation, mildly seasonal site in Costa Rica than at comparable low-
lands sites, and ascribes this pattern to a greater biomass and richness of her-
bivorous insects being associated with an increased harvestable primary productivity
of this site. Lowland habitats have relatively lower harvestable productivity
and fewer herbivore species, and such conditions may select for increased
generalist behavior in carnivorous predators and parasitoids. Thus when tropical
forests are cleared and replaced by fields or monoculture food crop plantings,
a greater incidence of population outbreaks of herbivorous species is pre-
dicted, since (1) the food supply is increased per unit area of habitat, (2) the
host plant in question may have few or no effective chemical defenses, and (3)
the carnivorous predator-parasitoid complex may be less effective in controlling
such populations.
The behavior, feeding physiology, and life history characteristics of car-
nivorous predators and parasitoids might be adapted to the exploitation of many
kinds of prey or low densities of anyone kind of prey, so that considerable time
might be needed to reestablish an equilibrium between abundance of herbivores
and such biotic control agents in disturbed or agricultural habitats dominated by
monocultures or simple polycultures of palatable crop species. Although some
tropical crop species have vesicular-arbuscular mycorrhizal associations with
390 CHAPTER 9

their root systems, which augments growth rates,(693) such systems and other
factors such as fertilizer application may do relatively little to offset the losses in
fitness due to intense herbivore attack.(57) Janzen(57) discusses various aspects of
the inability of monoculture plantings of palatable crops in the nonseasonal or
mildly seasonal tropics to protect themselves against herbivorous insects. Selec-
tive breeding within crop species reduces the spectrum of variation in chemical
defenses that normally occurs in wild breeding populations of the species, reduc-
ing its defenses against herbivores. (57) Pesticide applications in the tropics may
kill off what biotic control agents there are to exploit targeted "pest" species,
thus only intensifying the abundance of the herbivores.
Janzen(692) has delineated some of the major features of why it is difficult to
grow food crops in the tropics, and the reader is referred to that discussion.
Carroll(694) has pointed out that tropical agroecosystems may be very complex in
terms of the kinds of interactions among plants, insects, and microorganisms,
and predicting levels of harvest yields requires a knowledge of these components
and how they fit together. Such considerations indicate a need for long-term, in-
depth studies on the details of agroecosystems. Carroll,(694) in his examination of
the corn-bean-squash agroecosystem, points out that harvest yield will be a func-
tion of subtle interactions of predator-prey and host-pathogen competition, and
mutualisms among the organisms in such systems. Although considerable con-
troversy exists as to the points of origin for several food crop species in the New
World tropiCS,(695) apparently cultivation of such plants is very ancient, suggest-
ing considerable opportunity for an evolutionary buildup of complex com-
munities of organisms, including insects, centered around these plants as re-
sources. (694)
The conversion of tropical forests to pastures and open savannas involves
the erosion of the soil,(696) setting the stage for colonization by relatively few
species such as some grasses. Tropical grasslands in tum may have a very high
harvestable productivity promoting large biomasses of various taxa of herbivor-
ous insects.(697) Such communities can be very stable even though less diverse
than other communities in tropical habitats. The clearing of lowland tropical dry
forests in Central America to establish cattle pastures has provided a new envi-
ronment for the diversification of swollen-thorn Acacia species and establish-
ment of continuous popUlations of these plants. (372) Such changes in the geo-
graphical distribution of these plants result in changes in the community or guild
of insects feeding on them, not to mention the distribution of Pseudomyrmex
ants. In some Temperate Zone agricultural habitats dominated by a single crop
species, increased densities of the plant result in a lowering of the number of
herbivorous insect species associated with the crop. (698) Dense plantings are
characterized by a greater number of herbivore insect species and few carnivor-
ous predator species, and similar patterns are expected for food crop
monocultures in the tropics. (57)
INSECT SPECIES IN AGRICULTURAL HABITATS 391

The establishment of successional habitats in the lowland tropics in Africa

has provided the opportunity for ecological diversification in wood- and debris-
feeding insects such as termites. In some regions, termites form a major compo-
nent of the insect biomass in agricultural and secondary successional habitats and
in which they increase plant productivity by enriching the soils with nutri-
ents. (699.700) Mielke(70o) points out that indigenous peoples of termite zones in
Africa have coexisted well with these insects, although approaches to modem
agricultural technology include programs aimed at eradication of termites.
Again, as with Janzen's discussion/ 57 ) there appears to be a conflict with attempt-
ing to convert tropical habitats into massive food-generating machines for the
world. In Africa, the extensive clearing of tropical forests and the resulting
establishment of mixed forest and savanna ecosystems has resulted in new as-
semblages of insects such as butterflies exploiting these new environments. (701)
In fact, it is suggested that some butterflies appear to be better suited for the
exploitation of resources in such ecosystems than they were in their original
environments,(701) but such conclusions are difficult to confirm without an
analysis of the whole communities involved. Weedy gardens in the tropics are
believed to be suitable habitats for many butterfly species, although such open
systems of "farming" butterflies do not necessarily take into account the inter-
play of genetic structure of populations in relation to optimal distributions of
resources and effects of mortality factors. Clearly we need more data on the
long-range effects of such open systems on the maintenance of butterfly diversity
in selected habitats.
The absence of severe annual changes in weather conditions in many re-
gions, the ability for plant species to accumulate in certain kinds of tropical soils,
and favorable, long growing seasons together provide a huge obstacle for
monoculture and simple polyculture for tropical food crops, when viewed in
terms of herbivorous insects exploiting wild and introduced plant species. On a
different level, material from earlier chapters has made us realize that tropical
pollination systems in natural habitats have evolved in response to a variety of
selection pressures influencing the spatial distribution of a plant population,
optimal reproductive periods, and availability of pollinators. Much of that dis-
cussion focused on the often specialized coevolved associations between plants
and pollinators, even though some pollinating insects such as stingless bees may
be fairly opportunistic in foraging tactics. As a result of selective inbreeding and
agronomic requirements centered around harvesting, the pollination systems of
tropical crops are often specialized and the challenge becomes the need to find
the optimal pollinating insect.(702) JanzenC42 ) believes that animals pollinating
monocultures of plants in the tropics are drawn from the communities of pol-
linators associated with adjacent natural forests, and since the crop cannot move
away from the site, it depends upon some portion of these communities for
reproduction. The matter becomes critical if the object of harvesting is fruit or seed.
 392 CHAPTER 9

Many current and past field studies of insect pollination in the tropics have
emphasized the often great specificity between pollinators and plant species in
natural forest ecosystems.<334l On the other hand, some flowers attract a variety
of opportunistic insect visitors such as stingless bees. A good example of our
present lack of information on the pollination systems of some rain forest under-

TABLE 9.1
Amount of Pollen on the Bodies of Insects Captured on Pigeon Pea Flowers
in India a

With (+) Mean

and Number Mean number of percentage ~
without (-) of insects Mean number of pigeon pea pigeon pel
Species pollen loads examined pollen grains pollen grains pollen

Hymenoptera
Apis dorsata Fab. + 11 3,045 3,045 100
2 1,250 1,250 100
Apis florea Fab. + 2 250 250 100
Megachile bicolor (Fab.) + 9 82,334 81,056 98
4 2,875 1,500 69
Megachile lanata (Fab.) + 9 107,333 47,389 75
3 1,334 1,167 67
Megachile conjuncta Smith + 6 51,500 41,417 86
0 0
Megachile spp. + 12 39,666 39,000 94
8 7,875 7,000 83
Anthophora sp. 16,500 16,500 100
Xylocopa spp. + 2 50,500 49,500 98
3 2,000 1,667 91
Polistes sp. 11 1,090 1,045 99
Delta sp. 2 2,000 50 27
Rhynchium brunnem Fab. 3 1,500 1,333 92
Diptera
Musca lusoria Wied. 6 1,500 1,250 88
Carcelia sp. 0 0
Lepidoptera
Catochrysops cnejus Fab. 15 534 467 87
Lampides boeticus L. 15 167 100 93
Syntomis cyssea Stoll. 1 500 500 100
Terias hecabe 3 833 833 100
Homoptera
Clavigrella horrens D. 2 2,000 1,750 90
Dysdercus sp. 0 0

"From Williams. 17· "

INSECT SPECIES IN AGRICULTURAL HABITATS 393

TABLE 9.2
Foraging Behavior of Different Species of Bees on "Normal" Pigeon
Pea Flowers in Indian

Species

Megachile Xylocopa
A. dorsata A. florea spp. spp.

Number of bees observed 26 61 81 20

Percentage of bees with pollen loads 92 67 78 50
Mean number of flowers visited per plant 1.9 1.8 2.7
Mean number of plants visited per row 1.4 1.4 1.3
Mean number of plants moved per plant change 1.2 1.8 1.1
Mean number of rows moved per row change 1.4 1.4 3.5
Time spent per flower visit( s) 14.7 25.6 3.8 3.6

"From Williams."·"

story plants is illustrated by a recent studies of Soderstrom and Calderod703 ) that

reported discovering two species of phorid and two species of cecidomyiid flies
that were the possible pollinators of certain grasses in Brazil, Colombia, Panama,
and Venezuela. Their data require a reexamination of traditional views of pollina-
tion systems in tropical grasses and other plant groups as well. When a plant
species is grown in monoculture, it may attract a wide range of insect visitors,
simply because it provides a concentrated and large resource adjacent to a
mixed-forest environment. Depending upon the floral design and physiology of
that particular plant species, some degree of specificity displayed will be in terms
of which insect visitors actually pollinate the flowers. For example, Williams(704)
censused the great variety of insects visiting the flowers of the crop species
pigeon pea, C. cajan, in India and discovered that although the flowers attract
many different insect taxa, a smaller portion bear pollen grains of this plant
species, and an even smaller percentage of these insects, mostly bees such as
Megachile, actually cross-pollinate the flowers (Table 9.1). Williams's data
illustrate the high specificity that can exist for cross-pollinating insects in a
tropical food crop species, even though the flowers attract a variety of insects.
Interestingly, Williams noted an abundance of thrips on the flowers of this
species and cites other reports in which thrips were found to have exceedingly
high loads of pigeon pea pollen on their bodies. Yet the possible role of thrips in
the pollination of pigeon pea remains obscure. Williams also demonstrated that
some species of Megachile bees discriminate between cultivars and wild forms
of pigeon pea, which differ in flower structure (Table 9.2).
394 CHAPTER 9

9.2. INSECT POLLINATION IN AN AGRICULTURAL HABITAT IN THE

TROPICS: THE CASE OF CACAO

In previous discussion of tropical seasonality and pollination systems (Chap-

ters 7 and 8), there was a discussion of cacao pollination. Cacao as a cultivar in
both the New and Old World tropics (although the origin is New World) falls into
the class of very low fruit set compared with the proclivity of the mature trees to
produce flowers. General reviews of the biology of cultivated T. cacao are given
in Cuatrecasas(S9S) and Purseglove.(644) The emphasis in previous chapters con-
cerning cacao has been the possible role of pollination as a limiting factor in fruit
set (Fig. 8.4) and the effects of strongly seasonal patterns of rainfall on asyn-
chrony between flowering and peak activity of pollinating midges. It was
stressed that cacao in its original wild state as an understory tree of lowland
tropical rain forests in South America had a patchy spatial distribution along
streams with several trees to a patch and that an equilibrium was established
between abundances of flowers and pollinating insects through time. Such an
equilibrium is not unexpected in natural tropical forest ecosystems, as illustrated
by recent studies on the pollinator specificity and abundance in relation to fruit
set in large tropical trees. (117)
From a consideration of recent thought and data on the evolution of breeding
systems in tropical trees,<638) it would appear that cacao, in its wild state, opted
through evolutionary time for low fruit set relative to flower production, with
individual trees being physiologically limited for low fruit set due to other
phenotypic constraints affecting overall fitness in the rain forest environment.
The breeding system of the tree was a self-incompatible system in which tiny
certatopogonid midges and possibly other insects exchanged pollen between
individuals within a clump, perhaps between nearby clumps in the forest under-
story. According to one view, (394) the cauliflorous condition of cacao and other
tree species of tropical rain forest understories is believed to be due to physiolog-
ical constraints. Although cacao as a cultivar experiences considerable damage
from herbivorous insects and pathogens in some regions,(70S) the chemical de-
fense systems of the foliage of this tree species are generally known to be well
developed and probably very functional in the original habitat.
Although such a species, therefore, stands a good chance for agronomic
success in monoculture or simple polyculture (with crop shade trees such as
Hevea rubber), Old and New World cacao plantations are plagued today with
diseases and defoliation generated by insects, the result of colonization by insects
from surrounding forest habitats. (705) Overall, cacao has been reasonably success-
ful as a major cash crop grown in simple polyculture. Efforts to increase fruit set
in cacao plantations should include the detailed study of the breeding systems in
both plantation and natural habitats for this tree species. This effort is necessary
to determine whether some proportion of the trees in an area are functionally
INSECT SPECIES IN AGRICULTURAL HABITATS 395

andromonoecious, being pollen donors for other trees. Relatively few data exist
on the intertree variation in fruit set in cacao in either wild or cultivated cacao
stands, so that we have little information about the breeding system of the tree
species. Once the breeding system is determined, including the disclosure of
whether or not there is functional andromonoecy (see Chapter 8 for a discussion
of this topic for cacao), meaningful statements about the effects of cultivation on
the pollination system of cacao will be possible.
Pollinator populations may be limited in cacao plantations by the low abun-
dance of suitable breeding sites for the midges (see the review of Winder<588».
This argument is based on the observation that experimentally-placed breeding
substrates of these insects results in a patchy but rapid increase in the larval
populations of the midges.(S90) We have few data on the life tables of pol-
linating midges, although the high densities of predatory arthropods and Anolis
lizards in some cacao plantations(47S) and the observation that salticid spiders
successfully capture adult midges on cacao trees (A. M. Young, unpublished
data), indicate that various biotic mortality agents may regulate midge popula-
tions. The courtship and mating behavior of these midges (several genera
and species; see Table 9.3) is probably not characterized by large mating
swarms. Thus adult population densities are suspected of being low, and adults
are not abundant at cacao flowers.(652) Young(S90) noted in a preliminary way that
patches of breeding substrates with ant colonies generally lack ceratopogonid
midge larvae (at least later instars), while patches with high densities of midge
larvae lack ants.
Leston<476) has described the "ant mosaic" associated with cacao planta-
tions in Ghana. Ants and other predatory arthropods readily invade cacao planta-
tions from adjacent mixed-species forests, as these polyculture agricultural
habitats contain both food supplies and nesting sites. The rate of invasion of
cacao plantations by ants, for example, will be a function of (1) resident ant
species in the surrounding forest and the degree to which these species are in
competition for resources with one another; (2) the spectrum of intercolony
variation within a species in the surrounding habitat, so that some colonies may
be more prone to move into agricultural habitats than others(706); and (3) the
suitability of the agricultural habitat in terms of prey, other resources, and nest-
ing sites. In this regard, a significant factor in the agricultural habitat is the
amount of shade cover and the degree of shade tolerance in potential invading ant
species. Similar arguments can be generated for the communities of other ar-
thropods, and their associated diseases in the case of vectors, that may invade
from the surrounding habitat. Leston(705) showed for cacao trees interspersed with
wild shade trees in Ghana that the crop canopy insect fauna is a subset of the
fauna associated with the forest canopy, having the same dominant ant species
and some fugitive species of ants from surrounding savanna habitat. Leston
suggested that changes in the light regime are the primary factor resulting in
 TABLE 9.3
The Distribution and Abundance of Various Genera and Species of Ceratopogonid Midges in Two Cacao Plantations
(La Tigra and EI Uno) in Northeastern Costa Rica a

Number of Individuals

Natural
ground Box Simulated Banana
Plantation Genera Species Flowers Foliage litter" litter bromeliads stems Total

La Tigra Forcipomyia n. sp. nr. argenteola Macfie 3 3 33 146 213

Forcipomyia sexvittata Wirth 7 10 123 140
Forcipomyia cinctipes group 3 5 7 20 35
Forcipomyia species-l 3 4 8
Forcipomyia species-2 2 3
Forcipomyia species-3 5 7
Forcipomyia species-4 4 10 14
Forcipomyia new species 20 20
Forcipomyia species-5 2 4 6
Dasyhelea borgmeieri Wirth 3 5
& Waugh
Dasyhelea williamsi Wirth 6 6
& Waugh
Dasyhelea species-l I 2 3
Atrichopogon species-l 2 2 6
JII ••• :_I~ _____ ~
... ......... : ...... .,
- 7
 Stilobezzia species-2 ,t.

Fittkauhelea amazonica Wirth 4 6

& Blanton
Culicoides hylas Macfie 7 12
Culicoides glabellus Wirth 3
& Blanton
Culicoides diabolicus Hoffman
Total individuals: 14 o 368 519
% of total sample: 2.7% 0% 70.8%
EI Uno Forcipomyia genualis (Lw.) I 18 43
Forcipomyia cinctipes group 2 II
Forcipomyia species-I 2
Forcipomyia species-2 7
Dasyhelea borgmeieri Wirth
& Waugh
Culicoides hylas Macfie 2 8
Culicoides foxi Ortiz 3
Total individuals: 3 o 20 81
% of total sample: 3.7% 24.6%

"From A. M. Young.cs"'a,
• Specimens were obtained by collecting from cocoa flowers and foliage, and sorting tbrough rotten organic substrates (August 1978-October 1979). Approximately 15% of total field
time was devoted to examining flowers, another 10% to sweeping the foliage of cacao trees, and 60% to examining samples of rotten organic substrates (ground litter, litter in simulated
bromeliads, and banana stems).
'Controls.
398 CHAP;rER 9

shifts in the insect fauna in such agricultural habitats. Forest degradation is

followed by an almost immediate reduction in herbivorous Coleoptera, which
can no longer maintain breeding populations on a scattered food base, and
complete removal of forest results in the establishment of savanna insect faunas
associated with cacao. Such degraded forests are rapidly colonized by herbivor-
ous forms such as Lepidoptera, then followed by carnivorous predators. The
overall species richness of degraded forests is well below that of natural
forests. (705)
Janzen<42) believes that monoculture forest habitats, presumably naturally
occurring ones, draw their pollinators from surrounding mixed-species forests,
and this may also be true for agricultural habitats. The ceratopogonid midges in
general require moist conditions for breeding, and adults (Fig. 8.5), owing to
their small body size, also position themselves within the moist regions of verti-
cal and horizontal gradients in air humidity in different kinds of habitats. Little is
known about the complete feeding habits and nutritive requirements of adults for
genera and species assumed or proven to be pollinators of cacao.(588.589) The
studies of Downes on Nearctic ceratopogonids clearly point to specialized forms
of pollen-feeding in adult females of some genera and species. (245) Yet the niche
breadth with respect to pollen types exploited within a habitat is not known for
any species anywhere. Richards(394) alludes to the pollination of some tropical
rain forest understory trees and shrubs by nonspecialized dipterans among other
nonhymenopterous insects, yet in-depth studies of the tenacity for certain forms
to pollinate only certain tree species, and the mechanisms underlying the interac-
tion, are not known for most understory trees, although there is considerable
interest in this area. (42) The effectiveness of anyone species of ceratopogonid
midge for pollinating cacao will be, in part, a function of the physiological niche
breadth for food resource types and the realized or ecological niche breadth at
anyone time and within a particular natural or agricultural habitat. A common
shade tree for cacao is Hevea rubber, whose flowers are also pollinated by these
midges. The relationships between the pollinating midge communities associated
with the overs tory of rubber and the understory of cacao in such a simple
polyculture have not been studied.
In the case of the well-known cultivar Vanilla jragrans (Orchidaceae),
individual vines have a significantly higher growth rate in plantations on Puerto
Rico where the overstory consists of wild trees left there when the land was
underbrushed than in similar plantations where all of the vegetation was cleared
and shade trees were replanted. (707) Puerto Rican vanilleries utilize legume trees
such as Erythrina and Inga to provide the shade necessary for suitable growth of
the vanilla vines, but because some species of planted shade trees are very
susceptible to varieties of root rot, highest yields of beans are obtained with
natural shade cover left for shading vanilla vines. (708) In a manner similar to that
discussed by Leston(70S) for optimal cacao production in Ghana, retention of a
INSECT SPECIES IN AGRICULTURAL HABITATS 399

natural forest overstory when planting vanilla or cacao functions to keep down
herbivorous insect populations and other organisms that may otherwise reduce an
introduced shade cover to a leafless condition very rapidly, perhaps attacking the
crop plants themselves. Natural shade cover may also promote the activity of
pollinating insects such as bees by preserving a portion of the habitat required for
nesting, optimal thermoregulatory conditions, and retention of landmarks used in
marking home ranges. The amount and kind of shade cover over cacao in Ghana
has been shown to be a major factor in determining the intensity of attack by
capsids on cacao trees.(70S) When cash crop species are relatively inedible to
many herbivores, and when the shade is provided by a natural overstory, the least
chance appears to exist for serious outbreaks of pest insects on the crops.
Planted shade trees may not respond in the same manner to seasonal cycles in
rainfall as do species of the natural forest and therefore may go deciduous and
injure the crop population through increased desiccation. (707) This effect may
increase the susceptibility of individual trees or vines to attack by certain groups
of insects.
Even if cacao trees tum out to utilize a breeding system characterized by
functional andromonoecy, one must ask questions concerning the adaptive role
of such a system within the context of the natural habitat of this rain forest tree
species. The answers to such questions will not be possible without in-depth
knowledge of the population dynamics of the pollinating insects and their natural
history. If the average cacao plantation does not provide adequate abundance of
suitable resources and physical conditions required for optimal population
growth of these pollinators throughout most of the year, and if replenishing
pollinator populations annually from adjacent forest habitats is not possible, fruit
set in cacao may be lowered even further for individual trees functioning as
pollen acceptors under the assumption of functional andromonoecy or other
breeding systems. If conditions within plantations can be demonstrated to limit
midge populations and if fruit set is lowered because of such effects, questions
can then be asked about the optimal design of cacao plantations as a means of
increasing harvestable productivity (beans) without necessarily increasing the
area of terrain occupied by plantations. In most existing plantations, flowers
losses are very high, as indicated in a recent census of flower drop in a Costa
Rica plantation (Table 9.4). Flower drop in T. cacao refers to the loss of buds or
open flowers on healthy trees, generally the result of pollination failure or other
factors (e.g., removal by Atta and Ectatomma ants). Two different plantations
within the same region but differing in terms of complexity of shade cover
exhibit different phenologies of flowering throughout the year (Fig. 9.1). The El
Uno plantation has a monoculture canopy of Hevea rubber that goes partially
deciduous during the short dry season of this region, causing an increase in the
amount of sunlight hitting the cacao trees below, thereby resulting in a burst of
flowering at the beginning of the rainy season (Fig. 9.1). The more complex
 TABLE 9.4
Some Estimates of Flower Drop for Cacao Trees in the La Tigra Plantation in Northeastern Costa Rica a•b

Number of marked
Initial positions Number of new
Tree Number of buds (pins) present pods Number of flowers! Number of empty Flower drop
no. markedb later" presentd buds present positions (loss) (%)

320 310 3 15 282 About 95%

2 560 534 10 523 About 99%
3 285 280 5 23 252 About 97%
4 483 476 9 17 450 About 95%

aFrom Young.(S908)
"The position of individual buds on each of four healthy trees was marked by gently inserting a colored pin next to each bud and recording of pins used per tree; pins were inserted finnly
to minimize their loss by falling out. but none were inserted at the base of a bud so as to minimize injury to the bud cushion.
cThe initial marking was done on 23 March 1979, and the census of pins, and presence or absence of new pods or new flowers, was taken on 27 June 1979. Slight losses in pins, through
falling out, etc., were expected and found.
"New pods present included healthy pods and aborted pods present at locations of pins on each of the four trees censused. Trees censused were similar in size and healthiness, and all
were located within 10 m of the others.
INSECT SPECIES IN AGRICULTURAL HABITATS 401

natural canopy of the La Tigra plantation, consisting of at least eight species of

forest trees, sustains considerable flowering throughout the year because the trees
are patchy in distribution and exhibit differential effects to dry periods (Fig. 9.1).
Pronounced differences in flowering patterns between different cacao plantations
in the same general climatic region of the wet tropics differentially affect insect
pollinator populations and fruit set.
Within small samples of mature and flowering trees in each plantation, peak
fruit set occurs in the first month of the rainy season, slightly sooner in the La
Tigra plantation, presumably because flower production is already high at this
time while lagging in EI Uno (Fig. 9.2). Increased fruit set at this time may be
due to increased abundance of pollinating midges in both plantations resulting
from wetter conditions favoring their breeding and survival. The effects of sea-
sonality discussed here are minimal in the sense that these plantations occur in a
region of premontane tropical rain forest where the dry season each year is
relatively short and erratic in length and severity. Severe tropical dry seasons
(e.g., in Ghana; see Chapter 7) may reduce midge populations so drastically that
there is considerable asynchrony between pollinators and flowers at the begin-
ning of the rainy season.
The two plantations are strikingly different in terms of the laundry list of
pollinating midge genera and species, and numbers of individuals, collected in
each (Table 9.3). The greater species richness and abundance of midges in the La
Tigra plantation is suggested(590) to be the result of more suitable resources and
increased habitat complexity (litter substrates). The existing mosaic of different
kinds of rotting organic debris present in a cacao plantation will have a profound
effect on the species of ceratopogonid midges present since different groups
exhibit different preferences for breeding sites.(653) Young(590a) conducted a pre-
liminary field study in which sets of known breeding substrates of these midges
(see WindetS89 ) for a review) were distributed in both plantations in equal num-
bers and with similar patterns of distribution. After remaining undisturbed and
being allowed to decay for several months, the substrates were censused for
larvae (only larger instars censused) and pupae of midges. The details are sum-
marized in Young,(590a) but it is interesting to note here that although the results
are tentative, midge abundance (for a few species only) is increased significantly
in rotting slices of banana trunk in the La Tigra plantation but not for any
treatment in the EI Uno plantation (Table 9.5). Furthermore, mature cacao trees
adjacent to patches of experimentally placed banana trunk slices have increased
fruit set (Table 9.6), although once again these findings are at best tentative
owing to small sample sizes. Young(590a) suggests that ecologically more com-
plex cacao plantations may support a greater species richness of pollinating
midges and the abundance of some species may be limited by available breeding
sites, perhaps as a result of competition among species. To what extent midge
populations are kept down due to predation and parasitism is not known at
this time.
 720

E 630
.§
co 540
.S!
i'u 450
i!!
Go 360
'"
:E
g 270
::.
180

90

1800

1600
/\,
1400 EL UNO PLANTATION ! \\
, ,
1200 ,,: \,
,
,/ ~, ,,
1000 ; \
:
,,
~,
,
800
!
,,
\.,,
.. 600

I 400
/
,/ \
\
o~ 200
"'\'" _________ /) V
....
-g
O~~~==~==E=~-L-L~--L-i--L-L~
-g
~ 4000
o
:0 LA TIGRA PLANTATION
~ 3500

Z" 3000
,\ /\
2500 /
",,
\,
\J
// \\\' ! \""""
2000 /
..............

1500

1000

500

oL--L~ __~~-L__L-~-L~__~-L~__L-~
ASONDJFMAM AS
1978 1979
Successive Months

FIGURE 9,1. Monthly phenological patterns of rainfall and production of flowers (buds and open
flowers) in an open cacao plantation (La Tigra) and shady plantation (Ei Uno) in Costa Rica. Broken
lines represent buds, and solid lines represent open flowers. The same number (20) of marked trees
were censused in each plantation to generate these data. [From A. M. Young. 1. Appl. Ecol. 19:46-
67 (1982). Copyright 1982 by the British Ecological Society.J
INSECT SPECIES IN AGRICULTURAL HABITATS 403

It has also been observed that predominantly one species of stingless bee,
Trigona jaty, collects large quantities of pollen from cacao flowers but does not
contribute to pollination and fruit set.(40S) In fact, the bee might be robbing pollen
that would otherwise be used both as a reward (food) and source of pollination by
ceratopogonid midges. The extent to which such bees take pollen from cacao
trees will be in part a function of the proximity of other suitable pollen supplies in

100
EL UNO COCOAJ=J~RI\I1
----.,,
I
I
90 , NO.trees: 28
/'
I
I
Total no. small pods \

,
/
80
I
,
,
I
I including aborted: 752
" " ,,
I
I
\ --- Healthy small pods ,,
II I

,,
70 I , I
" and aborted pods ,-----, I

60 \ - Healthy small pods ,, \I

I
,, \I
50 I

"
"\, /"
I

40 I I
I

30 \ /
I
I
/
III
1::1
0
20
\ I
,I
",.,'....... "
I
/l. I I

" ,,,
I
tV I
0 I "",//
(J 10 '/
0
U
"0 0
t 100 ~AIlCiF!AJ:OCQA FARM
.c
E 90 No. trees: 28
:::J
z Total no. small pods
80 including aborted: 767

70

60

50 ',
I,

,I ,,
I I

"
40 \,
,,
30 ,/
, \
,
,, ',
20 'v'

10
A S 0 N 0 J F M A M J J A S
1978 1979
Successive Months

FIGURE 9.2. Phenological patterns of fruit production in two cacao plantations within the same
general climatic zone of Costa Rica. Broken lines represent healthy small pods and aborted pods, and
solid lines represent healthy small pods. Overall production of small and aborted fruit (pods) is
strikingly similar for these samples between the two plantations, although a seasonal pattern is
somewhat more evident in the exposed plantation (La Tigra). [From A. M. Young, J. Appl. Ecol.
19:46-47 (1982). Copyright 1982 by the British Ecological Society.]
404 CHAPTER 9

the surrounding habitat. For example, in another cacao plantation in Costa Rica,
strips of blooming Schistocarpha oppositifolia (Compositae) attract large num-
bers of T. jaty but the bees do not visit, to any appreciable extent, cacao flowers
adjacent to these patches of weedy species. A critical bit of information needed to
determine whether pollen-robbing by such bees in cacao groves contributes to
losses in fruit set is the determination of ecological fruit set versus physiological
fruit set capacities of individual trees. While the studies described above are
preliminary, they do suggest that the degree of successful pollination in a cash
tropical crop such as cacao depends in part upon (1) the floristic complexity of
the shade cover and indirect effects of this on insect popUlations and (2) the
effects of surrounding habitats on the influx of insects influencing pollination
within the agricultrual habitat.

TABLE 9.5
The Distribution and Abundance of Forcipomyia Midges (Larvae and
Pupae) in Various Treatments of Arboreal and Ground Litter Breeding
Substrates in the La Tigra and EI Uno Cacao Plantations in
Northeastern Costa Rica a

Total volume Total number of

of substrate Forcipomyia Number of Forcipomyia
sampled on larvae + pupae per unit volume,
Treatments each census· in all censuses all censuses

La Tigra
Natural cacao leaf
litter (control)C 1.20 m3 51 About 11 midges/m3
Thickened ground cacao
leaf litter<l 2.00 m3 57 About 7 midges/m3
Arboreal cups of litter" 1.52 m3 0 0
Rotting banana stem slices! 0.14 m3 334 About 119 midges/m3
El Uno
Natural cacao leaf
litter (control)" 1.20 m3 16 About 3 midges/m3
Thickened ground cacao
leaf litter' 2.00 m3 25 About 3 midges/m3
Arboreal cups of litter" 1.52 m3 0 0
Rotting banana stem slices 0.14 m3 20 About 7 midges/m3

a From Young. (590a)

"There were four censuses of natural litter, thickened litter, and arboreal cup litter, and two of banana litter.
'Samples took 10 collections of leaf litter, each 2 m x 2 m x 3 em.
dSamples were obtained from square wooden frames, 2 m X 2 m X 20 em deep, replicated 10 times, and
placed beneath cacao trees; all replicates were then censused at each census and the litter replaced. The litter
was replaced after each census.
e Twenty plastic cups were placed in 18 cacao trees and filled with rotting cacao leaves, rotting banana stem
fragments, or rotting cacao pods. The litter was replaced after each census.
fO ne hundred ninety-six slices of banana stems were distributed among seven plots for each experiment.
INSECT SPECIES IN AGRICULTURAL HABITATS 405

TABLE 9.6
The Effects of Different Rotten Organic Substrates of Pollinating Midges
on Fruit Set in Cacao Trees in Two Plantations (La Tigra and EI Uno)U

Number of trees Total number of new Mean (± S.D.)

Treatment examined fruits present number of new fruits

La Tigra
Rotten banana stems 14 61 4.4 ± 3.9
Control 14 18 1.3 ± 1.6
Box litter 20 24 1.1 ± 0.8
Simulated bromeliads 26 33 1.7 ± 1.4
Control plots of
natural litter 24 27 1.4± 1.2
El Uno
Rotten banana stems 14 18 1.8 ± 1.4
Control 14 22 1.6 ± 1.3
Box litter 20 19 1.5 ± 1.4
Simulated bromeliads 24 20 1.3± 1.4
Control plots of
natural litter 24 24 1.5 ± 1.2

aFrom A. M. Young. (59Oa)

'Controls were deliberately increased to maximize taking into account any effects of inter-tree variation in fruit
set. For a two-tailed test with 95% confidence limits, the only significant difference was between the mean
fruit set of trees in plots of rotten banana stems and the two groups of control trees in La Tigra (t = 2.69,
p = 0.001; t = 3.86, p = 0.001). All trees were censused in September 1979.

9.3. GRAZING AND RELATED ACTIVITiES BY INSECTS IN

AGRICULTURAL HABITATS IN THE TROPICS

In Temperate Zone monoculture fields, periodic grazing is essential for the

maintenance of stable populations of herbivorous Coleoptera such as
chrysomelids, because the maturation of host plants is accompanied by a reduc-
tion in foliage quality as food for these insects as the growing season pro-
gresses. (709) Grazing stimulates the growth of new meristem tissues, a high-
quality resource for these herbivores_ In strongly seasonal tropical areas, heavily
grazed pastures or mowed habitats may also enhance the maintenance of insect
popUlations associated with the plants and, to some degree, the carnivorous
predators and parasitoids exploiting them. In the mildly seasonal or nonseasonal
wet tropics grazing, effects may be more dispersed throughout the year. In some
Temperate Zone guilds of closely related species of leaf hoppers associated with
Kentucky bluegrass, effects of mowing on population structure within a species
depend upon whether the species is a fugitive; different species exhibit different
responses_ (710)
406 CHAPTER 9

The studies of Scriber and FeenyC477) with selected Temperate Zone lepid-
dopterans illustrate the selective force of larval feeding efficiency in the regula-
tion of host plant selection in herbivorous insects in gener,al. Host plant choice
will be a function of the ease with which the insect can extract the nutrients from
the target tissues once the plant is located in the habitat. Ease of extraction of
harvestable nutrients will be a function of, among other things, physical tough-
ness of the tissues as a challenge to the feeding process, the kinds and concen-
trations of defensive compounds in the tissues, the water content of the tissues,
and the kinds and concentrations of nutrients in the tissues. As an individual plant
grows, all of these factors are expected to change to some degree, and the initial
condition of the plant and its temporal changes through the growing season
determine the quality of the plant as a food base to insects; quality is expected to
be different for different insects feeding on the plant. From the viewpoint of the
insect species, suitable or optimal quality is interpreted as the ease with which the
harvested plant material can be converted into new insects, either directly
(mobilization of nutrients for production of eggs, courtship traits, etc.) or indi-
rectly (use of nutrients to provide energy for avoiding carnivorous predators and
parasitoids, for locating suitable shelters, etc.). In Neotropical butterflies such as
Morpho and Parides it is not uncommon to observe gravid females placing
viable eggs on exceedingly small seedlings of host plants in natural habitats (A.
M. Young, unpublished observation),(311J host plant individuals too small to
allow complete development of the larvae. Apparently such plants are very
nutritive or nutrients are easily obtained from them, and the oviposition behavior
is correlated to allow exploitation of such plants. A critical piece of data lacking
is the proportion of such plants used by the average gravid female butterfly in a
particular habitat (compared with oviposition on mature or large plants).
Scriber and Feeny(477) observed that Rattus philenor larvae feeding on small
seedlings of Aristolochia in Texas pine forests actively move among individual
plants as they are depleted. Dispersal of this kind may also occur in other butter-
flies and is not necessarily related to density of insects on host plants. Rather,
dispersal oflarvae, usually low densities per plant in the Neotropical species studied,
is a response to decline in quality of individual host plants as they grow. The short
annual growing season in the temperate zones imposes a progressive change on
host plants in terms of the factors influencing suitability as food for herbivorous
insects. An analogous condition is predicted for the strongly seasonal lowland
tropics, where the annual growing is also shortened by the intervening dry season
(analogous in effect, for a large part of the plant communities, to the northern
winter). In the tropical rain forest regions, the growing season is extended
throughout the year for most plant species, and the depletion of suitable tissues
(i.e., new meristems) is expected to be slower than in seasonal regions. Not all
plant species will respond to monthly changes in rainfall patterns, and each
species will also have a breeding system necessitating some energetic tradeoffs
INSECT SPECIES IN AGRICULTURAL HABITATS 407

with vegetative growth tactics. The more-or-Iess continuous rainfall pattern of

many tropical rain forest localities ensures that new vegetative growth will occur,
even at minimal levels, throughout most of the year. In strongly seasonal re-
gions, moist forest refugia may continue to exhibit vegetative growth among
individual plant species even though surrounding vegetation has gone deciduous.
The former situation, mentioned in Chapter 7, provides a nucleus of food
supplies for some herbivorous insects during the long and severe dry season of
such regions.
In subtropical or mild temperate regions such as southeast Australia, annual
environmental conditions may regulate potential pest populations by a multi-
plicity of density-dependent and density-independent factors. Detailed studies of
the pasture scarab, Aphodius tasmaniae, (MaelzdS77,71l-713») provide an example
of how an insect becomes a pest species in monoculture or simply polyculture
agricultural habitats, with some implications for similar phenomena in the
tropics. This medium-sized beetle feeds in the larval stage on the root systems of
grasses and clover, and pest conditions have resulted from improved pastures
being established following European settlements in this region: improved pas-
tures mean the expansion of existing pasturelands and the addition of clover as a
food source for livestock. What has really happened is that not only has the food
base for A. tasmaniae been progressively increased in total size over the post set-
tlement years but the food quality for the insect has been improved: both this
species and an allied pest species, A. howitti, exhibit a preference for feeding on
clover, and growth rates of immature stages are faster on clover than on
grassesY12,714l Clover was introduced as a food plant for livestock and was
rapidly exploited by the beetles as well. High-density patches of the larvae of
both species impede the carrying capacity of pastures for livestock, and the
addition of fertilizers to the soil has been of great benefit to the beetles in
improving the quality of the food supply. Because the larvae are effective herbi-
vores of the root systems of pasture clover and grasses, the harvestable pro-
ductivity of pastures for livestock dwindles in areas where infestations are high.
Patches of high infestation tend to occur in the same areas in both beetle species
in successive years, perhaps as the result of unfavorable weather conditions
impeding oviposition-related flights of mated female beetles to colonize other
areas. The popUlations of both species are greatly influenced by prevailing
weather conditions, especially the overall climatic pattern of monthly amounts of
rainfall.
High rainfall during the summer months promotes the gro';"th of the insect
popUlations through increased growth of pasture host plants and sufficient
moisture to trigger flight activity in the beetles. A. howitti appears to have a
polymorphism underlying oviposition behavior; a large batch of eggs is deposited
by the average female before feeding on dung, and a second batch i~. deposited
after feeding on dung. Dung feeding is followed by considerable flight activity
408 CHAPTER 9

before the second and smaller batch of eggs is oviposited into cracks in the soil
around healthy clover plants. The presence of dung, however, is not an absolute
for survival as A. howitti can occur at high densities in ungrazed turf. It appears
that the introduction of grazing into the original habitat of the beetle, and the
development of dung heaps as a patchy but often abundant resource, has altered
the feeding and oviposition behavior of A. howitti and to some extent of A.
tasmaniae as well: fresh dung heaps may be nutritively rich resources for the
development of early larval instars and ovaries, if the adult females obtain nu-
trients such as nitrogen, which is then used in egg production in a manner some-
what analogous to the exploitation of pollen nitrogen in some Neotropical but-
terflies. (3 J)
The predilection of larvae for clover roots may also represent a more effi-
cient feeding strategy(477) in that nitrogen may be concentrated as a resource for
rapid mobilization to build larval and future ovariole proteins. Both beetles are
univoltine, and years of high abundance are years of heavy and steady summer
rains but light or no winter rains. Winter rains promote the development of
pathogenic fungal populations that kill off the larval stages. During a "good"
year, populations are distributed in a somewhat patchy manner, with the greatest
densities of larvae occurring in pastures about 2-3 years old and with dense
vegetation including clover. Areas of high and low densities are usually in-
terspersed, and although there is some evidence that intraspecific competition
among A. tasmaniae larvae occurs in high-density patches, there is little evi-
dence that such behavior generates a density-dependent mortality shadow for the
populations. But Came(714) shows that considerable density-dependent mortality
exists among larvae of A. howitti and that a variety of other biotic control agents
(i.e., fungal pathogens, predation by carabid beetle larvae, predation by birds
and thymnid wasps) and abiotic factors (desiccation and drowning) generate
density-independent mortality in the larval populations. For both species the
distribution of suitable oviposition sites is considered a limiting factor since
gravid females exhibit considerable site selection when ovipositing.
Pastures in the wet tropical lowlands may have high harvestable prod-
uctivity above ground and undoubtedly provide a food base for a variety of
pasture scarabs . Yet variations in annual abundance of such insects are predicted
to be low, and overall population densities of immature stages will also be low:
the imminently wet conditions of such regions probably kill off large numbers of
immature stages either directly through drowning or through promoting various
entomophagous pathogens in the soil. Spatial distributions of popUlations or
subpopulations for a species are expected to be very patchy due to differential
mortality and not necessarily due to quality of resources such as host plants. Even
though the introduction of livestock into wet tropical lowlands in recent years in
Central America may provide an expanded but limited food or oviposition base
for such insects, some studies indicate that cow pats are ephemeral and patchy
INSECT SPECIES IN AGRICULTURAL HABITATS 409

