QUANTITATIVE

ECOLOGY -
A NEW UNIFIED
APPROACH

Clarence Lehman, Shelby Loberg, & Adam

Clark
University of Minnesota
 University of Minnesota
Quantitative Ecology - A New Unified
Approach

Clarence Lehman, Shelby Loberg, & Adam

Clark
This text is disseminated via the Open Education Resource (OER) LibreTexts Project (https://LibreTexts.org) and like the hundreds
of other texts available within this powerful platform, it is freely available for reading, printing and "consuming." Most, but not all,
pages in the library have licenses that may allow individuals to make changes, save, and print this book. Carefully
consult the applicable license(s) before pursuing such effects.
Instructors can adopt existing LibreTexts texts or Remix them to quickly build course-specific resources to meet the needs of their
students. Unlike traditional textbooks, LibreTexts’ web based origins allow powerful integration of advanced features and new
technologies to support learning.

The LibreTexts mission is to unite students, faculty and scholars in a cooperative effort to develop an easy-to-use online platform
for the construction, customization, and dissemination of OER content to reduce the burdens of unreasonable textbook costs to our
students and society. The LibreTexts project is a multi-institutional collaborative venture to develop the next generation of open-
access texts to improve postsecondary education at all levels of higher learning by developing an Open Access Resource
environment. The project currently consists of 14 independently operating and interconnected libraries that are constantly being
optimized by students, faculty, and outside experts to supplant conventional paper-based books. These free textbook alternatives are
organized within a central environment that is both vertically (from advance to basic level) and horizontally (across different fields)
integrated.
The LibreTexts libraries are Powered by NICE CXOne and are supported by the Department of Education Open Textbook Pilot
Project, the UC