resources.(71S) Because this resource is very temporary the movement patterns of

individual beetles among a collection of cow pats is best described as stochas-
tiC,(71S) since the quality distribution attached to the resource is changing rapidly.
Some studies have also indicated that different kinds of habitats in the tropics
influence the colonization of dung and carrion by coprophagous scarab beetles
near Leticia, Amazonas, Colombia, in two different years. A significant discov-
ery was that the coprophagous scarab fauna of terra firma was considerably
richer, in terms of number of species and individuals, than that of flooded areas,
although deforested areas in the terra firma had fewer species. The species
associated with dung in deforested areas tends to be different from those of other
habitats where diversity is high. Because the scarab fauna associated with dung
in forest clearings is reduced, dung persists longer and provides a more stable
microhabitat for various other insects such as flies, enhancing the opportunities
for disease organisms such as helminths to build up and infect humans in such
areas.
In the strongly seasonal lowland tropics of areas such as Central America,
historically speaking, livestock agriculture has existed for much longer periods
than in wet areas and one might expect that dung-exploiting insects such as
scarab beetles might have more diverse communities, since such communities
have had more time to develop, and the density and overall abundance of live-
stock are greater. In the strongly seasonal tropics, where pasturelands prevail, the
ecological conditions underlying the dynamics of pasture-inhabiting herbivorous
insects such as pasture scarab beetles approaches those of the system described in
the studies of Maelzer and Carne(711-713) for southeastern Australia: the annual se-
vere and long dry season imposes a density-independent regulation on the popu-
lations of these insects if their host plants experience a reduction in quality as a
food base for these insects and the life stages of these insects experience massive
mortality from desiccation and other abiotic effects. Likewise, the often heavy
rains associated with the early phase of the wet season may kill off immature
stages for some forms. The relationship between seeds in cow pats and the kinds
of insects associated with them remains to be studied. The overall reduction in
the annual growing season in such areas, viewed from the standpoint of grasses
and other forage providing a food base for immature stages of scarabs in pas-
tures, will result in seasonal fluctuations in the population sizes and spatial
properties of popUlations for these insects. Shifting livestock distributions be-
tween seasons will also change the abundance and density of cow pats as a
resource for a variety of insects: cow pats may exhibit considerably different
compositions of nutritive substances, including partially digested or undigested
seeds, and thereby provide an index of seasonal changes in availability of seeds
and other resources derived from plant species in grazed habitats.
Although some detailed studies of Neotropical scarabs associated with dung
have been made,(SS8) as well as natural-history studies of individual species, there
410 CHAPTER 9

have been few published attempts to determine the role of these insects in
degrading cow pats in tropical ecosystems and to analyze cow pats as a source of
food for other insects. Tropical cow pats may turn out to be interesting monitors
of interactions between some herbivorous insects and seeds if such forms exploit
the seeds after they have passed through the digestive tracts of wild and domesti-
cated vertebrates. The tropical dung scarab Onthophagus gazella of Africa and
Asia has recently been introduced into the American southwest to control horn
and face flies breeding in dung.(7l7) These beetles are effective in burying large
quantities of fresh dung, thereby lowering the survival of the flies which breed in
it on the ground surface; these parasitic forms cannot then reach the surface and
infest livestock. At the same time, the beetles increase soil fertility through their
processing of the buried dung. Tropical coprophagous scarabs are effective agents
for reducing dung-breeding parasite of humans and livestock. The degree to
which anyone species is effective in doing so in its indigenous habitats will be a
function of the number of species occupying pastures.
The degree to which a host plant species for a phytophagous insect is
restricted in distribution will be a determining factor in the extent to which the
plant population becomes saturated with the herbivore population, assuming an
absence of effective biotic control agents on the insect population. If the insect is
good at dispersing, and if the plant population is relatively restricted in distribu-
tion, a high probability exists that the herbivore will attain outbreak conditions.
Such an effect will be pronounced if the herbivore is very host specific, or if
alternative host plant species are relatively scarce in the region. Island popula-
tions of host plant species are very susceptible to such effects since their distri-
butions are very restricted. At the same time, deliberate introductions of a biotic
control agent to limit the popUlations of the herbivore for a plant species targeted
for economic use will be effective, especially if the generation time of the control
agent is considerably less than that of the herbivore. Oryctes rhinoceros is a
scarab beetle pest of the coconut palm in Papua, New Guinea, and the larvae in
natural popUlations in Western Samoa are killed off in large numbers by a
baculovirus disease. Deliberate introduction of larvae infected with the disease
into Papua, New Guinea, where the beetle still has a relatively limited distribu-
tion, quickly led to effective control ofthe insect in some areas, presumably from
the spread of the virus into the resident breeding population from the introduced
propagules.(7lS) Newly ec10sed adult beetles have the virus, and these can then be
released into other areas of the island where the beetles are abundant.
Secondary successional habitats and agricultural habitats such as pastures
and monoculture plantations of cacao, bananas, and coffee may be stable envi-
ronments in the tropics in the sense that there is little turnover in the composition
of the vegetation over short periods of time. Harvestable productivity may also
be high in some secondary habitats and pastures, providing enough plant biomass
to maintain a high species richness of herbivorous insects. The patch structure of
INSECT SPECIES IN AGRICULTURAL HABITATS 411

simple polyculture agricultural habitats such as cacao shaded by rubber or an

assortment of wild trees or coffee shaded with legume trees such as Inga repre-
sents a predictable environment in terms of providing food and shelter for insect
species. Some forms of husbandry ensure that such resources are maintained in
these agricultural habitats.
Several genera and species of Neotropical cicadas undergo adult
emergences around the trunks of Inga trees in Costa Rican coffee plantations:
study plots with individual shade trees have far greater numbers of cicada nym-
phal skins than similar plots containing only coffee bushes (A. M. Young,(78iaJ;
see Young 037J for a list of these cicada species). A typical pattern is that of the
medium-sized cicada Fidicina semilata in the coffee farms at Naranjo in the
Cordillera Central northwest of San Jose in Costa Rica: nymphal skins are found
beneath large individuals of various species of Inga planted to shade coffee
bushes, and few or none are found around coffee bushes away from these trees
(Table 9.7). The trees are planted in even rows. The data also suggest that only a
small portion of the trees have nymphal skins associated with them, so that these
cicada populations are not saturating the agricultural habitat. This region of Costa
Rica possesses a landscape mosaic in which there are expansive coffee planta-
tions and small strips of natural forest along streams. The cicadas are probably
moving into the coffee habitats through a courtship-oviposition-nymphal feed-
ing association with the Inga trees, although the evidence is still incomplete.
There is a striking pattern of many Neotropical cicadas emerging at large legume
trees in both natural forest and agricultural habitats in different climatic zones of
the country, (37) suggesting a possible "primary" association of the nymphs with
the root crowns of these trees. This is not to say that nymphs do not feed solely
on the roots of these trees, since they undoubtedly encounter palatable root
systems associated with many different plants in natural forest habitats in particu-
lar. Use of the term primary in this context implies that (1) Neotropical cicadas
were originally associated with legume trees in the evolution of tropical forests
and what we are looking at is an archaic coevolved association, or (2) contempo-
rary selection favors such associations with legume trees because they provide an
efficient food source for developing cicadas. As cicadas have incredibly long life
cycles, presumably as the result of exploiting generally nutrient-poor plant saps,
selection may favor the exploitation of those root crowns that increase feeding
efficiency, even if just a small amount. Doing so reduces the age of first repro-
duction and maximizes fitness under individual selection.
White and Strehl(366J have emphasized that the long life cycles of cicadas are
mostly the result of nymphs' feeding on nutrient-poor xylem fluids; the rate of
growth is very low, and several years may be required for complete nymphal
development. The developmental time of any species of tropical cicada is not
known. That the striking pattern of nymphal skins of many Central American
species is associated with large legume trees in a variety of habitats, including
 TABLE 9.7
Collections of Nymphal Skins of the Cicada F. semi/ata Associated with /nga Shade Cover Trees in a Large Coffee
Plantation at Naranjo, Alajuela Province, Costa Rica a

Distribution of nymphal skins b

Coffee Total
Number of DBH Number of bushes Coffee nymphal skins
Collection Tree trees range tree Range within 16 m' bushes
date species censused (cm) trunks (trees) of a tree 16 m' Ground n 00 Total

29 December 1nga aff. jissicalyx 17 20-30 12 0-2 74 5 4 55 40 95

1972 Inga galdmanii 3 25-35 4 0-1 25 0 19 II 30
Inga leptalaba 6 20-25 10 0-2 3 0 2 10 5 15
II February I. aff. fissicalyx 17 20-30 3 0-1 6 0 5 5 9 14
1973 I. galdmanii 3 25-35 0 0 0 0 0
I. leptalaba 6 20-25 0 0 0 0 0
17 April I. aff. fissicalyx 17 20-30 6 0-1 0 0 0 4 2 6
1973 I. galdmanii 3 25-35 0 0 0 0 0
I. leptalaba 6 20-25 0 0 0 0 0
28 April I. aff. fissicalyx 17 20-30 0 0 0 0 0
1973 I. galdmanii 3 25-35 0 0 0 0 0
I. leptalaba 6 20-25 0 0 0 0 0

"Total area is 416 m' with a 4 m x 4 m plot around each tree. Area estimate excludes areas checked beyond 16 m'.
b Density of nymphal skins: 0.40/m'.
INSECT SPECIES IN AGRICULTURAL HABITATS 413

planted shade trees in some agricultural habitats, suggests that the root systems of
such trees contain better nutrients for nymphs than other tree species. The
nitrogen-fixing properties of legume roots results in higher concentrations of
amino acids and other nitrogen-containing substances being found in legume root
systems, and root systems of other tropical trees with nitrogen-fixing nodules
may also have higher concentrations of amino acids and other nitrogen com-
pounds. Cicada nymphs feed primarily on nitrogen-containing substances instead
of carbohydrates. <366,495) Selection would favor a cicada species in tree-species-
rich habitats to feed on those trees with the greatest concentrations of required
nutrients in the root systems, and many Neotropical Leguminosae are good
candidates. The interesting question develops as to whether legume trees were
the original host trees of cicadas in tropical forests or if strong contemporary
selection favors such coevolved interactions. Tropical cicadas may have "pri-
mary" host tree association, since the likelihood is very high, especially in
tropical forests, that nymphs encounter the root systems of many different plants
and that many of these may be suboptimal in terms of nutrients available to
developing cicadas. In cases where high densities of nymphal skins are as-
sociated with nonleguminous tree species, for example, Z. smaragdina in groves
of Goethalsia meiantha (Tiliaceae) in secondary succession in Central American
rain forests,(7l) it would be interesting to determine if the root systems of these
tree species have nitrogen-fixing nodules associations.
We also stress that other phenotypic constraints operate in cicada popula-
tions to determine host tree associations: the environmental cues used by gravid
female cicadas in finding suitable oviposition sites. The studies of Lloyd and
White(719) on North American periodical cicadas have shown that females of
these species require specific kinds of oviposition sites and that eggs die in
incorrect ones. Several Neotropical cicadas oviposit in dead tissues on understory
trees and palms in forest habitatsyo,7o Toughness of the wood, bark, and pith are
some determinants of oviposition site and hatching success. (719) If a habitat is
altered in such a way that suitable oviposition sites are no longer available to
cicadas, existing generations in the ground represented by feeding nymphs may
be the last generations to emerge in that habitat. The root systems for developing
nymphs may be the same, or nymphs may switch to other species, but oviposi-
tion site deprivation, through a change in the environment (e.g., flooding of a
lowland stream clears away understory vegetation) can lead to a local extinction
of the cicada population in that habitat or that portion of the habitat.
Alternatively, legume root crowns may' 'work" the soil in such a way as to
increase its suitability for developing cicadas, and more so than around other
trees, increasing the survivorship of nymphs. Furthermore, the increased logging
activities in lowland and premontane tropical wet forest zones in Central and
South America act as a potent form of artificial selection in leaving behind
softwood trees such as legumes. Thus cicadas may be forced to exploit legumes
 414 CHAPTER 9

FIGURE 9.3. Cicadas associated with mountain rain forest in Costa Rica (Bajo la Hondura zone).
(a) C. postica; (b) P. biolleyi resting on moss-covered tree trunks in this forest habitat; (c) nymphal
skins of P. biolleyi (left) and C. postica (right). The scale is in centimeters.

as a response to a shrinking supply of trees in tropical forests. Finally, it is

important to recognize the role of physical features of root tissues and bark in
determining the feeding preferences of nymphs.
Our factual data base for the feeding behavior of Neotropical cicada nymphs
is scanty at this time, and data are lacking on the egg-adult developmental time
for any species. The montane tropical rain forest cicada P. biolleyi may have as
INSECT SPECIES IN AGRICULTURAL HABITATS 415

.,

FIGURE 9.3. (Continued)

many as five distinct size classes of nymphs in the soil at the same time, suggest-
ing a developmental time of at least 5 years. (36) This cicada shares the same
forest habitat with another species, Car.ineta postica, and the two species are
readily distinguished in the field both as adults and as final nymphal skins (Fig.
9.3). Neotropical cicadas are nonperiodical, and since emergences occur each
year, there is great turnover in genetic structure from year to year and populations
tend to be similar in size in successive years, especially in the wet lowlands.(70,711
Although evidence exists that some species in the premontane tropical rain forest
zone of northeastern Costa Rica are invading secondary successional habitats,<711
some of these species still occupy primary forests. In some secondary habitats,
emergences occur near trees and shrubs other than Leguminosae. The invasion of
some secondary habitats in the tropics may involve a switch to nonleguminous
root crowns as the primary host association of nymphs. Physiologically and
behaviorally such transitions, in response to selection created by loss of primary
forest habitat and the original host associations of nymphs with leguminous root
crowns, mayor may not involve a qualitative change in the nature of the assumed
coevolved association: this will depend upon the degree of similarity or dissimi-
larity between the original host tree and the newly colonized one in terms of those
physical and chemical properties related to gustation and feeding success by
nymphs. Some lowland and premontane rain forest cicadas are associated with
the margins of forest habitats, in some ways perhaps making them preadapted to
invade earlier successional stages in the same region when the carrying capacity
416 CHAPTER 9

of the original habitat becomes constrictive in terms of maintaining breeding

populations for these insects.
The deliberate planting of lnga in coffee plantations in Central America
may have provided the ideal ecological stepping stone for some cicadas to main-
tain breeding populations as the original forest environment diminished from
human activities. In the central highlands of Costa Rica and also in the major
coffee-growing districts, a dominant stream edge tree species is the legume Zygia
longifolia, and several cicada species emerge at these large trees.(138) The associ-
ation of nymphs with the root crowns of these abundant trees, and the evolution
of oviposition tactics to permit the gravid females to place eggs at or near such
trees, may have been two principal adaptations that paved the way for invading
the leguminous overstory of coffee plantations by cicadas. Nymphal skin
densities beneath overstory lnga trees tend to be lower than for the large Zygia
trees along streams adjacent to such plantations, and the adult population in the
plantation for each species tends to be more dispersed than at stream edges. The
stream edge habitat in such regions consists primarily of the Zygia trees, in an
even row down each side of the stream, with an occasional individual of a
different species interspersed. When the original vegetation cover of the sur-
rounding hills was cleared to plant coffee, cicada populations might have been
compressed into the stream edge environment because the surviving host trees
occurred there. Under such conditions, avian predation, effects of floods on
nymphs and emerging adults, might have been some of the forms of selection
favoring the eventual invasion of newly established coffee plantations once the
lnga trees reached a certain size.
The transition to exploiting the coffee plantation environment may have
involved a change in population structure, nainely, one that emphasized a disper-
sal of reproductive effort over the abundant supply of lnga trees: the average
gravid female of a species such as F. semilata may distribute her egg clutches
over several lnga trees in the coffee plantation, while in the stream edge envi-
ronment there would be a greater tendency to place multiple egg clutches in the
same or a few closely spaced Zygia trees. Also, in the coffee plantation, there
would be a reduced tendency for gravid females to converge on an individual
tree, because the environment is not very patchy in terms of oviposition sites.
Such oviposition tactics would result in scattered reproductive effort in the coffee
plantation and at the same time many empty host tree patches for each generation
because the resource base is large and uniform in spatial distribution. The ob-
served spatial patchiness of nymphal skins associated with overstory trees in
coffee plantations may also result from differential survival effects on eggs and
nymphs. Our present knowledge of Neotropical cicadas is mostly at the level of
knowing some portion of geographical ranges, habitat associations of adults and
nymphs, associations of nymphal skins with certain tree species, and data on
annual peak periods of adult emergences. We greatly need to study the detailed
INSECT SPECIES IN AGRICULTURAL HABITATS 417

natural history of individual species to determine some of the ecological traits

underlying population structure in relation to habitat type.
In the premontane tropical rain forest zone of northeastern Costa Rica,
plantations of black pepper (Piper nigrum) less than 15 years old have uniform
borders made of plantings of madero negro, or Gliridicia nigra, a small legume
tree. Within the past 5 years these trees have been used as major chorusing sites
for cicadas such as F. pronoe, although nymphal skins have not yet been found
beneath these trees (A. M. Young, unpublished observation). What might be
happening here is that we are observing the early stages of colonization of these
legumes as a resource by these cicadas. Alternatively, the cicadas may be using
the trees merely as chorusing sites and not as oviposition sites and root crowns
for nymphs, High densities of adult F. pronoe (as many as 30 in a 3-m-tall tree at
one time) have been seen and courtship observed. The black-pepper habitat is
relatively new in this region of Costa Rica and the deliberate planting of madero
negro trees as perimeters to the plantings may represent a situation somewhat
analogous to the early cultivation of coffee in Costa Rica more than 200 years
ago. But apparently, if the breeding hypothesis is correct, a first generation of
cicadas has not yet been produced on these trees.
Some species of nymphalid butterflies such as Metamorpha and Anartia
have successfully invaded coffee plantations in Central America where larval
host plants occur as brush borders along the perimeters. (317,322) But at least
species of Anartia are extremely labile and are known to be colonists of a variety
of weedy habitats in the American tropics.(72Ol Yet we expect that the total
butterfly and moth faunas of agricultural habitats such as coffee plantations will
be considerably depauperate relative to adjacent secondary successional habitats.
Tropical agricultural habitats screen out through ecological time most of the
herbivorous insects occupying natural habitats, an exception being the highly
palatable food crops exploited by a variety of herbivorous insects. But even
under these conditions, the diversity of such communities is less than in other
habitats. The toxic foliage of cacao, coffee, and other tropical cash crops effec-
tively screens out many insects, and those that occupy these environments are
usually exploiting overstory or weedy species associated with perimeters and
gaps. In many cases, weedy species will be very patchy spatially, limiting
herbivore populations in these agricultural habitats. Many agricultural habitats
are unsuitable for maintaining populations of carnivorous' predators and
parasitoids that exploit herbivorous insects, and in such instances, herbivore
populations may build up to greater densities than normally encountered in other
kinds of tropical habitats. Once again, we need to conduct comparative field
studies on the natural history and population dynamics of selected herbivorous
insect species in agricultural and natural habitats in the tropics.
Borrowing from the concept of local and geographical effects influencing
the community of pollinating insects associated with plant species in tropical
418 CHAPTER 9

forests (Chapter 8), other kinds of insect-plant coevol ved associations may follow
similar patterns. One example, gradually unfolding as the result of recent field
work in Costa Rica, is the interactions of cicadas with legume tree species
in tropical forests. The cicada-legume interaction, postulated on the basis
of the observation that nymphal skins of various species are more abundant
near large legume trees in different kinds of tropical forests than around other
large legume trees in the same habitats, appears to have both local and re-
gional (geographical) traits. At the base of this discussion is a point already
discussed in this chapter, namely, the belief that cicadas in tropical forests have a
primary host tree, usually a legume, that is the place where eggs are placed in the
environment and where the nymphs grow faster as the result of tapping into
nutrient-rich xylem fluids of legume root systems. The question is then the
examination of local and geographical patterns of numbers of cicada species and
numbers of legume tree species exploited under the assumption of feeding as-
sociation of nymphs with these trees. The following tentative patterns emerge.
1. The lowland and premontane tropical rain forests dominated by large P.
macroloba have high densities of several genera and species of medium-to-
large-sized cicadas; fewer species and lower densities are found on other legumes
with a more patchy spatial distribution than Pentaclethra. The high densities of
adult Pentaclethra in these forests promote high densities of cicadas associated
with them.
2. On the eastern slopes of the Cordillera Central above about 800 m, the
density of Pentaclethra drops off markedly and forests contain large trees be-
longing to several leguminous genera and species, each having a patchy spatial
distribution. The result is low densities of nymphal skins per tree and fewer
cicada species occurring at the same tree species and same individual trees in the
forest. Cicada populations are more evenly distributed, at least in terms of
nymphal skins through the forest. Under these conditions, the carrying capacity
of the environment for the same cicada species found in the adjacent lowlands
(e.g., as F. mannifera) is defined in terms of the population exploiting a broad
range of legume tree species. The nymphs may be exhibiting an expanded form
of monophagy as found in the lowlands for specialist feeding on Pentaclethra
and a few other legumes. Then again, the legumes may be interpreted as the same
resource from the viewpoint of the cicada species in different geographical re-
gions. The mid-elevation cicada community contains fewer species and the popu-
lation structure of each species is far more diffuse than for similar species in
similar lowland habitats.
3. On the western side of the Cordillera Central, cicada species are
specialists on the legume tree Z. longfolia and have secondarily, in some cases,
moved to Inga planted in coffee plantations as a shade cover. There are fewer
species in such regions than in the regions mentioned in (1) and (2) above. In
areas within this region where forested valleys remain, the cicada species are
INSECT SPECIES IN AGRICULTURAL HABITATS 419

associated with forest legume tree species and densities of nymphal skins are
low, approaching values seen for the mid-elevations on the eastern slopes of the
Cordillera Central. The diffuse spatial distribution of legume tree species in some
of these habitats promotes low densities of cicadas per unit area.
4. In the dry lowlands of northwestern Costa Rica, cicada species emerge in
low densities per tree around large legume trees with patchy distribution in forest
remnants. The pattern is similar to that seen in other regions where legume tree
species are low in density and several genera and species are widely scattered in
the forest habitat. The total number of genera and species of cicadas is low but
approaches that of western slopes of the Cordillera Central.
The low densities of nymphal skins for a cicada species around individual
large legume trees are believed to be the result of several factors, including
gravid females dispersing reproductive effort over a greater number of individual
trees in response to a small average patch size, competition among nymphs of
different species at the same trees, and differential survival of nymphs at dif-
ferent trees. Large patches of Pentaclethra present gravid female cicadas with a
large "target" for egg placement (in the understory beneath the host trees) and
also provide nymphs of several species with a large primary food base, minimiz-
ing competition and any deleterious effects of a limited food supply. Under the
assumption that many Neotropical cicadas are legume specialists as feeding
nymphs (although not exclusively so), the distribution of cicada populations in
tropical forests and the maintenance of local and regional patterns of cicada
diversity are determined by the spatial distribution and relative abundance of
large legume trees. The occurrence of secondary and agricultural habitats in
many regions of the tropics has added additional levels of environmental
heterogeneity for cicadas. In some cases, the populations of some species, often
only a fraction of the total species of the region, can make the transition to plant
species in these new habitats, but the end result is predicted to be a gradual
colonization of these habitats at a rate determined for each species by many
factors, including (l) the rate of removal of the original forest cover, (2) the
availability of suitable new host plants in the new habitat(s), (3) if (2) is low, the
rate at which a species can evolve the ability to exploit new host plants of a
different physiological makeup with regard to feeding and ovipositing cues, and
(4) the interactions among species if several are making such a transition at the
same time or doing so at different rates. The availability of suitable oviposition
sites in the new habitat will also be a potentially limiting factor in cicada coloni-
zation. The predilection of many species to chorus in sunshine, and the mobility
of the adults, promote the colonization of secondary and agricultural habitats by
tropical cicadas.
Xylem feeding by cicada nymphs on tree roots 066.495) probably precludes
direct contact with toxic secondary substances but may result in the operation of
other forms of feeding specificities by nymphs. The marked, though not exclu-
420 CHAPTER 9

sive, assoCIatIOn of cicada nymphal skins with the immediate environs of

canopy-size legume trees in different habitats in the tropics suggests that cicadas
survive better in association with these trees than with others. Possibly cicada
nymphs obtain a greater amount of nutrients from the roots of legume trees, as a
result of the nitrogen-fixing association of the root systems in many groups of
legumes, or perhaps the association and other factors of the roots make for better
soil conditions for cicada nymphs. Perhaps the tissues of legume root systems are
easier to penetrate by cicada nymphs than other root systems. The selective
lumbering of hardwoods in tropical forests excludes the legume trees, making
them a more abundant resource for cicadas. Natural catastrophes such as land-
slides and floodings provide opportunities for legume tree species to colonize
new habitats in the early stages of succession. A combination of physiological
properties of legumes and ecological factors promoting their high abundance in
many tropical habitats (including natural and man-induced changes in the envi-
ronment) may promote the exploitation of legume tree root crowns by the
nymphs of most species of tropical cicadas.

9.4. TROPICAL INSECT PEST POPULATIONS

Many tropical food crops are beleaguered with herbivorous insects that from
the viewpoint of humans can be considered "pests," since their feeding activity
removes large amounts of harvestable tissues that would otherwise be available
for human consumption. Janzen(57) has discussed some of the reasons why the
regulation of insect pests of such crops is almost a fruitless task. The long
evolutionary history of herbivorous insects adapting to the defensive machinery
of plants in tropical habitats makes them ideal for exploiting the often inbred and
low-resistance crop species: selection would favor such a niche transition as more
and more forests are cleared because (1) the crop is often more palatable; (2) the
insect may extract nutrients more efficiently from the crop; (3) the density of the
crop is high if in monoculture or simple polyculture and therefore more accessi-
ble to the herbivore communities associated with surrounding vegetation; (4) the
crop becomes a predictable resource through successive growing seasons,
thereby stabilizing the populations of individual herbivore species associated
with it, and a period of strong directional selection favoring the switch to the crop
species (or, alternatively, nonevolutionary physiological response if the crop is
similar to the wild host plant species from the insect's viewpoint) will be fol-
lowed by stabilizing selection to maximize exploitation of the crop species; and
(5) there may be a higher fitness for the insect species in leaving behind in its
natural habitat a complex of carnivorous predators and parasitoids so that larger
popUlations build up on the expanded and accessible food base in the agricultural
habitat. It is the habitat switch and weakened defensive systems of crop species
that are primary factors in causing the insect to become a pest species. Coupled
INSECT SPECIES IN AGRICULTURAL HABITATS 421

with the long growing season and high density of the crop species, these factors
provide ideal conditions for such insects to realize the carrying capacity of the
agricultural habitat. Tropical crop species may provide sufficient resources for a
variety of insects without resulting in interspecific competition, and harvestable
producti vity is usually high in such plants, especially if production of foliage has
been maximized in the breeding program.
Levins and Wilson(72ll show that damage to crops resulting from herbivory
is not a linear function of the size of the herbivore population, since in some
cases low densities of herbivores are sometimes beneficial to the plant species,
stimulating the growth of meristem. But high densities usually result in a lower-
ing of yield per unit time or total yield. As shown in some field studies, such as
that of Hinckley<722) on the coconut rhinoceros beetle, O. rhinoceros, fitness of
individuals within a population of an actual or potential pest species will vary
considerably over different kinds of resource patches, generating considerable
heterogeneity in conditions promoting outbreaks locally. Thus herbivory by a
single species will vary considerably in space and time, and the intensity of such
feeding on a crop will be a function of many factors. Levins and Wilson(72J) stress
that a pest species will damage a particular crop only to the extent allowable
by constraints of community structure. They enumerate the following kinds of
interactions of the pest insect with other organisms and the environment at the
community level:
1. Different parasitoids or predators may utilize the same host at different
stages of their life cycle.
2. Different parasitoids or predators may utilize the same host in different
microclimates with or without host movement between sites.
3. Natural enemies of a pest species may also use alternate hosts, thus af-
fecting natural enemy abundance and variability.
4. The control organisms themselves may have natural enemies (predators,
parasitoids, pathogens).
5. A cluster of pest species may interact via common predators and food
competition in such a way that controlling one of them may increase another and
not reduce damage.
6. Agronomic interventions in the field-releasing of control agents,
applying of pesticides, clearing or encouraging vegetation at the edges, etc.-are
undertaken as a response to the stage of the system. The pathways and para-
meters linking observations of the system and intervention are themselves part
of the system and affect its dynamics.
7. There is a nested hierarchy of control agents each of which becomes
primary at different population densities of host.
The above sorts of interactions are expected to occur in tropical agricultural
habitats perhaps to a greater degree than in many Temperate Zone habitats
because the growing season is generally longer, giving more time for interactions
422 CHAPTER 9

to occur, and also because the pool of potentially interacting species is generally
greater. Carroll(694) has called attention to the complex sets of interactions be-
lieved to take place in tropical agroecosystems. Under point 6 above, tropical
examples are probably not controlled in any long-term, effective manner by
application of pesticides. (723) But in terms of various levels of ecological interac-
tions among organisms in tropical agroecosystems, one might anticipate that
there will be many such interactions, and ones changing through time (i.e.,
points 1-5 and 7 in the above list). Levins and Wilson<72!) discuss the role of the
regional mosaic of crop lands and natural forest patches in contributing to the
likelihood that some insect species become pests in some regions. In the lowland
and mid-elevation tropics, there is continual and rapid turnover of habitats into
agroecosystems, small or large, and such a mosaic is very dynamic.
The clearing of the lowland and mid-elevational regions of the wet tropics
today imposes a dynamic state of change on insect communities associated with
forest environments: colonizations and local extinction of species populations in
different habitats, natural and agricultural, alter the spatial distributions of insect
species. If natural habitat refugia do not remain, insects will shift to crop species.
A shift to a crop as a host plant will be a function of food abundance and quality
in existing refugia (fragments of the original habitat), popUlation growth re-
quirements of the insect species, migratory or dispersal ability of individuals in
the popUlation, and distances between refugia and crop stands. Polyphagous
species utilizing several different wild host plants in refugia may move in
stepping-stone manner to the crop stands and probably stand a better chance for
successful colonization of crop species because of relatively greater feeding
flexibility than strictly monophagous species. Other resources required by each
insect species (e.g., nectar sources in the case of adult butterflies) will also
contribute to the process of defining the spatial distribution of the insect popula-
tion between refugia and crop stands.
The regional approach emphasizes the association of an insect species with
both sorts of habitats, natural or agricultural, and the degree to which both kinds
of environments are necessary in joint distribution to provide sufficient carrying
capacity for the species. Interwoven with such effects will be the influences of
local community structure and organization upon whether or not a particular
species makes the ecological shift to an agricultural habitat, exploiting the crop
species or other species (weeds or overstory trees) in it. Selection for dispersal
into an agricultural habitat from a diminishing natural habitat and the likelihood
for successful colonization of the new environment depend upon the kinds of
interactions among species in the local community in the natural habitat and the
kind of community in the agricultural habitat.
A tradeoff exists: adapt to high density of the same species or adapt to
competition with other species. If the colonist species leaves behind biotic con-
trol agents that keep its popUlation well below the carrying capacity of the new
INSECT SPECIES IN AGRICULTURAL HABITATS 423

environment, both kinds of competition may prevail, depending upon the popula-
tion growth characteristics of the colonist species and the amount and quality of
the new food base. If the food base is very large and of high quality [i.e., can be
harvested efficiently, compared with the wild host plante s)] and the species is
essentially K -selected, colonization will be relatively free from competition of
both kinds. Yet if other insect species are already associated with the crop host
plant, even if exploiting different parts or different individuals within the crop
patch, there may be a very subtle form of indirect competition since resources are
being depleted. If the colonist is an r-selected form and the food base is still
large, colonization may be successful. If the food base is small (i.e., if the crop
patch is small and different crop patches are widely separated and interspersed
with other vegetation) intraspecific competition may occur as population den-
sities rise on the crop plants. The negative effects of such competition may be
offset in tropical regions by continual growth of the plants in the long growing
season of mildly seasonal and nonseasonal regions. Interspecific competition
may be intense if the crop species already has a high herbivore load and the
colonist species hasn't the kind of genotypes best suited to counter such effects.
Although for most food crops in monoculture the food base will usually be
large, fluctuations in quality may be frequent. Success in colonization, or promo-
tion of rapid local extinctions within the crop stand and refugia mosaic, will also
be molded by climatic factors(72)) and a variety of agronomic practices affecting
the spacing of the crop and its arrangement relative to other vegetation patches. It
is known that Temperate Zone complexes of herbivorous insects associated with
a particular kind of host plant are sensitive to a broad range of biotic and abiotic
control factors either lessening or enhancing food abundance and quality or
acting as outright mortality agents on the insects.(724) In the lowland and mid-
elevation wet tropics, such effects may be generally constant throughout the
year, with the possible exception of the last 2 months of the rainy season when
rains are heaviest.
Guilds or complexes of herbivores associated with certain plants in strongly
seasonal tropical lowland will show periodic reductions in some or all species
within guilds or complexes in relation to detrimental effects on plants and insects
promoted by seasonal cycles (e.g., Fig. 9.4). This model (Fig. 9.4) assumes that
many tropical environments are fairly nonseasonal and wet. In the temperate
zones, comparable species or herbivores, one exploiting a crop, the other a wild
plant species in the same climatic regime, both exhibit an initial overshoot in
abundance at the beginning of the growing season, following a brief population
growth phase associated with the appearance of new meristems. The crop herbi-
vore builds up and maintains a larger breeding population during the annual
growing season than the herbivore exploiting a wild plant species, because the
crop plant species is usually far more abundant per unit area, thereby providing
a larger resource base for the insect population, and because a very abundant crop
424 CHAPTER 9

",,,-..............
,/ \\,
I . . . ...
,i . . . ""----.. . . . . .------.. . . . . . .
.......... _--............
1 TEMPERATE CROP PEST -_

W
,! ' '\,\
N
en !
I TEMPERATE FI ELD HERBIVORE
\ ,
Z
o
~ ,,i/ \
..J
;:)
---...................._-_ ...............-----_..._--------_ .. _---- ..............----.

j/
a.
oa. ",',......

" LOWLAND I MID-ELEVATION TROPICAL CROP PEST

(RAIN FOREST)

LOWLAND IMID-ELEVATION TROPICAL HERBIVORE

(RAIN FOREST)

ANNUAL GROWING SEASON

FIGURE 9_4_ A hypothetical model for annual patterns of population growth and maintenance for
herbivorous insects associated with forest, field, and crop habitats in the temperate zones, and
lowland and mid-elevation tropics.

pest may satiate some or all of its predators. The herbivore associated with the
wild plant is expected to experience more biotic interactions with other ar-
thropods (competitors, predators) than the crop pest, and the plant will have a
more patchy distribution. These ecological factors interact to keep the herbivore
population, in this situation, at a lower level of numerical abundance. In both
situations, popUlation growth and maintenance are compressed into the growing
season, with popUlations falling off rapidly with onset of winter. In the tropics,
the situation is about the same in terms of relative abundance between herbivores
associated with crops and those on wild host plants. But since biotic controls are
predicted to be even greater in lowland and mid-altitude habitats in the tropics,
there is a greater absolute difference in abundance of the herbivore between crop
and natural habitats once populations stabilize within the census history. It is
difficult to generalize about such effects for all insects and plants in a particular
region or habitat since each species will have a different evolutionary history and
different ecological requirements for maintaining breeding populations. The
complex, often subtle multitude of biotic and abiotic factors affecting each indi-
vidual within a single population of an insect species requires detailed natural
history studies in order to determine the spectrum of variation in the ways the
INSECT SPECIES IN AGRICULTURAL HABITATS 425

population as a whole responds to changes in habitat and resource availability

(e.g., Fig. 9.5).
There have been, of course, attempts at developing mutant genetic lines of
targeted pest species(72S) as one approach to changing the impact of the insect on
the crop or livestock, but such an approach is unlikely to be fruitful and promis-
ing for tropical regions, since the host organism itself will respond to such
changes (i.e., perhaps become more vulnerable when the selection pressure is
relaxed and thereby increase its susceptibility to another pest species). The high
species densities of insects in many tropical habitats increase the chances of a

200 ~(1-20 SPECIES/HOUR)

150

100

50

en 200
HYMENOPTERA ( 0 -10 SPECIES/ HOUR)
!!:! ( EXCL. ANTS)
uW 150
a.
en
..J 100
..J
<t 50
en
~
::l
o~--~====~--------~----------------------------~
1800
o 1200
~
o
~ 1000

U.
o 800
en
ffi AOO
III
~
::l
Z 50 COLEOPTERA (8-20 SPECIES/HOUR)

AO

30

20

10

930 AM 1030 1130 1230 130 230 330 A30 530 900

HOURS OF THE DAY

FIGURE 9.5. Changes in the abundances for all species combined in four different groups of insects
exploiting experimental arboreal baits in tropical rain forest over approximately 12 hr. Also given are
ranges of total number of species seen hourly within these groups.
426 CHAPTER 9

resource being filled if another species ceases to use it or is deterred from using
it. In some Temperate Zone crops, somewhat successful attempts have been
made to control pest insects with insecticide applications at times other than
flowering as a means of reducing the mortality of bees and other pollinators of
such crops. <7(3) But the same bean crop as in the study of Bardner et al. (723)
located in the lowland Central American rain forest zone would support a greater
number of herbivore species than in their English field, and the populations
would be high throughout most of the year as a result of the long growing season.
The flowering regime would also bt< different. Together, these conditions buffer
the herbivore guild against an insecticide, since the insects remain abundant
throughout most of the year (i.e., lack the seasonal cycling of Temperate Zone
crop systems) and any complex of carnivorous predators and parasites associated
with them would also be knocked down by the treatment. The result would be an
increase in the herbivore popUlations as the growing season advances. Popula-
tions of the crazy ant, Anoplolepis longipes, in the Seychelles can be temporarily
controlled by specific kinds of toxic baits that are effective in the short run,(706)
but long-term solutions to the control of such an organism are probably lacking.
Such an ant species produces many colonies per unit time as an introduced
species, and the manpower required to check for every colony in an area, even on
an island, would be high. Short-term local control measures may be effective,
but eradication of the species in the long term is not a likely prospect. The
spillover effects of toxic baits on other members of the local arthropod com-
munities have not been studied.
From detailed studies of the interactions of parasitoids with pest-labeled
herbivorous insects in the temperate zones ,<726-728) it is seen that density-
dependent regulation of pests by parasitoids or predators works best at low
densities of the host species, i.e., during periods of popUlation census when the
insect is not behaving as a pest. Parasitism tends to become obscured as an
effective mechanism for regulating populations when densities soar in some
systems studied.<7(8) Seasonal fluctuations in the abiotic components of the envi-
ronment also affect the effectiveness of parasitism in regulating a host popula-
tion.<7(7) More than one density-dependent mechanism may regulate the same
population. In the humid or wet tropics in particular, densities of herbivore
species in natural habitats may be kept low through the operation of several
different forms of extrinsic-control agents such as parasitoids. Coupled with the
energy sink associated with the physiology of detoxification and processing the
defensive substances in host plants, these factors keep the herbivore populations
generally low through successive generations. To handle the defensive chemistry
of a host plant may require an energy allotment then made inaccessible for
making more eggs. Birth rate is lowered through evolutionary time as the
coevolutionary interaction of the insect and its host plant continues and changes.
Such an effect, coupled with the often scattered or spatially patchy distributions
INSECT SPECIES IN AGRICULTURAL HABITATS 427

of host plant populations, ensures that the density of individuals of a given

species per host plant individual or clump remains low. Parasitoids and carnivor-
ous predators may effectively keep numbers well below the carrying capacity of
the plant population when densities are low. In the strongly seasonal lowland
tropics, the dry season adds a dimension of density-independent mortality to both
herbivores and their carnivores and parasitoids, excluding, of course, popula-
tions thriving in moist refugia. When the habitat is converted into a highly
palatable monoculture of a crop species, the herbivores's phenotypic constraints
associated with overcoming the defenses of a host plant are relaxed, and the
thereby increasing energy available for reproduction through evolutionary time,
promotes high population growth on the expanded food base. Population densities
of the herbivore soar to levels that promote a complex of carnivores and
parasitoids to exploit the herbivore on its new food base. However, these biotic
control agents cannot regulate the populations. In the high-altitude regions of the
tropics, the rigors of the physical climate would keep herbivore populations
down throughout the growing season and abiotic mortality during the dry season
would be high.
 CHAPTER 10

BIOGEOGRAPHICAL
AND REGIONAL
EVOLUTIONARY-
ECOLOGICAL EFFECTS ON
THE MAINTENANCE OF
TROPICAL INSECT FAUNAS
A Brief Perspective

A central theme in the development of a unified approach to understanding the

high species richness, population dynamics, and habitat associations of insects in
the tropics in the foregoing chapters has been the effects of local selection
pressures on individuals in the populations of single species, and the interactions
among species as the basis for coevolved sets of interactions underlying the
composition of species assemblages, guilds, or communities.

10.1. HISTORICAL EVENTS AND ECOLOGICAL PROCESSES

One might be tempted to conclude that the distribution patterns of species in

particular communities can be explained fully in terms of contemporaneous
ecological effects. Yet there must also be a historical or evolutionary perspective
to explaining the associations of insects with certain regions, habitats, and com-
munities in the tropics. Thus one also might conduct an exhaustive field study of
the ecological properties of a group of sap-feeding insects such as plant hoppers
in a particular climatic zone of South America, determining that a high degree of
host plant specificity is a key factor in accounting for the species composition of
the group in that region. However, this is not an explanation as to why these
species occur in the region, and to what degree have regional or biogeographical

429
430 CHAPTER 10

effects molded the initial occurrence of these fonns in the region. The fact that
different regions of the world have experienced different geological histories,
and the absence of some major groups of insects from some Old and New World
tropical regions or relatively poor representation of other groups, is an indication
of the role of biogeographical events in detennining the broad distributions of
insect faunas, tropical and extratropical. The general observation that some re-
gions of the terrestrial tropics contain more species than others, and more than do
Temperate Zone regions, is really a question of the interplay of biogeographical
events with the dynamics of local communities. The sum total of such interac-
tions is that regions of the tropics with high species richness are characterized by
faunas with higher rates of speciation or reduced rates of extinction once popula-
tions of each species come into some sort of equilibrium with resources in the
habitat. Tropical regions with great species richness are also considered those
with benign and predictable environmentsY04)
The ecological question incorporates an attempt to account for how each
species in the local community maintains a breeding population in tenns of how
each species exploits resources. The evolutionary question incorporates such
infonnation but goes further: it attempts to explain why these species occurred
there in the first place. Geographical distributions of single species or whole
groups are explained in tenns of ecological and broad-scale climatic and other
environmental patterns, and how such patterns change through evolutionary
time, causing changes in the distributions of species and in the morphology,
behavior, and ecology of species. Tropical regions consist of mosaics of different
topographic and climatic regions. The large amount of heterogeneity in the
climate, vegetation, topography, and soils(729) of South America points to the role
of regionality in determining the distributions of organisms over the land mass.
The recognition of distinctive soil regions(729) indicates that plant communities
will also be very distinctive among different regions, in tum influencing the
distribution of herbivorous insects associated with individual plant species or
plant families. Thus we might expect that tropical mountain tops above 4000 m
will have very different insect faunas from intennediate elevations of about 1000
m, and that lowland dry zones will support different insect faunas from lowland
tropical rain forest zones. Superimposed upon these vertical and horizontal pat-
terns of environmental heterogeneity over large areas are regular and irregular
changes induced by shifting climatic conditions, regional catastrophes such as
earthquakes and volcano eruptions, and the impact of humans in clearing large
expanses of forest for agriculture and lumbering. It is also recognized that natural
geological cycles in the earth result in changing distributions of tropical rain
forests, expanding and contracting in a kind of pulsation cycle corresponding to
cycles of climatic change through geologic time. We are referring to a variety of
regional and biogeographical effects that operate in different time frames, in-
fluencing distributions of organisms in different ways.
EFFECTS ON MAINTENANCE OF TROPICAL INSECT FAUNAS 431

There have been numerous in-depth studies on the systematics and biogeo-
graphical origins of specific kinds of insect groups in tropical regions(730) as well
as attempts to explain phylogenetic patterns of specific groups in terms of discer-
nible regions within the New and Old World Tropics.(731) Such surveys have
emphasized the great differences in richness of floras and faunas among distinc-
tive regions in the tropics and the correlations of such patterns with abiotic
factors such as changes in topography and soil profiles. (731) These studies have
provided the framework for both ecological and biogeographical studies in tropi-
cal regions. (732) Thus as developed below, it is believed that cycles of wetness
and aridity during the Pleistocene promoted cycles of expansion and contraction
in Amazonian rain forests. Insect species trapped in forest islands(469,623--626,733)
must have experienced drastic changes in selection pressures associated with
distribution and availability of resources, in tum affecting population size and
degree of ecological specialization among genotypes within the local popula-
tion. (734.735) Such isolation followed by periods of new contact helps to explain
the present-day distribution of species within different groups throughout much
of tropical South America. Brown ,(622-624) for example, has discovered many
distinct refugia of Heliconius butterflies in South America and emphasized the
role of historical events such as the pulsing cycles of climatic change, in deter-
mining the overall evolution and distributions of species. We need to recognize
such events and weigh this information with an examination of current ecological
properties associated with the popUlations of individual species, their geograph-
ical distributions, their habitat associations, and the spectrum of resources
utilized (e.g., larval host plants) in attempting to erect a model for the local and
regional distribution of the group. Vane-Wright(736) has emphasized the impor-
tance of ecological factors in the speciation of butterflies, especially in the
tropics, where more intense patterns of spatial and temporal heterogeneity of re-
sources appear to be promoting ecological diversification in local communities.
Vane-Wright stresses the added role of behavioral mechanisms such as mimicry in
promoting species richness in tropical butterflies.
Cycles of climatic changes influencing the size and distribution of suitable
habitats for particular species, as exemplified by Quaternary refugia model for
Amazonian floras and faunas, determine the degree to which the gene pools of a
species are isolated at a given locality and time. Any breeding population of an
insect species will contain different kinds of genotypes, some of which are
adapted to conditions in the present environment, others of which promote the
colonization or migration to new environments. Shifts in directional selection
will determine in part the distribution of such classes of genotypes in the popula-
tion. Range extensions for a species, following the appearance of increased
suitable habitat, such as during a period of rain forest expansion under the
geological model, are accompanied by a reduction in cohesiveness of the breed-
ing population and the increase in frequency of genotypes capable of local or
432 CHAPTER 10

extended migration to locate and colonize new resource patches. The environ-
ment, under such conditions, may be encountered as a series of disruptive
changes in physical and biotic conditions, or as a gradual (clinal) shift in such
properties. The degree to which a species in the process of range extension
encounters other populations of the same species, or populations of closely
related or ecologically similar species, will influence the rate at which the species
undergoes range extension. Disruptive environmental gradients may present the
challenge of new sets of resources and can function as ecological filters in
blocking range extension if the species is not able to exploit them.
Brittnacher et al. (737) noted considerable geographical variation in the elec-
trophoretic analyses of 16 loci in several species of the butterfly genus Speyeria
in California. They tentatively conclude that selection may be maintaining the
observed patterns of genetic divergence within each species over the geograph-
ical regions surveyed. Species widely distributed over altitudinal gradients in the
tropics are expected to be experiencing very different forms of selection pressure,
both within and between selection established by abiotic and biotic components
ofthe environment. As Brittnacher et al. (737) indicate, the degree to which a local
breeding popUlation or unit of a species is residential, that is, less prone to
dispersal, the greater the differentiation locally in genetic structure from popula-
tion to popUlation. Janzen(46) has called attention to the divergence in ecological
characteristics of tropical mountains versus lowlands. Such divergences probably
result within a species found in both areas, in either (1) expression of a flexible
response system to environmental changes at the physiological or behavioral
levels, particularly if there is considerable dispersal of individuals over the grad-
ient, (2) local genetic (evolutionary) diversification, particularly when dispersal
is reduced or absent and (3) a combination of (1) and (2). Attempts have been
made to explain the geographical distribution of whole butterfly communities in
terms of limitations in geographical range imposed by local, specialized habitat
requirements in mountain regions.(738) Local specialized ecological conditions
affecting a species result in ecological divergence and geographical patchiness.
Both abiotic and biotic factors will determine the spatial distribution of an
insect species, both locally and regionally, although different forms of these
environmental factors may be involved. For a given species, a generalized
phenotype will result from regional influences, e.g., abiotic factors such as
ambient temperature, humidity, and light. A variety of other abiotic factors will
also influence the distribution of a species over large areas. A region can also be
viewed as the collection of different habitats encompassing large land areas
horizontally (e. g., the tropical rain forest flood plain region of northeastern Costa
Rica) or vertically (e.g., a transect through an altitudinal gradient extending from
the top of the Peruvian Andes down into the Amazonian rain forest) .The distribu-
tion of a butterfly species, for example, over a region will be a function of
primarily (1) the ability of the species to withstand possibly marked changes in
EFFECTS ON MAINTENANCE OF TROPICAL INSECT FAUNAS 433

physical features of the environment across the region surveyed, (2) the degree to
which larval and adult resources are distributed across the region in question, (3)
the types of interactions with other consumer species including potential compe-
titors and predators, and (4) historical events. The distribution of the species
across the region is a dynamic condition in which one or more distinct breeding
populations experience a changing set of selection pressures arising from interac-
tions with the species in the habitat or habitats occupied by the population(s).
Thus one expects that the actual boundaries of the species range are changing. If
the species is successful in occupying an increasing number of habitats through
colonizing episodes, there will be a range extension if the original populations
remain intact as well. In addition to a host of abiotic factors, the primary biotic
factor in the case of insects for determining the geographical range of the species
in question is the distribution of food sources. Other biotic factors, such as
interactions among species within communities and trophic interactions, will be
characteristic of the particular habitats occupied by the species. The broad geo-
graphical distribution of the species thus becomes a matter of determining the
past and present ranges of suitable environments to be occupied by the species.
This is the historical component of species richness in the tropics.
Prolonged periods of drought in some subtropical regions can result in very
different responses by butterfly populations for a species widely distributed, with
some populations going extinct, others remaining stable in size, and still others
actually increasing in size.<73S) The crucial factor appears to be the response
patterns of both the insect and its host plants to drought conditions. Response
patterns of this kind vary considerably with locality and habitat conditions.(735)
Janzen(46) suggests that organisms experiencing a severe tropical dry season
should be better adapted to crossing a dry zone from a wet zone in the tropics
than a similar organism occupying a rather nonseasonal wet zone. Changes in
temperature along a gradient are probably a more effective barrier to dispersal for
organisms in the tropics than absolute height of a mountain on which the gradient
occurs.(67) Such a phenomenon helps to explain the frequently observed high
degree of endemism in tropical mountain top insect faunas.(739) Various orders of
insects often show high percentages of endemism (greater than 60%) for
mountain top assemblages in the tropics. (740) Such patterns are explained in part
by the distribution of suitable resources for individual genera and species and
also by the effective barriers to dispersal subsequent to colonization due to
temperature and temperature-related effects creating physical barriers to move-
ment (e. g., moisture gradients). In the temperate zones, insects experience
mountains as less of a temperature gradient than in the tropics because these
species are already adapted to severe seasonal fluctuations in temperature and
other abiotic factors, lessening the effect of spatial gradients in temperature.(46)
Thus endemism is probably greater in insect faunas along altitudinal gradients in
the tropics than along comparable gradients in the temperate zones. If tempera-
434 CHAPTER 10

ture zones are equated with regions, the tropics tend to have a greater diversity of
geographical zones generating a high regional diversity of insects and other
organisms. The ecological factors associated with specific differences between
habitats making up the geographical range of the species constitute the ecological
component for explaining species richness, and much of the preceding chapters
were devoted to this aspect of tropical insect diversity. The following sets of
ecological factors are operative:
1. Regional
Broad abiotic factors: temperatures, humidity, light, soil cover, etc.
Biotic factors: distribution and abundance of food resources (i.e., food
plants for phytophagous insects). Together, these factors select for a
generalized phenotype capable of occupying a range of environments
or habitats.
2. Local (Habitat)
Specific sets of physical features: local soil nutrient profiles, topographic
effects on moisture and light, abundance of shelters and other re-
sources affecting fitness.
Biotic components of the niche: food and other resources, predators and
parasites, competing species. Together, these factors mold the life
history pattern of the species in each habitat, determining the patterns
of phenotypic allocation of resources to mating, foraging, escaping
from predators and parasites, and buffering against physiologically
stressful changes in temperature, moisture, etc.
Both general sets of factors determine the strategies of environmental re-
sponse given by Slobodkin and Rapoport(99) and discussed in Chapter 1. As
discussed in Chapter 5, it is also important to consider that patterns of environ-
mental response often differ greatly between different habitats in a region.
White C7411 has considered this phenomenon as the "area-effect" component of
speciation. Different habitats produce different selection pressures, and when a
species has a continuous population over a large area, and considerable genetic
and phenotypic variation exists in the population, speciation proceeds by
localized area effects in which popUlations or subpopulations associated with
different habitats diverge phenotypically. The vicariance model (refugia or core
area) of speciation essentially predicts that when a species with a large, continuous
breeding popUlation eventually becomes fragmented due to the formation of new
barriers over the region (e.g., geological, climatic), the resulting, isolated sub-
populations may diverge into different sub specific forms and perhaps into new
species. Thus the vicariance model of speciation is a logical extension of area
effect speciation. Biogeographers working in the tropics are becoming increas-
ingly interested in this approach to understanding the dynamics of speciation.
Thus one envisions that biogeographical patterns of particular groups of species
EFFECTS ON MAINTENANCE OF TROPICAL INSECT FAUNAS 435

or individual species are really the result of evolutionary or historical events

affecting regional distribution,(742) and the local effects of how various sets of
ecological factors in distinct habitats determine the tenacity for a population to
maintain itself through time. Surveys of insect groups along altitudinal gradients
in tropical countries (e.g., Mani and Singh(740») reveal discontinuities in faunas as
a function of elevation and past geological events affecting migration and distri-
bution of species. Yet it is necessary to look beyond such data and attempt to
examine the question of abrupt or gradual changes in species composition and
richness for particular groups on a regional basis.(743) Some species of alpine
Lepidoptera exhibit considerable geographical variation,(44) presumably the re-
sult of differing selection pressures and overall environmental conditions along
altitudinal gradients. Alternatively, the recent studies of Noonan on the carabid
beetle, Notiobia peruviana, in the Andes of South America, support the view of
vicariance origin of new subspecific forms. Noonan censused populations of this
beetle over about a 4000-m-altitudinal transect from sea level to the high slopes
of the Andes and found evidence to support the vicariance model of speciation in
this insect. (745)
In this chapter an emphasis is placed on the ecological properties and distri-
bution patterns of insects along altitudinal gradients in the tropics, with further
emphasis on the Neotropics. There is an attempt to explain the adaptive re-
sponses of insects to altitudinal gradients as well as a brief examination of
biogeographical effects as related to the evolutionary differentiation of high
elevation insect faunas from the tropical lowlands.

10.2. ALTITUDINAL AND OTHER MACROGRADIENTS IN TROPICAL

REGIONS

MacArthui l6 ) stressed that one must consider the time scale when attempt-
ing to account for changes in the flora and fauna of a region. For example, the
time scale of thousands of years encompasses the Pleistocene with fluctuating
periods of dryness and wetness over large areas such as Amazonian rain forest.
The studies of Vanzolini and Williams(746) and Haffd469 ) imply the existence of
core areas or refugia in which tropical rain forest remained during periods of
dryness. Many such core areas of forest islands were scattered over large regions
of South America, especially below 900 m. The inference of core areas or refugia
arises from present-day patterns of evolutionary radiation in different groups of
plants and animals, including insects,(622-624) and from paleontological data. Core
areas remained wet during periods of Pleistocene and post-Pleistocene dryness and
functioned as reservoirs for many species. Different cores had different ecological
conditions, presumably since communities became fragmented as forests con-
tracted in size and patches. Physical conditions may have been uniform in each
436 CHAPTER 10

core area, but conditions varied considerably between core areas, promoting
evolutionary radiation as the result of different forms of selection pressure to
physical and biotic components of core areas (e. g., the vicariance model).
The existence of specific core areas cannot be proven but can be strongly
infected from existing patterns of distributions of species or groups of species.
The studies of Darlingtod747.748) stress the relationship between size of a region
and the number of species found there: on the time scale of millions of years,
continental areas with mild climates accumulated more species than other
areas, and such tropical regions presumably had more successful colonists to
invade other regions. Competitive superiority within communities is held to
evolve most frequently in larger regions where communities are larger by
virtue of occupying a greater area of habitat(s), and such species may include
good colonists for island invasion. Yet selection for genotypes to be optimal
colonists does not necessarily include competitive superiority ,(87) since newly
arriving colonizing species on islands often do poorly if the island already
contains ecologically similar species that are good competitors.(9Jl Whether we
refer to actual archipelagos(742) or ecological islands on mainlands ,(93.749) the
initial invasion of a colonizing species into the island communities does not
necessarily imply superiority in becoming integrated into the existing com-
munities. Success will be largely a function of the existing species composi-
tion of the community<750l and the degree to which competitive genotypes
are being selected for. Thus in the inferred cycles of expansion and contraction
of Amazonian rain forests the original pattern of high species diversity ac-
cumulated in the lowlands of tropical South America became subdivided into
refugia periodically, and each contraction phase generated a size distribut!on
of such refugia, with species richness and community organization and complex-
ity varying as a function of sampling error and size of the core area.
Larger core areas probably had a greater number of species and presumably
experienced a greater amount of evolutionary radiation for particular taxa.(751) I
refer to such patterns as an "unstable ecological mosaic" in the sense that each
core area was different in size, species composition, community structures, and
physical conditions from others, and that each must have passed through a period
of nonequilibrium communities following its formation in the cycles of climatic
change. Possibly such communities, because they represented sampled portions
of other communities that evolved in the intact forest environments, and because
of insufficient time, never attained equilibrium conditions in the core area state.
Yet the studies of Brown(62:?:-624) indicate that the existence of core areas is
implied in distributional patterns of groups of closely related species such as
Heliconius or ithomiine butterflies.
MacArthur<16) asserts that the high species richness of insects and other
invertebrates in the tropics results from (1) their ability to maintain denser popu-
lations than vertebrates, (2) lower extinction rates, (3) small body size allowing
EFFECTS ON MAINTENANCE OF TROPICAL INSECT FAUNAS 437

speciation in relatively small portions of habitats, and (4) physiologicallimita-

tions of invertebrates to invade Temperate Zone areas, thereby lowering immi-
gration rates. MacArthur states that invertebrates in the tropics exhibit a high
proneness toward speciating and that many forms are highly endemic to certain
regions. Distribution patterns of endemic species of insects commonly seen along
altitudinal gradients in the tropics are best interpreted as the result of steep
temperature and moisture gradients blocking immigration among zones.
Ability for an insect species to successfully invade another zone on an
altitudinal or horizontal gradient will also depend upon the community being
invaded. If a community has saturated the resource base within a particular zone
on the gradient, invasion by a colonist is improbable.(1S) Such effects may be
seen at the level of whole communities or faunas: where the assemblages meet
along the gradient, a narrow zone of chaos(1S) in the replacement area of the
environment acts as a form of diffuse competition and the structure and organiza-
tion of each community is maintained. The monarch butterfly may represent a
species in some form of diffuse competition with the queen at montane overwin-
tering sites in Mexico: the monarch enters the habitat temporarily but does not
reproduce there, while the queen may do so. At the same time, the monarch is
only a transient member of Temperate Zone communities, since it is not adapted
to withstand the physical rigors of the northern winter. The monarch shuttles
between two very different kinds of communities, one tropical and one temper-
ate, but reproduces only in one, and generally does best where the queen is
absent. Truly residential lowland tropical populations of the monarch are pre-
sumably of sufficiently low density (immature stages) so that competition with
the queen is probably low. The original selection pressures favoring the evolution
of the migratory behavior in the monarch presumably consisted of a mixture of
competition from the queen under certain conditions, effects of seasonality on
larval and adult resources in dry forest regions of Central America, and the great
evolutionary radiation of the larval host plants into the Temperate Zone of North
America.
Yet the monarch is the exception in the sense that one can recognize whole
insect faunas restricted to different regions in which these biogeographical realms
coincide with present or past distinct physical boundaries of continents and
islands, and areas within continents. Within such realms, it is necessary to
discern subregions containing distinctive assemblages of species. Going from the
lowland (0-900 m) tropical rain forest on the eastern slopes of the Andes in
South America to the snow line at mountain tops (about 4500 m) entails passing
through several distinctive vegetational and abiotic zones ,(739) each with a distinct
set of insect species and with some degree of overlap in these faunas, especially
at the lower elevations. Mani(739) provides considerable detailed examples of the
insect groups found in the Andes at various elevations and, in some cases,
attempts to explain biogeographical origins.
438 CHAPTER 10

Current explanations for the origins of tropical biotas stress the apparent
gene flow that occurred between populations on different tropical continents into
the mid-Tertiary as the result of plate tectonics.(733) The advent of a modem
theory of island biogeography04!l together with inferences about the effects of
Pleistocene climates on Amazonian rain forest vegetationl 469 ,623,733) has portrayed
the effects of mainland (continental) island habitats (core areas or refugia) on the
present-day distribution of flora and fauna in the Neotropical region. Various types
of coevolution between plants and insects must have occurred in these forest en-
vironments, most notably (1) the evolution of herbivore defense systems in plants,
(2) partitioning of nectar and pollen sources through specialized pollination sys-
tems, and (3) the evolution of mutual isms between ants and plants, the latter
including the elaboration of extrafloral nectaries. Furthermore, the structure and
organization of insect communities in tropical forest habitats, particularly in
lowland (tierra caliente) and mid-elevation (tierra templada) zones, and the
accumulation and cyclic fragmentation of whole insect faunas (biogeographical
effects) during the Pleistocene, have also been a major factors in determining the
geographical distribution of insect species within tropical regions.
Prance(733) has also emphasized the geological differentiation of Amazonia
into distinct forest types, based largely upon the ecological constraints arising
from the structure and composition of soils in different regions and along topo-
graphic gradients. These forest types vary in terms of degree of flooding and
composition of the vegetation, establishing distinct areas for the accumulation of
different insect communities. Soil type, slope, and elevation, coupled with the
effects of flooding, have created the mosaic of forest types in Amazonia, of
which the terra firma forest represents the major portion and one of the richest in
terms of floral composition.(733) Some forest types, because of their relative
isolation within others, promote considerable endemism, also the result of these
types of habitats being scarce and patchy in distribution. Brown and Benson(752)
discuss the impact of the camp ina forest habitat on endemism in Heliconius
butterflies in Amazonia.
Prance(733) accounts for four distinct dry periods during the Pleistocene in
which the Amazonian forest region was reduced, each time to a different degree.
Such pulsations of forest reduction created a complex series of responses in
insect species found in the region, conceptualized as periods of evolutionary
radiation followed by reduction in species richness and increased endemism.
Thus many species today have seemingly puzzling disjunct geographical or re-
gional distributions that can be explained by the refugia or core area model.
Rather than observing that a species has a rather transitional pattern of dif-
ferentiation along horizontal or vertical gradients over large areas, the more
characteristic pattern is that of disjunct populations, a condition explained by the
refugia model. Some or all of these populations for a given species reflect the
location of core areas during the recent geological past. Thus Amazonia and
EFFECTS ON MAINTENANCE OF TROPICAL INSECT FAUNAS 439

other regions within the Neotropical region, including Mexico and Central
America, are viewed as highly heterogeneous regions in the sense that they
contained numerous core areas that regulated the subsequent pattern of range
expansion throughout the region. The problem for the present-day biogeographer
is to reconstruct the spatial (geographical) routes followed by individual species
or groups of species (genus, tribes, etc.) before, during, and following the
existence of the refugia, and the amount of time involved to reach a certain kind
of distributional pattern.
The biogeographer approaches such questions through the lenses of pulsa-
tions in climate and resulting effects on the size and composition of habitats
occupied by the species being studied. Historical events are therefore held to be
of prime significance in the determination of geographical distributions of
species,(753) but it is also necessary to consider contemporary or ecological
phenomena in interpreting such patterns. The refugia model generates a complex
fabric of species distributions in which allopatric popUlations arise for each
species and different spheres of selection pressures promote eventual specia-
tion.(754) Brown and Benson(752) describe the subspecific differentiation of
Heliconius hermathena into four distinct and separate savannalike habitats sur-
rounded by forests. Prance(733) emphasizes the role of historical factors in the
shaping of forest types and their distribution in Amazonia. Yet it is necessary
also to view distributions of species in terms of ecological processes such as seed
predation, pollination requirements, and niche availability. These topics were
examined in previous chapters.
The message here is that historical and ecological effects, at the time the
species are being studied, must be taken into account together to explain existing
patterns of species. The historical component concerns the impact of past major
climatic changes on floras and faunas, while the ecological component relates to
interactions between species in communities and across trophic associations in
tropical ecosystems. But the ecological component is not exclusive of the histori-
cal: specific kinds of ecological interactions during the past also contribute to the
present-day ecological traits of individual species.
Fittkau(22-24) notes the great abundance of insects and other arthropods found
in South America and discusses the evolutionary development of these faunas as
a function of geographical isolation from other regions, promoting endemism.
South America is divided into two major regions: the tropical Guianan-Brazilian
region and the more temperate Andean-Patagonian region.(75S) This tropical re-
gion, encompassing lowlands and mid-elevation zones, possesses a fauna overall
that appears to be derived from that of Africa,(75S) while the temperate region
bears closer similarities to that of Australia and New Zealand. Ecologically, the
tropical faunas are adapted to forest environments while those of the temperate
zones are adapted to open savanna or prairielike environments. It is the tropical
region that contains the greatest number of species, including insects, while the
440 CHAPTER 10

Temperate Zone has fewer species. Both regions contain high percentages of
endemics within different groups, although the Temperate Zone of South
America is considerably lower in species numbers than other comparable Tem-
perate Zone regions.(24) As discussed by Fittkau(24) and Prance,(733) it is the
tropical region that exhibits subregional diversification into distinctive faunistic
areas, largely the result of geological history and resulting topographic and soil
profiles. Prance(733) outlines in detail these subregions in terms of forest types in
Amazonia.
Geological and paleontological data indicate that South America, part of
Central America, and the Antilles were isolated from North America by a
Pacific-Atlantic marine channel during the Tertiary.(24) The land bridge of Cen-
tral America seen today arose in the Pliocene, about 2-3 million years ago and
served as a vehicle of faunal exchanges between the two continents. Evidence
from arthropods for such exchange is seen in the ectoparasites of some mam-
mals.(m Contemporary mountain tops in Central America disclose ecologically
isolated and diversified Nearctic forms as islands in the otherwise tropical low-
lands. During the Tertiary some parts of present-day Central America were
"tropical sections" of the Nearctic region.(24)
Illies(756) discusses the biogeography and ecology of some freshwater insects
in tropical mountains and lowlands, emphasizing that aquatic communities are
more buffered against fluctuating ecological conditions over short periods of time
and thereby provide a more accurate view of past biogeographical events of
larger time scales, accounting for present-day distribution patterns. The Plecop-
tera and other groups attain their greatest diversity as cold-adapted (oligos-
tenothermal) forms in tropical regions of South America but are poorly repre-
sented in the temperate region of that continent. Some 44 genera and 13 species
occur in the Neotropical region, with the Andean mountains alone accounting for
31 genera at the northern and southern ends and these faunas being separated by
the desert barrier of northern Chile and western Peru. Ten other genera are
endemic to mountains of eastern Brazil and two others to mountains of Guiana
and eastern Venezuela,(25,756) Illies(25.756) makes an interesting biogeographical
comparison of the montane Plecoptera faunas of South America and Europe. Both
regions contain a high number of monotypic genera, 15 of 32 genera in the Neo-
tropical region and 9 of 29 in Europe, and there are similar numbers of genera
with 2-3 species and very few with 4-5 species in both regions. The European
fauna also contains a higher number of species than the Neotropical region. Sim-
ilar geological or historical events have apparently determined the faunistic
complexity of Plecoptera in both regions: alternating patterns of climate during
the Ice Age promoted the interplay of isolation and subsequent breakdown of
distribution barriers after the advance and retreat of glaciers.
The European fauna in particular is characterized by many endemic species
restricted to certain mountain regions, but so is the Andean fauna.(756) Illies(756)
EFFECTS ON MAINTENANCE OF TROPICAL INSECT FAUNAS 441

suggests that high mountain areas of Europe are extremely isolated and are
positioned in an east-west axis, that is, against the axis of glacial pulsations and
parallel to the leading edge of glaciation. During periods of glaciation each of
these mountain areas was included in the movement patterns of the aquatic fauna,
but following periods of glacial retreat each of these mountain areas became a
refugium in its elevated parts resulting in isolated populations and evolution of
endemic species. But in the Andes, the mountains are positioned north to south,
the same direction as the axis of glacial movement in the Quaternary. Thus every
glacial advance thrusted the fauna toward the equator along the slopes of the
Andes, and during periods of glacial retreat the fauna recolonized extensive areas
to the south, and to a lesser degree the north, and recovered its original distribu-
tion patterns without long-term isolation. In a comparative sense, the Andes
provided insufficient opportunities to attain the degree of endemism seen in the
mountains of Europe, even though pronounced cycles in climate during the
Quaternary resulted in the appearance and dissolution of forest refugia. Illies(756)
maintains that the European montane fauna is oversaturated with species and
represents a unique situation from the biogeographical standpoint. Lowland
Neotropical areas contain a greater number of species of aquatic insects but fewer
genera, (756) indicating a strong ecological component to explaining the species
richness of these communities. Ecological processes such as competition and
niche diversification may be promoting considerable speciation in the lowland
tropics. Mani(739) provides a detailed discussion of the tropical mountain envi-
ronment as affecting the vertical distributions of terrestrial insects, stressing the
evolutionary limitations of cold, dry conditions on the upper limits of faunas and
also as affecting the availability of food resources.
The bulk of high-altitude insects, defined here as those forms above the
natural forest line, are scavengers or omnivores on a variety of organic debris,
and food chains are expected to be relatively simple compared with those in the
terrestrial and arboreal communities at lower elevations. The tropical mountain is
analogous to the high latitudes in that harsh conditions tend to select for
generalist species, dampen overall rates of speciation, and allow successful
species to develop rather large annual populations. Diurnal, rather than seasonal,
cycles in temperature, together with local topographic and climatic effects such
as rain and wind shadows, provide the major physical features of the high-
altitude environment of tropical regions influencing the distribution of insect
species. A predominance of some forms such as carabid beetles above 2000 m
reflects the suitability of such regions for species that exploit a variety of similar
and abundant resources (decaying vegetation, seasonally available seeds of
grasses, and limited predation on other invertebrates). Body-size gradients along
altitudinal transects in the mountains of tropical regions are believed to be the
result of reduced dessication stress, and perhaps physiological stress associated
with reduced rates of growth and less nutritive food resources at higher eleva-
442 CHAPTER 10

tions. Carabid beetle body size shows a gradual decrease in the mean for a given
species with increased elevation.(739) High-altitude insects in general appear to be
specialists in the sense that they occupy a stressful-environment zone that
excludes, primarily for physiological reasons rather than ecological saturation of
communities, many of the forms integrated into lowland and mid-elevation
communities. Yet within the zone, the species must clearly be generalist in terms
of niche effects: such organisms must be able to switch to different kinds of food
resources as they become available, seasonally or otherwise. Harsh environments
select for generalist species. (104) The reader is encouraged to examine the detailed
treatise of Mani(739) pertaining to the biogeographical and ecological properties of
insects at high altitudes in tropical regions.

10.3. NOTES ON ALTITUDINAL GRADIENTS IN THE TROPICS

The study of Silberglied et al. (720) of the butterfly genus Anartia in the
American tropics provides an interesting example of geographical differentiation
of an ancestral form into distinct evolutionary lineages. Silberglied et al. (720)
suggest three distinct phylogenetic lineages to Anartia, one of the most wide-
spread and common nymphalid genera over lowland rain forest, lowland dry
forest, and mid-elevation moist or wet forest habitats in Central and South
America and the Caribbean. Anartiafatima and A. amathea are a pair of sister
species that evolved from an originally widespread ancestor when populations of
this form were isolated during "the Tertiary subsidences of the Panamanian
isthmus, after which time the distinctive colors and patterns of the two species
evolved. "(720) As with other butterfly groups, there has been secondary contact
between the species in Panama in recent geological history. A second lineage
comprises the sister species of A. chrysopeles and A. lytrea of the Caribbean
islands, with distinctive isolation on islands producing the present-day species.
A. jatrophe (a third lineage) is a widespread species exhibiting considerable
geographical variation but without speciation taking place, because adults are
mobile and gene flow is high. mo )
Anartia butterflies in general possess ecological traits making many popula-
tions representative of colonizing episodes in secondary tropical habitats where
the butterflies exploit weedy species of Acanthaceae as larval host plants. (322,720)
Another system deserving of similar analysis is the apparently sister pair of
species of the genus Metamorpha, another nymphalid Neotropical group as-
sociated with weedy Acanthaceae and disturbed habitats.(294,317) In Costa Rica,
there appears to be some altitudinal separation of M. stelenes and M. epapha, the
latter species being common above 500 m on the eastern slopes of the Cordillera
Central and the former species abundant in the lowlands (A. M. Young, unpub-
lished data, 1969-1980). Both species are capable of exploiting each other's
EFFECTS ON MAINTENANCE OF TROPICAL INSECT FAUNAS 443

larval host plants (several genera and species of Acanthaceae, including some
also used by Anartia). At about 220 m in northeastern Costa Rica there is a
contact zone between the species, but with M. stelenes the more abundant.
Metamorpha in Central America is also a successful colonizer of coffee planta-
tions where the larval host plants survive beneath and along the borders of coffee
bushesY16) Detailed morphological, behavioral, and ecological studies of both
species, in the spirit of Silberglied et ai., (720) are needed to determine how
closely related the species are (wing colors and patterns differ greatly between
the two species, much more so than between Anartia species), and to what
degree populations are isolated from one another along an altitudinal gradient or
along horizontal gradients. Both species share several general features of their
life cycles and behavior with Anartia. Young and Muyshondt(316) noted consid-
erable geographical and local variation in wing color patterns and intensity of
colors in M. steienes, which appears to be the more widely distributed ofthe two
species in Central America. Tentatively it is suggested that M. epapha represents
the geographical differentiation of the genus into mid-elevation and lower
montane tropical rain forests in Central America and that M. steienes is the
lowland form, bearing mimetic resemblance to the he Ii coniine Philaethria dido,
also abundant in lowland tropical rain forests.(300) Patterns of evolutionary dif-
ferentiation of the genus Morpho along altitudinal gradients in Central and South
America(I76) also deserve detailed study to determine the probable pathways to
speciation.
The survey studies of mid-elevation and montane butterfly faunas in some
parts of Peru(757,758) suggest the need for more exhaustive studies on the distribu-
tion and ecology of individual genera and species. Although Mani(739) states that
our knowledge of Andean Lepidoptera is good,(739) few data exist on the ecology
of individual species and biogeographical distributions, save for the recent
studies of Shapiro. (75,}-761l Lamas(757) has made a survey of coastal-region
Lepidoptera in Peru west of the Andes, and such information is useful in deter-
mining the uniqueness of the endemic Andean fauna. In addition to survey-type
field studies on insect faunas across altitudinal gradients in the tropics, we need
detailed studies of single genera and species (e.g., Silberglied et ai. (720) used
such an approach for asking questions about biogeographical patterns of Anartia
in general) with an attempt to determine the environmental factors that influence
distribution over the gradient, the degree of isolation among popUlations of a
species, and the changes in communities of insects and other arthropods along
the gradient. The recent study of Noonan, (745) a detailed examination of historical
and ecological factors that determined the present distribution of forms of N.
peruviana Deyean, represents one of the few studies of its kind.
One major factor influencing altitudinal changes in distribution of insect
species in the tropics is the changes in the flora, since a majority of insects are
associated with vegetation or its decomposition products. Studies (e.g., Buckley
444 CHAPTER 10

et al. (762») have clarified the role of stressful environmental factors on the
physiology of tropical upper montane rain forest plants. Gentry(763) discusses the
transitions in plant communities along elevation gradients in Central and South
America, emphasizing the higher species richness of premontane tropical rain
forest than other formations such as tropical moist forest. Janzen et al. (412)
conducted sweep-sample studies of insects and other arthropods along an altitud-
inal transect in the Venezuelan Andes (200- to 3600-m gradient in secondary
vegetation) and found that the greatest number of insect species occurs at inter-
mediate elevation (about 1000 m) and that species richness declines with increas-
ing elevation beyond this point, with some species becoming very agundant at
the higher elevations. Janzen et al. (412) also found a pattern of reduction in mean
body size of insects with increasing elevation. Interestingly, Janzen et al. found
that Diptera and parasitic Hymenoptera are not reduced in species richness with
the proportion in reduction found in other groups, a pattern perhaps related to the
ecological requirements and body-size distributions of these two groups: both
groups are characterized by relatively small body size and therefore perhaps
would not respond as much to increasing elevation as would other groups in
which body-size range was much greater. Furthermore the larvae of both groups
may be generalists on a variety of resource types, making these insects ecologi-
cally flexible to thrive over an altitudinal gradient. Mani(739) suggests that
montane insect faunas, especially those of the upper limits, do not exhibit adapta-
tions to seasonal changes in climate as do lowland forms. Buskirk and Bus-
kirk,(523) however, found considerable change in the species diversity of insects
with the annual seasonal cycle in lower montane rain forest in Costa Rica. Upper
montane areas are presumably less seasonal than lower montane areas in the
tropics. In Papua, New Guinea, species richness of moths declines drastically
from 220 to 2800 m, and any higher than this there is no evidence of seasonal
changes in moth abundance or diversity.(513)
The colonization of high elevations in tropical regions by various groups of
insects is a function of the amount of genetic variation in already established
populations, especially if an ancestral or contemporary form is highly adaptable
to lower elevations. The evolutionary history of the species and the amount of
time passed since the montane habitat developed are additional factors influenc-
ing colonization. Noonan(745) discovered that N. peruviana occurs from sea level
to above 4100 m in South America and has large amounts of morphological
variation, and that the observed pattern of variation is accountable in terms of the
vicariance model. Mani(739) has stressed the strong transient component to the
high-altitude insect faunas resulting from wind patterns and topography. Yet
within the transect of about 500 to 2000 m, distinctive changes in the resident
communities will be associated with living vegetation, litter cover, and soil. The
studies of Campbell and Deam(76S) with grasshoppers in the mountains of Au-
stralia reveal that species occupying high-elevation regions are successful col-
EFFECTS ON MAINTENANCE OF TROPICAL INSECT FAUNAS 445

onists because a portion of the phenotype was capable of responding to the

distinctive ecological conditions of the higher areas. Clearly, as their studies
show, a great many morphological characters remain invariant over altitudinal
transects, while others show clinal variation. Therefore, it is important to distin-
guish between the portion of the existing genotype capable of responding to new
environmental conditions that may be associated with a range extension over an
elevation gradient and that portion that does not change.
Parsons and Bock(766) have discussed the relative capacities of various
species of Drosophila over temperate and tropical regions of Australia to with-
stand very different physical conditions of the environment. They have em-
phasized the local differentiation of rain forest forms to be sensitive to desicca-
tion problems, while the more cosmopolitan forms exhibit intermediate levels of
preference in such gradients as ambient temperature and humidity. Their surveys
of different forest habitats within the humid tropical zone of North Queensland
revealed that the more floristically diverse forest habitat associated with basaltic
soils supported a greater species diversity of Drosophila subgenera. The basaltic
forest is characteristic of lowland tropical forest in this region of Australia, while
higher-elevation forests are on granitic soils. (766) The range in altitudes studied
was from less than 400 m to 1600 m. Parsons and Bock(766) report the six
distinctive kinds of forest found along this gradient. They point out that all of the
major subgenera (tive) of Drosophila are represented by several species along
this gradient from the humid tropical zone, while the nontropical fauna of south-
ern Australia is characterized by the predominance of one subgenus. The tropical
Drosophila fauna is maintained by different subgenera and species exploiting
different kinds of food resources.(766) Miskimen(490) studied the relationship be-
tween altitude, time of flowering, and appearance of soldier beetle populations
on flowers along an altitudinal transect in Colombia. Timing of flowering is
increasingly postponed with increasing altitude for the various Compo sitae used
by soldier beetles (Chauliognathus proteus), and the beetles appear pro-
gressively later on the flowers with increasing altitude. In populations found in
the cool, moist highlands, beetles are darker in color than in the lowland tropical
rain forest zone, presumably to increase heat absorption under cooler conditions.
In moving from the tropical lowlands into upper montane rain forest, insect
species richness bulges at mid-elevations and then drops off, as seen for com-
munities in secondary vegetation.(412) Decrease in species richness is the result of
the interplay of (1) increasingly stressful thermal and moisture conditions with
higher altitudes and (2) changes in the distribution and abundance of resources.
Thus zones of upper montane forest on tropical mountain tops are similar to some
Temperate Zone areas at the higher latitudes. Within the transect from lowland
tropical rain forest to upper montane forest, herbivorous insects probably remain
as specialists on certain groups of plants and the trophic organization of com-
munities is not expected to change. An overall lowering of species richness in
446 CHAPTER 10

some groups above mid-elevation habitats may allow some species to become
capable of thriving in these zones. Beyond the upper-montane forest zone, insect
species richness probably drops off even further, and a shift in community
structure, with an emphasis on scavenger and decomposer species, takes place.
The feeding habits of high-altitude insects, defined as forms living above 4000 m,
have been discussed at length by Mani.(739)
Wille(67 ) discovered that populations of the dung beetle Megathoposoma
candezei, a species associated with human feces, decline markedly in size below
600 m in Costa Rica. In regions of the tropics where species richness is high in
the lowlands, the populations of individual insect species for some trophic posi-
tions and form some groups of insects in certain kinds of habitats may be
considerably smaller than populations of the same species at higher elevations,
where corresponding patterns of species richness may be reduced. In order to
examine such relationships between species richness for, say, the communities of
dung beetles, and the sizes of populations of individual species along altitudinal
gradients in the tropics, it is necessary to determine whether an optimal habitat or
microhabitat for each species changes in frequency along the gradient range
occupied by the species as well as the size and complexity of the community
along the gradient. If a certain kind of pasture habitat preferred by a particular
species of dung beetle occurs in greater frequency at higher altitudes than in the
adjacent lowlands, the beetle may become more abundant at higher altitudes,
especially if other species of that community drop out along the gradient. Alter-
natively, if the preferred habitat does not increase in abundance but the beetle
does, the explanation is related to changes in the community arising from other
effects of altitude.

10.4. BIOGEOGRAPHICAL PATTERNS OF THE ORIGINS OF

TROPICAL INSECT FAUNAS

Of profound ecological significance in determining the composition and

structural diversity of tropical floras and faunas have been the interglacial and
glacial cycles of the Quaternary. (20],469.768) It now appears that the distribution of
species richness in tropical insects has been generated largely by the distribution
of plant communities and the cyclic effects of expanding and contracting forest
refugia and other forms of vegetation (e.g., savanna patches) within the lowland
and lower montane tropical rain forest zones of Central and South
America.(733,769) Communities or some other unit of whole assemblages of or-
ganisms are believed to oscillate in diversity and structure as a function of such
geological cycles, but an additional perturbation within the present glacial-
interglacial climatic window is man's alteration of the vegetation cover in the
tropics. (770) Such a perturbation, widespread and expanding, may result in a
EFFECTS ON MAINTENANCE OF TROPICAL INSECT FAUNAS 447

further decline of species richness than otherwise anticipated as a natural result of

the glacial cycle.(770) The impact of man on the mammalian faunas of southern
Africa in the geological past(770 attests to the diversity-lowering impact of human
activities on faunas.
The generally rich biota of Amazonia is believed to have been the result of
the periodic pulsation of the physical environment arising from cycles of climatic
change within the past 3 million years(769): the periodic fragmentation or
vicariance of forests into restricted and isolated refuges (core areas), each within
a pulse and among successive pulses having a distinct sphere of selection pres-
sures and intensity of genetic drift effects. Given sufficient time, the refugia
hypothesis states that such a pulsation of habitats would result in the evolutionary
radiation of new species or subspecific forms. Depending upon the size and
proximity of different refugia during anyone pulse, subsequent periods offorest
expansion (amalgamation of refugia) would give rise to patterns of endemism
and clinal variation. The pulsing of the overall physical environment in the
Neotropical region probably resulted from Pleistocene and possibly past-
Pleistocene climatic changes.
Livingstone(769) discusses the lack of an adequate pollen record to prove the
existence of refugia cycles in the Neotropical region, but others(469.62r624) have
presented convincing though inconclusive evidence from animals, including but-
terflies. Suspected or determined tropical centers of endemism are inconsistent
with marked and recent changes in the Neotropical flora as revealed by pollen
analysis, and it is unlikely that all endemics have evolved since the beginning of
the present interglacial period about 10,000 years ago.(769) It is also unlikely that
rates of range extension of endemics from refugia or core areas are about the
same for all species. The localized distribution of many plant endemics as seen
today suggests that their popUlations were perhaps too small to show up in pollen
analysis, even though present.(769) The Neotropical region is characterized by
both lowland (Amazonian) refugia and montane refugia on the slopes of the
Andes, with the axis of migration of groups in the latter being vertical along
temperature and moisture gradients resulting from the glacial-interglacial cy-
cles.(469) Lowland refugia were characterized primarily by lateral or horizontal
expansion or contraction. The high Andes became available for colonization by
plants and animals since the end of the Tertiary and with considerable altitudinal
and high-altitude lateral differences at different locations along the Andes
chain.!2ot> High-altitude environments provided refugia conditions promoting geo-
graphical isolation and evolutionary radiation, and plant species experienced
large-range extensions . 'during glacial periods when vegetation zones were low-
ered. "(200 Simpson(201) discusses the differences in pathways of range extension
and differentiation in different parts of the Andes, but all areas had interglacial
periods characterized by the isolation and differentiation of species. As revealed
by the studies of Brown,(623-626) Brown and Benson,(204) Benson,(282) and
448 CHAPTER 10

Turner,(621,772) the effects of shifting distributions of lowland tropical rain forest

on the population dynamics of larval host plants (Passifloraceae) have been a
major factor in understanding the present-day geographical distribution of
heliconiine butterflies, most notably Heliconius, in Central and South America.
Similar effects are expected for other groups of plant-associated insects in both
lowland and montane regions of the tropics.
The widely accepted view that favorable ecological conditions in the low-
land tropics promote high species richness(lS) is reconcilable with the concepts of
uneven distributions of species or groups of species in accordance with the forest
refugia model. On the one hand, the ecological mechanisms promoting high
diversity in the tropics (predation, competition, niche diversity, etc.) provide the
local opportunity for species to diversify, given a sufficiently complex habitat or
collection of habitats and isolation from other such collections of habitats. Yet
superimposed upon the trend for the promotion of species richness through local
ecological mechanisms is the idea of discontinuities in environment, com-
munities, and individual species ranges due to cyclic geological (historical)
events such as climatic changes promoting Quaternary forest refugia in
Amazonia and in the Andes.(469) As discussed in Chapters 5, 7, and 8 the
explanation of how high species richness for particular taxa is maintained in the
tropics involves an analysis of local ecological processes determining the struc-
ture and dynamics of guilds, species assemblages associated with specific kinds
of resource patches, and whole communities and the resulting trophic interac-
tions between different communities forming an ecosystem. Yet we also need to
consider the regional patterns of species distributions and recognize that such
distributions are also a function of regional environmental or geological events
promoting or discouraging the associations of complex species systems. Thus the
fact that the Neotropical region contains a very high number of endemic sphingid
moth genera and species(7S1) is explained in part by the ecological refugia model,
which accounts for why genera and species are diverse in some regions and not in
others. But there is also the ecological explanation: sphingid species within an
area partition the environment in distinct ways (seasons, larval host plants, adult
hosts).
Turner 772 ) has outlined the major pathways of adaptive radiation of the
heliconiine butterflies, particularly Heliconius, that occurred in South America
during Pleistocene periods when forest refugia were in existence. Species pairs
such as H. melpomene and H. cydno apparently diverged as distinct species in
different, isolated forest refugia, and subsequently, when the forests expanded
again, developed overlapping ranges so that today one species (H. cydno) is
essentially a forest species while the other occurs at forest edges and in secondary
succession. (329) Periods when forests expanded provided the opportunity for con-
siderable range expansion of these butterflies, as seen today in the widespread
distribution of some species over Central and South America. Many other
EFFECTS ON MAINTENANCE OF TROPICAL INSECT FAUNAS 449

species, however, exhibit restricted distributions in South America, believed to

be the result of forest refuge cycles. Some areas support a very high number of
species, apparently as the result of past geographical expansion of the heliconiine
faunas of four or five surrounding and now-obliterated forest refugia.(622-624)
Concurrent with such effects is the likely differentiation of larval host plants
(Passifloraceae in this case) within different forest refugia, so that when faunas
and floras subsequently expanded, areas of overlap received a high diversity of
host plant species as well. (329) Gilbert and Smiley(329) stress the need to separate
out such historical effects from ecological factors in attempting to explain local
patterns of heliconiine butterfly diversity.
Brown et al. (622) provide data on the distribution of races in the butterfly
genus Heliconius in South America that support previous studies regarding the
hypothesis of forest refugia in the recent geological past determining the patterns
of race formation in other organisms.(469,62L772l The various races of H. mel-
pomene and H. erato exhibit rigid parallel variation throughout South America
as do 10 other similarly patterned species, as the result of strong stabilizing
selection favoring Mullerian mimicry. The striking discovery here is that the
butterflies are largely confined to forest refugia of the recent geological past and
that the precise mimetic patterns evolved in different refugia are expected to be
different because ecological conditions are not the same in all refugia. The
existence of mUltiple species Mullerian mimicry complexes, and the phenotypic
differences among species or races of butterflies in the mimicry complexes of
different original refugia, reflects the isolation of the refugia in the past and does
not necessarily reflect a pattern of geographical variation sensu stricto. (622) The
studies of Heliconius and ithomiine butterflies in South America(622-626) demon-
strate how varying ecological conditions among different isolated refugia in the
geological past have shaped the evolution of individual species, and whole
mimicry complexes. (72 ) Brown(622-624) recognizes five to nine endemic centers of
mimetic color patterns for Heliconius and ithomiines, reflecting long evolution-
ary periods of specialized ecological adaptation to local environmental factors
(soil types, host plants, etc.). Complex "subcenters," as many as seven on the
slopes of the Andes also occur, showing high numbers of endemic species, ex-
plained by ecological differences rather than geographical patterns per se. (622-624)
Brown(623J also discusses that some species are good at dispersal, thus
breaking down any previously existing discontinuous patterns of distribution
conforming to the expected patterns from the forest refugia model. These species
pass through barriers effective for other species, becoming' 'corridors for disper-
sal" and resulting in such species having less distinct boundaries associated with
environmental barriers. The problem is complex because different species of
Heliconius will exhibit different responses to the same barriers, a function of
behavioral patterns as well as population growth parameters also being different.
BroWn<622,623) noted an inverse relationship between diversity of species and
450 CHAPTER 10

degree of endemism, suggesting that contemporary ecological rather than histori-

cal factors often determine patterns of co-occurrence among species. Thus the
so-called paleoecological refuge model cannot be used to explain all distri-
butional patterns, even when the organisms are species that exhibit some patterns
conforming to the inferred past distribution of forest refugia.
Brown(478) has analyzed the geographical patterns of subspecies and species
of the ithomiine butterfly genera Melinaea and Mechanitis, showing that mime-
tic affinities and patterns of localization are clear for Melinaea and less so for
Mechanitis. A detailed summary of the analysis of the evolutionary radiation of
Heliconius butterflies in relation to the forest refugia model is given in
Brown,(468) in which it is recognized that some species and races are explained by
the inferred positions of forest refugia and core areas with geographical isolation
producing these forms, while others have arisen as the direct result of secondary
contact and creation of zones of hybridization following isolation in refugia.
The examination of the geographical distribution patterns of subspecies and
species of Heliconius, together with information on life cycles, larval host plant
needs, and behavior,(203) represents a synthetic approach to understanding the
role of biogeographical and ecological factors in molding distributions of species
and races spatially. In the case of many species of Heliconius, the maintenance
of local and geographically varying polymorphisms within a species has been due
to strong local stabilizing selection favoring Mullerian mimetic association with
other species and ithomiines.(204) Thus Heliconius numata is unpalatable and
highly polymorphic over much of its extensive geographical range in South
America, with as many as seven distinct morphs co-occurring in the same popu-
lation. Such a pattern of local variation is not expected for a Mullerian mimic
since strong stabilizing selection would favor one phenotype or form to be
associated with mimics of other species. But Brown and Benson(204) explain this
unusual condition by the existence of many species of abundant ithomiines, also
unpalatable, but with local habitats having sufficiently abundant species to inter-
act mimetically with the various morphs of H. numata. Thus the complex
polymorphism of H. numata is maintained by the very heterogeneous as-
semblages of ithomiine butterflies. The system provides an example of how a
rare species of unpalatable Mullerian mimic is maintained through polymorphism
favored by interaction with abundant and unpalatable ithomiines.
Although not all groups of insects attain their greatest taxonomic diversity in
the tropical regions, as exemplified by Michener's recent biogeographical
analysis ofbees,(J67) that most insects do attain great species and generic richness
in the tropics is likely the result both of historical and of contemporary ecological
conditions. The most convincing piece of evidence relating to why insects are so
diverse in the tropics has to do with the great diversity of tropical floras and the
timing of major periods of evolutionary radiation in relation to insect evolution.
The tropics support the greatest diversity of pollinating organisms and flower-
EFFECTS ON MAINTENANCE OF TROPICAL INSECT FAUNAS 451

visiting organisms of any region, a factor that may contribute to the lower
representation of bees in the tropics (many plants are pollinated in the tropics by
butterflies, moths, flies, beetles, bats, hummingbirds, mice, etc.). Yet insects
also represent the largest group of herbivores in the tropics. Small body size,
high population growth rates, short generation times, genetic and phenotypic
flexibility-these are probably some of the major properties of insects making
them successful herbivores in the tropics. A plethora of studies of plant-insect
interactions in the tropics in recent years has demonstrated the role of such
interactions in shaping the structure of tropical plant communities. And although
hypotheses such as the paleoecological refuge model of regional variations in the
generic, species, and subspecific richness of tropical insect (and other organisms)
faunas help to explain why such taxonomically rich areas of the tropics are often
discontinuous geographically, equal emphasis, in attempting to explain the great
taxonomic diversity of most insect groups in the tropics, should be given to the
analysis of contemporary ecological interactions, particularly through the lenses
of present-day and past coevolutionary properties of species interactions.

10.5. SOME MECHANISMS OF ECOLOGICAL AND EVOLUTIONARY

DIFFERENTIATION

Mayt 773 ) discusses the role of genetic and phenotypic variation in determin-
ing the ability of a species to thrive in a range of environments. A consideration
of why species of tropical insects are widespread across certain kinds of en-
vrionmental gradients while others, even ecologically similar ones, are far more
restricted in distribution, necessitates an understanding of the relationship be-
tween adaptive or phenotypic flexibility within breeding populations of the
species, the range of environments available for colonization, and the evolution-
ary history of the species. The ability of a species to become a resident of a given
community within a particular habitat will determine the overall geographical
distribution of the species, or its distribution relative to core areas or refugia if
that sort of analysis is appropriate. The reason for this statement is that the
geographical range of a species is really a compilation of different habitats
capable of being occupied by the species to varying degrees. The degree (i.e., a
percentage rating of successful colonizing episodes might be one measure of
"degree' ') to which a species becomes a member of a given community within a
potential or realized geographical range depends upon (1) the constellation of
physical envrionmental parameters affecting the ability of different genotypes to
feed, mate, and survive, (2) the interactions of the species with others in the
community, including competitors, predators, and mutualists, and (3) the
amount of genetic and phenotypic flexibility represented by the range of indi-
viduals in the population as a means of "fitting into" the community. The degree
452 CHAPTER 10

to which the niche of the invading species is already fully occupied by another
species, the degree to which the incumbent species is a specialist for that niche,
and the degree to which the invading species is a specialist for the same niche
will determine the ability for the species to enter into the community. The
success of invading species may also be a function of the relationship between
competitive ability and density or abundance of participating species: sometimes
competitive ability is inversely related to density of competing species.(81,82)
Heterozygous loci have been considered the basic mechanism by which a
species can adapt to changing conditions of the environment, both spatially and
temporally. (773) Constant environments may result in a gradual loss of
heterozygosity in Drosophila, (774) although it is not known if, in general, popula-
tions in uniform tropical environments are essentially homozygous. The problem
of defining a uniform or constant environment is largely one of the units of
measurements chosen: the environment might appear to be uniform overall, but
finer resolution may reveal patterns of microheterogeneity. (775) Zones of hybridi-
zation between species may provide new genetic variation that permits one or the
other of the species to enter into a new adaptive zone. (76) When formerly allopat-
ric species come into secondary contact, distinct phenotypic patterns will be
expressed within that part of the geographical range where the contact is occur-
ring(776); such patterns can usually only be interpreted through the inference of
secondary contact, which mayor may not eventually result in parapatric specia-
tion. Given sufficient genetic variation, selection can modify ecological traits
such as competitive ability or developmental time(232,777J as well as the ability to
adapt to specific temperature conditions. (86 ) The data of Levins(386) illustrated
how different species of Drosophila partition the environment in Puerto Rico in
terms of differential genetic and behavioral preferences for ambient temperature.
Levins showed that broad-niched species are successful in adapting to a range of
temperature conditions by high individual flexibility rather than by the genetic or
evolutionary divergence of populations across the range of environments studied.
Widely separated breeding popUlations of the same insect species are not neces-
sarily genetically different,(764) presumably since behavioral and physiological
systems of flexibility buffer populations against environmental fluctuations.
The pattern of resources in the environment, both temporally and spatially,
will determine the degree to which two or more ecologically similar species are
juxtaposed into competition if the resource is limited in supply, the two or more
species are specialists on it, and the geographical ranges of the species overlap.
Where many ecologically similar species co-occur in the same area, as seen for
many groups of insects in the tropics, it is necessary to examine patterns of niche
partitioning arising from competition and related biotic interactions and those
arising from the interaction of phenotypic constraints within local populations of
each species with physical features of the environment.
EFFECTS ON MAINTENANCE OF TROPICAL INSECT FAUNAS 453

Thus the evolution of thermal acclimation mechanisms in subtropical or

tropical Drosophila(386) may be an adaptive response to the phsyical environment
having little or nothing to do with interactions among species. If a species is
capable of exploiting a range of resources locally or regionally, the observed
restriction of the species to a portion of the resources may have resulted from
"interference" competition, in which the presence of other species on some of
the resource patches leads to an active displacement among species resulting in
microallopatry ,(778) instead of direct exploitation-type competition,(779) in which
each species actively removes resources that would otherwise be used by the
other species.
The structuring of insect communities in the tropics is highly dependent
upon the organization of guilds of closely related genera or species exploiting the
same resources or a narrow spectrum of resources within habitats. Some tropical
communities are characterized by the co-occurrence of several closely related
species on the same resources, and such assemblages(780) may be organized by
the level of competitive interaction between species if resources are limiting and
populations approach the carrying capacity. Mayt773 ) discusses the various kinds
of mechanisms that allow several closely related species to co-occur without the
formation of inviable hybrids from cross-matings. In some insects, reproductive
isolation is mediated by visual discrimination systems precluding copula-
tions,(781) while in others ecological mechanisms and other forms of behavioral
mechanisms reduce the likelihood of cross-matings. Much of the integrity of
species richness in tropical insect communities depends upon the effective opera-
tion of reproductive isolation mechanisms in areas where species are sympatric or
microsympatric.
It is the interplay of ecological interactions between species in contemporary
time, incorporation of such ecological traits into whole communities over evolu-
tionary time, and the impact of geological history that determines gradients in
species richness in the tropics. When environments are stressful or rigorous in
terms of physical fluctuations in certain parameters or rather uniform conditions
of a harsh nature, biotic communities are thought to be physically controlled,004)
a result of which is the variable success for invading species to establish them-
selves in such communities.(632) Such situations may be generally characteristic
of the very high latitudes and also high altitudes in the tropics. When environ-
mental conditions are benign and predictable, there may be an evolutionary
accumulation of highly coevolved species in local communities, as exemplified
by the association of insects with the forest understory tree T. cacao, but with an
overall increase in species richness being asymptotic,(634) One of the major
, 'problems" with the control of insect pests on tropical crops such as cacao and
coffee is the continual availability of a large pool of species associated with
various kinds of natural vegetation, providing for replacement of specific pests as
454 CHAPTER 10

they are "eliminated" through ineffective (in the long term) pest control pro-
grams. Some regions of the tropics tend to accumulate a greater number of
species of insects than expected in terms of predicted levels of species saturation
in communities: such areas are ideal candidates for continual outbreaks of insect
species on crops (Amazonia, Southeast Asia, tropical Australia). It is the ex-
ploitation by native, coevolved insect herbivores that limits the distribution and
abundance of native plant species (populations).
Thus the geographical distribution of plant species can be greatly influenced
by the intensity of herbivory from insect species that are specialized feeders on
particular plants. In the mid-elevation and lowland tropics, such effects are
expected to be intense for most dicotyledons. Both insect and vertebrate herbi-
vore species may exhibit considerable variation in plant diet along altitudinal
gradients in the tropics(782) not necessarily related to competitive exclusion or
ecological release in certain zones along the gradient. Plant communities in
montane localities in the tropics are sometimes greatly influenced by periodic
natural catastrophes over long periods of time (i.e., landslides and volcanic
eruptions) that cause large-scale changes in the floral composition of forests and
postponing the natural sequence of succession. (783)
It is therefore the highly-coevolved interaction of insects and other herbi-
vores with plants, highly coevolved with certain groups, geological and climatic
history, and the periodic reoccurrence of regional catastrophes, that determines
the composition of plant communities within the montane section of altitudinal
gradients in the tropics. Other ecological processes, such as interspecies competi-
tion and character displacement, place upper limits on the overall species rich-
ness found in a certain kind of habitat,(lS) especially when populations of compo-
nent species within a given trophic level are approaching equilibrium conditions.
Switches in host plant exploitation patterns within a widely distributed insect
herbivore species along an altitudinal gradient or other form of environmental
gradient may be promoted by shifts in community structure, emphasizing or
de-emphasizing competition, shifts in distribution and abundance of the host
plant, changes in the physical environment affecting the activity patterns of the
insect, and large-scale alterations in the habitat.
Whether or not the insect can successfully switch to another host plant
species is a function of the degree to which feeding behavior is regulated primar-
ily by phenotypic flexibility (behavioral, physiological) or by genotype. The
existence of a "host selection gene" may determine feeding flexibility in the
sense that the homozygous condition may promote feeding on only one kind of
host plant, the alternate homozygous condition (assuming a two-allele system)
favoring feeding on the second kind of plant, and the heterozygous condition
favoring the flexible feeding condition. <390 An insect popUlation may contain
different "plant feeding" genotypes in the sense that each genotype is capable of
promoting feeding on a particular plant species. When an introduced plant
EFFECTS ON MAINTENANCE OF TROPICAL INSECT FAUNAS 455

species (an exotic) is juxtaposed in the same region or habitat with other native
species, genotypes associated with the native species may shift to the introduced
species if it is closely related to the native species. G. L. Bush suggests the occur-
rence of "host plant races" that may differ, in the specific case of apple fruit flies
associated with fruit tree species, in terms of evolutionary divergence in the
intestinal bacteria flora in the guts of fly larvae. In some groups of Old World
tropical ants, polymorphic genera such as Dorylus in the subfamily Dorylinae
exhibit considerably higher behavioral flexibility than closely related
monomorphic genera such as Aenictus in terms of the degree of specialized
foraging behavior or lack thereof.(430) Species of Aenictus are specialized pred-
ators on other ants in tropical habitats. (430) Thus for tropical insects in different
trophic roles, polymorphism tends to result in more flexible behavior in terms of
the breadth of resources utilized, but such patterns are expected to vary consider-
ably in different communities, habitats, and along environmental gradients over
large areas of terrain, because selection pressures will also vary.

10.6. SOME EXAMPLES OF DIFFERENTIAL ADAPTATION TO

TROPICAL LOWLAND AND MONTANE AREAS BY INSECTS

The Neotropical carabid beetles (Carabidae) have provided the opportunity

for the development of a rather comprehensive model explaining the complex
relationships between evolutionary radiation within the family, dispersal, and
climatic shifts.(417l Historically, carabids were of particular interest to biogeo-
graphy because of the tendency toward flightlessness in some mid-elevation,
montane, and high-latitude forms. (753) Flightlessness in carabids is believed to be
related to the degree of stability of particular geographical regions, since oth-
erwise this phenotypic trait would lead to extinction in regions of environmental
instabili ty .(417)
Flightlessness is found in insects occupying some isolated but stable habitats
such as mountain tops. The likelihood for the evolution of flightlessness in
insects under such conditions may involve a switch in the allocation of resources
for other phenotypic traits, particularly those related to reproductive potential.
An insect population may segregate out from each generation a certain portion of
"dispersal prone" and "nondispersing" genotypes. As the number of
dispersal-prone genotypes is reduced by emigration and if no new immigrants
arrive in the population, there is a net decrease in the abundance of dispersing
individuals and population size declines.<7(3) The result is an increase in the
proportion of nondispersing genotypes, and if the environment is stable, direc-
tional selection favors the nondispersing genotypes and subsequent stabilizing
selection promotes the evolution of phenotypic traits that optimize adaptation to
the local and stable environment. There is a switch in the allocation of energy to
456 CHAPTER 10

the development of flight systems to other traits such as reproduction if selection

favors the evolution of phenotypic traits that optimize local adaptation. Depend-
ing upon the degree of isolation, the size of the stable environment, the size and
distribution of the insect population, and the distribution and abundance of re-
sources, selection may favor the retention of some dispersal-prone genotypes,
since under some conditions local dispersal would be selected for. The retention
of such genotypes, even at low frequency, would provide the mechanism for the
species to emigrate should the environment suddenly change. Slatet424) presents
evidence from African populations of lygaeid bugs that flightless forms within
flight-polymorphic species are more frequently encountered in stable habitats.
More recently Noonan(74S) found that individuals of N. peruviana have short
wings where habitats are favorable and continuous and longer wings where
otherwise. The data support the "checkerboard" hypothesis of Darlington.(784)
Thiele(78S) has summarized in some detail the major adaptive features of
carabid beetles, with an emphasis on the relation of body form to foraging habits,
habitat associations, and trophic positions in communities. The great evolution-
ary and ecological diversification of the group, particularly in the tropical re-
gions, attests to the versatility of carabids to exploit a range of environments. The
carabids are hypothesized to have arisen in the Mesozoic with a "rapid pulse of
radiation and diversification" with the early lineages being associated with river
and stream edge habitats.(417)
The periodic and predictable cycles of inundation in Amazonian forests
provided subsequent selection for invasion in evolutionary time of mesic forests,
leading to another pulse of adaptive radiation, and other pulses in diversification
paralleled stages in the evolution of tropical forests, notably in terms of vertical
structuring of forests and the invasion of arboreal microhabitats (and canopies)
by carabids.(417) Subsequent geological events altering the climate of Amazonia,
such as cycles of Pleistocene climatic change, coupled with the dispersal abilities
of lineages, resulted in complex geographical distribution patterns. (417) Thus the
general pathway of carabid beetle evolution in South America consisted of an
origin in tropical wet lowlands where the insects, as ecological generalists, were
placed in competition with ecologically similar but unrelated forms such as ants,
with subsequent radiation and diversification primarily into (1) mesic forests with
exploitation of arboreal microhabitats and from here into mid-elevation (high-
land) and montane forests and mountain tops, and (2) as specialists into forest
environments at other latitudes from the wet lowlands and also into other kinds of
habitats. Erwin(417) points out that carabid beetles are most diverse generically in
the tropical lowlands and mesic forests, with greatest species richness occurring
in specialized environments such as mountain tops and tropical forest canopies
and in north and south cold temperate habitats. Erwin's model for the evolution
and diversification of carabid beetles stresses the concept of generic and species
richness being driven by fluctuations in the physical environment, coupled with
EFFECTS ON MAINTENANCE OF TROPICAL INSECT FAUNAS 457

ecological processes in local communities, resulting in strong local selection

pressures that leads to within- and between-habitat diversification in taxa. His
model may be more generally applicable for other groups of insects with defini-
tive Neotropical origins but with subsequent highly cosmopolitan distributions of
genera and species.
Erwin's model accounts for the tropics being rich in genera and species of
carabid beetles, even though populations may be very patchy in space, particu-
larly in lowland and mid-elevation zones. Darlington<753) recognized the impor-
tance of local ecological interactions affecting carabids as a major source of
selection pressure in terms of how different genera and species respond to tem-
perature and humidity gradients in tropical wet forests. Darlington<753) originally
developed an application of the taxon cycle to the evolution of carabid beetles in
the tropics, stressing the origin of forms in the tropical wet lowlands with good
powers of dispersal into other habitats (as discussed much later by Erwin(417) and
the eventual evolution of other dominant forms that subsequently colonize other
habitats or reinvade lowland areas where previous forms have become extinct.
Both Darlington<753) and Erwin(417) have developed conceptual models to explain
the geographical distribution, local and regional diversity, and patterns of
morphological, behavioral, and ecological differentiation as related to patterns of
climatic change and changes in forest communities.
Jeanne(410) has found that ant species in the tropics exhibit a great degree of
microhabitat specificity and that many species in lowland rain forests are canopy
specialists. Such results reveal that ants have been a major source of selection
pressure for other groups of insects, e.g., carabid beetles, in similar mi-
crohabitats. Tropical soil arthropod faunas are very complex in terms of species
richness,(786) more so than in the Temperate Zone. Erwin<417) mentions the impact
of ants in tropical forests on the evolution of carabid beetles, and one might go so
far as to say that in the lowland tropics, where there is a high level of generic and
species richness for many groups of invertebrates, there are probably many
examples of the "joint evolution" of whole groups such as ants and carabid
beetles: the ants are a selection pressure in the sense of competing for the same
microhabitats as carabids, and the carabid beetles are a likely source of selection
pressure for ants, although probably less so than vice versa. The high species
number of carabids at higher altitudes in the tropiCS(417) and the corresponding
decline in ant species along such gradients may be indicative of a compensatory
replacement of whole groups in the context that carabids have been more suc-
cessful at colonizing high-altitude habitats than ants
As for most groups of insects in the tropics, we need to examine the
adpative significance of subtle differences in morphology, behavior, and ecolog-
ical traits between taxa of carabid beetles occupying the same communities and
to extend such an approach to the analysis of how such communities change
geographically.(787) The fact that carabids exhibit gradual reductions in mean
458 CHAPTER 10

body size from the lowlands to mountain tops in the tropics(739) has to be exam-
ined in terms of the relationship of body size, energy allocation strategies, and
effects of the physical environment and the interactions with other organisms that
occur along such gradients. Explaining gradients in body size strictly in terms of
physiological constraints associated with temperature and moisture does not re-
ally elucidate the adaptive significance of such morphological patterns to interac-
tions with other organisms in communities. If carabid beetle and other insect
popUlations on tropical mountain tops are characterized by a relatively larger size
(abundance) than popUlations of the same or closely related species at lower
elevations, we need to critically examine the strategy of resource allocation in
such differing popUlations. This strategy in each community relates to the distri-
bution and abundance of resources; the quality of. resources as affecting fitness;
and the selection pressures arising from competition, predation, and other
ecological interactions in each habitat. It is necessary to determine the role of
heterogeneous gradients of abiotic factors along with biotic factors in determin-
ing the local assemblages of groups such as carabids and how these factors
change along altitudinal gradients. Specialized feeders such as termites in the
lowland tropics, insects for which a specialized feeding habit and moisture con-
straints may limit geographical range,(788) playa major role in nutrient cycling.
The degree to which an insect entering into a new adaptive zone such as a
tropical mountain top can differentiate in response to local selection pressures is a
function of the degree of genetic and phenotypic flexibility within the po pula-
tiod 789 ) and the amount of time that has passed since the invasion occurred.
Perhaps some of the most interesting data on such properties of adaptive re-
sponse, and constraints upon it, come from the recent studies of Shapiro(472,759-
761l on Andean pierine butterflies. As Shapiro(472) points out, little is known about
the ecology and evolutionary history of the unique genera of pierines associated
with the high Andes. Shapiro(759) has been most interested in examining the
hypothesis that the Andean pierine fauna was derived from Nearctic stock during
the Pleistocene glaciation period. From his studies of life histories in some
selected genera in the Andes(472,761l he has discovered some evidence suggesting
divergence from the Nearctic stock. Most interesting, however, is the relation-
ship between geographical region and the capacity of pierines to exhibit wing
color pattern polyphenisms in response to changes in photoperiod regimes.o 60l
Most extratropical pierine butterflies are multivoltine and exhibit distinctive
seasonal phenotypes and a facultative diapause, both controlled by photoperiod.
Some univoltine and monophenic Holarctic species exhibit polyphenisms when
experimentally induced under laboratory conditions, a phenomenon Shapiro calls
"latent polyphenisms." Such phenotypic flexibility suggests that these species
are derived from multivoltine and polyphenic ancestors. C76Ol
In the Andes of South America, some endemic pierine buttertlies are both
multivoltine and monophenic but cannot be induced to exhibit polyphenisms under
EFFECTS ON MAINTENANCE OF TROPICAL INSECT FAUNAS 459

experimental conditions as can some monophenic Holarctic forms.(759) Yet it is

believed that these Andean forms are derived from Nearctic stock during the
Pleistocene. The lack of phenotypic flexibility suggests that either they have
secondarily lost the ability to exhibit plasticity in response patterns to photo-
period or they never had the condition which calls for a reexamination of the
hypothesis that these pierines are derived from Nearctic forms.
Inherent in the first explanation is the idea that since Pleistocene glaciation
is relatively recent, the time for complete differentiation from the Nearctic stock
should have been sufficient and that some portion of the phenotype should
exhibit relict traits such as latent polyphenism in response to changes in photo-
period. Since the apparent invasion of the Andes by pierines occurred at about
the same time as the invasion of the Andes by their larval host plants, the
Cruciferae,(790l one is tempted to accept the first explanation, although more data
are needed on the ecology of Andean pierines and that of other groups of insects
with similar evolutionary and biogeographical histories.
The examination of phenotypic traits of butterflies such as voltinism and life
history and of responses of the phenotype to environmental factors such as
photoperiod is a methodology that can be used to assess the degree of evolution-
ary and ecological differentiation in other groups of insects along latitudinal and
altitudinal gradients, with an emphasis on comparing temperate and tropical
faunas. Scotf 79Jl has discussed the apparent, recent partial differentiation of the
butterfly fauna of the Antilles, with derivations from Central and South America
and similarities with that of Florida. Comparative studies, in the spirit of
Shapiro,(760) on the divergence in ecological traits and phenotypic flexibility or
lack thereof, between the island and mainland popUlations for some of these
genera and species would be particularly instructive in explaining the success of
invasions into the archipelago.
The determination of the evolutionary pathways of divergence in relation to
geographical patterns,(792) across either latitudinal or altitudinal gradients (both
situations exposing species or whole complexes of species to sets of very
heterogeneous environmental conditions) and including biotic and abiotic com-
ponents of ecosystems, constitutes the approach needed to explain not only the
origins of tropical insect faunas but indeed the origins of such groups on a world
wide basis. Both latitudinal and altitudinal gradients expose species or whole
complexes of species to heterogeneous environmental conditions, resulting in
some phenotypic traits being greatly modified while others remain stable (relict).
Natural-history and ecological studies emphasizing the subtle differences in
form, behavior, and physiological response systems among co-occurring groups
such as a species complex, supplement biogeographical analyses in undertaking
such a task. Yet the data base from such an approach for most groups of insects in
the tropics is very meager. We need detailed data on trophic organization in
ecosystems, including the determination of rather precise resources needed by
460 CHAPTER 10

species in certain habitats at a given time of the year, as exemplified by

Greathead.(793) Little is known about the impact of vertebrates on insect popula-
tions, and the data of Hespenheide(794) indicate that such predation can be heavy
in some habitats. Such studies are needed to explain short-term fluctuations in
tropical insect communities and how, along with fluctuations in the physical
environment, specific kinds of biotic interactions change or remain stable geo-
graphic all y .
Much of the foregoing discussion, and the preceding chapters, emphasizes
the need for biologists to consider both contemporary ecological factors and the
historical or biogeographical effects in understanding the distribution and abun-
dance of insects in the tropics. A comprehensive understanding of both the broad
geographical distributions of species or whole groups of species, and the distribu-
tion patterns of species associated with different habitats, in the tropics, requires
the total view, namely, that of the ecologist considering the historical perspective
and the biogeographer gathering as much ecological data as possible.
 REFERENCES

I. R. C. Lewontin, The Genetic Basis of Evolutionary Change. Columbia University Press, New
York (1974).
2. R. C. Lewontin and J. L Hubby. A molecular approach to the study of genic heterozygosity in
natural populations. II. Amount of variation and degree of heterozygosity in natural populations
of Drosophila pseudoobscura. Genetics 54:595-609 (1966).
3. R. Levins, Evolution in Changing Environments. Princeton University Press. Princeton, N.J.
(1968).
4. R. Patrick, A discussion of the results of the Catherwood Expedition to the Peruvian headwaters
of the Amazon, Verh. Int. Ver. Theor. Angew. Limnol. 15:1084-1090 (1964).
5. R. Patrick. The Catherwood Foundation Peruvian Amazon expedition, Monogr. Acad. Nat.
Sci. Philadelphia (1966).
6. P. R. Ehrlich and P. H. Raven, Butterflies and plants: A study in coevolution, Evolution
18:586-608 (1964).
7. D. R. Paulson, Oviposition in the tropical dragonfly genus Micrathyria (Odonata, Libel-
lulidae), Tombo 12: 12-16 (1969).
8. A. E. Emerson, Termite studies in the Belgian Congo, Rapp. Ann. Inst. Rech. S. Afr. Com-
mis .• 149-159 (1949).
9. A. E. Emerson, Geographical origins and dispersions of termite genera, Fieldiana Zool.
37:465-521 (1955).
10. W. L Peters and G. F. Edmunds, Jr., Revision of the generic classification of the eastern
hemisphere Leptophlebiidae (Ephemeroptera), Pac. Insects 12: 157-240 (1970).
11. E. G. Linsley, Ecology of Cerambycidae. Ann. Rev. Entomol. 4:99-138 (1959).
12. J. A. L Watson, Taxonomy, ecology. and zoogeography of dragonflies (Odonata) from the
north-west of western Australia, Aust. 1. Zool. 17:65-112 (1969).
13. O. Park, Pselaphid beetles of Jamaica, Trans. Am. Microsc. Soc. 79:5-25 (1960).
14. R. H. MacArthur, Patterns of species diversity, Bioi. Rev. 40:510-533 (1965).
15. R. H. MacArthur, Patterns of communities in the tropics, Bioi. J. Linn. Soc. 1: 19-30 (1969).
16. R. H. MacArthur, Geographical Ecology: Pattern and Distribution of Species. Harper & Row,
New York and London (1972).
17. Th. Dobzhansky, Evolution in the tropics, Am. Sci. 38:209-221 (1950).
18. E. J. Corner, in: Evolution as a Process (J. Huxley, A. C. Hardy, and E. B. Ford, eds.), pp.
34-45, G. Allen, London, (1954).
19. W. H. Gotwald, Jr., Comparative morphological studies of the ants, with particular reference
to the mouthparts (Hymenoptera: Formicidae), Cornell Univ. Agric. Exp. Stn. Mem. No. 408
(1969).
20. C. D. Michener. Observations on the nests and behavior of Trigona in Australia and New
Guinea (Hymenoptera, Apidae), Am. Mus. Novit. No. 2026 (1961).
21. A. Wille, Behavioral adaptations of bees for pollen collecting from Cassia flowers, Rev. Bioi.
Trap. San Jose Costa Rica 11:205-210 (1963).

461
462 REFERENCES

22. E. J. Fittkau, Remarks on limnology of central-Amazon rain-forest streams, Verh.Int. Theor.

Angew. Ver. Limnol. 15:1092-1096 (1964).
23. E. J. Fittkau, Artenmannigfiiltigkeit amazonischer Lebensriiume aus okologischer Sicht,
Amazoniana 4:321-340 (1973).
24. E. J. Fittkau, in: Biogeography and Ecology in South America (E. J. Fittkau, J. Iilies, H.
Klinge, G. H. Schwabe, and H. Sioli, eds.), Dr. W. Junk, The Hague (1969).
25. J. Iilies, The invertebrate fauna of the Huallaga, a Peruvian tributary of the Amazon River,
from the sources down to Tingo Maria, Verh. Int. Theor. Angew. Verein. Limnol. 15: 1077-
1083 (1964).
26. c. W. Rettenmeyer, Observations on the biology and taxonomy of flies found over swarm raids
of army ants (Diptera: Tachinidae, Conopidae), Univ. Kans. Sci. Bull. 42:993-1066 (1961).
27. C. W. Rettenmeyer, The diversity of arthropods found with Neotropical army ants and observa-
tions on the behavior of representative species, Proc. North Cent. Branch, Entomol. Society
Am. 17:14-15 (1962).
28. S. P. Hubbell, Tree dispersion, abundance, and diversity in a tropical dry forest, Science
203: 1299-1309 (1979).
29. D. H. Janzen, Herbivores and the number of tree species in tropical forests, Am. Nat.
104:501-529 (1970).
30. P. D. Hurd, Jr., and E. G. Linsley, A classification of the Squash and Gourd Bees Peponapis
and Xenoglossa (Hymenoptera: Apoidea), Univ. Calif. Publ. Entomol., Vol. 62 (1970).
31. L. E. Gilbert, Pollen feeding and reproductive biology of Heliconius butterflies, Proc. Natl.
Acad. Sci. U.S.A. 69:1403-1407 (1972).
32. E. G. Linsley and R. L. Usinger, Insects of the Galapagos Islands, Proc. Calif. Acad. Sci.
33: 113-196 (1966).
33. D. H. Janzen, Coevolution of mutualism between ants and acacias in Central America, Evolu-
tion 20:249-275 (1966).
34. D. R. Paulson, New records of Bahamian Odonata, Q. 1. Fla. Acad. Sci. 29:97-110 (1966).
35. A. J. Kohn and G. H. Orians, Ecological data in the classification of closely related species,
Syst. Zool. 11: 119-127 (1962).
36. J. Reiskind, Ant-mimicry in Panamanian clubionid and salticid spiders (Araneae: Clubionidae,
Salticidae), Biotropica 9: 1-8 (1977).
37. T. Erwin and L. V. J. M. Erwin, Relationships of predaceous beetles to tropical forest wood
decay. Part II. The natural history of Neotropical Eurycoleus macularis Chevrolat (Carabidae:
Lebiini) and its implications in the evolution of ectoparasitoidism, Biotropica 8:215-224
(1976).
38. B. Holldobler, Canopy orientation: A new kind of orientation in ants, Science 210:86-88
(1980).
39. H. Takaoka, Pathogens of blackfly larvae in Guatemala and their influence on natural popula-
tions of three species of Onchocerciasis vectors, Am. 1. Trap. Med. Hyg. 29:467-472 (1980).
40. D. H. Janzen, Euglossine bees as long-distance pollinators of tropical plants, Science
171:203-205 (1971).
41. D. H. Janzen, in: Taxonomy and Ecology (V. H. Heywood, ed.), pp. 201-211, Academic
Press, New York (1973).
42. D. H. Janzen, Promising directions of study in tropical animal-plant interactions, Ann. Mo.
Bot. Card. 64:706-736 (1977).
43. G. H. Orians, The number of bird species in some tropical forests, Ecology 50:783-801
(1969).
44. C. W. Myers, The ecological geography of cloud forest in Panama, Am. Mus. Novit. No. 2396
(1969).
45. W. Schwerdtfeger (ed.), Climates of Central and South America, Elsevier, New York (1976).
46. D. H. Janzen, Why mountain passes are higher in the tropics, Am. Nat. 101:233-249 (1967).
REFERENCES 463

47. D. H. Janzen, Why are there so many species of insects?, Proc. Int. Congr. Entomol. 15:84-
94 (1976).
48. P. M. Marsh, Descriptions of new Braconidae (Hymenoptera) parasitic on the potato tuber-
worm and on related Lepidoptera from Central and South America, J. Wash. (D.C.) Acad. Sci.
69:12-17 (1979).
49. J. H. Vandermeer, Niche theory, Ann. Rev. Eco!. Syst. 3:107-132 (1972).
50 .. N. L. Stanton, Patterns of species diversity in temperate and tropical litter mites, Ecology
60:295-304 (1979).
51. R. L. Wenzel and V. J. Tipton (eds.), Ectoparasites of Panama, Field Museum of Natural
History, Chicago (1966).
52. D. Lack, The Natural Regulation of Animal Numbers, Clarendon Press, Oxford (1954).
53. D. Lack, Ecological Isolation in Birds, Blackwell, Oxford. (1971).
54. R. C. Lewontin and M. J. D. White, Interaction between inversion polymorphisms of two
chromosome pairs in the grasshopper, Moraba scurra. Evolution 14:116-129 (1960).
55. S. Wright, On the roles of directed and random changes in gene frequency in the genetics of
populations, Evolution 2:279-294 (1948).
56. S. Wright, Evolution and the Genetics of Populations. Vol. 2, University of Chicago Press,
Chicago (1969).
57. D. H. Janzen, Interfield and interplant spacing in tropical insect control, in: Proc. Ann. Tall
Timbers Conf. Eco!. Anim. Contr. Habitat Manage. pp. 1-6, Tall Timbers Research Station,
Tallahassee (1972).
58. Thomas Park, Beetles, competition, and populations, Science 138:1369-1375,(1962).
59. D. R. A. Wharton, J. E. Lola, and M. L. Wharton, Population density, survival, growth, and
development of the American Cockroach, J. Insect Physiol. 13:699-716 (1967).
60. D. Pimentel, Animal population regulation by the genetic feed-back mechanism, Am. Nat.
95:65-79 (1961).
61. J. Le Gay Brereton, in: The Evolution of Living Organisms, Symposium of Royal Society of
Victoria, Melbourne, Australia, 1959 (1962).
62. J. Haffer, Speciation in Amazonian forest birds, Science 165: 131-137 (1969).
63. E. R. Pianka, Latitudinal gradients in species diversity: A review of concepts, Am. Nat.
100:33-46 (1966).
64. R. T. Paine, Food web complexity and species diversity, Am. Nat. 100:65-76 (1966).
65. J. H. Connell, Diversity in tropical rain forests and coral reefs, Science 199: 1302-1310 (1978).
66. J. H. Connell and E. Orias, The ecological regulation of species diversity, Am. Nat. 98:399-
414 (1964).
67. M. L. Cody, A general theory of clutch size, Evolution 20: 174-184 (1966).
68. R. E. Ricklefs and G. W. Cox, Taxon cycles in the West Indian avifauna, Am. Nat. 106: 195-
219 (1972).
69. H. Groot, The reinvasion of Colombia by Aedes aegypti: Aspects to remember, Am. J. Trop.
Med. Hyg. 29:330-338 (1980).
70. A. M. Young, Cicada ecology in a Costa Rican tropical rain forest, Biotropica 4:152-159
(1972).
71. A. M. Young, Environmental partitioning in lowland tropical rain forest cicadas, J. N.Y.
Entomol. Soc. 88:86-101 (1980).
72. D. H. Janzen, Sweep samples of tropical foliage insects: Effects of seasons, vegetation types,
elevation, time of day, and insularity, Ecology 54:687-708 (1973).
73. R. F. Denno, Ecotope differentiation in a guild of sap-feeding insects on the salt marsh grass,
Spartina patens, Ecology 61:702-714 (1980).
74. Th. Dobzhansky and C. Pavan, Local and seasonal variations in relative frequencies of species
of Drosophila in Brazil, J. Anim. Eco!. 19: 1-14 (1950).
464 REFERENCES

75. A. B. Da Cunha and Th. Dobzhansky, A further study of chromosomal polymorphism in

Drosophila willistoni in its relation to the environment, Evolution 8: 119- 134 (1954).
76. R. C. Lewontin and L. C. Birch, Hybridization as a source of variation to new environments,
Evolution 20:315-336 (1966).
77. A. M. Young, Notes on the behavioral ecology of Perrhybris lypera (Pieridae) in northeastern
Costa Rica, 1. Lepid. Soc. 34:36-47 (1980).
78. R. C. Lewontin, The units of selection, Ann. Rev. Ecol. Syst. 1: 1-18 (1970).
79. M. J. Wade and D. E. McCauley, Group selection: The phenotypic and genotypic differentia-
tion of small populations, Evolution 34:799-812 (1980).
80. E. O. Wilson, The Insect Societies, Belknap Press, Harvard University Press, Cambridge,
Mass. (1971).
81. F. J. Ayala, Dynamics of populations: Factors controlling population growth and population
size in Drosophila pseudoobscura and D. melanogaster, Ecology 48:67-75 (1967).
82. F. Ayala, Competition between species: Frequency dependence, Science 171:820-824 (1971).
83. W. W. Anderson, Genetic equilibrium and population growth under density-regulated selec-
tion, Am. Nat. 105:489-498 (1971).
84. R. C. Lewontin, The effects of population density and competition on viabiiity in Drosophila
melanogaster, Evolution 9:27-41 (1955).
85. R. C. Lewontin, The adaptations of populations to varying environments, Cold Spring Harbor
Symp. Quant. Bioi. 22:395-408 (1957).
86. I. M. Lerner, Resistance of populations to change, Proceedings of the 1956 International
Genetics Symposia Cytol. Suppl. pp. 520-521 (1957).
87. R. C. Lewontin, in: The Genetics of Colonizing Species (H. G. Baker and G. L. Stebbins,
eds.), pp. 77-94, Academic Press, New York (1965).
88. A. M. Young, Notes on the biology of Hypothyris euclea (Lepidoptera: Nymphalidae:
Ithomiinae) in Costa Rica, Pan-Pac. Entomol. 53: 104-1 13 (1976).
89. A. M. Young, Weather and the regulation of Hypothyris euclea (Lepidoptera: Nymphalidae:
Ithomiinae) populations in northeastern Costa Rica. 1. Lepid. Soc. 33:68-69 (1979).
90. F. J. Ayala, Evolution of fitness. I. Improvement in the productivity and size of irradiated
populations of Drosophila serrata and Drosophila birchii, Genetics 53: 883-895 (1966).
91. E. O. Wilson, The nature of the taxon cycle in the Melanesian ant fauna, Am. Nat. 95: 169-193
(1961).
92. M. Gadgil and W. H. Bossert, Life historical consequences of natural selection, Am. Nat.
104:1-24 (1970).
93. D. H. Janzen, Host plants as islands in evolutionary and contemporary time, Am. Nat.
102:592-595 (1968).
94. E. Mayr, Animal Species and Evolution, Belknap Press, Harvard University Press, Cambridge,
Mass. (1963).
95. W. G. Wellington, Individual differences as a factor in population dynamics: The development
of problem, Can. 1. Zool. 35:293-323 (1957).
96. W. G. Wellington, The synoptic approach to studies of insects and climate, Ann. Rev. En-
tomol. 2:143-162 (1957).
97. W. C. Allee, A. E. Emerson, O. Park, T. Park, and K. P. Schmidt, Principles of Animal
Ecology, Saunders, Philadelphia (1949).
98. P. M. Tuskes and L. P. Brower, Overwintering ecology of the monarch butterfly, Danaus
plexippus L., in California, Ecol. Entomol. 3: 141-153 (1978).
99. L. B. Siobodkin and A. Rapoport, An optimal strategy of evolution, Q. Rev. Bioi. 49: 181-200
(1974).
100. J. Grossfield, Species differences in light-influenced mating behavior in Drosophila, Am. Nat.
104:307-309 (1970).
REFERENCES 465

101. N. Smythe, Relationships between fruiting seasons and seed dispersal methods in a neotropical
forest, Am. Nat. 104:25-35 (1970).
102. P. Feeny, Defensive ecology of the Cruciferae, Ann. Mo. Bot. Gard. 64:221-234 (1977).
103. R. Levins, Theory of fitness in a heterogeneous envrionment. II. Developmental flexibility and
niche selection, Am. Nat. 97:75-90 (1963).
104. L. B. Slobodkin and H. L. Sanders, in: Diversity and Stability in Ecological Systems (G. M.
Woodwell and H. H. Smith, eds.), pp. 82-95, Brookhaven Symp. Bioi. No. 22, Brookhaven
National Laboratory, Brookhaven, N.Y. (1969).
105. K. A. J. Wise and J. L. Gressitt, Far southern animals and plants, Nature (London) 207: 101-
102 (1965).
106. F. R. Schram, Arthropods: A convergent phenomenon, Fieldiana Geol. 39:61-108 (1978).
107. A. Wille, The phylogeny and relationships between the insect orders, Rev. Bioi. Trop. 8:93-
123 (1960).
108. H. E. Hinton, On the origin and function of the pupal stage, Trans. R. Entomol. Soc. London
99:395-409 (1948).
109. C. Istock, in: Evolution of Insect Migration and Diapause (H. Dingle, ed.), pp. 171-190,
Springer-Verlag, Berlin, (1978).
110. F. M. Carpenter, Geological history and evolution of the insects, Proc. Int. Congr. Entomol.
15:63-70 (1976).
111. J. Maynard Smith, Optimization theory in evolution, Ann. Rev. Ecol. Syst. 9:31-56 (1978).
112. D. E. McCauley and M. J. Wade, Group selection: The genetic and demographic basis for the
phenotypic differentiation of small populations of Tribolium castaneum, Evolution 34:813-
821 (1980).
113. S. L. Swihart, Neural adaptations in the visual pathway of certain heliconiine butterflies, and
related forms, to variations in wing coloration, Zoologia 52: 1-14 (1967).
114. T. Eisner, in: Chemical Ecology (E. Sondheimer and J. B. Simeone, eds.), pp. 157-217,
Academic Press, New York (1970).
115. D. M. Wilson and R. R. Hoy, Optomotor reaction, locomotory bias, and reactive inhibition in
the milkweed bug Oncopeltus and the beetle Zophobas, Z. Vgl. Physiol. 58:136-152 (1968).
116. D. H. Janzen, Synchronization of sexual reproduction of trees within the dry season in CentraJ
America, Evolution 21:620-637 (1967).
117. G. W. Frankie, in: Coevolution of Animals and Plants (L. E. Gilbert and P. H. Raven, eds.),
pp. 192-209, University of Texas Press, Austin (1975).
118. G. W. Frankie, H. G. Baker, and P. A. Opler, Comparative phenological studies of trees in
tropical wet and dry forests in the lowlands of Costa Rica, 1. Ecol. 62:881-919 (1974).
119. E. B. Edney, Water balance in desert arthropods, Science 156:1059-1066 (1967).
120. G. Fraenkel and M. Blewett, The utilization of metabolic water in insects, Bull. Entomol. Res.
35: 127-139 (1944).
121. M. J. Berridge and B. L. Gupta, Fine-structural changes in relation to ion and water transport in
the rectal papillae of the blowfly, Calliphora, J. Cell Sci. 2:89-112 (1967).
122. H. A. Schneiderman and L. I. Gilbert, Control of growth and development in insects, Science
143:325-333 (1964).
122a. V. B. Wigglesworth, The Life Of Insects, World Press, Cleveland, Ohio (1964).
122b. C. M. Williams, Physiology of insect diapause. IV. The brain and prothoracic glands as an
endocrine system in the Cecropia silkworm, Bioi. Bull. 103: 120-138 (1952).
123. L. C. G. Simoes and G. Uemura, Hormonal control in insect development, Cienc. Cult.
(Maracaibo) 32:331-335 (1979).
124. A. Krishnakumaran, S. J. Berry, H. Oberlander, and H. A. Schneiderman, Nucleic acid
synthesis during insect development. II. Control of DNA synthesis in the Cecropia silkworm
and other saturniid moths, 1. Insect Physiol. 13: 1-57 (1967).
466 REFERENCES

125. A. Spielman and V. Skaff, Inhibition of metamorphosis and of ecdysis in mosquitoes, J.Insect
Physiol. 13: 1087-1095 (1967).
126. P. F. Roseler, I. Roseler, and A. Strambi, The activity of corpora allata in dominant and
subordinate females of the wasp Polistes gallicus, Insectes Soc. 27:97-107 (1980).
127. S. H. Clark and A. P. Platt, Influence of photoperiod on development and larval diapause in the
Viceroy Butterfly, Limenitis archippus, J. Insect Physiol. 15:1951-1957 (1969).
128. S. D. Beck, Insect Photoperiodism, Academic Press, New York (1968).
129. K. D. Kappus and C. E. Venard, The effects of photoperiod and temperature on the induction
of diapause in Aedes triseriatus (Say), J. Insect Physiol. 13: 1007-1019 (1967).
130. V. B. Wigglesworth, The Principles of Insect Physiology, 7th ed., Chapman and Hall, London
(1972).
131. Y. C. Siew, The endocrine control of adult reproductive diapause in the chrysomelid beetle
Galeruca tanaceti (L.). II, J. Insect Physiol. 11:463-479 (1965).
132. S. L. Goldson and R. M. Emberson, Relict diapause in an introduced weevil in New Zealand,
Nature (London) 286:489-490 (1980).
133. S. R. Sims, Diapause dynamics and host plant suitability of Papilio zelicaon (Lepidoptera:
Papilionidae), Am. Midi. Nat. 103:375-384 (1980).
134. D. H. Janzen and T. W. Schoener, Differences in insect abundance and diversity between
wetter and drier sites during a tropical dry season, Ecology 49:96-110 (1968).
135. A. M. Young, The popUlation biology of Neotropical cicadas. III. Notes on the behavioral
natural history of Pacarina cicadas in some Costa Rican grasslands, Entomol. News 85:239-
256 (1974).
136. A. M. Young, The popUlation biology of neotropical cicadas. I. Emergences of Procollina and
Carineta in a mountain forest, Biotropica 7:248-258 (1975).
137. A. M. Young, Notes on the faunistic complexity of cicadas (Homoptera: Cicadidae) in northern
Costa Rica, Rev. Bioi. Trop. 24:267-279 (1976).
138. A. M. Young, Habitat and seasonal relationships of some cicadas (Homoptera: Cicadidae) in
central Costa Rica, Am. MidI. Nat. 103:155-166 (1980).
139. A. M. Young, Seasonal adult emergences of cicadas (Homoptera: Cicadidae) in northwestern
Costa Rica, Milwaukee Publ. Mus. Contrib. Bioi. Geol. No. 40 (1981).
140. H. Dingle, Life history and popUlation consequences of density, photoperiod, and temperature
in a migrant insect, the milkweed bug Oncopeitus, Am. Nat. 102: 149-163 (1968).
141. H. Dingle, Migration strategies in insects, Science 175: 1327-1335 (1972).
142. H. Dingle, in: Evolution of Insect Migration and Diapause (H. Dingle, ed.), pp. 254-276,
Springer-Verlag, Berlin (1978).
143. C. G. Johnson, The Migration and Dispersal of Insects by Flight, Methuen, London (1969).
144. E. O. Wilson and W. H. Bossert, A Primer in Population Biology, Sinauer Associates,
Stamford, Conn. (1971).
145. T. R. E. Southwood, Migration of terrestrial arthropods in relation to habitat, Bioi. Rev.
37: 171-214 (1962).
146. R. B. Root and S. J. Chaplin, The life-styles of tropical milkweed bugs. Oncopeltus (Hemip-
tera: Lygaeidae) utilizing the same hosts, Ecology 57: 132-140 (1976).
147. A. M. Young, Notes on a migration of Uraniafulgens (Lepidoptera: Uraniidae) in Costa Rica,
J. N.Y. Entomol. Soc. 78:60-70 (1970).
148. W. Stubblebine, J. H. Langenheim, and D. Lincoln, Vegetative response to photoperiod in the
tropical leguminous tree Hymenaea courbaril L., Biotropica 10:18-29 (1978).
149. G. W. Ankersmit and P. L. Adkisson, Photoperiodic responses of certain geograprucal strains
of Pectinophora gossypiel/a (Lepidoptera), J. Insect Physiol. 13:553-564 (1967).
150. A. M. Young, Temporal selection for communicatory optimization: The dawn-dusk chorus as
an adaptation in tropical cicadas, Am. Nat. 117:826-829 (1980).
REFERENCES 467

151. W. H. Whitford, E. Depree, and P. Johnson, Foraging ecology of two Chihuahuan desert ant
species: Novomessor cockerelli and Novomessor albisetosus, Insectes Soc. 27: 148-156
(1980).
152. H. A. Hespenheide, Dispersion and the size composition of the aerial insect fauna, Ecol.
Entomol. 2: 139-141 (1977).
153. W. F. Snow, Field estimates of the flight speed of some West African mosquitoes, Ann. Trop.
Med. Parasitol. 74:239-242 (1980).
154. J. A. Downes, Arctic insects and their environment, Can. Entomol. 96:279-307 (1964).
155. D. R. Oliver, Adaptations of Arctic Chironomidae, Ann. Zool. Fenn. 5:111-118 (1968).
156. A. M. Young, Some differences between temperate and tropical populations of Monarch
(Danaus plexippus) and Queen (Danaus gilippus) butterflies (Lepidoptera: Danaidae), En-
tomol. News 85: 116-126 (1974).
157. J. A. Mutchmor, Temperature adaptation in insects, in: Molecular Mechanisms oJ Temperature
Adaptation, pp. 165-176, American Association for the Advancement of Science,
Washington, D.C. (1967).
158. G. Alexander, Cold hardiness in overwintering juvenile grasshoppers, Entomol. News
78:147-154 (1967).
159. J. A. Mutchmor and A. G. Richards, Low temperature tolerance of insects in relation to the
influence of temperature on muscle apyrase activity, J. Insect Physiol. 7: 141-158 (1961).
160. W. Nachtigall and D. M. Wilson, Neuro-muscular control of dipteran flight, J. Exp. Bioi.
47:77-97 (1967).
161. N. S. Church, Heat loss and the body temperatures of flying insects. 1. Heat loss by evaporation
of water from the body, J. Exp. Bioi. 37:171-185 (1960).
162. N. S. Church, Heat loss and the body temperatures of flying insects. II. Heat conduction within
the body and its loss by radiation and convection, J. Exp. BioI. 37: 186-213 (1960).
163. A. E. Kammer, Motor patterns during flight and warm-up in Lepidoptera, J. Exp. Bioi.
48: 89- 109 (1968).
164. J. E. Rawlins, Themoregulation by the Black Swallowtail Butterfly, Papilio polyxenes
(Lepidoptera: Papilionidae), Ecology 6:345-357 (1980).
165. B. Heinrich, Temperature regulation in the bumblebee Bombus vagans: A field study, Science
175: 185-187 (1972).
166. B. Heinrich and P. H. Raven, Energetics and pollination ecology, Science 176:597-602
(1972).
167. C. D. Michener, Biogeography of the bees, Ann. Mo. Bot. Gard. 66:277-347 (1979).
168. J. S. Weir, The effect of temperature variations on queen oviposition and colony foundation in
Myrmica, J. Insect Physiol. 1:352-360 (1958).
169. P. L. Miller, Factors altering spiracle control in adult dragonflies: Water balance, J. Exp. Bioi.
41:331-343 (1964).
170. P. L. Miller, Factors altering spiracle control in adult dragonflies: Hypoxia and temperature, J.
Exp. Bioi. 41:345-357 (1964).
171. K. P. Lamb, Some effects of fluctuating temperatures on metabolism development, and rate of
population growth in the cabbage aphid, Brevicoryne brassicae, Ecology 42:740-745 (1961).
172. K. P. Lamb, Water loss from aphids at moderate temperatures, J. Insect Physiol. 9:703-711
(1963).
173. B. J. Wall, Effects of dehydration and rehydration on Periplaneta americana, J. Insect
Physiol. 16: 1027 - 1042 (1970).
174. C. Papageorgis, Mimicry in neotropical butterflies, Am. Sci. 63:522-532 (1975).
175. A. M. Young, The comparative ethology and ecology of several species of Morpho butterflies
in Costa Rica, Stud. Neotrop. Fauna 8:17-50 (1973).
468 REFERENCES

176. A. M. Young and A. Muyshondt, Ecological and geographical expansion in tropical butterflies
of the genus Morpho in evolutionary time, Rev. Bioi. Trop. 20:231-263 (1972).
177. T. Pliske, Courtship behavior and use of chemical communication by males of certain species
of ithomiine butterflies (Nymphalidae: Lepidoptera), Ann. Entomol. Soc. Am. 68:935-942
(1975).
178. T. Pliske, Pollination of pyrrolizidine alkaloid-containing plants by male Lepidoptera, Envi-
ron. Entomol. 4:474-479 (1975).
179. T. Pliske, Attraction of Lepidoptera to plants containing pyrrolizidine alkaloids, Environ.
Entomo!' 4:455-473 (1975).
180. L. P. Brower and M. A. Jones, Precourtship interaction of wing and abdominal sex glands in
male Danaus butterflies, Proc. R. Entomol. Soc. London Ser. A 40:147-151 (1965).
181. R. 1. Vane-Wright, Pre-courtship activity and a new scent organ in butterflies, Nature (London)
239:338-340 (1972).
182. J. R. Aldrich, M. S. Blum, and S. S. Duffey, Male specific natural products in the bug,
Leptoglossus phyllopus: Chemistry and possible function, 1. Insect Physiol. 22: 1201-1206
(1976).
183. J. R. Aldrich and M. S. Blum, Aposematic aggregation of a bug (Hemiptera: Coreidae): The
defensive display and formation of aggregations, Biotropica 10:58-61 (1978).
184. R. Thornhill, Sexual selection within mating swarms of the lovebug, Plecia nearctica (Dip-
tera: Bibionidael, Anim. Behav. 28:405-412 (1980).
185. R. Thornhill, Competition and coexistence among Panorpa scorpionflies (Mecoptera: Panor-
pidae), Eco!. Monogr. 50: 179-197 (1980).
186. A. M. Young, Community ecology of some tropical rain forest butterflies, Am. Midi. Nat.
87: 146-157 (1972).
187. A. M. Young, Feeding behavior of Morpho butterflies in a seasonal tropical environment, Rev.
Bioi. Trop. 23: 101-123 (1975).
188. A. M. Young, The evolution of eyespots in tropical butterflies in response to feeding on rotting
fruit: An hypothesis, 1. N.Y. Entomoi. Soc. 87:66-77 (1979).
189. A. M. Young, The interaction of predators and "eyespot butterflies" feeding on rotting fruits
and soupy fungi in tropical forests: Variations on a theme developed by the Muyshondts and
Arthur M. Shapiro, Entomol. Rec. 1. Var. 90:63-70 (1980).
190. A. M. Young and J. H. Thomason, Demography of a confined popUlation of the butterfly
Morpho peleides during a tropical dry season, Stud. Neotrop. Fauna 9: 1-34 (1974).
191. B. W. Cornaby, Carrion reduction by animals in contrasting tropical habitats, Biotropica
6:51-63 (1974).
192. M. G. Naumann, Swarming behavior: Evidence for communication in social wasps, Science
189:642-644 (1975).
193. A. M. Young, Attacks by the army ant Eciton burchelli on nests of the social paper wasp
Polistes erythrocephalus in northeastern Costa Rica, 1. Kans. Entomol. Soc. 52:759-768
(1979).
194. M. J. West-Eberhard, The evolution of social behaviour by kin selection, Q. Rev. Bioi.
50:1-33 (1975).
195. R. M. Crewe and M. S. Blum, Alarm pheromones of the Attini: Their phylogenetic signifi-
cance, 1. Insect Physiol. 18:31-42 (1972).
196. M. S. Blum, R. M. Crewe, W. E. Kerr, L. H. Keith, A. W. Garrison, and M. M. Walker,
Citral in stingless bees: Isolation and functions in trail-laying and robbing, 1. Insect Physiol.
16: 1637-1648 (1970).
197. C. R. Carroll and D. H. Janzen, Ecology offoraging by ants, Ann. Rev. Ecol. Syst. 4:231-258
(1973).
198. L. M. Riddiford and C. M. Williams, The effects of juvenile hormone analogues on the
embryonic development of silkworms, Proc. Natl. A cad. Sci. U.S.A. 57:595-601 (1967).
REFERENCES 469

199. F. Engelmann, Endocrine control of reproduction in insects, Ann. Rev. Entomol. 13: 1-26
(1968).
200. W. K. Reisen and M. Aslamkhan, A release-recapture experiment with the malaria vector,
Anopheles stephensi Liston, with observations on dispersal, survivorship, population size,
gonotrophic rhythm and mating behaviour, Ann. Trap. Med. Parasitol. 73:251-269 (1979).
201. B. B. Simpson, Pleistocene changes in the flora of the high tropical Andes, Paleobiology
1:273-291 (1975).
202. A. H. Ehrlich and P. R. Ehrlich, Reproductive strategies in the butterflies. I. Mating frequency,
plugging, and egg number, J. Kans. Entomol. Soc. 51:666-697 (1978).
203. K. S. Brown, Jr., and O. H. H. Mielke, The heliconians of Brazil (Lepidoptera: Nymphalidae).
II. Introduction and general comments, with a supplementary revision of the tribe, Zoologica
1972:1-40 (1972).
204. K. S. Brown, Jr., and W. W. Benson, Adaptive polymorphism associated with multiple
Mullerian mimicry in Heliconius numata (Lepid. Nymph.), Biotrapica 6:205-228 (1974).
205. K. S. Brown, Jr., and J. V. Neto, Predation on aposematic ithomiine butterflies by tanagers
(Pipraeidea melanonota), Biotropica 8: 136-141 (1976).
206. L. P. Brower, W. H. CaJvert, L. E. Hedrick, and J. Christian, Biological observations on an
overwintering colony of Monarch butterflies (Danaus plexippus. Danaidae) in Mexico. J.
Lepid. Soc. 31:232-242 (1977).
207. A. M. Young, Wing coloration and reflectance in Morpho butterflies as related to reproductive
behavior and escape from avian predators, Oecologia 7:209-222 (1971).
208. A. J. Alexander, A study of the biology and behavior of the caterpillars and emerging but-
terflies of the subfamily Heliconiinae in Trinidad, West Indies. l. Some aspects of larvaJ
behavior, Zoologia 46:1-24 (1961).
209. J. M. Ringo, Sexual behavior in Hawaiian Drosophila (Diptera: Drosophilidae), J. N.Y.
Entomol. Soc. (Abstract) 88:68 (1980).
210. H. R. Hermann, Crespuscular and nocturnal activities of Paraponera clavata (Hymenoptera:
Formicidae: Ponerinae), Entomol. News 86:94-98 (1975).
211. A. M. Young, Notes on the foraging of the giant tropical ant Paraponera clavata (Formicidae:
Ponerinae) on two plants in tropical wet forest, J. Ga. Entomol. Soc. 11:41-51 (1976).
212. A. M. Young and H. R. Hermann, Notes on foraging of the giant tropical ant Paraponera
clavata (Hymenoptera: Formicidae: Ponerinae), J. Kans. Entomol. Soc. 53:35-55 (1980).
213. E. S. McCluskey, Circadian rhythms in male ants of five diverse species, Science 150: 1037-
1039 (1965).
214. T. J. Walker, Cryptic species among sound-producing ensiferan Orthoptera (Gryllidae and
Tettigoniidae), Q. Rev. Bioi. 39:345-355 (1964).
215. A. M. Young, Observations on the aggregation of adult cicadas (Homoptera: Cicadidae) in
tropical forests, Can. J. Zool. 58:711-722 (1980).
216. H. G. Baker and I. Baker, in: Taxonomy and Ecology (V. H. Heywood, ed.), pp. 243-264,
Academic Press, New York (1973).
217. H. G. Baker and I. Baker, in: Coevolution of Animals and Plants (L. E. Gilbert and P. H.
Raven, eds.), pp. 100-140, University of Texas Press, Austin (1975).
218. H. G. Baker, P. A. Opler, and I. Baker, A comparison ofthe amino acid complements of floral
and extrafloral nectars, Bot. Gaz. (Chicago) 139:322-332 (1978).
219. W. L. Brown, Jr., and E. O. Wilson, The evolution of the dacetine ants, Q. Rev. Bioi.
34:278-294 (1959).
220. A. M. Young, Possible evolution of mutualism between Mechanitis caterpillars and an ant in
northeastern Costa Rica, Biotropica 10:77-78 (1977).
221. A. M. Young and M. W. Moffett, Studies on the population biology of the tropical butterfly
Mechanitis isthmia in Costa Rica, Am. Midi. Natl. 101:309-319 (1979).
222. A. M. Young and M. W. Moffett, Behavioral regulatory mechanisms in populations of the
470 REFERENCES

butterfly Mechanitis isthmia (Lepidoptera: Nymphalidae: Ithomiinae) in Costa Rica: Adapta-

tions to host plants in secondary and agricultural habitats, Dtsch. Entomol. Z. 26:21-38
(1979).
223. R. W. Matthews, Nesting biology of the social wasp Microstigmus comes (Hymenoptera:
Sphecidae, Pemphredoninae), Psyche 75:23-45 (1968).
224. R. W. Matthews, Microstigmus comes: Sociality in a sphecid wasp, Science 160:787-788
(1968).
225. L. L. Rockwood, Distribution, density, and dispersion of two species of Atta in Guanacaste
Province, Costa Rica, J. Anim. Eco!. 42:803-817 (1973).
226. L. L. Rockwood, The effects of seasonality on foraging in two species of leaf-cutting ants
(Atta) in Guanacaste Province, Costa Rica, Biotropica 7:176-193 (1975).
227. J. O. Schmidt, in: Pesticides and Venom Neurotoxicity (D. L. Shankland, R. M. Hollingworth,
and T. Smyth, eds.), pp. 247-263, Plenum, New York (1978).
228. H. R. Herman and M. S. Blum, The morphology and histology of the hymenopterous poison
apparatus. 1. Paraponera clavata (Formicidae), Alln. Entomol. Soc. Am. 59:397-409
(1966).
229. A. M. Young, A human sweat-mediated defense against multiple attacks by the wasp Polybia
diguetalla in northeastern Costa Rica, Biotropica 10:73-74 (1977).
230. A. Wille and C. D. Michener, The nest architecture of stingless bees with special reference to
those of Costa Rica (Hymenoptera: Apidae), Rev. Bioi. Trop. Spec. Suppl. 21(1) (1973).
231. T. Park, E. V. Miller and C. Z. Lutherman, Studies in population physiology. IX. The effect of
imago population density on the duration of the larval and pupal stages of Tribolium conjitsum
Duval, Ecology 20:365-373 (1939).
232. P. S. Dawson, Developmental rate and competitive ability in Tribolium. II. Changes in compe-
titive ability following further selection for developmental time, Evolution 21:292-298 (1967).
233. F. J. Ayala, M. L. Tracey, L. G. Barr, and J. G. Ehrenfeld, Genetic and reproductive
differentiation of the subspecies, Drosophila equilloxalis caribbensis, Evolution 28:24-41
(1974).
234. W. C. Allee, Measurement of environmental factors in the tropical rain forest of Panama,
Ecology 7:273-302 (1926).
235. C. H. Fraser-Rowell, Experiments on aggregation of Phymateus purpurascens (Orthoptera,
Acrididae, Pyrgomorphinae), J. Zool. 152: 179-193 (1967).
236. J. Myers and J. N. M. Smith, Head flicking by tent caterpillars: A defensive response to
parasite sounds, Call. J. Zool. 56:1628-1631 (1978).
237. D. Otte and A. Joern, Insect territoriality and its evolution: Population studies of desert
grasshoppers on creosote bushes, J. Anim. Ecol. 44:29-54 (1975).
238. C. T. Brues, The specificity of food plants in the evolution of phytophagous insects, Am. Nat.
58:127-144 (1924).
239. V. G. Dethier, Evolution of feeding preferences in phytophagous insects, Evolutioll 8:33-54
(1954).
240. V. G. Dethier, in: Chemical Ecology (E. Sondheimer and J. B. Simeone, eds.), pp. 83-102,
Academic Press, London (1970).
241. V. G. Dethier, Evolution of receptor sensitivity to secondary plant substances with special
reference to deterrents, Am. Nat. 115:45-66 (1980).
242. G. Fraenkel, Insect nutrition, J. R. Coli. Sci. 13:59-69 (1943).
243. G. Fraenkel, The raison d 'etre of secondary plant substances, Science 129: 1466-1470 (1959).
244. R. T. Yamamoto and G. S. Fraenkel, The specificity of the tobacco hornworm, Protoparce
sexta, to solanaceous plants, Anll. Entomol. Soc. Am. 53:503-507 (1960).
245. J. A. Downes, The food habits and description of Atrichopogoll pollillivorus sp.n. (Diptera:
Ceratopogonidae), TrailS. R. Entomol. Soc. LOlldoll 106:439-453 (1955).
REFERENCES 471

246. L. M. Schoonhoven and V. G. Dethier, Sensory aspects of host-plant discrimination by

lepidopterous larvae, Arch. Neerl. Zool. 16:497-530 (1966).
247. L. M. Schoonhoven, in: Insect/Plant Relationships (H. F. van Emden, ed.), pp. 87-99,
Blackwell, Oxford (1973).
248. P. Jaisson, Environmental preference induced experimentally in ants (Hymenoptera: For-
micidae), Nature (London) 286:388-384 (1980).
249. B. Heinrich, The effect of leaf geometry on the feeding behaviour of the caterpillar of Manduca
sexta (Sphingidae), Anim. Behav. 19:119-124 (1971).
250. P. W. Miles, Interaction of plant phenols and salivary phenolases in the relationship between
plants and Hemiptera, Entomol. Exp. Appl. 12:736-744 (1969).
251. M. H. Sweet, On the original feeding habits of the Hemiptera (Insecta), Ann. Entomol. Soc.
Am. 72:575-579 (1979).
252. C. W. Schaefer and R. O'Shea, Host plants of three coreine tribes (Hemiptera: Heteroptera:
Coreidae), Ann. Entomol. Soc. Am. 72:519-523 (1979).
253. J. W. Webb, Life history and population dynamics of tree-dwelling Homoptera, J. N.Y.
Entomol. Soc. (Abstract) 88:80 (1980).
254. R. Levins and R. MacArthur, An hypothesis to explain the incidence of monophagy, Ecology
50:910-911 (1969).
255. W. W. Benson, Evidence for the evolution of unpalatability through kin selection in the
Heliconiinae (Lepidoptera), Am. Nat. 105:213-226 (1971).
256. D. A. Levin, The chemical defenses of plants to pathogens and herbivores, Ann. Rev. Ecol.
Syst. 7: 121-159 (1976).
257. G. A. Rosenthal and D. H. Janzen (eds.), Herbivores: Their Interaction with Secondary Plant
Metabolites, Academic Press, New York (1979).
258. P. Feeny, Seasonal changes in oak leaf tannins and nutrients as a cause of spring feeding by
winter moth caterpillars, Ecology 51:565-581 (1970).
259. P. P. Feeny, K. L. Paauwe, and N. J. Demong, Flea beetles and mustard oils: Host plant
specificity of Phyl/ofreta cruciferae and P. striolata adults (Coleoptera: Chrysomelidae), Ann.
Entomol. Soc. Am. 63:832-841 (1970).
260. J. M. Erickson and P. Feeny, Sinigrin: A chemical barrier to larvae of the black swallowtail
butterfly, Papilio polyxenes, Ecology 55: 103-111 (1974).
261. S. S. Rehr, D. H. Janzen, and P. P. Feeny, L-Dopa in legume seeds: A chemical barrier to
insect attack, Science 181:81-82 (1973).
262. S. S. Rehr, P. P. Feeny, and D. H. Janzen, Chemical defence in Central American non-ant
acacias, J. Anim. Ecology 42:405-416 (1973).
263. S. S. Rehr, E. A. Bell, D. H. Janzen, and P. P. Feeny, Insecticidal amino acids in legume
seeds, Biochem. Syst. 1:63-67 (1973).
264. R.1. Krieger, P. P. Feeny, and C. F. Wilkinson, Detoxication enzymes in the guts of caterpil-
lars: An evolutionary answer to plant defenses?, Science 172:579-581 (1971).
265. N. E. Stamp, Egg deposition patterns in butterflies: Why do some species cluster their eggs
rather than deposit them singly?, Am. Nat. 115:367-380 (1980).
266. A. W. Ghent, A study of the group-feeding behaviour of larvae of the Jack Pine Sawfly,
Neodirprion pratti banksianae Roh, Behaviour 16: 110-148 (1960).
267. P. D. Coley, Effects of leaf age and plant life history patterns on herbivory, Nature (London)
284:545-546 (1980).
268. M. J. Raupp and R. F. Denno, Interactions among willow herbivores as a determinant of
feeding and ovipositional patterns of the imported willow leaf beetle, Plagiodera versicolora
Laich. (Coleoptera: Chrysomelidae) J. N.Y. Entomol. Soc. (Abstract) 88:67 (1980).
269. C. R. Carroll and C. A. Hoffman, Chemical feeding deterrent mobilized in response to insect
herbivory and counter adaptation by Epilachna tredecimnotata, Science 209:414-416 (1980).
472 REFERENCES

270. A. M. Young, The biology of the butterfly Aeria eurimedea agna (Nymphalidae: Ithomiinae:
Oleriini) in Costa Rica, J. Kans. Entomol. Soc. 51: 1-10 (1978).
271. L. P. Brower and J. V. Z. Brower, Birds, butterflies, and plant poisons: A study in ecological
chemistry, Zoologica 49: 137-159 (1964).
272. L. E. Gilbert and P. R. Ehrlich, The affinities of the Ithomiinae and Satyridae (Nymphalidae),
J. Lepid. Soc. 24:297-300 (1970).
273. A. M. Young, On the evolutionary distance between asclepiadaceous-feeding danaids and
apocyanaceous-feeding ithomiids, J. Res. Lepid. 18:251-254 (1980).
274. D. J. Futuyuma, Food plant specialization and environmental predictability in Lepidoptera,
Am. Nat. 110:285-292 (1976).
275. A. M. Young, Evolutionary responses by butterflies to patchy spatial distributions of resources
in tropical environments, Acta Biotheor. 29:37-64 (1980).
276. D. F. Rhoades, in: Herbivores-Their Interaction with Secondary Plant Metabolites (G. A.
Rosenthal and D. H. Janzen, eds.), pp. 3-47, Academic Press, New York (1979).
277. L. B. Brattsten, C. F. Wilkinson, and T. Eisner, Herbivore-plant interactions: Mixed-function
oxidases and secondary plant substances, Science· 196: 1349-1352 (1977).
278. V. G. Dethier, Egg-laying habits of Lepidoptera in relation to available food, Can. Entomol.
91:554-561 (1959).
279. V. G. Dethier, Food-plant distribution and density and larval dispersal as factors affecting
insect populations, Can. Entomol. 91:581-596 (1959).
280. R. Straatman, Notes on certain Lepidoptera ovipositing on plants which are toxic to their
larvae, J. Lepid. Soc. 16:99-103 (1962).
281. M. S. McClure, Biology of Erythroneura lawsoni (Homoptera: Cicadellidae) and coexistence
in the sycamore leaf-feeding guild, Environ. Entomol. 3:59-68 (1974).
282. W. W. Benson, Resource partitioning in passion vine butterflies, Evolution 32:493-518
( 1978).
283. W. W. Benson, K. S. Brown, Jr., and L. E. Gilbert, Coevolution of plants and herbivores:
Passion flower butterflies, Evolution 29:659-680 (1975).
284. A. Muyshondt, Notes on the life cycle and natural history of butterflies of EI Salvador. II.
Anaea (Zaretis) itys (Nymphalidae), J. Lepid. Soc. 27:294-302 (1973).
285. A. Muyshondt, Notes on the life cycle and natural history of butterflies of EI Salvador. III.
Anaea (Consul)fabius (Nymphalidae), J. Lepid. Soc. 28:81-89 (1974).
286. A. Muyshondt, Notes on the life cycle and natural history of butterflies of EI Salvador. V.
Anaea (Memphis) morvus Boisduvali (Nymphalidae), J. Lepid. Soc. 29:32-39 (1975).
287. A. Muyshondt, Notes on the life cycle and natural history of butterflies of EI Salvador. VI.
Anaea (Memphis) pithyusa (Nymphalidae), J. Lepid. Soc. 29: 168-175 (1975).
288. A. Muyshondt, A. M. Young, and A. Muyshondt, Jr., The biology of the butterfly Dione juno
huascama (Nymphalidae: Heliconiinae) in EI Salvador, J. N.Y. Entomol. Soc. 81:137-151
(1973).
289. A. M. Young, Mimetic associations in natural populations of tropical butterflies. I. Life history
and structure of a tropical dry forest breeding popUlation of Battus polydamus, Rev. Bioi. Trop.
19:210-240 (1971).
290. A. M. Young, On the life cycle and natural history of Hymenitis nero (Lepidoptera:
Ithomiidae) in Costa Rica, Psyche 79:284-294 (1972).
291. A. M. Young, Breeding success and survivorship in some tropical butterflies, Oikos 23:318-
326 (1972).
292. A. M. Young, Notes on the life cycle and natural history of Dismorphia virgo (Pieridae:
Dismorphiinae) in Costa Rica, Psyche 79:165-178 (1972).
293. A. M. Young, A contribution to the biology of Itaballia caesia (Lepidoptera: Pieridae) in a
Costa Rican mountain ravine, Wasmann J. Bioi. 30:43-70 (1972).
REFERENCES 473

294. A. M. Young, The ecology and ethology of the tropical nymphaline butterfly, Victorina
epaphus. I. Life cycle and natural history, 1. Lepid. Soc. 26: 155-170 (1972).
295. A. M. Young, Notes on the biology of the butterfly, Heliconius cydno (Lepidoptera:
Heliconiinae) in Costa Rica, Wasmann 1. BioI. 31:337-350 (1973).
296. A. M. Young, Notes on the life cycle and natural history of Parides arcas mylotes
(Papilionidae) in a Costa Rican premontane wet forest, Psyche 80: 1-21 (1973).
297. A. M. Young, Notes on the biology of Phycioies (Eresia) eutropia (Lepidoptera: Nym-
phalidae) in a Costa Rican mountain forest, 1. N.Y. Entomol. Soc. 81:87-100 (1973).
298. A. M. Young, The life cycle of Dircenna relata (Ithomiidae) in Costa Rica, 1. Lepid. Soc.
27:258-267 (1973).
299. A. M. Young, On the biology of Hamadryas februa (Lepidoptera: Nymphalidae: Nym-
phalinae) in Guanacaste, Costa Rica, Z. Angew. Entomol. 75:380-393 (1974).
300. A. M. Young, Further observations on the natural history of Philaethria dido (Lepidoptera:
Nymphalidae: Heliconiinae), 1. N.Y. Entomol. Soc. 82:30-41 (1974).
301. A. M. Young, A natural historical account of Oleria zelica pagasa (Lepidoptera: Ithomiidae) in
a Costa Rican mountain rain forest, Stud. Neotrop. Fauna 9:123-140 (1974).
302. A. M. Young, Notes on the biology of Pteronymia notilla (lthomiidae) in a Costa Rican
mountain forest, 1. Lepid. Soc. 28:257-268 (1974).
303. A. M. Young, Notes on the natural history of rare Adelpha butterfly (Lepidoptera: Nym-
phalidae) in Costa Rican high country, 1. N.Y. Entomol. Soc. 82:235-244 (1974).
304. A. M. Young, On the biology of Godyris zavaleta caesiopicta (Lepidoptera: Nymphalidae:
Ithomiinae), Entomol. News 85:227-238 (1974).
305. A. M. Young, Observations on the life cycle of Heliconius hecale zuleika (Hewitson) in Costa
Rica, Pan-Pac. Entomol. 51:76-85 (1975).
306. A. M. Young, Notes on the life cycle of the butterfly Hypanartia kerfersteini (Nymphalidae:
Nymphalinae: Nymphalini) in Costa Rica, Brenesia 9:61-69 (1976).
307. A. M. Young, Studies on the biology of Heliconius charitonius L. (Nymphalidae:
Heliconiinae) in Costa Rica, Pan-Pac. Entomol. 52:291-303 (1976).
308. A. M. Young, Studies on the biology of Parides iphidamas (Papilionidae: Troidini) in Costa
Rica, 1. Lepid. Soc. 31:100-108 (1977).
309. A. M. Young, Notes on the defoliation of coconut palm (Cocos nucifera) by the butterfly
Opsiphanes quiteria quirinus (Lepidoptera: Brassolidae) in northeastern Costa Rica, Dtsch.
Entomol. Z. 24:353-365 (1977).
310. A. M. Young, Notes on the biology of the butterfly Hypoleria cassotis (Bates) (Nymphalidae:
Ithomiinae) in northeastern Costa Rica, Brenesia 10:97-108 (1977).
311. A. M. Young, Studies on the interactions of Morpho peleides (Morphidae) with Leguminosae,
1. Lepid. Soc. 32:65-74 (1978).
312. A. M. Young, Disappearances of eggs and larvae of Heliconius butterflies (Nymphalidae:
Heliconiinae) in northeastern Costa Rica, Entomol. News 89:81-89 (1978).
313. A. M. Young, Spatial properties of niche separation among Eueides and Dryas butterflies
(Lepidoptera: Nymphalidae: Heliconiinae) in Costa Rica, 1. N.Y. Entomol. Soc. 86:2-19
(1978).
314. A. M. Young, Over-exploitation of larval host plants by the butterflies Heliconius cydno and
Heliconius sapho (Lepidoptera: Nymphalidae: Heliconiinae: Heliconiini) in Costa Rica?, 1.
N.Y. Entomol. Soc. 88:217-227 (1980).
315. A. M. Young, Notes on the seasonal distribution of Anaea butterflies (NymphaJidae) in tropical
dry forests, Acta Oecol. 2: 17-30 (1981).
316. A. M. Young and A. Muyshondt, The biology of Morpho peleides in Central America,
Caribb. 1. Sci. 13:1-49 (1973).
317. A. M. Young and A. Muyshondt, Ecological studies of the butterfly Victorina stelenes
474 REFERENCES

(Lepidoptera: Nymphalidae) in Costa Rica and EI Salvador, Stud. Neotrop. Fauna 9:155-176
(1974).
318. A. M. Young and A. Muyshondt, Studies on the natural history of Central American butterflies
in the family cluster Satyridae-Brassolidae-Morphidae (Lepidoptera: Nymphaloidea). III. Op-
siphanes tamarindi and Opsiphanes cassina in Costa Rica and EI Salvador, Stud. Neotrop.
Fauna 10: 19-56 (1975).
319. A. M. Young and A. Muyshondt, Biology of Morpho polyphemus in EI Salvador, J. N.Y.
Entomol. Soc. 80: 18-42 (1972).
320. C. Wiklund, Oviposition preferences in Papilio machaon in relation to the host plants of the
larvae, Entomol. Exp. Appl. 17:189-198 (1974).
321. C. Wiklund, The evolutionary relationship between adult oviposition preferences and larval
host plant range in Papilio machaon L., Oecologia 18: 185-197 (1975).
322. A. M. Young and D. Stein, Studies on the evolutionary biology of the neotropical nymphalid
butterfly Anartia fatima in Costa Rica, Milwaukee Publ. Mus. Contrib. Bioi. Geol. No.8
(1976).
323. A. M. Young, Oviposition of the butterfly Battus belus varus Kollar (Papilionidae:
Papilioninae: Troidini), J. Lepid. Soc. 33:56-57 (1979).
324. A. M. Young, Oviposition by Heliconius hecale (Nymphalidae: Heliconiinae) on a grass
inflorescence in Costa Rica, J. Lepid. Soc. 34:66-68 (1980).
325. J. M. Pires, Th. Dobzhansky, and G. A. Black, An estimate of the number of species of trees in
an Amazonian forest community, Bot. Gaz. (Chicago) 114:467-477 (1953).
326. L. L. Rockwood, The effect of defoliation on seed production of six Costa Rican tree species,
Ecology 54: 1363-1369 (1973).
327. D. E. Breedlove and P. R. Ehrlich, Coevolution: Patterns of legume predation by a Iycaenid
butterfly, Oecologia 10:99-104 (1972).
328. K. Nakano, A model for the problem ofthe effects of defoliation on plant growth, Jpn. J. Ecol.
27:227-247 (1977).
329. L. E. Gilbert and J. T. Smiley, in: Diversity of Insect Faunas (L. A. Mound and N. Waloff,
eds.), pp. 89-104, Blackwell, London (1978).
330. B. J. Rathcke and R. W. Poole, Coevolutionary race continues: Butterfly adaptation to plant
trichomes, Science 187: 175-176 (1975).
331. A. M. Young, The rearing of Morpho peleides (Morphinae) on peanuts, J. Lepid. Soc.
28:90-99 (1974).
331a. A. M. Young, Notes on the natural history of Morpho granadensis polybaptus Butler
(Lepidoptera: Nymphalidae: Morphinae), and its relation to that of Morpho peleides limpida
Butler, J. N.Y. Entomol. Soc. 90: in press (1982).
332. D. H. Janzen, Tropical blackwater rivers, animals, and mast fruiting by the Dipterocarpaceae,
Biotropica 6:69-103 (1973).
333. D. McKey, P. G. Waterman, C. N. Mbi, J. S. Gartlan, and T. T. Struhsaker, Phenolic content
of vegetation in two African rain forests: Ecological implications, Science 202:61-64 (1978).
334. L. Van Der Pijl, Evolutionary action of tropical animals on the reproduction of plants, Bioi. 1.
Linn. Soc. 1:85-96 (1969).
335. K. Faegri and L. Van Der Pijl, The Principles of Pollination Ecology, Pergamon Press, New
York (1971).
336. D. J. Carter and J. C. Demming, Azygophleps albovittata Bethune-Baker (Lepidoptera: Cos-
sidae) attacking groundnuts in northern Nigeria, with descriptions of the immature and imaginal
stages, Bull. Entomol. Res. 70:399-405 (1980).
337. G. S. Hartshorn, in: Tropical Ecological Systems: Trends in Terrestrial and Aquatic Research
(F. B. Golley and E. Medina, eds.), pp. 41-51, Springer-Verlag, New York (1975).
338. L. E. Gilbert, in: Coevolution of Animals and Plants (L. E. Gilbert and P. H. Raven, eds.), pp.
210-240, University of Texas Press, Austin (1975).
REFERENCES 475

339. W. W. Benson, Natural selection for mullerian mimicry in Heliconius erato in Costa Rica,
Science 176:936-939 (1972).
340. D. Simberloff, Species turnover and equilibrium island biogeography, Science 194:572-578
(1976).
341. R. H. MacArthur and E. O. Wilson, The Theory of Island Biogeography, Princeton University
Press, Princeton, N.J. (1967).
342. L. P. Brower, J. V. Z. Brower, and C. T. Collins, Experimental studies of mimicry. 7.
Relati ve palatability and Mullerian mimicry among neotropical butterflies of the subfamily
Heliconiinae, Zoologica 48:65-84 (1963).
343. M. Rothschild, in: Insect/Plant Relationships (H. F. van Emden, ed.), pp. 59-83, Blackwell,
Oxford (1972).
344. M. Rothschild and D. N. Kellett, Notes on the reactions of various predators to insects storing
heart poisons (cardiac glycosides) in their tissues, J. Entomo!' Ser. A 46(2): !O3-11 0 (1972).
345. M. Rothschild, J. Von Euw, and T. Reichstein, Some problems connected with warningly
coloured insects and toxic defense mechanisms, Mitt. Basler Afrika Bibliogr. 4~: 135-158
(1972).
346. M. Rothschild, J. von Euw, and T. Reichstein, Cardiac glycosides (heart poisons) in the
polka-dot moth Syntomeida epilais Walk. (Ctenuchidae: Lep.) with some observations on the
toxic qualities of Amata (=Syntomis) phegea (L.), Proc. R. Soc. London Ser. B 183:227-247
(1973).
347. F. H. Pough and L. P. Brower, Predation by birds on Great Southern White butterflies as a
function of palatability, sex, and habitat, Am. Midi. Nat. 98:50-58 (1977).
348. L. P. Brower and C. M. Moffitt, Palatability dynamics of cardenolides in the Monarch But-
terfly, Nature (London) 249:280-283 (1974).
349. L. P. Brower, J. V. Z. Brower, and J. M. Corvino, Plant poisons in a terrestrial food chain,
Proc. Natl. Acad. Sci. U.S.A. 57:893-898 (1967).
350. L. P. Brower, W. N. Ryerson, L. 1. Coppinger, and S. C. Glazier, Ecological chemistry and
the palatability spectrum, Science 161: 1349-1351 (1968).
351. L. P. Brower, M. Edmunds, and C. M. Moffitt, Cardenolide content and palatability of a
popUlation of Danaus chrysippus butterflies from West Africa, J. Entomol. 49: 183-196
(1975).
352. L. P. Brower, P. B. McEvoy, K. L. Williamson, and M. A. Flannery, Variation in cardiac
glycoside content of monarch butterflies from natural popUlations in Eastern North America,
Science 177:426-429 (1972).
353. S. S. Duffey and G. G. E. Scudder, Cardiac glycosides in North American Asclepiadaceae, a
basis for unpalatability in brightly coloured Hemiptera and Coleoptera, J. Insect Physiol.
18:63-78 (1972).
354. F. A. Vaughan, Effect of gross cardiac glycoside content of seeds of common milkweed,
Asclepias syriaca, on cardiac glycoside uptake by the milkweed bug Oncopeltusfasciatus, J.
Chern. Eco!. 5:89-100 (1979).
355. A. M. Young, Notes on the interaction of the Neotropical bug Paryphes blandus Horvath
(Hemiptera: Coreidae) with the vine Anguria warscewiczii Hook. f. (Cucurbitaceae), Brenesia
17:27-42 (1980).
356. D. F. Owen, Tropical Butterflies, Clarendon Press, Oxford (1971).
357. D. A. Clark, Age- and sex-dependent foraging strategies of a small mammalian omnivore, J.
Anim. Ecol. 49:549-563 (1980).
358. B. C. Campbell and S. S. Duffey, Tomatine and parasitic wasps: Potential incompatibility of
plant antibiosis with biological control, Science 205:700-702 (1979).
359. J. H. Vandermeer, Relative isolation and seed predation in Calliandra grandiflora, a
mimosaceous legume from the highlands of Guatemala, Biotropica 6:267-268 (1974).
476 REFERENCES

360. J. M. Kenyi, Seasonal defoliation of thickets by larvae of Belenois aurota, Afr. J. Eco!.
18:123-126 (1980).
361. S. D. Hendrix, An evolutionary and ecological perspective of the insect fauna of ferns, Am.
Nat. 115:171-196 (1980).
362. D. H. Janzen, Seed-eaters versus seed size, number, dispersal and toxicity, Evolution 23: 1-27
(1969).
363. D. H. Janzen, Escape of Cassia grandis L. beans from predators in time and space, Ecology
52:964-979 (1971).
364. D. H. Janzen, Escape of juvenile Dioclea megacarpa (Leguminosae) vines from predators in a
deciduous tropical forest, Am. Nat. 105: 97 -112 (1971).
365. D. H. Janzen, Escape in space by Sterculia apetala seeds from the bug Dysdercusfasciatus in a
Costa Rican deciduous forest, Ecology 53:350-361 (1972).
366. J. White and C. E. Strehl, Xylem feeding by periodical cicada nymphs on tree roots, Ecol.
Entomol. 3:323-327 (1978).
367. R. F. Denno, M. J. Raupp, D. W. Tallamy, and C. F. Riechelderfer, Migration in heterogene-
ous environments: Differences in habitat selection between wing forms of the dimorphic plant-
hopper, Prokelisia marginata (Homoptera: Delphacidae), Ecology 61:859-867 (1980).
368. S. McNeill and T. R. E. Southwood, in: Biochemical Aspects of Plant and Animal Coevolution
(1. B. Harborne, ed.), pp. 77-98, Academic Press, London (1978).
369. D. Horsfield, Relationships between feeding of Philaenus spumarius (L.) and the amino acid
concentration in the xylem sap, Ecol. Entomol. 2:259-266 (1977).
370. D. P. Janos, Mycorrhizae influence tropical succession, Biotropica 12 (Supplement):56-64
(1980).
371. V. C. Sgarbieri, Estudo do conteud6 e de algumas caracteristicas das proteinas em sementes de
plantas da familia Leguminosae, Cienc. Cult. (Sao Paolo) 32:78-84 (1980).
372. D. H. Janzen, Swollen-thorn acacias of Central America, Smithson. Contrib. Bot. No. 13
(1974).
373. D. H. Janzen, Protection of Barteria (Passifloraceae) by Pachysima ants (Pseudomyrmecinae)
in a Nigerian rain forest, Ecology 53:885-892 (1972).
374. D. H. Janzen, Allelopathy by myrmecophytes: The ant Azteca as an allelopathic agent of
Cecropia, Ecology 50:147-153 (1969).
375. D. H. Janzen, Dissolution of mutualism between Cecropia and its Azteca ants, Biotropica
5: 15-28.
376. S. Risch, M. McClure, J. Vandermeer, and S. Waltz, Mutualism between three species of
tropical piper (Piperaceae) and their ant inhabitants, Am. Midi. Nat. 98:433-444 (1977).
377. B. L. Bentley, The protective function of ants visiting the extrafloral nectaries of Bixa orellana
(Bixaceae), J. Ecol. 65:27-38 (1977).
378. K. H. Keeler, The extrafloral nectaries of Ipomoea leptophylla (Convolvulaceae), American
Journal of Botany 67:216-222 (1980).
379. K. H. Keeler and R. B. Kaul, Morphology and distribution of petiolar nectaries in Ipomoea
(Convolvulaceae), Am. J. Bot. 66:946-952 (1979).
380. S. Koptur, Facultative mutualism between weedy vetches bearing extrafloral nectaries and
weedy ants in California, Am. J. Bot. 66: 1016-1020 (1979).
381. R. Levins, in: Genetics and Evolution Symposium (E. T. Drake, ed.), pp. 325-340, Yale
University Press, New Haven, Conn. (1968).
382. J. S. Jones and R. F. Probert, Habitat selection maintains a deleterious allele in a heterogeneous
environment, Nature (London) 287:632-633 (1980).
383. R. C. Richmond, in: Ecological Genetics: The Interface (P. F. Brussard, ed.), pp. 127-142,
Springer-Verlag, New York (1978).
384. R. H. Richardson and J. S. Johnston, Ecological specialization in Hawaiian Drosophila.
Habitat selection in Kipuka Puaulu, Oecologia 21: 193-204 (1975).
REFERENCES 477

385. K. Y. Kaneshiro, Ethological isolation and phylogeny in the plantibia subgroup of Hawaiian
Drosophila, Evolution 30:740-745 (1976).
386. R. Levins, Thermal acclimation and heat resistance in Drosophila species, Am. Nat.
103:483-499 (1969).
387. J. R. Powell, Genetic polymorphisms in varied environments, Science 174: 1035-1036 (1971).
388. J. F. McDonald and F. J. Ayala, Genetic response to environmental heterogeneity, Nature
(London) 250:572-574 (1974).
389. C. E. Taylor and J. R. Powell, Microgeographic differentiation of chromosomal and enzyme
polymorphisms in Drosophila persimilis, Genetics 85:681-695 (1977).
390. J. S. Jones, Can genes choose habitats?, Nature (London) 286:757-758 (1980).
391. G. L. Bush, Sympatric host race formation and speciation in frugivorous flies of the genus
Rhagoletis (Diptera, Tephritidae), Evolution 23:237-251 (1969).
392. G. B. Johnson, Relationship of enzyme polymorphism to species diversity, Nature (London)
242: 193-194 (1973).
393. M. Edmunds, Polymorphism in the mimetic butterfly Hypolimnas misippus L. in Ghana,
Heredity 24:281-302 (1969).
394. P. W. Richards, The Tropical Rain Forest, Cambridge University Press, Cambridge (1952).
395. L. R. Holdridge, Life Zone Ecology, Tropical Science Center, San Jose, Costa Rica (1967).
396. L. R. Holdridge, W. C. Grenke, W. H. Hatheway, T. Liang, and J. A. Tosi, Jr., Forest
Environments in Tropical Life Zones: A Pilot Study, Pergamon Press, Oxford (1971).
397. J. L. Wolfson, Oviposition response of Pieris rapae to environmentally-induced variation in
Brassica nigra, Entomo!' Exp. App!. 27:223-232 (1980).
398. H. B. N. Hynes, Notes on the aquatic Hemiptera-HeteropteraofTrinidad and Tobago, B.W.I.,
with a description of a new species of Martarega B. White (Notonectidae), Trans. R. Entomol.
Soc. London 99:341-360 (1948).
399. J. L. Gressitt, J. Sedlack, and J. J. H. Szent-Ivany, Flora and fauna on backs of large Papuan
moss-forest weevils, Science 150: 1833-1835 (1965).
400. R. Silberglied and A. Aiello, Camouflage by integumentary wetting in bark bugs, Science
207:773-775 (1980).
401. L. F. Jiron and R. Zeledon, EI genero Anastrepha (Diptera; Tephritidae) en las principales
frutas de Costa Rica y su relaci6n con pseudomiasis humana, Rev. Bioi. Trop. 27: 155-161
(1979).
402. J. H. Bruga and R. J. Olembo, Plant food preferences of ,some sympatric drosophilids of
tropical Africa, Biotropica 3: 151-158 (1971).
403. P. Walter, Comportement de recherche et d 'exploitation d'une masse steriora1e chez quelques
coprophages Afro-tropicaux (Co!., Scarabaeidae), Ann. Soc. Entomol. Fr. 16:307-323 (1980).
404. J. C. Schuster, Saltatorial Orthoptera as common visitors to tropical flowers, Biotropica
6: 133-140 (1974).
404a. H. Fraser Rowell, Food plant specificity in neotropical rain-forest acridids, Entomol. Exp.
Appl. 24:451-462 (1978).
405. A. M. Young, The ineffectiveness of the stingless bee Trigona jaty (Hymenoptera: Apidae:
Meliponinae) as a pollinator of cocoa (Theobroma cacao L.), 1. Appl. Eco!. 18: 149-155
(1981).
406. P. G. Bystrak and W. W. Wirth, The North American species of Forcipomyia, subgenus
Euprojoannisia (Diptera: Ceratopogonidae), U.S. Dept. Agr. Tech. Bull. No. 1591 (1978).
407. A. M. Laessle, A micro-1imnological study of Jamaican bromeliads, Ecology 42:499-517
(1961).
408. M. H. Robinson and B. Robinson, Associations between flies and spiders: Bibiocommensalism
and dipsoparasitism, Psyche 84:150-157 (1977).
408a. A. M. Young, Feeding and oviposition in the giant tropical damselfly Megaloprepus
coerulatlls (Drury) in Costa Rica, Biotropica 12:237-239 (1980).
478 REFERENCES

409. D. M. Windsor, Nesting association between two neotropical polybiine wasps (Hymenoptera,
Vespidae), Biotropica 4:1-3 (1972).
410. R. L. Jeanne, A latitudinal gradient in rates of ant predation, Ecology 60: 1211-1224 (1979).
411. R. K. Colwell, Competition and coexistence in a simple tropical community. Am. Nat.
107:737-760 (1973).
412. D. H. Janzen, M. Ataroff, M. Farinas, S. Reyes, N. Rincon, A. Soler, P. Soriano, and M.
Vera, Changes in the arthropod community along an elevational transect in the Venezuelan
Andes, Biotropica 8: 193-203 (1976).
413. D. J. Futuyma and F. Gould, Associations of plants and insects in a deciduous forest, Ecol.
Monogr. 1979:33-50 (1979).
414. W. W. Murdoch, F. C. Evans, and C. H. Peterson, Diversity and pattern in plants and insects,
Ecology 53: 819-828 (1972).
415. P. Becker, Island colonization by carnivorous and herbivorous Coleoptera, J. Anim. Ecol.
45: 893-906 (1976).
416. T. W. Schoener and D. H. Janzen, Notes on environmental determinants of tropical versus
temperate insect size patterns, Am. Nat. 102:207-224 (1968).
417. T. L. Erwin, in: Carabid Beetles: Their Evolution, Natural History, and Classification (T. L.
Erwin, G. E. Ball, D. R. Whitehead, and A. Halpern, eds.), pp. 539-592, Dr. W. Junk The
Hague. (1979).
418. R. A. Beaver, Host specificity of temperate and tropical animals, Nature (London) 281: 139-
141 (1979).
419. M. A. Sharp and D. R. Parks, Plant resources and butterfly habitat selection, Ecology
55:870-875 (1974).
420. M. D. Rausher, Larval habitat suitability and oviposition preference in three related butterflies,
Ecology 60:503-511 (1979).
421. P. M. Dolinger, P. R. Ehrlich, W. L. Fitch, and D. E. Breedlove, Alkaloid and predation
patterns in Colorado lupine populations, Gecologia 13: 191-204 (1973).
422. F. S. Chew, Coevolution of pierid butterflies and their cruciferous foodplants. II. The distribu-
tion of eggs on potential foodplants, Evolution 31:568-579 (1977).
423. J. A. Slater, Lygaeid bugs (Hemiptera: Lygaeidae) as seed predators of figs, Biotropica
4: 145-151 (1972).
424. J. A. Slater, The incidence and evolutionary significance of wing polymorphism in Iygaeid
bugs with particular reference to those of South Africa, Biotropica 9:217-229 (1977).
425. J. H. Lawton, in: Diversity of Insect Faunas (L. A. Mound and N. Waloff, eds.), Blackwell,
London (1978).
426. U. Irmler and K. Furch, Production, energy and nutrient turnover of the cockroach Epilampra
irmleri Rocha e Silva & Aguiar in a Central-Amazonian inundation forest, Amazoniana
6:497-520 (1979).
427. J. C. Miller, Niche relationships among parasitic insects occurring in a temporary habitat,
Ecology 61:270-275 (1980).
428. E. O. Wilson, A social ethogram of the neotropical arboreal ant Zacryptocerus varians (Fr.
Smith), Anim. Behav. 24:354-363 (1976).
429. W. H. Gotwald, Jr., Predatory behavior and food preferences of driver ants in selected African
habitats, Ann. Entomol. Soc. Am. 67:877-886 (1974).
430. W. H. Gotwald, Jr., Behavioral observations on African army ants of the genus Aenictus
(Hymenoptera: Formicidae), Biotropica 8:59-65 (1976).
431. W. H. Gotwald, Jr., Trophic ecology and adaptation in tropical Old World ants of the subfam-
ily Dorylinae (Hymenoptera: Formicidae), Biotropica 10: 161-169 (1978).
432. M. Yasuno, The study of the ant population in the grassland at Mt. Hakkoda. II. The distribu-
tion pattern of ant nests at the Kayano grassland, Sci. Rep. Tohoku Univ. Ser. 4 30:43-55
(1964).
REFERENCES 479

433. L. K. Johnson and S. P. Hubbell, Aggression and competition among stingless bees: Field
studies, Ecology 55:120-127 (1974).
434. L. C. Cole, The population consequences of life history phenomena, Q. Rev. Bioi. 29: 103-137
(1954).
435. J. H. Connell, in: The Dynamics of Populations (P. J. Den Boer and G. R. Gradwell, eds.), pp.
298-312, Oosterbeek, Utrecht, The Netherlands (1971).
436. D. H. Janzen, When is it coevolution?, Evolution 34:611-612 (1980).
437. R. B. Root, Organization of a plant-arthropod association in simple and diverse habitats: The
fauna of collards (Brassica oleracea), Eco!. Monogr. 43:95-124 (1973).
438. Y. Ito, A. Shibazaki, and O. Iwahashi, Biology of Hyphantria cunea Drury (Lepidoptera:
Arctiidae) in Japan. IX. Population dynamics, Res. Populo Eco!. 11:211-228 (1969).
439. H. G. Andrewartha and L. C. Birch, The Distribution and Abundance of Animals, University
of Chicago Press, Chicago, (1954).
440. L. B. Siobodkin, F. E. Smith, and N. G. Hairston, Regulation in terrestrial ecosystems, and the
implied balance of nature, Am. Nat. 101:109-124 (1967).
441. M. Lloyd, Self-regulation of adult numbers by cannibalism in two laboratory strains of flour
beetles (Tribolium castaneum), Ecology 49:245-259 (1968).
442. A. Milne, The natural control of insect populations, Can. Entomol. 84:193-213 (1957).
443. Y. Ito, Factors that affect the fluctuations of animal numbers, with special reference to insect
outbreaks, Bull. Natl. Inst. Agric. Sci. (Japan) Ser. C 13:57-89 (1961).
444. L. M. Cook, K. Frank, and L. P. Brower, Experiments on the demography of tropical but-
terflies. I. Survival rate and density in two species of Parides, Biotropica 3: 17-20 (1971).
445. R. Silberglied and A. Aiello, Disruptive coloration in butterflies: Lack of support in Anartia
fatima, Science 209:617-619 (1980).
446. W. W. Murdoch, Population stability and life history phenomena, Am. Nat. 100:5-11 (1966).
447. D. Pimentel, The influence of plant spatial patterns on insect populations, Ann. Entomol. Soc.
Am. 54:61-69 (1961).
448. S. Wright, General, group and special size factors, Genetics 17:603-619 (1932).
449. S. Wright, Tempo and mode in evolution: A critical review, Ecology 26:415-419 (1945).
450. R. A. Fisher and E. B. Ford, The spread of a gene in natural conditions in a colony of the moth
Pana.xia dominula L., Heredity 1:143-174 (1947).
451. P. R. Ehrlich and L. E. Gilbert, Population structure and dynamics of the tropical butterfly
Heliconius ethilla, Biotropica 5:69-82 (1973).
452. L. M. Cook, E. W. Thomason, and A. M. Young, Population structure, dynamics and disper-
sal of the tropical butterfly Heliconius charitonius, J. Anim. Ecol. 45:851-863 (1976).
453. A. M. Young, A communal roost of the butterfly Heliconius charitonius L. in Costa Rican
premontane tropical wet forest, Entomol. News 89:235-243 (1978).
454. A. M. Young and J. H. Thomason, Notes on communal roosting of Heliconius chariton ius in
Costa Rican high country, J. Lepid. Soc. 29:243-255 (1975).
455. A. M. Young and M. E. Carolan, Daily instability of communal roosting in the neotropical
butterfly Heliconius charitonius (Lepidoptera: Nymphalidae: Heliconiinae), J. Kans. Entomol.
Soc. 49:346-359 (1976).
456. D. F. Owen and D. O. Chanter, Population biology of tropical African butterflies. Sex ratio and
genetic variation in Acraea encedon, J. Zool. 157:345-374 (1969).
457. W. K. Reisen and P. F. L. Boreham, Host selection patterns of some Pakistan mosquitoes, Am.
J. Trop. Med. Hyg. 28:408-421 (1979).
458. J. T. Landahl and R. B. Root, Differences in the life tables of tropical and temperate milkweed
bugs, genus Oncopeltus (Hemiptera: Lygaeidae), Ecology 50:734-737 (1969).
459. R. A. Fisher, The Genetic Theory of Natural Selection, Clarendon Press, Oxford (1930).
460. U. Irrnler, Population-dynamics and physiological adaptation of Pentacomia egregia Chaud.
(Col. Cicindelidae) to the Amazonian inundation forest, Amazoniana 4:219-227 (1973).
480 REFERENCES

461. W. M. Schaffer, Optimal reproductive effort in fluctuating environments, Am. Nat. 108:783-
790 (1974).
462. j. T. Giesel, Reproductive strategies as adaptations to life in temporally heterogeneous envi-
ronments, Ann. Rev. Ecol. Syst. 7:57-79 (1976).
463. J. A. Wiens, Population responses to patchy environments, Ann. Rev. Ecol. Syst. 7:81-120
(1976).
464. J. A. Wiens, On competition and variable environments, Am. Sci. 1977:590-597 (1977).
465. R. H. MacArthur and R. Levins, The limiting similarity, convergence and divergence of
coexisting species, Am. Nat. 101:377-385 (1967).
466. E. L. Charnov and W. M. Schaffer, Life-history consequences of natural selection: Cole's
result revisited, Am. Nat. 107:791-793 (1973).
467. R. H. MacArthur and E. R. Pianka, On optimal use of a patchy environment, Am. Nat.
100:603-609 (1966).
468. K. S. Brown, Jr., Geographical patterns of evolution in Neotropical Lepidoptera. Systematics
and derivation of known and new Heliconiini (Nymphalidae: Nymphalinae), J. Entomol. Ser.
B 44:201-242 (1975).
469. J. Haffer, in: The South American HerpetoJauna: Its Origin, Evolution, and Dispersal (W. E.
Duellman, ed.), pp. 107-140, Museum of Natural History, University of Kansas Monograph
No.7, Lawrence (1979).
470. B. A. Fuseini and R. Kumar, Ecology of cotton stainers (Heteroptera: Pyrrhocoridae) in
southern Ghana, Bioi. J. Linn. Soc. 7:113-146 (1975).
471. H. A. Hespenheide, Ecological inferences from morphological data, Ann. Rev. Ecol. Syst.
4:213-229 (1973).
472. A. M. Shapiro, The life history of an Equatorial montane butterfly, Tatochila xanthodice
(Lepidoptera: Pieridae), J. N.Y. Entomol. Soc. 86:51-55 (1978).
473. R. C. Fisher, A study in insect multiparasitism. II. The mechanism and control of competition
for possession of the host, J. Exp. Bioi. 38:605-628 (1961).
474. L. L. Rockwood, Seasonal changes in susceptibility of Crescentia alata leaves to the flea
beetle, Oedionychus sp., Ecology 55: 142-148 (1974).
475. R. M. Andrews, Evolution of life histories: A comparison of Anolis lizards from matched
island and mainland habitats, Breviora, No. 454, 1-51 (1979).
476. D. Leston, Entomology of the cocoa farm, Ann. Rev. Entomol. 15:273-294 (1970).
477. J. M. Scriber and P. P. Feeny, Growth of herbivorous caterpillars in relation to feeding
specialization and to the growth form of their food plants, Ecology 60:829-850 (1979).
478. K. S. Brown, Jr., Geographical patterns of evolution in Neotropical Lepidoptera: Differentia-
tion of the species of Melinaea and Mechanitis (Nymphalidae, Ithomiinae), Syst. Entomol.
2: 161-197 (1977).
479. J. R. G. Turner, Experiments on the demography of tropical butterflies. II. Longevity and
home-range behavior in Heliconius erato, Biotropica 3:21-31 (1971).
480. S. B. J. Menken, Allozymes and the population structure of Yponomeuta cagnagellus (Hubner)
(Lepidoptera), Neth. J. Zool. 30:228-242 (1980).
481. R. Levins and R. H. MacArthur, Maintenance of genetic polymorphism in a spatially
heterogeneous environment: variation on a theme by Howard Levene, Am. Nat. 100:585-590
(1966).
482. O. Ha1kka, in: Diversity oj Insect Faunas (L. A. Mound and N. Waloff, eds.), pp. 41-55,
Blackwell, London (1978).
483. U. N. SafrieI and U. Ritte, Criteria for the identification of potential colonizers, Bioi. J. Linn.
Soc. 13:287-297 (1980).
484. P. W. Price, Colonization of crops by arthropods: Nonequilibrium communities in soybean
fields, Environ. Entomol. 5:605-611 (1976).
REFERENCES 481

485. J. N. Thompson and M. F. Willson, Disturbance and the dispersal of fleshy fruits, Science
200: l161-1J63 (1978).
486. A. W. Harvey, Hybridization studies in the Schistocera americana complex. I. The specific
status of the Central American Locust, Bioi. J. Linn. Soc. 12:349-355 (1979).
487.P. DeBach, Weather and the success of parasites in population regulation, Canadian En-
tomologist 97:848-863 (1965).
488. H. J. L. Heesen, Egg production of Pterostichus oblongopunctatus (Fabricius) (Col.,
Carabidae) and Philonthus decorus (Gravenhorst) (Col., Staphylinidae), Neth. J. Zool.
30:306-314 (1980).
489. R. D. Alexander, Life cycle origins, speciation, and related phenomena in crickets, Q. Rev.
Bioi. 43:1-41 (1968).
490. G. W. Miskimen, Environmental factors affecting soldier beetle distribution and coloration in
Colombia, Biotropica 4:85-92 (1972).
491. L. A. Lacey and J. D. Charlwood, On the biting activities of some anthropophilic Amazonian
Simulidae (Diptera), Bull. Entomol. Res. 70:495-509 (1980).
492. H. Ebert, On the frequency of butterflies in eastern Brazil with a list of the butterfly fauna of
Pecos de Caldas, Minas Gerais, J. Lepid. Soc. Suppl. 3:1-48 (1969).
493. H. K. Clench, A list of the butterflies of Andros, Bahamas, Ann. Carnegie Mus. 46: 173-193
(1977).
494. D. A. T. Baldry and D. H. Molyneux, Observations on the ecology and trypanosome infections
of a relict population of Glossina medicorum Austen in the Komoe Valley of Upper Volta,
Ann. Trop. Med. Parasitol. 74:79-91 (1980).
495. W. W. K. Cheung and A. T. Marshall, Water and ion regulation in cicadas in relation to xylem
feeding, J. Insect Physiol. 19:1801-1816 (1973).
496. U. Irmler, Die Struktur der Carabiden- und Staphylinidengesellschaften in zentralamazonischen
Uberschwemmungswaldern, Amazoniana 6: 30 1-326 (1978).
497. U. Irmler, Abundance fluctuations and habitat changes of soil beetles in Central Amazonian
inundation Forests (Coleoptera: Carabidae, Staphylinidae), Stud. Neotrop. Fauna Environ.
14: 1-16 (1979).
498. H. Puig, Production de litiere en foret guyanaise: Resultats preliminaires, Bull. Soc. Hist. Nat.
Toulouse 115:338-346 (1980).
499. G. W. Frankie, P. A. Opler, and K. S. Bawa, Foraging behaviour of solitary bees: Implications
for outcrossing of a neotropical forest tree species, J. Ecol. 64: 1049- 1057 (1976).
500. Th. H. Fleming, Do tropical frugivores compete for food?, Am. Zool. 19: 1157-1J 72 (1979).
SOl. R. I. M. Dunbar, Demographic and life history variables of a population of Gelada Baboons
(Theropithecus gelada), J. Anim. Ecol. 49:485-506 (1980).
502. T. Alerstam and P. H. Enckell, Unpredictable habitats and evolution of bird migration, Oikos
33:228-232 (1979).
503. G. P. Waldbauer, in: Evolution of Insect Migration and Diapause (H. Dingle, ed.), pp.
127 -144, Springer-Verlag, Berlin (1978).
504. K. Vepsalainen, in: Evolution of Insect Migration and Diapause (H. Dingle, ed.), pp. 218-
253, Springer-Verlag, Berlin (1978).
505. C. A. Istock, The evolution of complex life cycle phenomena: An ecological perspective,
Evolution 21:592-605 (1967).
506. S. Masaki, in: Evolution of Insect Migration and Diapause (H. Dingle, ed.), pp. 72-100,
Springer-Verlag, Berlin (1978).
507. H. A. Mooney, P. R. Ehrlich, D. E. Lincoln, and K. S. Williams, Environmental controls on
the seasonality of a drought deciduous shrub, Diplacus aurantiacus and its predator, the
Checkerspot Butterfly, Euphydryas chalcedona. Oecologia 45: 143-146 (1980).
482 REFERENCES

508. P. J. Den Boer, Exclusion or coexistence and the taxonomic or ecological relationship between
species, Neth. 1. Zool. 30:278-306 (1980).
509. J. J. M. Van Alphen and H. Boer, Avoidance of scramble competition between larvae of the
Spotted Asparagus Beetle, Crioceris duodecimpitnctata L. (Chrysomelidae) by discrimination
between unoccupied and occupied Asparagus berries, Neth. 1. Zool. 30: 136-143 (1980).
510. H. Wolda, Seasonal fluctuations in rainfall, food and abundance of tropical insects, 1. Anim.
Eco!. 47:369-381 (1978).
511. H. Wolda, Fluctuations in abundance of tropical insects, Am. Nat. 112:1017-1045 (1978).
512. H. Wolda, Seasonal fluctuations in rainfall, food and abundance of tropical insects, 1. Anim.
Eco!. 47:369-381 (1978).
513. P. D. N. Hebert, Moth communities in montane Papua New Guinea, 1. Anim. Ecol. 49:593-
602 (1980).
514. D. F. Owen, Species diversity and seasonal abundance in tropical Sphingidae, Proc. R.
Entomol. Soc. London Ser. A 44:162-168 (1969).
515. D. F. Owen and D. O. Chanter, Species diversity and seasonal abundance in tropical
Ichneumonidae, Oikos 21: 142-144 (1970).
516. D. F. Owen, J. Owen, and D. O. Chanter, Seasonal changes in relative abundance and
estimates of species diversity in a family of tropical butterflies, Oikos 23:200-205 (1972).
517. G. G. Montgomery (ed.), The Ecology of Arboreal Folivores, Smithsonian Institution Press,
Washington, D.C. (1978).
518. H. W. Bates, The Naturalist on the River Amazons: A Record of Advemures, Habits of
Animals, Sketches of Brazilian and Indian Life, and Aspects of Nature under the Equator,
during Eleven Years of Travel, J. Murray, London (1892).
519. C. S. Elton, Conservation and the low population density of invertebrates inside neotropical
rainforest, Bioi. Conserv. 7:3-15 (1975).
520. H. M. Brasell, G. L. Unwin, and G. C. Stocker, The quantity, temporal distribution and
mineral-element content of litterfall in two forest types at two sites in tropical Australia, 1.
Eco!. 68:123-139 (1980).
521. A. K. Singh and R. S. Ambasht, Production and decomposition rate of litter in a teak (Tectona
grandis) plantation at Varanasi (India), Rev. Ecol. Bioi. Sol 17: 13-22 (1980).
522. J. A. Winder, Cacao pollination: Microdiptera of cacao plantations and some of their breeding
places, Bull. Entomol. Res. 61:651-655 (1972).
523. R. E. Buskirk and W. H. Buskirk, Changes in arthropod abundance in a highland Costa Rican
forest, Am. Midi. Nat. 95:288-298 (1976).
524. D. Denlinger, The developmental response of flesh flies (Diptera: Sarcophagidae) to tropical
seasons: Variation in generation time and diapause in East Africa, Oecologia 35: 105-\ 07
(1978).
525. D. L. Denlinger, Seasonal and annual variation of insect abundance in the Nairobi National
Park, Kenya, Biotropica 12: 100-106 (1980).
526. M. Bates, Observations on climate and seasonal distribution of mosquitoes in eastern Colom-
bia, 1. Anim. Ecol. 14: 17-25 (1945).
527. A. M. Young, Feeding and oviposition in the giant tropical damselfly Megaloprepus
coerulatus (Drury) in Costa Rica, Biotropica 12:237-239 (1980).
528. C. S. Pittendrigh, The Bromeliad-Anopheles-Malaria complex in Trinidad. I. The Bromeliad
flora, Evolution 2:60-89 (1948).
529. H. T. Dalmat, Ecology of simuliid vectors of onchocerciasis in Guatemala, Am. Midi. Nat.
52:175-196 (1954).
530. H. T. Dalmat, Biology and control of simuliid (Diptera) vectors of onchoceriasis in Central
America, Proc. 1m. Congr. Entomol. 3:517-533 (1958).
REFERENCES 483

531. H. Sioli, General features of the limnology of Amazonia. Verh. Intern. Verein. Limnol.
15: 1053-1058 (1965).
532. J. Bowden, Weather and the phenology of some African Tabanidae, J. Entomol. Soc. South.
Afr. 39:207-245 (1976).
533. W. A. Lamborn, Some features in the life history of Tabanidae in Nyasaland, EastAfr. Med. J.
15:46-50 (1938).
534. J. P. Dempster, The population dynamics of grasshoppers and locusts, Bioi. Rev. 38:490-529
(1963).
535. W. A. Sodeman, Jr. G. E. Rodrick, and A. L. Vincent, Lampyridae larva: A natural predator
of schistosome vector snails in Liberia, Am. J. Trop. Med. Hyg. 29:319 (1980).
536. J. 1. Petersen, H. C. Chapman, O. R. Willis, and T. Fukuda, Release of Romanomermis
culicivorax for the control of Anopheles albimanus in EI Salvador. II. Application of the
nematode, Am. J. Trop. Med. Hyg. 27: 1268-1273 (1978).
537. M. J. Klowden and A. O. Lea, Blood meal size as a factor affecting continued host-seeking by
Aedes aegypti (L.), Am. J. Trop. Med. Hyg. 27:827-831 (1978).
538. J. A. Seawright, D. G. Haile, M. G. Rabbani, and D. E. Weidhaas, Computer simulation of
the effectiveness of male-linked trans locations for the control of Anopheles albimanus
Wiedemann, Am. J. Trop. Med. Hyg. 28:155-160 (1979).
539. J. T. Whitlaw, Jr., and B. N. Chaniotis, Palm trees and Chagas' disease in Panama, Am. J.
Trop. Med. Hyg. 27:873-881 (1978).
540. R. C. Rainy, Weather and movements of locust swarms: A new hypothesis, Nature (London)
168: 1057-1060 (1951).
541. E. S. Brown and G. Swaine, New evidence on the migration of moths of the African Ar-
myworm, Spodoptera exampta (Wlk.) (Lepidoptera, Noctuidae), Bull. Entomol. Res.
56: 671-684 (1966).
542. J. J. Matthee, The induction of diapause in eggs of Locustana pardalina Walker (Acrididae) by
high temperature, J. Entomol. Soc. South. Afr. 41:25-30 (1978).
543. W. L. Nutting, in: Biology oJ Termites, Vol. I (K. Krishna and F. M. Wessner, eds.), pp.
233-282, Academic Press, New York (1969).
544. W. Leuthold and O. Bruinsma, Pairing behavior in Hodotermes mossambicus (Isoptera),
Psyche 84:109-119 (1977).
545. S. M. Orner and J. L. Coudsley-Thompson, Dry season biology of Anopheles gambiae Giles in
the Sudan, Nature (London) 217:879-880 (1968).
546. M. J. Morris, The influence of constant and changing photoperiods on imaginal diapause in the
red locust (Nomadacris septemJasciata Serv.), J. Insect Physiol. 11: 1105-1119 (1965).
547. E. J. Usua, Induction of diapause in the maize stem borer, BusseolaJusca, Entom. Exp. Appl.
16:322-328 (1973).
548. D. B. Carlisle, P. E. Ellis, and E. Betts, The influence of aromatic shrubs on sexual maturation
in the desert locust Schistocera gregaria, J. Insect Physiol. 11: 1541-1558 (1965).
549. D. C. D. Happold, Seasonal distribution of adult dragonflies at Khartoum, Sudan, Rev. Zool.
Bot. Afr. 77:50-61 (1968).
550. D. C. D. Happold, Dragonflies of Jebel Marra, Sudan, Proc. R. Entomol. Soc. London
41:87-91 (1966).
551. P. S. Corbet, The life-histories of two summer species of dragonfly (Odonata: Coenagriidae),
Proc. Zool. Soc. London 128:403-418 (1957).
552. P. S. Corbet, Temperature in relation to seasonal development of British dragonflies (Odonata),
Proc. Int. Congr. Entomol. 2:755-758 (1958).
553. P. S. Corbet, Notes on the insect food of the Nile crocodile in Uganda, Proc. R. Entomol. Soc.
London Ser. A 34:17-21 (1959).
484 REFERENCES

554. D. R. Paulson and C. E. Jenner, Population structure in overwintering larval Odonata in North
Carolina in relation to adult flight season, Ecology 52:96-107 (1971).
555. T. T. Macan, The Odonata of a Moorland fishpond, Int. Rev. Gesamten Hydrobiol. 49:325-
360 (1964).
556. T. W. Donnelly, The Odonata of Dominica British West Indies, Smithson. Contrib. Zool. No.
37 (1970).
557. D. A. Maelzer, The effect of temperature and moisture on the immature stages of Aphodius
tasmaniae Hope (Scarabaeidae) in the lower south-east of south Australia, Aust. J. Zool.
9: 173-202 (1961).
558. G. Halffter and E. G. Matthews, The natural history of dung beetles of the subfamily
Scarabeinae (Coleoptera, Scarabaeidae), Folia Entomol. Mex. Nos. 12-14 (1966).
559. G. S. Rajulu and G. Krishnan, The epicuticle of millipedes belonging to the genera Cin-
galobolus and Aulacobolus with special reference to seasonal variations, Z. Naturforsch.
23b:845-851 (1968).
560. R. J. Hoffman, Environmental control of seasonal variation in the butterfly Colias eurytheme:
Effects of photoperiod and temperature on pteridine pigmentation, J. Insect Physiol.
20: 1913-1924 (1974).
561. R. 1. Hoffman, Environmental uncertainty and evolution of physiological adaptation in Colias
butterflies, Am. Nat. 112:999-1015 (1978).
562. A. M. Young, Aggregative behavior of Anartia fatima (Nymphalidae: Nymphalinae) in
Guanacaste Province, Costa Rica during the dry season, J. Lepid Soc. 33:58-60 (1979).
563. T. C. Emmel and C. F. Leck, Seasonal changes in the organization of tropical rain forest
butterfly popUlations in Panama, J. Res. Lepid. 8: 133-152 (1970).
564. A. M. Young, D. Tyrrell, and D. MacLoed, Entomophthora echinospora, a fungus pathogenic
on the tropical cicada, Procollina biolleyi, J. Invertebr. Palhol. 21:87-90 (1972).
565. P. de T. Alvim, Eco-physiology of the cacao tree, Proc. Conf. Int. Rech. Agron. Cacao,
Abidjan, 23-35 (1967).
566. F. I. O. Nwoke, Effects of number of photoperiodic cycles on induction and development of
flowers and fruits in Corchorus olitorius L. (var. oniyaya Epen.), Ann. Bot. 45:569-576
(1980).
567. A. Raw, Seasonal changes in the numbers and foraging activities of two Jamaican Exomalopsis
species (Hymenoptera, Anthophoridae), Biotropica 8:270-277 (1976).
568. R. L. Jeanne, Aspects of the biology of Stelopolybia areata (Say) (Hymenoptera: Vespidae),
Biotropica 5: 183-198 (1973).
569. G. W. Frankie and R. Coville, An experimental study on the foraging behavior of selected
solitary bee species in the Costa Rican dry forest (Hymenoptera: Apoidea), J. Kans. Entomol.
Soc. 52: 591-602 (1979).
570. C. D. Michener, M. L. Winston, And R. Jander, Pollen manipulation and related activities and
structures in bees of the family Apidae, Univ. Kans. Sci. Bull. 51:575-601 (1978).
571. R. A. Schlising, Sequence and timing of bees foraging in flowers of Ipomoea and Aniseia
(Convolvulaceae), Ecology 51:1061-1067 (1970).
572. H. G. Baker, Studies in the reproductive biology of West African Rubiaceae, J. West Afr. Sci.
Assoc. 4:9-24 (1958).
573. H. G. Baker, Evolutionary mechanisms in pollination biology, Science 139:877-883 (1963).
574. P. Feinsinger, Ecological interactions between plants and hummingbirds in a successional
tropical community, Ecol. Monogr. 48:269-287 (1978).
575. A. M. Young, Notes on a community ecology of adult sphinx moths in Costa Rican lowland
tropical rain forest, Caribb. J. Sci. 12:151-163 (1972).
576. S. Laroca and O. H. H. Mielke, Ensaios sobre ecologia de comunidade em Sphingidae na
Serra Do Mar, Parana, Brasil (Lepidoptera), Rev. Brasil BioI. 35:1-19 (1975).
REFERENCES 485

577. E. S. Morton, On the evolutionary advantages and disadvantages of fruit eating in tropical
birds, Am. Nat. 107:8-22 (1973).
578. P. Utrio and K. Eriksson, Volatile fermentation products as attractants for Macrolepidoptera,
Ann. Zool. Fenn. 14:98-104 (1977).
579. J. A. Derr, Coevolution of the life history of a tropical seed-feeding insect and its food plants,
Ecology 61:881-892 (1980).
580. P. M. Barrer and J. M. Cherrett, Some factors affecting the site and pattern of leaf-cutting
activity in the ant Alta cephalotes L., J. Entomol. Ser. A 47:15-27 (1972).
581. J. M. Cherrett, The foraging behaviour of Atta cephalotes L. (Hymenoptera, Formicidae). I.
Foraging pattern and plant species attacked in tropical rain forest, J. Anim. Ecol. 37:387-403
(1968).
582. B. Shorrocks, The distribution and abundance of woodland species of British Drosophila
(Diptera: Drosophilidae), J. Anim. Ecol. 44:851-864 (1975).
583. H. L. Carson and W. E. Johnson, Genetic variation in Hawaiian Drosophila. I. Chromosome
and allozyme polymorphism in D. setostimentum and D. ochrobasis from the island of Hawaii,
Evolution 29: 11-23 (1975).
584. S. B. Pipkin, R. L. Rodriguez, and J. Leon, Plant host specificity among flower-feeding
Neotropical Drosophila (Diptera: Drosophilidae), Am. Nat. 100: 135-156 (1966).
585. E. Duffey, Ecological studies on the Large Copper Butterfly Lycaena dispar Haw. batavus
Obth. at Woodwalton Fen National Nature Reserve, Huntingdonshire, J. Appl. Ecol. 5:69-96
( 1968).
586. A. M. Young, Mimetic association in natural populations of tropical papilionid butterflies, J.
N.Y. Entomol. Soc. 79:210-244 (1971).
587. L. G. Saunders, Methods for studying Forcipomyia midges with special reference to cacao-
pollinating species (Diptera, Ceratopogonidae), Can. J. Zool. 37:33-51 (1959).
588. J. A. Winder, Recent research on insect pollination of cocoa, Cocoa Growers' Bull. 26: 11-19
(1977).
589. 1. A. Winder, Cocoa flower Diptera; Their identity, pollinating activity and breeding sites,
PANS 24:5-18 (1978).
590. A. M. Young, Etudes experimentales sur Ie terrain quant a I'ecologie du comportement et a la
biologie du peuplement des chironomides pollinisants du cacaoyer (Diptera: Ceratopogonidae)
au Costa Rica, Conference Internationale sur la Recherche Cacaoyere, Douala, Cameroun
(1979).
590a. A. M. Young, Effects of shade cover and availability of midge breeding sites on pollinating
midge populations and fruit set in two cocoa farms, J. Appl. Ecol. 19:46-67 (1982).
591. J. G. Myers, Notes on wild cacao in Surinam and British Guiana, Kew Bull. 1930: 1-10 (1930).
592. F. Hardy, The chemical and ecological researches on cacao, Trop. Agric. (Trinidad) 12:175-
178 (1935).
593. F. Hardy, Some aspects of cacao soil fertility in Trinidad, Trop. Agric. (Trinidad) 13:315-317
(1936).
594. A. Thomas, The dry season in the Gold Coast and its relation to the cultivation of cacao, J.
Ecol. 20:263-269 (1932).
595. J. Cuatrecasas, Cacao and its allies-A taxonomic revision of the genus Theobroma, Contrib.
U.S. Natl. Mus. 35:379-614 (1964).
596. B. Haines and R. B. Foster, Energy flow through litter in a Panamanian forest, J. Ecol.
65: 147-155 (1977).
597. M. J. N. Okwakol, Estimation of soil and organic matter consumption by termites of the genus
Cubitermes, Afr. J. Ecol. 18: 127-131 (1980).
598. An. A. Federov, The structure of the tropical rain forest and speciation in the humid tropics, J.
Ecol. 54: I-II (1966).
486 REFERENCES

599. T. Uchida an~ T. Nakashima, Some observations on the population density of the spruce
barkbeetle, Ips typographus Linne, during the period of the outbreak in forests devastated by
typhoon of September, 1954, in Hokkaido, Res. Bull. Coil. Exp. For. Hokkaido Univ.
21:149-168 (1961).
600. R. W. Thorp, Structural, behavioral, and physiological adaptations of bees (Apoidea) for
collecting pollen, Ann. Mo. Bot. Gard. 66:788-812 (1979).
601. R. L. Dressler, Pollination in euglossine bees, Evolution 22:202-210 (1968).
602. S. Salas and L. F. Jiron, Simbiosis entre cochinillas de cola (Homoptera: Margarodidae) y otros
insectos. II. Estructura del sistema y significado ecologico, Brenesia 10/11:57-64 (1977).
603. H. E. Evans and R. W. Matthews, Observations on the nesting behavior of South American
sand wasps (Hymenoptera), Biotropica 6: 130-134 (1974).
604. R. E. Coville and P. L. Coville, Nesting biology and male behavior of Trypoxylon (Trypar-
gilum) tenoctitlan in Costa Rica (Hymenoptera: Sphecidae), Ann. Entomol. Soc. Am.
73: 11O-119 (1980).
605. W. J. Mattson and N. D. Addy, Phytophagous insects as regulators of forest primary produc-
tion, Science 190:515-522 (1975).
606. S. T. A. Pickett, Non-equilibrium coexistence of plants, Bull. Torrey Bot. Club 107:238-248
(1980).
607. D. C. Culver, Analysis of simple cave communities: Niche separation and species packing,
Ecology 51:949-958 (1970).
608. D. C. Culver, Species packing in Caribbean and north temperate ant communities, Ecology
55:974-988 (1974).
609. J. R. Karr, Structure of avian communities in selected Panama and Illinois Habitats, Ecol.
MOllogr. 41:207-233 (1971).
610. D. H. Morse, The insectivorous bird as an adaptive strategy, Ann. Rev. Ecol. Syst. 2:177-200
(1971).
611. P. W. Price, Insect Ecology, John Wiley, New York (1975).
612. G. E. Hutchinson, Homage to Santa Rosalia, or why are there so many kinds of animals?, Am.
Nat. 93:145-159 (1959).
613. A. M. Young, Notes on a population outbreak of the beetle Uroplata sp. (Coleoptera:
Chrysomelidae) on the tree Bunchosia pilosa (Malpighiaceae) in Costa Rica, J. N.Y. Entomol.
Soc. 87:289-298 (1979).
614. D. R. Strong, Jr., Rolled-leaf hispine beetles (Chrysomelidae) and their Zingiberales host
plants in Middle America, Biotropica 9:156-169 (1977).
615. J. A. Slater, Monocots and chinch bugs: A study of host plant relationships in the lygaeid
subfamily Blissinae (Hemiptera: Lygaeidae), Biotropica 8: 143-165 (1976).
616. D. E. Wilson and D. H. Janzen, Predation on Scheelea palm seeds by bruchid beetles: Seed
density and distance from the parent palm, Ecology 53:954-959 (1972).
617. B. A. Drummond, III, Comparative ecology and mimetic relationships of ithomiine butterflies
in eastern Ecuador, Ph.D. dissertation, University of Florida, Gainesville (1976).
618. S. Iwao, Some effects of grouping in lepidopterous insects. L'effect de groupe chez les
animaux, Colloq. Int. C. NI. R. S. 173:185-212 (1967).
619. K. Fujisaki, Breakup and re-formation of colony in the first-instar larvae of the Winter Cherry
Bug, Acanthocorus sordidus Thunberg (Hemiptera: Coreidae), in relation to the defense
against their enemies, Res. Populo Bioi. (Kyoto) 16:252-264 (1975).
620. B. J. Rathcke, Competition and coexistence within a guild of herbivorous insects, Ecology
57:76-87 (1976).
621. J. R. G. Turner, Adaptive radiation and convergence in subdivisions of the butterfly genus
Heliconius (Lepidoptera: Nymphalidae), Zool. J. Linn. Soc. 58:297-308 (1976).
REFERENCES 487

622. K. S. Brown, Jr., P. M. Sheppard, and J. R. G. Turner, Quarternary refugia in tropical

America: Evidence from race formation in Heliconius butterflies, Proc. R. Soc. London Ser. B
187:369-378 (1974).
623. K. S. Brown, Jr., Geographical patterns of evolution in neotropical Lepidoptera. Systematics
and derivation of known and new Heliconiini (NymphaJidae: Nymphalinae), J. Entomol. Ser.
B. 44:201-242 (1976).
624. K. S. Brown, Jr., in: Biogeographie et Evolution en Amerique Tropicale (H. Descimon, ed.),
pp. 118-160, Publications du Laboratoire de Zoologie de I 'Ecole NormaJe Superieure, Paris
(1976).
625. K. S. Brown, Jr., Centros de evolucao, refugios quaternarios e conservacao de patrimonios
geneticos na regiao neotropical: Padroes de diferenciacao em Ithomiinae (Lepidoptera: Nym-
phalidae), Acta Amazonica 7:75-137 (1977).
626. K. S. Brown, Jr., Ecologia Geografica E Evolucao Nas Florestas Neotropicais-Part VI Na
Serie "Padroes Geograj1cos De Evolucao Em Lepidopteros Neotropicais", Universidade Es-
tadual de Campinas, Sao Paulo, Brazil (1979).
627. T. G. Whitham, Habitat selection by Pemphigus aphids in response to resource limitation and
competition, Ecology 59: 1164-1176 (1978).
628. T. G. Whitham, The theory of habitat selection: Examined and extended using Pemphigus
aphids, Am. Nat. 115:449-466 (1980).
629. R. E. Silberglied and O. R. Taylor, Jr., Ultraviolet reflection and its behavioral role in the
courtship of the sulfur butterflies Colias eurytheme and C. philodice (Lepidoptera, Pieridae),
Behav. Ecol. Sociobiol. 3:203-243 (1978).
630. J. R. G. Turner, in: Evolutionary Biology, Vol. 10 (M. K. Hecht, W. C. Steere, and B.
Wallace, eds.), Plenum, New York (1977).
631. D. J. Stradling, The nature of the mimetic patterns of the brassolid genera, Caligo and
Eryphanis, Ecol. Entomol. 1: 135-138 (1976).
632. D. Pimentel, Population ecology of insect invaders of the maritime provinces, Can. Entomol.
98:887-894 (1966).
633. R. L. Metcalf, R. A. Metcalf, and A. M. Rhodes, Cucurbitacins as kairomones for diabroticite
beetles, Proc. Natl. Acad. Sci. U.S.A. 77:3769-3772 (1980).
634. D. R. Strong, Rapid asymptotic species accumulation in phytophagous insect communities:
The pests of cacao, Science 185:1064-1066 (1974).
635. R. T. Holmes, J. C. Schultz, and P. Nothnagle, Bird predation on forest insects: An excJosure
experiment, Science 206:462-463 (1979).
636. D. W. Inouye and O. R. Taylor, Jr., A temperate region plant-ant-seed predator system:
Consequences of extra floral nectar secretion by Helianthella quinquenervis, Ecology 60: 1-7
(1979).
637. A. M. Young, Interactions of Phileathria dido (Heliconiinae) and Victorina steneles (Nym-
phalinae) at Stachytarpheta flowers in Costa Rica: Evidence against mimetic association, Acta
Bioi. Venez. 8:1-17 (1972).
638. A. M. Young, Mimetic associations in populations of tropical butterflies. II. Mimetic interac-
tions of Battus polydamus and Battus bellus, Biotropica 4: 17-27 (1972).
638a. K. S. Bawa, Breeding systems of tree species of a lowland tropical community, Evolution
28:85-92 (1974).
639. K. S. Bawa and P. A. Opler, Dioecism in tropical trees, Evolution 29:167-179 (1975).
640. D. A. Levin, in: Topics in Plant Population Biology (c. T. Solbrig, S. Jain, G. B. Johnson,
and P. H. Raven, eds.), pp. 131-156, Columbia University Press, New York (1979).
641. D. A. Levin and W. W. Anderson, Competition for pollinators between simultaneously flower-
ing species, Am. Nat. 104:455-467 (1970).
488 REFERENCES

642. D. W. Schemske, Pollinator specificity in Lantana camara and L. trifolia (Verbenaceae),

Biotropica 8:260-264 (1976).
643. E. M. Barrows, Floral maturation and insect visitors of Pachyptera hymenaea (Bignoniaceae),
Biotropica 9: 133-134 (1977).
644. J. W. Purseglove, Tropical Crops: Dicotyledons. John Wiley, New York (1968).
645. J. P. Parvais, P. de Reffye, and P. Lucas, Observations sur la pollinisation libre chez Theo-
broma cacao: Analyse matbemathique des donnees et modelisation, Cafe Cacao The 21:253-
262 (1977).
646. F. Massaux, C. Tchiendji, C. Misse, and B. Decazy, Etude du transporte du pollen de cacaoyer
par marquage au 32p, Cafe Cacao The 20:163-172 (1976).
647. T. Kaufmann, Preliminary observations on cecidomyiid midge and its role as a cocoa pollinator
in Ghana, Ghana J. Agric. Sci. 6:193-198 (1973).
648. T. Kaufmann, Ecology and behavior of cocoa pollinating Ceratopogonidae in Ghana, W.
Africa, Environ. Entomol. 4:347-351 (1975).
649. F. Privat, Les Bromeliaces, lieu de developpement de quelques insectes pollinisateurs des
fleurs de cacao, Brenesia 16: 197-212 (1979).
650. D. R. Glendinning, Natural pollination of cacao, Nature (London) 193: 1305 (1962).
651. P. Dessart, Contribution a I 'etude des Ceratopogonidae (Diptera), Bull. Agric. Congo Beige
52:525-540 (1961).
652. S. de J. Soria, and W. W. Wirth, Ciclos de vida dos polinizadores do cacaueiro Forcipomyia
spp. (Diptera, Ceratopogonidae) e algumas anotacoes sobre 0 comportamento das larvas no
laboratorie, Rev. Theobroma 5:3-22 (1975).
653. S. de J. Soria, W. W. Wirth, and H. A. Besemer, Breeding places and sites of collection of
adults of Forcipomyia spp. midges (Diptera, Ceratopogonidae) in cacao plantations in Bahia,
Brazil: A progress report, Rev. Theobroma 8:21-29 (1978).
654. D. Fish and S. de J. Soria, Water-holding plants (phytotelmata) as larval habitats for
ceratopogonid pollinators of cacao in Bahia, Brazil, Rev. Theobroma 8: 133-146 (1978).
655. J. De la Cruz and S. de 1. Soria, Estudio de fluctaciones de polinizacion del cacao por las
mosquitas Forcipomyia spp. (Diptera, Ceratopogonidae), en Palmira, Valle, Colombia, Acta
Agron. (Palmira) 23: 1-17 (1973).
656. F. W. Cope, The mechanism of pollen incompatability and compatability in Theobroma cacao
L., Heredity 17: 157-182 (1962).
657. A. F. Posnette, Natural pollination of cocoa, Theobroma leiocarpa Bern. on the Gold Coast II,
Trop. Agric. (Trinidad) 19:188-191 (1942).
658. A. F. Posnette, Pollination of cacao in Trinidad, Trop. Agric. (Trinidad) 21: 115-118 (1944).
659. D. Leston, Entomology of the cocoa farm, Ann. Rev. Entomol. 15:273-294 (1970).
660. D. G. Gibbs and D. Leston, Insect phenology in a forest cocoa-farm locality in West Africa,
J. Appl. Ecol. 7:519-548 (1970).
661. D. G. Gibbs, A.-D. Pickett, and D. Leston, Seasonal popUlation changes in cocoa capsids
(Hemiptera, Miridae) in Ghana, Bull. Entomol. Res. 58:279-293 (1968).
662. W. M. Wheeler, Studies ofneotropical ant-plants and their ants, Bull. Mus. Compo Zool. Harv.
Univ. 110(1) (1942).
663. B. Hocking, in: Coevolution of Animals and Plants (L. E. Gilbert and P. H. Raven, eds.), pp.
78-90, University of Texas Press, Austin (1975).
664. J. B. Free, Insect Pollination of Crops, Academic Press, London (1970).
665. E. L. Chamov, J. Maynard Smith, and J. J. Bull, Why be an hemaphrodite?, Nature (London)
263:125-126 (1976).
666. T. 1. Givnish, Ecological constraints on the evolution of breeding systems in seed plants:
Dioecy and dispersal in gymnosperms, Evolution 34:959-972 (1980).
REFERENCES 489

667. M. R. Mesler, J. D. Ackerman, and K. L. Lu, The effectiveness of fungus gnats as pollinators,
Am. J. Bot. 67:564-567 (1980).
668. H. G. Baker, Evolution in the tropics, Biotropica 2: 101-111 (1970).
669. J. Roughgarden, Resource partitioning among competing species-A coevolutionary approach,
Theor. Populo Bioi. 9:388-424 (1976).
670. A. H. Gentry, Bignoniaceae of southern Central America: Distribution and ecological speci-
ficity, Biotropica 8: 117-131 (1976).
671. E. M. Barrows, Nectar robbing and pollination of Lantana camara (Verbenaceae), Biotropica
8: 132-135 (1976).
672. E. ·M. Barrows, Robbing of exotic plants by introduced carpenter and honey bees in Hawaii,
with comparative notes, Biotropica 12:23-29 (1980).
673. E. O. Wilson, Patchy distribution of ant species in New Guinea rain forests, Psyche 65:26-38
(1958).
674. A. M. Young, Nectar robbing of "four-o-clocks," Mirabilis jalapa (Nyctaginaceae) by the
social paper wasp Epipona querini (Vespidae: Polistinae: Polybiini) in Costa Rica, Biotropica
12:222-223 (1980).
675. L. A. McDade and S. Kinsman, The impact of floral parasitism in two neotropical
hummingbird-pollinated plant species, Evolution 34:944-958 (1980).
676. C. E. King, Resource specialization and eqUilibrium popUlation size in patchy environments,
Proc. Natl. Acad. Sci. U.S.A. 68:2634-2637 (1971).
677. W. W. Anderson, Genetic equilibrium and population growth under density-regulated selec-
tion, Am. Nat. 105:489-498 (1971).
677a. A. M. Young, Nectar and pollen robbing of Thunbergia grandiflora by Trigona bees in Costa
Rica, Biotropica 14: (in press) (1982).
678. D. W. Roubik, Competitive interactions between neotropical pollinators and Africanized honey
bees, Science 201:1030-1032 (1978).
679. D. H. Janzen, Tropical agroecosystems, Science 182: 1212-1219 (1973).
680. A. J. Nicholson, The balance of animal popUlations. J. Anim. Eco!. 2:138-178 (1933).
681. L. C. Birch, The intrinsic rate of natural increase of an insect popUlation, J. Anim. Eco!.
17: 15-26 (1948).
682. J. S. Denslow, in: Tropical Succession, Vol. 12, (J. Ewel, ed.), pp. 47-55, Biotropica Suppl.
( 1980).
683. J. O. Tahvanainen and R. B. Root, The influence of vegetational diversity on the population
ecology of a specialized herbivore, Phyllotreta cruciferae (Coleoptera: Chrysomelidae),
Oecologia 10:321-346 (1972).
684. R. G. Cates and G. H. Orians, Successional status and the palatability of plants to generalized
herbivores, Ecology 51:410-418 (1975).
685. R. S. Miller, Pattern and process in competition, Adv. Eco!. Res. 4:1-74 (1967).
686. R. S. Miller, in: Diversity and Stability in Ecological Systems, pp. 63-70, Brookhaven Sym-
posia in Biology (1969).
687. J. N. Raybould and H. K. Mhiddin, Studies on the immature states of the Simulium neavei
Roubaud complex and their associated crabs in the Eastern Usambara Mountains in Tanzania.
III. Investigations on development and survival and their relevance to control, Ann. Trap. Med.
Parasitol. 72:177-187 (1978).
688. L. E. Gilbert and M. C. Singer, Butterfly ecology, Ann. Rev. Eco!. Syst. 6:365-398 (1975).
689. L. E. Gilbert, Butterfly-plant coevolution: Has Passiflora adenopoda won the selectional race
with heliconiine butterflies?, Science 172:585-586 (1971).
690. P. P. Calvert, Studies on Costa Rican Odonata. II. The habits of the plant-dwelling larva of
Mecistogaster modestus, Entomol. News 22:402-411 (1910).
490 REFERENCES

691. J. H. Vandermeer, J. Addicott, A. Andersen, J. Kitasako, D. Pearson, C. Schnell, and H.

Wilbur, Observations on Paramecium occupying arboreal standing water in Costa Rica, Ecol-
ogy 53:291-293 (1972).
692. D. H. Janzen, The unexploited tropics, Bull. Ecol. Soc. Am. 51:4-7 (1970).
693. D. S. Hayman, Mycorrhiza and crop production, Nature (London) 287:487-488 (1980).
694. C. R. Carroll, Biotic interactions in tropical agroecosystems, Am. Zool. 19: 1057-1064 (1979).
695. C. B. Heiser, Jr., Cultivated plants and cultural diffusion in nuclear America, Am. Anthropol.
67:930-949 (1965).
696. R. Daubenmire, Some ecologic consequences of converting forest to savanna in northwestern
Costa Rica, Trop. Ecol. 13:31-51 (1972).
697. L. K. Vats and J. S. Singh, Population, biomass and secondary net production of above ground
insects in a tropical grassland, Trop. Ecol. 19:51-64 (1978).
698. D. Pimentel and F. A. Stone, Evolution and population ecology of parasite-host systems, Can.
Entomol. 100:655-662 (1968).
699. M. B. Usher, Studies on a wood-feeding termite community in Ghana, West Africa, Biotropica
7:217-233 (1975).
700. H. W. Mielke, Termitaria and shifting cultivation: The dynamic role of the termite in soils of
tropical wet-dry Africa, Trop. Ecol. 19: 117-122 (1978).
701. D. F. Owen, Adaptation to man-made environments and phenology in Precis butterflies, Rev.
Zool. Bot. Afr. 83:361-373 (1971).
702. J. B. Free and I. H. Williams, The Pollination ojCrops by Bees, Apimondia and International
Bee Research Association Special Publication, Bucharest (1977).
703. T. R. Soderstrom and C. E. Calderon, Insect pollination in tropical rain forest grasses, Bio-
tropica 3: 1-16 (1971).
704. I. H. Williams, Behaviour of insects foraging on pigeon pea (Cajanus cajan (L.) Millsp.) in
India, Trap. Agric. (Trinidad) 54:353-363 (1977).
705. D. Leston, The natural history of some West African insects. 14. Entomological changes
accompanying degradation of the forest, Entomol. Mon. Mag. 115: 105-112 (1979).
706. T. Lewis, J. M. Cherrett, I. Haines, B. Haines, and P. L. Mathias, The crazy ant (Anoplolepis
longipes (Jerd.) (Hymenoptera, Formicidae) in Seychelles, and its chemical control, Bull.
Entomol. Res. 66:97-111 (1976).
707. E. H. Medina, The value of utilizing existing shade in the growing of vanilla, J. Agric.
40: 117-124 (1943).
708. C. Pennington, F. A. Jimenez, and T. Theis, A comparison of three methods of vanilla culture
in Puerto Rico, Turrialba 4:79-87 (1954).
709. R. W. Smith and J. B. Whittaker, Factors affecting Gastrophysa viridula populations (Co-
leoptera: Chrysomelidae) in different habitats, J. Anim. Eco!. 49:537-548 (1980).
710. R. A. Bean and R. F. Denno, The effects of mowing on the guild of sap-feeding insects
associated with Kentucky bluegrass, J. N.Y. Entomol. Soc. (Abstract) 8:36-37 (1980).
7 lOa. J. M. Scriber and P. P. Feeny, New foodplant and oviposition records for Battus philenor
(Papilionidae), J. Lepid. Soc. 30:70-71 (1976).
711. D. A. Maelzer, The effect of other organisms on the survival rate of Aphodius tasmaniae Hope
(Scarabaeidae) in the lower south-east of South Australia, Aust. J. Zool. 10:617-629 (1962).
712. D. A. Maelzer, The emergence as a pest of Aphodius tasmaniae Hope (Scarabaeidae) in
pastures in the lower south-east of South Australia, Aust. J. Zool. 10:95-112 (1962).
713. D. A. Maelzer, Rainfall and changes in abundance of Aphodius tasmaniae Hope (Coleoptera:
Scarabaeidae) in the lower south-east of South Australia, Aus!. J. Zool. 12:263-278 (1964).
714. P. B. Came, An ecological study of the pasture scarab Aphodius howitti Hope, Aus!. J. Zool.
4:259-314 (1956).
REFERENCES 491

715. I. Hanski, Movement patterns in dung beetles and in the dung fly, Anim. Behav. 28:953-964
(1980).
716. H. F. Howden and V. G. Nealis, Effects of clearing in a tropical rain forest on the composition
of the coprophagous scarab beetle fauna (Coleoptera), Biotropica 7:77-83 (1975).
717. V.S.D.A., Beetle biocontrol in livestock diseases, U.S.D. Agric. Res. 29:15 (1980).
718. B. D. Gouck, Release and establishment of the baculovirus disease of Oryctes rhinoceros (L.)
(Coleoptera: Scarabaeidae) in Papua New Guinea, Bull. Entomo!' Res. 70:445-453 (1980).
718a. A. M. Young, Notes on seasonality and habitat associations of tropical cicadas (Homoptera:
Cicadidae) in premontane and montane tropical moist forests in Costa Rica, J. N.Y. Entomo/.
Soc. 89:123-142 (1981).
719. M. Lloyd and J. White, On the oviposition habits of 13-year vs. 17-year periodical cicadas of
the same species, J. N.Y. Entomo!' Soc. 84:148-155 (1976).
720. R. Silberglied, A. Aiello, and G. Lamas, Neotropical butterflies of the genus Anartia: Sys-
tematics, life histories, and general biology (Lepidoptera: Nymphalidae), Psyche 86:219-260
(1979).
721. R. Levins and M. Wilson, Ecological theory and pest management, Ann. Rev. Entomol.
25:287-308 (1980).
722. A. D. Hinckley, Ecology of the coconut rhinoceros beetle, Oryctes rhinoceros (L.) (Coleop-
tera: Dynastidae), Biotropica 5: 111-116 (1973).
723. R. Bardner, K. E. Fletcher, and J. H. Stevenson, Pre-flowering and post-flowering insecticide
applications to control Aphis fabae on field beans: Their biological and economic effectiveness,
Ann. Appl. Bioi. 88:265-271 (1978).
724. N. Waloff and P. Thompson, Census data of populations of some leafhoppers (Auchenor-
rhyncha, Homoptera) of acidic grassland, J. Anim. Ecol. 49:395-416 (1980).
725. W. M. Rogoff, Behavior modification for insect management by competitive displacement,
Bull. Entomol. Soc. Am. 26:121-125 (1980).
726. T. Burnett, A model of host-parasite interaction, Proc. 10th Internat. Congress of Entomology,
Vol. 2, 679-686 (1958).
727. T. Burnett, Aspects of the interaction between a chalcid parasite and its aleurodid host, Can. J.
Zool. 45:539-578 (1967).
728. M. P. Hassell, Foraging strategies, population models and biological control: A case study, J.
Anim. Ecol. 49:603-628 (1980).
729. K. J. Beek and D. L. Bramo, in: Biogeography and Ecology in South America (E. J. Fittkau,
J. Iilies, H. Klinge, G. H. Schwabe, and H. Sioli, eds.), pp. 82-112, Dr. W. Junk, The Hague
(1968).
730. G. F. Edmunds, Jr., Phylogenetic biogeography of mayflies, Ann. Mo. Bot. Gard. 62:251-263
(1975).
731. R. Schuster, in: Biogeography and Ecology in South America (E. J. Fittkau; J. lilies, H.
Klinge, G. H. Schwabe, and H. Sioli, eds.), pp. 741-763, Dr. W. Junk, The Hague (1968).
732. G. H. Schwabe, in: Biogeography and Ecology in South America (E. J. Fittkau, J. Illies, H.
Klinge, G. H. Schwabe, and H. Sioli, eds.), pp. 113-136, Dr. W. Junk, The Hague (1968).
733. G. T. Prance, The origin and evolution of the Amazon flora, Interciencia 3:207-222 (1978).
734. W. B. Heed, in: Ecological Genetics: The Interface (P. F. Brussard, ed.), pp. 109-126,
Springer-Verlag, New York (1978).
735. P. R. Ehrlich, D. D. Murphy, M. C. Singer, C. B. Sherwood, R. R. White, and I. L. Brown,
Extinction, reduction, stability, and increase: The responses of Checkers pot Butterfly (Euphyd-
ryas) populations to the California drought, Oecologia 46: 101-105 (1980).
736. R. I. Vane-Wright, in: Diversity of Insect Faunas (L. A. Mound and N. Waloff, eds.), pp.
56-70, Blackwell, London (1978).
492 REFERENCES

737. J. G. Brittnacher, S. R. Sims, and F. J. Ayala, Genetic differentiation between species of the
genus Speyeria (Lepidoptera: Nymphalidae), Evolution 32: 199-210 (1978).
738. W. Geiger, Observations eco-faunistiques sur les lepidopteres de la Tourbiere du Cachot (Jura
Neuchatelois). I. Methodes, faunistiques et caracteristiques de peuplement, Bull. Soc.
Neuchatel. Sci. Nat. 103: 11-27 (1980).
739. M. S. Mani, Ecology and Biogeography of High Altitude Insects, Dr. W. Junk, The Hague
(1968).
740. M. S. Mani and S. Singh, Entomological survey of Hitnilaya. Part XXVI. A contribution to our
knowledge of the geography of the high altitude insects of the nival zones from the north-west
Himalaya. Part 3, J. Bombay Nat. Hist. Soc. 59:77-99 (1962).
741. M. 1. D. White, Modes of Speciation, W. H. Freeman, San Francisco (1978).
742. H. Heatwole and R. Levins, Biogeography of the Puerto Rican bank: Species-turnover on a
small cay, Cayo Ahogado, Ecology 54:1042-1055 (1973).
743. J. Terbough, Distribution on environmental gradients: Theory and a preliminary interpretation
of distributional patterns in the avifauna of the Cordillera Vileabamba, Peru, Ecology 52:23-40
(1971).
744. A Trawoger, Der Alpenwollafter, Eriogaster arbusculae FRR. Ein Beitrag zur Kenntnis und
Erforschung der alpinen Schmetterlingsfauna (Insect: Lepidoptera, Lasiocampidae), Ber Nat.
Med. Ver. Innsbruck 64: 107-132 (1977).
745. Noonan, G. R., South American species of the subgenus Anisotarsus Chaudoir (Genus
Notiobia Perty, Carabidae, Coleoptera, Insecta). Part I. Systematics and natural history, Mil-
waukee Pub!. Mus. Contrib. Bioi. Geol. No. 44 (1981).
746. P. E. Vanzolini and E. E. Williams, South American anoles: Geographic differentiation and
evolution of the Anolis chrysolepis species group (Sauria, Iguanidae), Arq. Zool. (Sao Paulo)
19: 1-298 (1970).
747. P. 1. Darlington, Jr., Carabidae of mountains and islands: Data on the evolution of isolated
faunas, and on atrophy of wings, Ecol. Monogr. 13:37-61 (1943).
748. P. 1. Darlington, Jr., Australian carabid beetles, V. Transition of wet forest faunas from New
Guinea to Tasmania, Psyche 68: 1-26 (1961).
749. P. A. Opler, Oaks as evolutionary islands for leaf-mining insects, Am. Sci. 62:67-73 (1974).
750. V. F. Eastop, in: Diversity of Insect Faunas (L. A. Mound and N. Waloff, eds.), pp. 71-88,
Blackwell, London (1978).
751. H. Schreiber, Dispersal Centres of Sphingidae (Lepidoptera) in the Neotropical Region, Dr.
W. Junk, The Hague (1978).
752. K. S. Brown, Jr. and W. W. Benson, Evolution in modern Amazonian non-forest islands:
Heliconius hermathena, Biotropica 9: 95-117 (1977).
753. P. J. Darlington, Jr., Zoogeography: The Geographical Distribution of Animals, John Wiley,
New York (1957).
754. W. L. Brown, Jr., and E. O. Wilson, Character displacement, Syst. Zool. 5:49-64 (1956).
755. E. J. Fittkau, J. lilies, H. Klinge, G. H. Schwabe, and H. Sioli (eds.), Biogeography and
Ecology in South America, Vols. 1 and 2, Monogr. Bioi, 18-19: 1-946 (1969).
756. J. lilies, in: Biogeography and Ecology in South America (E. J. Fittkau, J. lilies, H. Klinge, G.
H. Scwabe, and H. Sioli, eds.), pp. 685-708, Dr. W. Junk, The Hague (1969).
757. G. Lamas, A preliminary check-list of the butterflies (Lepidoptera) of Peru west of the Andes,
Rev. Cienc. Universidad Nacional Mayor de San Marcos (Lima) 70:59-77 (1977).
758. G. Lamas, Mariposas y conservacion de la naturaleza en el Peru, Bol. Colon. Suiza Peru 1978:
61-66 (1978).
759. A. M. Shapiro, Evidence for obligate monophenism in Reliquia santamarta, a neotropical-
alpine pierine butterfly (Lepidoptera: Pieridae), Psyche 84:183-190 (1977).
760. A. M. Shapiro, Developmental and phenotypic responses to photoperiod and temperature in
REFERENCES 493

an Equatorial montane butterfly, Tatochila xanthodice (Lepidoptera: Pieridae), Biotropica

10:297-301 (1978).
761. A. M. Shapiro, The life history of Reliquia santamarta, a neotropical alpine pierine butterfly
(Lepidoptera: Pieridae), J. N.Y. Entomol. Soc. 86:45-50 (1978).
762. R. C. Buckley, R. T. Corlett, and P. J. Grubb, Are the xeromorphic trees of tropical upper
montane rain forests drought-resistant?, Biotropica 12: 124-136 (1980).
763. A. H. Gentry, Floristic knowledge and needs in Pacific tropical America, Brittonia 30: 134-
153 (1978).
764. R. D. Schrier, M. J. Cullenward, P. R. Ehrlich, and R. R. White, The structure and genetics of
a montane population of the Checkerspot Butterfly, Chlosyne palla. Oecologia 25:279-289
(! 976).
765. N. A. Campbell and J. M. Deam, Altitudinal variation in, and morphological divergence
between, three related species of grasshoppers, Praxibulus sp., Kosciuscola cognatus and K.
usitatus (Orthoptera: Acrididae), Aust. J. Zool. 28:103-118 (1980).
766. P. A. Parsons and I. R. Bock, The population biology of Australian Drosophila. Ann. Rev.
Eco!. Syst. 10:229-245 (1979).
767. A. Wille, Obsefvations on the behavior of a tropical rain forest dung beetle, Megathoposoma
candezei (Harold) (Coleoptera: Scarabaeidae), Rev. Bioi. Trop. 21:41-57 (1973).
768. F. Vuilleumier, Insular biogeography in continental regions. The northern Andes of South
America, Evolution 23:688-702 (1970).
769. D. A. Livingstone, History of the tropical rain forest, Paleobiology 6:243-244 (1980).
770. R. W. Fairbridge, Effects of Holocene climatic change on some tropical geomorphic processes,
Quat. Res. (N.Y.) 6:529-556 (1976).
771. R. G. Klein, Environmental and ecological implications of large mammals from Upper Pleis-
tocene and Holocene sites in southern Africa, Ann. S. Afr. Mus. 81:223-283 (1980).
772. J. R. G. Turner, in: Biogeographie et Evolution en Amerique Tropicale (H. Descimon, ed.),
pp. 98-117, Publications du Laboratoire de Zoologie de I 'Ecole Nonnale Superieure, Paris
(1976).
773. E. Mayr, Animal Species and Evolution. Harvard University Press, Belknap Press, Cambridge,
Mass. (1963).
774. R. C. Lewontin, Studies on heterozygosity and homeostasis. II. Loss of heterosis in a constant
environment, Evolution 12:494-503 (1958).
775. D. A. Roff, Population stability and the evolution of dispersal in a heterogeneous environment,
Oecologia 19:217-237 (1975).
776. A. M. Lucas, Clinal variation in pattern and colour in coastal popUlations of the butterfly
Tisiphone abeona (Donovan) (Lepidoptera: Satyrinae), Aust. J. Zool. 17:37-48 (1969).
777. T. Park and M. Lloyd, Natural selection and the outcome of competition, Am. Nat. 89:235-
240 (1955).
778. M. V. Brian, Exploitation and interference in interspecies competition, J. Anim. Ecol.
25:339-347 (1956).
779. D. H. Morse, Resource partitioning in bumble bees: The role of behavioral factors, Science
197:678-680 (1977).
780. R. E. Moreau, Ecological isolation in a rich tropical avifauna, J. Anim. Ecol. 17:113-126
(1948).
781. J. K. Waage, Reproductive isolation and the potential for character displacement in the dam-
selflies, Calopteryx maculata and C. aequabilis (Odonata: Calopterygidae), Syst. Zool.
24:24-36 (1975).
782. J. J. Pizzimenti and R. DeSalle, Dietary and morphometric variation in some Peruvian rodent
communities: The effect of feeding strategy on evolution, Bioi. J. Linn. Soc. 13:263-285
(1980).
494 REFERENCES

783. T. T. Veblen, F. M. Schlegel, and B. Escobar, Structure and dynamics of old-growth

Nothofagus forests in the Valdivian Andes, Chile, J. Ecol. 68:1-31 (1980).
784. P. J. Darlington, Jr., The carabid beetles of New Guinea. Part IV. General considerations;
analysis and history offauna; taxonomic supplement, Bull. Mus. Compo Zool. Harv. Univ. 142
(1971).
785. H. U. Thiele, Carabid Beetles in Their Environments. A Study on Habitat Selection by Adapta-
tions in Physiology and Behaviour, Springer-Verlag, Berlin (1977).
786. J. A. Wallwork, The Distribution and Diversity of Soil Fauna, Academic Press, New York
(1976).
787. A. S. Corbet, The distribution of butterflies in the Malay Peninsula (Lepid.), Proc. R. En-
tomol. Soc. London Ser. A 16:101-116 (1941).
788. T. Abe, Studies on the distribution and ecological role of termites in a lowland rain forest of
West Malaysia. 4. The role of termites in the process of wood decomposition in Pasoh Forest
Reserve, Rev. Ecol. Bioi. Sol 17:23-40 (1980).
789. H. L. Carson, D. E. Hardy, H. T. Spieth, and W. S. Stone, in: Essays in Evolution and
Genetics in Honor of Theodosius Dobzhansky, pp. 437-543, Appleton-Century-Crofts, New
York (1970).
790. P. H. Raven and D. I. Axelrod, Angiosperm biogeography and past continental movements,
Ann. Mo. Bot. Gard. 61:539-673 (1974).
791. 1. A. Scott, Biogeography of Antillean butterflies, Biotropica 4:32-45 (1972).
792. M. G. Emsley, Speciation in Heliconius (Lep. Nymphalidae): Morphology and geographic
distribution, Zoologica 50:191-254 (1965).
793. D. J. Greathead, Observations on two species of Systoechus (Diptera Bombyliidae) preying
on the Desert Locust, Schistocerca gregarid (Forskal) in eastern Africa, Entomophaga
3: 1-22 (1958).
794. H. A. Hespenheide, Selective predation by two swifts and a swallow in Central America, Ibis
117:82-99 (1975).
 INDEX

Abdominal hair pencils, 91 A nastrepha , 183

Acacia, 165-166,356-357,390 Anderson, W. W., 48
karroo, 117 Andes, 12,57,58,159,233,287,432,435,
Acanthaceae, 215, 267, 272, 442 437, 443-444, 447
Acizzia russellae, 117 Andira inermis, 360
Acraea encedon, 242 Andrewartha, H. G., 229
Acrididae, 310 Andromonoecism
Activating hormone, 71, 72 functional, 366-367, 395, 399
Adaptations to mating, 52 morphological,367
Adaptive radiation, 163, 350 Angiosperm host plants, 160
Aedes aegypti, 30 Anguria, 154
Aenictus, 455 warscewiczii, 88-89
Aeria eurimedea, 131 Anole lizards, 352
Africa, 3, 80, 82, 83, 179,213,257,258,261, Anolis, 395
290,310.321,323,332,335,364,374, Anopheline mosquito, 10
391,410,439 Anopheles, 34, 83, 311
Aggregative behavior, Ill, 154,258,320.347 gambiae, 312
Agriculture, 178, 430 Anoplolepis longipes, 426
livestock, 409 Antarctica, 57
monoculture, 210, 388, 389, 394,420 Antilles, 459
polyculture, 210, 389, 394, 420 Ant mosaic, 395
Agroecosystems, 280, 390, 422 Ants, 138, 303, 346, 382-385, 457
corn-bean-squash, 390 army, 93, 106,210,343
Aiello, A., 182 community shift. 278
Alcohols, 96 crazy, 426
Aldrich, J. R., 91, 154 dacetine, 105
Alkaloids. 120. 122 desert. 83
Allopatry, 262 distribution, 390
micro-, 262, 453 diversification of, 6, 9
Altitudes, high, 57, 231 driver, 213-214, 385
Amazon forest, 207, 249, 292, 308, 309. 327. foraging behavior, 96, 101, 210-213, 345,
333,337,431,432,436,438,447, 371, 377
456-457 ground-dwelling, 213
Anaea, 321 host plants. 166-169,356-357
(Memphis) morvus, 300.322 leaf-cutter, 110. 115,210-213. 329
(Zaretis) itys, 300, 321 mutualisms, 356-357
A nartia , 417, 442-443 nesting sites, 117,189-191
amathea, 442 predation on eggs, 156
chrysopeles, 442 venoms. 110
fatima, 267, 320 Aphelandra, 372
iytrea, 442 Aphids, 88. 352

495
496 INDEX

Aphodius Bawa, K. S., 357, 364, 366, 368

howitti, 407 Bees, 145, 215-216, 279, 300, 323-324, 326,
tasmaniae, 318, 407 383,450
Apocyanaceae, 131, 150,345 bumblebee, 86-87
Aposematism honeybee, 374
color, 150 orchid, 340
display, 154 pollination by, 340, 358-361, 367, 370, 393,
Archipelagos, 436 399
taxon cycle, 47, 165 robber, 96
Arctic region, 83, 84 solitary foraging, 325, 360
Area-effect, 434 stingless, 96,111,263,361,372,374,392,
Aristolochia 403
constricta, 348 Beetles, 150, 196, 233, 262, 328,451
as host plant, 203, 406 adaptations to desert environment, 70
Arthropods, 96, 320 coconut rhinoceros, 421
abundance, 304 cophrophagous scarabs, 409-410
evolutionary success, 62-67 density of, 346
foliage-inhabiting, 303 dung, 184, 318, 409-410, 446
as hosts, 377 flea, 350
predatory, 319, 395 habitat selection, 203, 216
seasonal habitat change, 69 hispine,344
Ascia monusta, 150 ladybird, 197
Asclepiadaceae, 131, 150 migration, 80, 308-309
Asia, 107,410 outbreaks, 318
Asilidae, 260 pasture scarabs, 318,407-409
Asynchrony, between plants and pollinators, as pollinators, 324, 360-361
333 rolled-leaf hispine, 343
Atta, 329, 399 seed-eating, 195
Attinae, 96 soldier, 445
Aulacobolus, 319 sprucebark, 339
Australia, 3, 301, 407, 445 Belostomidae, 11, 382
Avoidance behavior, of predators. 93 Belenois aurota, 159, 319
Ayala, F. J., 40, 43,111 Belgian Congo, 3
Azteca, 206, 357 Beltian bodies, 166
Azygophleps albovittata, 146 Benson, W. W., 132, 141, 148,292,349,350,
438-439,447
Bahamas, 8, 289 Bentley, B. L., 167
Baits, 426 Bignoniaceae, 361
artificial, 215, 380-385 Biogeographic events, 350, 430, 460
fruit, 183-184, 383,401 Biogeography, 2, 56, 148,258
Baker, H. G., 12, 104, 145, 216, 357, 362 Birch, L. c., 35, 229
Baker, I., 104, 145,216,362 Birds, 278
Barriers, evolutionary, 434 Bixa, 356
Barro Colorado Island, 126, 183, 327 orellana, 167
Barrows, E. M., 361, 371 Black pepper, 417
Barteria, 166 Black water, 308
Bates, M., 306 Blum, M. S., 91, 96,154
Battus, 126 Body size in insects, 198-199,278,285,304,
phile.nor, 406 329
INDEX 497

Bombyliidae, 310 Butterfly (cont' d)

Bossert, W. H., 48, 229 habitat selection, 203-205, 321, 417
Braconidae, 232, 235, 244, f.78 host plant interactions, 135,319,324,359
Brassolidae, 184, 353 mate recognition, 352-353
Brazil, 163, 292, 300 migration, 78, 80
Breeding sites, 364 monarch, 53-54, 294, 437
Breeding systems, 278, 357-358 mortality of, 235, 294
outcrossing, 87, 326, 340 oviposition, 126-132, 157, 181,204,231,
Breedlove, D. E., 140,203 289, 299
Brereton, J. Le Gay, 25 as pollinators, 279, 370, 451
British Isles, 8, 33 I population structures, 256
Bromeliads, 179, 186-189, 378 predation on, 244
artificial, 382 seasonal activity, 332
as breeding sites, 189, 307 sex ratios, 241-242
Brower, L. P., 54, 91, 98,150 speciation, 431
Brown, K. S., Jr., 98,150,292,431,436, thoracic temperatures, 84
438-439,447,449 troidine swallowtail, 348
Brown, W. L., Jr., 105 vertical structuring of, 89
Bruchidae, 164, 344 viceroy, 77
Brues, C. T., 114 whole faunas, 443, 459
Bugs Bystrak, P. G., 364
African cotton stainers, 257, 328
ambush,261 Cacao, 261, 394
blissine, 344 flowers, 362-363, 367
coccoid, 340 plantations, 186-188,262-264,332,380-
lovebugs, 9 I 382, 394-404
milkweed, 246-248 pollination, 363-367
plant hoppers, 163,429 shade cover, 399
seed-eating, 195 Caesaria, 321
spittle, 268 California, 77
tria to mine , 3 I 2 Caligo, 101
Bush, G. L., 176,455 Calpodes ethlius, 43,46
Buskirk, R. E., 303 Calvert, P. P., 378
Buskirk, W. H., 303 Cameroon, 142
Butterfly Campbell, B. C., 158
abundance, 235, 293, 321, 338 Canna edulis, 43
coloration, 150 Cannibalism, 231
color sensitivity, 68 Capparis tomentosa, 159,319
communal roosting in, 259, 267 Capsidae, 399
courtship behavior, 88-93, 97-100 Carabidae, 58, 157,201,210,237,298,308-
.daily activity pattern, 83, 101 309, 408, 442, 455-458
differentiation of species, 254 Cardiac glycosides, 120, 151
dispersal, 225 Caribbean islands, 167, 197
disruptive wing coloration, 235 Caribbean watershed, 289
distribution, 3, 4 Carin eta postica, 415
diversity, 39 I Carpenter, F. M., 63
feeding habits, 19, 116, 154-156,216,298, Carrion colonization, 409
328, 336, 353, 408 Carroll, C. R., 96, 168,213,262,390,422
generation time, 280 Cassia, 215
498 INDEX

Caste systems, 105 Cicada (cont' d)

Catastrophes nymphs, 162, 291. 414-420
natural, 227, 269, 296, 297, 343,420,430, oviposition, 413, 416
454 seasonal emergence, 78-79
fires, 34 Cicindelidae, 308
floods, 178,292,303,307-309,337,341, Cingalobolus, 319
388 Circadian rhythms, 287
hurricanes, 388 Climatic predictability, 351
mud slides, 388 Clinal shift, 432
typhoons, 339 Clinal variation, 447
Cecidomyiidae, 93, 382, 393 Clutch size, 27, 97, 114
Cecropia, 206, 357 Coaction, 376
Central America, 6, 74, 79, 80, 132, 147, 161. definition, 223
164,165,179,206,255,265,279,290, Cockroaches, 24, 70, 88,101, 111,207,309,
307,322,328,336,390,408,409,417. 382
437 molt induction, 73
Cerambycidae, 3 Coconut palm, 410
Ceratopogonidae. 93, 116, 186, 261, 303, Cody, M. L., 27, 97
332-333, 362, 382. 396-398 Coevolution, 25, 86,132-133,138,141,181,
Chagas' disease, 312 338,341,355,358,438
Chalcidae,43 Coffee plantations, 411
Chanter, D. 0., 312 Cole, L. c., 230
Chaplin, S. J., 81 Coleoptera, 32, 151,277,278,382,383,384,
Characters, 239, 254 398,405
displacement, 183, 259 Coley, P. D., 126, 146-147
divergence, 259 Colias eurytheme. 320
plesiomorphic, 117 Collembolans, 303, 335
symplesiomorphic, 117 Colombia, 30, 81, 306, 409
Chemical defenses, of plants, 117, 141, 182, Colonizing episodes, 41, 48, 176, 218-219,
202 224, 226, 229, 245, 270
Chew, F. S., 205 Colonizing species, 248, 257
Chironomidae, 84 Colony
Chirollomous. chromosomes, 34 behavior in, 105
Chromosome inversion frequencies, 177 mortality in, 106
Chrysomelidae, 32-33, 77,131,194,299,346, Colwell, R. K., 190-191,215
355,405 Commensal associations, 189
Cicada, 50, 83, 93, 97,101,114 Communal displays, 98
aggregations, 207 Communal feeding, 126
body size, 199-200 Communication
chorus, 50, 102-104,286,290,417 acoustical, 287
emergence, 289-291, 293, 313-317, 327, Community, 2,19.43,49,69-70,97,114,
343, 411,416 436
food of, 117, 413 biologically accommodated, 239. 255
generation time, 280 foliage-inhabiting, 290
habitat, 147, 184, 199-201, 208-209, 322- litter, 113, 335-336
323 physically controlled, 239. 255
host plants of, 161-163,323,413 pollinator, 157
life cycle length, 220 Competition, 26, 32,41. 43, 86, 96. 138,228,
niche overlap, 30 259-266, 355, 453
niche shift, 265 diffuse, 437
INDEX 499

Competition (cant' d) Cycles (cant' d)

exploitation-type, 453 morphisms, 249-250, 288. 296, 297
"interference", 453 population, 157,307,318,329,342
Competitive ability, 226 seasonal, 289, 299-322, 423
frequency-dependent, 343
Compositae, 404 DaCunha, A. B., 175
Connell, J. H., 26, 138, 178,338 Danaidae, 91,120,131
Cook, L. M., 294, 332 Danaus
Cordillera central, 418 chrysippus, 150
Core area, 434-436, 438, 447 p/exippus, 54-55. 294
Coreidae, 88, 91, 154 Darlington, P. J., 436, 456-457
Comer, E. J., 6 Darwin, C., 115
Corpora allata, 71-72 Darwin's theory of evolution. 38
Corpora cardiaca, 71 Dawson,P.S., III
Cossidae, 146 Defensive behavior, 63, 78, 242
Costa Rica, 16,30,41,52,78, 88, 102, 110, Defogging, 196
140,142,147,154,162,167,183,184, Defoliation, 137. 140, 158-160,298,319,341,
189,196,201,205,230,255,262,267, 394
269,275,282,289,292,293,303,307, Demes, 255, 266
313,321-323,326-327,329,331,344, selection, 288
346,348,361,371,372,378,381-383, Denlinger, D. L., 312
389,399,411,415,417,419,442 Denno, R. F., 32
Courtship patterns, 153 Density, 22, 40-43. 81-82, III, 114
Courtship pheromones, 91 dependence, 229, 230, 232-233, 311, 407,
Crabs. 377 408.426
Cretaceous. 63 independence, 226. 229, 232-233, 261,280,
Crewe, R. M., 96 282.299,339,348,388-389,407,409,
Crickets, 283. 297-298 427
Crops Denslow, J. 5., 376
cash, 389 Dethier, V. G., 115, 116,377
food, 14, 388-391. 420 Detoxification. 123
pests of, 424-427 enzymes of, 123, 131, 139
Cruciferae, 146, 205, 459 Development
Cryptic behavior, 120 and environmental cues, 72
Cryptic coloration, 91 hemimetabolous, 62
Cuatrecasas, J., 364, 394 holometabolous. 62
Cubitermes, 335 integration by hormones, 71-74
Cucurbitaceae, 8, 88, 258, 355 molt cycles, 71-72
Curcurbitacins, 355 rate of. 59. 61, 218-220
Curcurbita moschata, 131 Diapause. 71, 76-79,146,279,298,306,309.
Culex, 83 312,329
Culicidae, 93 facultative, 330
Cuticle, 84, 319 pupal, 330
Cyanogenic substances, 122 reproductive. 197. 268, 278. 288, 294, 300,
Cycles 322. 330
climatic, 431 Diel rhythms, 104
daily, 286-287, 290, 327 Dietary overlap, 330
fragmentation, 438 Diets. and diapause. 82
glacial, 446 Dimorphism, 296
of inundation, 456 sexual, color. 50
500 INDEX

Dingle, H., 79-81, 246 Drosophila (cont' d)

Dioecism, 364, 368 pseudoobscura, 35
Diplacus aurantiacus, 298 resource partitioning, 194
Diptera, 359, 378, 382, 444 simulans, 35
Dismorphia willis toni , 175
egg placement, 157 Drosophilidae, 189
virgo, 38 Drummond, B. A., III, 345
Dispersal, 5, 25, 49, 79, 80, 111,224-226,240, Duffey, S. S., 150, 158
257,269 Dysdercus, 257,328,329
barriers to, 433 bimaculatus, 329
genotypes, 456 jasciatus, 257-258
home range movements, 353, 370, 377 nigrojasciatus, 258
seasonal, 292 superstitiosus, 257
Dispersal agents, of seed orfruit, 138, 161,328,
341, 358 Ecdysone, 72
Dispersion, 61 Eciton, 93
definition, 225 Ecological barrier, 141, 265, 287, 346
Displacement, ecological, 453 Ecological colonization, 47, 49, 256
Distephana, 141 Ecological differentiation, local rain forest
Distribution forms, 432, 445
body size, 291 Ecological diversity, 100, 148,341,434
patch, 88, 266, 268, 328 Ecological release, 309
generalist, 349 Ecological screen, 358
specialist, 147, 148,368 Ecological shifts, 56, 163, 349, 422
structure, 135,256,266,295,341,377, Ecosystems
388 trophic organization of, 278
and spatial heterogeneity, 250, 254-255, 269 tropical, 13, 335
and temporal heterogeneity, 250, 268 Ectatomma, 169, 399
Diurnal cycle Ecuador, 93, 101, 233, 345
of activity, 10 1 Egg
of temperature, 83, 281 maturation, 362
Dobzhansky, T. H., 6, 26, 34, 175 predation on, 156,231
Dominica, 262 production, 59, 216
Dominance Ehrlich, A, H., 97
ecological, 138, 343 Ehrlich, P. R., 3, 97,131,140,156,182,203,
DNA synthesis 241. 332
during molt cycle, 72 Eisner, T., 96
L-dopa, 160 Elevational gradients, 180,239, 255, 286, 356
Dorylinae, 455 EI Salvador, 3 I 1, 320, 352
Dorylus, 455 Emergence
Downes, J. A., 83-84, 398 allochronic, 314, 324, 327, 343
Dragonflies, 11,88,225,316-318 cicada, 184-185,313-317,343
Drift, genetic, 239, 267-268, 288,447 polymodal, 295
Drosophila, 54, 98, 218, 445, 452-453 seasonal, 323
flower-feeding, 330 Emerson, A. E., 3
fungus-feeding, 331 Emigration, density-dependent, 3 I I
genetics, 34-35, 177 Emmel, T. c., 321
grimshawi, 98 Endemism. 48, 433, 447, 450
habitat, 174-175,330 Endocrine control systems, and social evolution.
persimilis, 175 73-74
population density, 40, 4 I, 183, 300 Endogenous activity patterns, 101
 INDEX 501

Endopterygota, 62 Feeding behavior, 71, 93-95, 202, 408

Energetics, foraging, 86 Feeding categories, 114-115
Energy budget, 96 Feeding habits, 63,101,105,164,257,320,
Environments 346
carrying capacity of, 175,226 facultative switches, 341
coarse-grained, 174, 252, 256, 258-260, obligatory switches, 341
295 Feeding stimulants, 355
fine-grained, 174, 252, 255, 256, 258-260 Feeding strategies, 256, 408
nonseasonal, 233 Feeny, P. P., 56, 123, 146, 157,264,406
physically controlled, 234 Ferns, as host plants. 160
predictable, 56, 58, 229 Fidicina
seasonal, 233 mannijera, 102, 322,418
spatial patchiness of, 250 pronoe, 102-103, 184,417
tropical lowland, 234-235 semilata, 411,416
unpredictable, 56, 58 sericans, 102-103, 162
Enzyme systems, 73, 256 Fig seeds, 206
Erickson, J. M., 123 Fisher's Fundamental Theorem of Natural Selec-
Erwin, T. L., 9, 201,456-457 tion, 219-221
Erythrina, 187-188,398 Fitness, 2,40,67, 120, 131, 154,219-221,
Esterifying acids, 146 245, 320, 322, 371
Ethiopia, 294 Fittkau, E. J., 6, 439, 440
Eueides Flacourticeae, 300
aliphera, 205 Flies, 93, 97,176,216,249,267,328,340,
iybia, 205 360-361,383,410,451,455
Eugenol, 383 black, 9, 311
Euglossinae, 86, 342 fungus gnats, 269
Euphydryas chaicedona, 298 tsetse, 290
Euprojoannisia, 362 Flight, heat production, 84-87
Euptychia, 348 Flightlessness, 455-457
Eurema daira, 78, 295, 296 Flight speeds. 83
Eusociality, 164 Florida, 8
Evapotranspiration, 286 Flower
Evolution, arthropod, 62-67 cauliflorous, 366
Evolutionary history, 433 color change syndrome, 361
Ewel, J., 303 drop, 299, 400
Exocrine control systems, 73 morphological specializations, 87, 358
Exopterygota, 62 pollination of, 69
Extinction, 156, 254, 256 production, 368
rates, 338 resources, 279
and species abundance, 6 Flowering, 275, 426
Exudates seasonal periodicity. 282, 323, 333
of homopterous insects, 190 Flowering strategies
Eyespot, markings, 98, 353 diel, 86, 104
mass, 324, 325-327, 360-361, 371
Facultative associations, 168 Foliage
Facultative shifts, 261 as food. 325
Famesol,72 quality, 405
Fecundity, 40,61,219,245 Folivores, 294, 299
definition, 218 Food, spermatophores as. 97
Fecundity schedule, age-specific, 217, 221 Foraging behavior, 164, 210, 287, 329, 370
Feeding aggregates, 91 density-dependent, 371
502 INDEX

Foraging behavior (cont' d) Gentry, A. H., 444

diversification of, 96 Geographical differentiation, 183, 443
diurnal pattern of, 10 I Geographical distribution, 454
Forcipomyia 364~365 Geographical patchiness, 432
Forest Geographical range, 77
Amazon inundation, 207, 292 Geographical variation, 150
canopy, 180,290,308,326 Geological histories, 430, 453
flowers, 10 Gerris, 296
specialists, 457 Ghana, 329, 395, 398, 401
islands, 431 Ghent, A. W., 126
light gap, 233, 278, 331, 339, 376 Gilbert,L.E.,8, 131, 132, 141,148,216,241,
moist pockets, 277 332, 345, 347-348, 350, 355
tropical dry, 7,266,279,282,301,321-324, Givnish, T. J., 368
329,331,390,430 Glaciers, 440
tropical rain, 7, 8,110,167,196,266,301, Gliridicia nigra, 417
313, 327, 332, 336, 430 Glossina medicorum, 290, 377
Foundress, 106 Goethalsia meiantha, 162,413
Four-o'clocks, 372-373 Gold Coast, 333
Frankie, G. W., 12,69, 145, 157,324-325, Gotwald, W. H., Jr., 6, 213, 385
327,357 Gradient
Fraser-Rowell, C. H., 111 altitudinal, 433, 435, 454, 459
French Guyana, 13, 292 body size, 441
Fruit c1inal, 279
removal, 278 ecotonal, 336
rotten, 336 geographic, 336, 432, 451
set, 332, 360, 367, 369, 394,399,401 sunlight, 195, 336
shadow of, 328 thermal, 195
Fruit-inhabiting flies, 176 Granadilla, 141
Fruit investment strategy, 368 Grasshopper, 112, 150, 221, 340,444
Fruiting Groundnut tree, 146
mast, 375 Guanacaste Province, 78, 102, 140, 154, 230,
seasonal, 327-328 275
Fungus gardens, 329 Guatemala, 307, 311
Guild, insect, 123, 235. 266, 336
Gadgil, M., 48, 229
Galapagos Islands, 8 Habitat, 5,49,56,67,97.104,178-195
Galeruca tanaceti, 77 associations, 431
Gene carrying capacity, 338
flow, 239 complexity, 262, 401
frequency, 177 correlation, 251
host selection, 454 definition of, 61, 178
pool, 22, 77 divergence, 176
Generalists, 57,140,201, 228,234-235,456 diversity, 8, 65,101,142,239.387
feeding habitats, 253 foraging, 105
habitat of, 210 invasion. 59, 70. 391
Genetic variation, 1,41,49,79,219,255,268, partitioning, 83
338, 340 selection, 174-178,202-207,216,250,345,
homozygosity, 34, 256 376
Genotype, 47, 48, 218 Habitat types
enzyme, 175 agricultural, 47. 165.301
frequencies, 182 arboreal, 309, 456
INDEX 503

Habitat types (cont' d) Heteroptera, 183

cow pats, 409-410 Hevea brasiliensis, 362, 394, 398, 399
desert, 57, 70, 74, 78 Hibiscus, 329
fields, 388-389 Hoffman, R. J., 320
garden, 321, 391 Holdridge, L. R., 217
grasslands, 390,393,407 Holldobler, B., 9
lowland hot and wet, 81, 83, 432 Homoptera, 30, 32, 115, 117, 120, 147, 163,
marginal, 163 191, 196, 269, 299, 342, 343
micro, 175-176,343 Host plant
pasture, 265, 410 defensive systems, 49, 357
riverine, 5 distribution, 410
secondary, 41 diversity, 350
successional, 391 egg placement on, 231
water, 5 insect colonization of, 49, 117, 132, 135,
Haffer, J., 435 142,323
Haldane, J. B. S., 43 larval, 195, 244, 294, 295, 321, 448
Halffter, G., 318 range, 344
Happold, D. C. D., 316 selection, 406
Hardy-Weinberg equilibrium, 39 specialization, 345
Hartshorn, G. S., 161,376 specificity, 176, 429
Hawaii, 98, 175, 177 variation, 150
Heinrich, B., 85-86,117,145 Hubbell, S. P., 7, 11,215
Heliconiines, 120, 169 Humic acid, 308
Heliconinae, 141,205,349-351 Humidity, and body water content, 70-71
Heliconius, 8,98, 100, 117, 132, 141, 154-156, Hummingbirds, 191,327
216,242,267,299,362,431,436,438, Hurd, P. D., Jr., 8
448-449,451 Hybridization, zones of, 450, 452
charitonius, 241-243, 332 Hymenaea courbaril, 82
cydno, 448 Hymenoptera, 105, 110, 215, 261, 382, 389
erato, 449 parasitic, 65, 305, 444
ethilla, 241, 332 social evolution of, 73
hermathena, 439 Hypercontingent species, 190-191, 194
melpomene, 448-449 Hyperodes bonariensis, 77
numata, 292 Hypothuris 121, 127,346-347
Hemiptera, 30, 32, 93, 117, 150, 151,258,269 euclea, 346
coloration, 154 colonizing episodes, 41-45
migration, 80
Herbivores, 8,17,32,49,97.114.115,137. Ichneumonidae, 312
138,140,146,149,231,275,295.337, lilies, J., 6, 440
338 India, 319
deterrants, 356 Indo-Malaysia, 179
guild, 227, 426 Inga 38, 157, 398, 418
and host plant interactions. 158-160,319, Inouye, D. W., 356
350, 410, 423 Insecticide application, 426
population size, 24-25 Interstitial species, 191
seasonal changes in populations, 76 Intertropical Convergence Zone, 316-317
Heritability, measure of, 220 Ipomoea, 326
Hermann, H. R., 101, 189, 211 Ips typographus, 233, 339
Hermaphroditism, 364, 367 Isoptera, 105
simultaneous, 369 Ithomiidae,90-92,98,12I,127,13I,146,292,
Hespenheide, H. A., 460 345, 352, 436, 449
504 INDEX

Jaisson, P., 117 Life spans, 242

Janzen, D. H., 8, 13, 14,22,26,32,49,64-67, Life style, definition, 7, 61
69,86,88,96,104,117, 123,138,140, Life tables, 2, 59, 61, 217, 224, 243, 395
142, 146, 156, 161, 162, 163, 165, 166, Limenitis archippus, 77
168, 177, 195-198,202,206,213,216, Limiting factors
217,230,262,266,275,277,279,291, for growth, 333
300,303,323,327,331,337,338,344, for host plants, 345
346,351,355-357,364,369,374-375, for insect populations, 339, 408
387-391,398,420,432,433 for pollination, 332, 368, 394
Jeanne, R. L., 190, 457 Lineages, phylogenetic, 442
Johnson, C. G., 80 Linsley, E. G., 3, 8
Johnson, L. K., 215 Lipids
Jones, M. A., 91 and diapause, 82
Jurassic, 62 in diet, 82, 169
Juvenile hormone, 72, 79 Litter
leaf, 292, 301, 303, 336
Kellett, D. N., 150 organisms, 335
Kenyi, J. M., 159,319 Lloyd, M., 413
Ketones, 96 Locusts, 79, 80, 82, 279, 310-312
Khartoum, 316-317 Longevity, 59, 73
Kinsman, S., 372 Lumbering, 420
Kohn, A. J., 9 as a perturbation, 178, 430
Krieger, R. 1., 131 Lycaena dispar, 331
Lycaenidae, 140
Lack, D., 19 Lygaeidae, 205,456
Lamas, G., 443
Lampyridae,311 MacArthur, R. H., 6, 9, 27, 49, 63, 74, 120,
Lantana camara, 371 148, 219, 295, 338, 346, 435; 436
Latitude, 79, 81, 82 McCauley, D. E., 40
Lauraceae, 300, 322 McDade, L. A., 372
Lawton, J. H., 277 Machaerium seemani, 269-270
Leaf hopper, 182, 342,405 McKey, D., 142
Leek, C. F., 321 Maelzer, D. A., 318. 407
Leguminosae, 117, 142, 163,413,419 Manduca sexta, 85
seeds, 163-164 Mating
Lepidoptera, 84, 99, 117, 127, 161, 195,216, strategies, 93, 96, 287, 340
227,244,262,268,277,288,312,319, Matthews, E. G., 318
347,358,361,377,398,435 Matthews, R. W., 110, 189
Lerner, 1. M., 41, 43 Mayflies, 377
Leston, D., 398 taxonomic diversity, 3
Levins, R., 56,121,175,177,295.421,452- Mayr, E., 49, 451, 453
453 Mechanitis, 121, 264-265, 450
Lewontin, R. c., 1,35,40,41,47,59. III. isthmia, 106-107, 141,231,346-347
218, 221, 228 Mecistogaster, 378
Liberia, 535 coerulatus, 307
Life cycle Megachile, 393
patterns, 62-63 Megaloprepus, 378
stages, 122, 156, 229, 280, 294 coerulatus, 378
Life history, 81, 217, 257, 289, 323, 353, 375 Melinaea, 451
synchronization, 283, 294 Meristems, 132, 169,203,260,277,279.293,
traits, 339 405,406
INDEX 505

Mesozoic, 62 Moilis (cont'd)

Metamorpha, 417 sphinx, 85, 87
epapha,442 spruce budworm, 137
stelenes, 320, 442 tent caterpillar, 137
Mexico, 43, 93, 98, 203, 361,437 tobacco homworm, 117
Michener, C. D., 6, III Mountain
Micrathyria, 3 insect faunas, 288, 432
Microsomal mixed-function oxidases, 123 tops, 70, 74, 84,430,433,455-458
Microstigmus comes, 110, 189 Mucuna, 160
Midges urens, 164, 270
pollinating, 261, 332-333, 367, 394- Mustard oil glycosides, 122
404 Mutchmor, J. A., 84
Migratory behavior, 54, 71, 80-82, 257, 295, Mutualism, 138, 165,202,210, 356
310-311,329,330 dissolution of, 206
Milkweed plants, 79, 81, 151,294 Muyshondt, A., 132, 156,443
Millipedes, 319 Mycorrhizal fungi, 163, 389
Mimicry, 98, 120,235,336,352-353,431,
449 Nairobi National Park, 305
Mirabilis jalapa, 372-373 Natural history studies, 195,229,257,260,
Mischocyttarus immarginatus, 189 301,346,399,424,459
Mites, nectarivorous, 191 Naumann, M. G., 93
Moffett, M. W., 126, 141, 150, 263 Nectandra, 322
Monoecism, 368 Nectar
Monophagy, 120, 132, 140, 176, 196,237,256, nutrient concentrations, 145, 216, 327, 361,
330,348 362
Monophenism, 458 rewards, 340, 370-372
Monoterpenes, 353 iliief, 194, 371-372
Monteverde cloud forest, 12, 327 time of peak production, 104
Moore, T. E., 290 Nectaries
Moraba, 221 extrafloral, 138, 167-170, 189,356-357
scura, 22 foliar, 166
Morphidae, 353 Nephila clavipes, 378
Morpho, 90-91, 98, 99, 116, 142, 184, 241, Nesting habits, 164, 189-190, 213, 340
406,443 Neto, J. Y., 150
achilles, 10 I Net reproductive rate, 219, 221, 245
amathonte, 98 New World tropics, 17
granadensis, 142, 348 New Zealand, 77
peleides Niche, 5, 237, 358
abundance, 322 breadili, 31,48,251,252,256
aggregates, 93-95 overlap, 194
daily activity, 10 I and physiological adaptations, 67-71
food of, 128, 184 shifts, 263
habitats, 270 Nigeria, 146, 166
host plants of, 142, 348 Nitrogen
phenotypic variation, 254-255 and egg production, 59
population dynamics, 240-241 in plant tissues, 117, 163,407,408,
sexual dimorphism, 50, 53 413
Mosquitoes, 306-307, 311, 323, 382 Nomadacris septempjasciata, 312
Moilis, 324-325, 346, 360,451 Noonan, G. R., 435, 443-444, 456
armyworm, 160 Norili America, 3, 137, 413
gypsy, 137 Notiobia peruviana. 435,456
506 INDEX

Novomessor Parasites, 19,48,65,78, 114,207-210,254,

albisetosus, 83 278,426
cockerelli, 83 population dynamics of, 76, 238, 243
Nymphalidae, 300, 320 Parasite types
ecto-,440
Oak trees, 157,275 endo-, 194, 261
Ocotea, 36, 38 k1epto-, 189
Odonata, 3, 5, 8, 17,93, 189,307,318,378 Parasitic ants, 231
Oleria zelica pagasa, 142 Parasitic flies, 16, 228, 230, 231
Oligocene, 63 Parasitic wasps, 16, 159,228,230,231,244,
Oligolectic behavior, 86 261
Onchocerciasis, 311 Parasitism
Oncopeltus, 69, 79-82, 246-247 floral, 371-372
Jasciatus, 80 muiti-, 261
Onthophagus gazella, 410 Parides. 50, 235, 294, 332, 406
Opler, P. A., 78, 295, 296 Park, 0., 5
Optimization theory, 67 Park, T., 24, 43,111
Optomotor response systems, 69 Pars intercerebral is , 71
Orchidaceae, 398 Paryphes blandus. 88-89, 154
Organization for Tropical Studies, 12 Passiflora
Orians, G. H., 9 as food, 350
Orthoptera, 101, 157, 184, 186, 196,251,269, vitifolia. 205
277,288,292,299,300,346,347 Passifloraceae, 141, 169, 351, 448
Olyctes rhinoceros. 410, 421 Patrick, R., 3
Osa Peninsula, 215. 293 Paulson, D. R., 3, 8
O'Shea, R .• 117 Pectinophora gossypiella, 82
Overwintering behavior, 294 Pemphigus. 352
Oviposition. 43, 93, 98. 126, 133, 169.299, Pelltaclethra. 30, 189,418
310 macroloba, 147,161,336,418
cluster, 126, 231. 259 Pentacomia egregia. 308
single, 347 Periplaneta americana. 24
sites, 413 Pem1ian, 62
Oviposition behavior. 3, 132, 135.348.407- Perrhybris. 38
408 lypera. 36, 98, 231
Owen, D. F., 312, 321 Peru, 443
Oxidases, mixed function, 123, 131, 154 Pesticide application, 390, 422
Peters, W. L., 3
Pachyptera hymenaea, 361 Petiolar cavities, 167
Pachysima, 166 Phasmids, 101
Paine, R. T., 26 Pheidole. 167
Palatability, 141, 150, 154,259 Phenolics, 122
Paltothyreus, 9 Phenotype
Panama, 12, 147, 299, 321, 327 definition, 34
Papilio flexibility of, 132, 255, 338, 341
machaon, 133 frequencies, 182
polyxenes, 85 seasonal, 458
zelicaon, 77 Pheromones
Papua New Guinea, 183,410, 444 alarm, 96
Paraponera. 9 sex, 120
clavata. 101, 110, 189, 211 Philaenus. 268
INDEX 507

Philaethria dido, 443 Polybia, 93, [10

Photoperiod, 73, 81, 82, 279, 312, 320 occidentalis, [89
and activity, 285, 290 Polybiine wasp, [89
and diapause, 79 Polymorphism, 34-35,175, [77,249,255,267,
and migration, 69 292, 296, 450, 455
Phymateus purpurascens, 112 Polyphagy, [2 \, 132, [40, 154, 176, 196, 231,
Pianka, E. R., 25, 27 237, 256, 348
Pieridae, 12, 36, 78, 98, 233, 296, 458- Polyphenisms
459 latent, 458-459
Pieris, 205 Pompilidae, 107
rapae, 181 Ponerinae, [10, 2[ I
Piegon pea, 393 Poole, R. W., [41
Pimentel, D., 24, 353 Population
Piperaceae, [67 census history, [5, 22-24, 233, 239, 276
Piper nigrum, 417 control mechanisms
P[antations, cacao, [86- [88 abiotic, [4, 228
Plant defense mechanisms, 13, 49, 7 [, [[ 7 , biotic, 14,217,224,228,232,234-235,
132, 137-[40, 145, [56, 182,243, 237,244,248,257,260,28[,288,297,
273-276, 337, 406 319,341,347,355,356,389,390,410,
Plate tectonics, 438 424
Platypodidae, 202 physiological, 313
Plecoptera, 440 differentiation, 40, 220, 254
Pleistocene, 25, 149,254,431,438,456,458- dynamics, 22, 54, 221, 226, 232, 240, 327
459 outbreaks, 137,238,281,318-319,335,339,
Pliske, T., 90-91, 120, 146 348,355,399,410,421,453
Polistes, 73,93, 106 regulation, 25, 229-239, 242
Pollen size, 240, 242-246, 271, 332, 377
balls, 362 stability, 241, 254
collecting, 326 structure, 96, 241, 256
donor, 370, 395 Populus, 352
feeding, 116, 215, 216, 241, 300, 398 Pough, F. H., 150
interference, 364 Prance, G. T., 438, 440
nitrogen, 408 Predation, 19,32,48,65,78,86,96-97,114,
production cost, 368 254
rewards, 371 Predator
robber, 371, 372, 403 avoidance, 287
starch content, 362 population dynamics of, 238, 243
Pollinating guild, 325, 370-371 seasonal changes in populations, 76
Pollination seeds, 160-164
cross-, 371 vertebrate, 235, 258
systems, 138, 149,323-327,329,340,358- Predator-parasitoid complex, 388
375, 39 [-394 Prey preferences, 340, 385
self-compatib[e, 360 Price, P. W., 273
self-incompatible, 360, 364, 394 Procollina biolleyi, 343,414-415
Pollinator Productivity
abundance, 332-333, 368 harvestab[e, 210, 2[4, 249, 250, 261,350,
asynchrony with host plant, 40[ 388, 399
avai[ability, 138, 364 hypothesis, 26
limitation of fruit set. 367 Protective coloration, 235
long-distance, 266, 342 Protective guard insects, 361
508 INDEX

Prothoracic gland, 71-72 Resource (conI' d)

Psephalidae, 5 patchy, 260
Pseudomyrmex, 165, 166,356-357,390 utilization, 28-29, 43
Psyllidae, 117 variety, 351
Pteridines, 320 Resource abundance, 207
Puerto Rico, 175,206, 398,452 Resource allocation, 229
Pulsations, glacial, 441 Rettenmeyer, C. W., 6
Puntarenas Province, 102 Rhoades, D. F., 133, 139
Purseglove, J. W., 366, 394 Rhodnius, 72, 77
Pursuit invitation behavior, 352 Rhythms
Pyrrolizidine alkaloids, 120, 146 activity, 100-101, 104
daily, 285-287
Quesada gigas, 102, 184-185,313-314 Richards, P. W., 179,366
Ringo, J. M., 98
Rainfall Risch, S., 167
and butterfly mortality, 235-236, 294 RNA synthesis, 72
and diapause, 79 Robinson, B., 189
and insect density, 348 Robinson, M. H., 189
seasonal patterns, 317, 318, 322 Rockwood, L. L., 110, 140,262, 329
Rain shadows, 362, 441 Roosting, communal, 259, 267
Range Root
geographical, 434, 452-454 clover, 407, 408
topographical, 77 crowns, 30, 117, 291,413,415
Range expansions, 180,431 systems, 117, 189,407
Rapoport, A., 54 Root, R. B., 81
Rathcke, B. J., 141 Rosenthal, G. A., 123, 140
Rausher, M. D., 203 Rothschild, M., 150
Raven, P. H., 3, 156, 182 Rowell, H. F., 186
Refugia Ruwenzori National Park, 319
evolutionary, 434
glacial forest, 350 Salticidae, 395
model, 439, 448-451 Sanders, H. L., 70
moist, 248,427 Sap-feeding insects, 163
moist forest, 78, 196, 275, 278, 288, 407 Satiation, of predators and parasites, 230
mountain forest, 148 Saturniidae, 295
natural habitat, 422 Satyridae, 91, 184, 320, 353
Quaternary, 351,431,441,446,448 Satyrinae, 348
riparian forest, 312 Savanna zone, 329
Region, definition, 254-255 Scavenging insects, 278
Rehr, S. S., 123, 160, 164,356 Scent organs, 153
Reproductive effort, 229 Schaefer, C. W., 117
Reproductive isolation, 182 Scheelea, 312
Reproductive output, 332 Schistocarpha oppositifolia, 186,404
Reproductive patterns of plants, 169 Schistocera gregaria, 312
Reproductive rate, 247 Schistosoma, 311
Reproductive value, 248 Schmidt, 1.0.,110
Resource Schoener, T. W., 88,195,196,198,217,230.
distribution, 65 275,291,331
limitation, 135, 260, 340 Schram, F. R., 62
partitioning, 194, 345, 349-351 Schwerdtfeger, W., 74
INDEX 509

Scolytidae, 202 Simberloff, D., 148

Scorpion flies, 377 Simpson, B. B .. 97,447
Scott, J. A., 459 Simulidae, Amazonian, 287
Scriber, 1. M., 264, 406 Simulium, 307
Scudder, G. G. E., 150-151 neavi, 377
Season Slash-and-bum clearing. 79
dry, 3,52, 69, 73, 74-75, 78, 79,81,82,86, Slater, 1. A., 206, 344, 456
146,244,248,250,261,267,278-280, Slobodkin, L. B., 54, 70
283,292,297,306-312,318-333,382, Smiley, 1. T., 141,345,347.348
389,401,409, 427 Smith, 1. M., 177
fauna density in, 180, 196-199,275 Smythe, N., 54, 327
growing, 73, 426 Social behavior, 63, 93, 96, 104-111
wet, 75, 78,197,248,261,269,277,286- Solanaceae, 345
289, 305 Solanum rugosum, 41, 346
and insect abundance, 275, 293, 297 Songs
Seasonal changes, in insect populations, 74, 75, species-specific, 101
299-322 timing of chorus, 102-104, 290
Seasonality, 12-13,26-27,57,64,96,138, South Africa, 117, 206
235,248,273-333,341,388,437 South America, 25, 74, 79, 101, 148, 179,255,
Seed 265,340,348,351,394,429,431,435,
predation, 160-164, 195,206,231,266,327, 437
329,339,341,357,369 Southwood, T. R. E., 80
production costs, 368 Specialists
Seed crop, 328 ecological, 57, 65,140,201,239,331,342.
production, 161 456
Seed dispersal agents, 10, 358-359 grain, 191
Seed mortality agents, 10 habitat of, 210
Selection, 41,54,59,217-221,239 sequential, 191
density-dependent, 295 Speciation, 27,49,430,431
directional, 36-38, 49, 238, 296, 353, 420, and species abundance, 6
431. 455-457 isolating mechanisms, 279
disruptive, 38-40, 267, 296 Species
frequency-dependent, 288 diversity, 278
individual, 221, 245 richness, 338, 384
K, 226-228,231,234-237,266 Speyeria, 432
models. 36-40 Sphecidae, 107, 340
r, 226-228,231,235,266 Spiders, orb weavers, 189, 378
seasonal, 289 Sphingidae, 86, 313, 325, 327, 359
stabilizing, 36-37, 56, 89, 169,338.353, Stanton, N. L., 16
420, 449-450, 455-456 Staphylinidae, 308-309, 383
units of, 220 Stem-boring guilds, 348
Sequestering mechanisms, 120-121, 151 Sterculia, 329
Sex ratios, 240-241 apetala, 328
Sexual selection, 96 Stradling, D. 1., 353
Seychelles, 426 Strehl, C. E., 162
Sgarbieri, V. c., 163 Strong, D. R., If., 344, 355
Shapiro, A. M., 12,233,443,458-459 Stubblebine, W., 82
Sheltering behavior, 352 Succession, 86, 339
Sida, 329 seasonal, 327-329
Silberglied. R., 183,442 stages, 388
510 INDEX

Surinam, 284 Toxic secondary substances, 146, 256, 258

Survivorship, 40, 240, 246 Transient species, 336
age-specific, 221 Trapline home range movements, 370-371
Swarm raids, 342 Tree falls, 178, 337, 388
Sweat, human, 108-110 Tribolium, 24,43, 231, 255
Swihart, S, L., 68 Trichome hairs
Swollen-thorn acacia, 165 evolution of, 141
Symphyla, 62 Trigona, 86, 186, 326
Synchronization foraging behavior, 215
of flowering, 86, 279, 292, 323, 368, 371 Julviventris, 372
of life cycles, 230, 283, 289, 308, 312, 323 nesting habits, 6
of populations, 290, 329 jaty, 186, 403-404
Trinidad, 183, 186, 332, 348
Tabanidae, 309-311 Trophic level, 251, 278, 335
Tabanus Trophic relationships, 6, 198
guineensis, 310 Trypoxy/on, 340
taeniola, 310 Turner, 1. R. G., 350, 352, 448
Tachinidae, 207, 231, 232, 235, 278 Tuskes, P. M .. 54
Tannins, 157, 273
Tanzania, 377 Uganda, 112,142,159,183.243.312,319,
Taxon cycle, 47-49, 165, 296, 457 377
Taylor, O. R., Jr., 356 Ultraviolet reflection cues. 352
Temperate Zone, 3, 6, 17, 32, 79, 80, 83-84, Upper Carboniferous. 62
100 Uraniafulgens, 81
changes in climate, 64 Usinger, R. L., 8
insect populations, I, 14, 15,75-80
Temperature
body, 81-88 Vandermeer, J. H., 159
daily variations in, 284-285 Van der Pijl, L.. 145,358
Tephritidae, 183 Vane-Wright, R.I., 91,431
Tennites, 105, 335, 391 Vanilla fragrans, 398
Tertiary, mid-, 438 Vanzolini, P. E., 435
Texas, 82 Vaughn, F. A., 151
Thasus acutangulus, 154 Vectors, disease, 335
Theobroma, wild species, 366 Venezuela, 82. 284
Theobroma cacao, 394, 399 Venoms. ant, 110
insects on, 453 Vepsalainen, K., 296
pollination of, 261, 325. 332-333. 362-365 Vespidae, 107.325
Thermal acclimation, 453 Vicariance model. 434
Thermoregulation. 85-86, 320, 352 Voltinism, 77.86. 137,276-277,318,408,
Thermoregulatory stress, 320 459
Thomason, J. H., 93, 240
Thornhill, R., 93 Wade, M. J .• 40
Thrips, 393 Waldbauer. G. P., 295
Thunbergia grandif/ora, 215,372 Walker, T. J., 101
Thyrnnidae,408 Wasps, 93, 111. 159. 189.249,300.323,324.
Tiliaceae, 162,413 345-347,360-361.383
Time stability hypothesis, 26 paper, 73,107,343,372
Tobago, 183 sand, 340
Tomatine, 158 sawflies, 126
INDEX 511

Watson, J. A. L., 3 Wing

Webb, J. W., 117 damage patterns, 98
Weevils evolution of, 63
Argentine stem, 77 muscle degeneration, 257
flightless, 183 pigments, 320
Wellington, W. G., 63 pockets, 91
Wenzel, R. L., 17 Wirth, W. W., 364
West Africa, ISO, 377 Wisconsin, 84, 176
West-Eberhard, M. J., 93 Wolda, H., 299
Western Samoa, 410 Wright, S., 239, 266
West Indies, 49, 206
White, J., 162,413 Young, A. M., 30, 36, 41, 50, 78, 83, 84, 93,
White, M. J. D., 221. 434 98,101-104,110,114,116,117,126,
White water, 308 131,132,135,141,142,156,157,161,
Wiens, J. A., 251, 253, 254 162, 186, 189,201,205,207,211,215,
Wigglesworth, V. B., 71, 77 240,255,263,289,290,291,300,307,
Wiklund, c., 135 313,320,323,328,332,340,343,345,
Wille, A., 6, 62, 111,215 346,348,372,395,401,417,442-443
Williams, C. M., 72
Williams, E. E., 435 Zacryptocerus varians, 213
Wilson, D. E., 344 Zaire, 184
Wilson, D. M., 69 Zammara
Wilson, E. 0.,40,47,93,104,105,107,148, smaragdina, 50-52, 161-162,343,413
164,213,219 smaragdula, 184,313,315,317
Wilson, M., 420 Zingiberales, 343-344
Winder, J. A., 395,401 Zophohas, 69
Windsor, D. M., 189 Zygia longifolia, 416,418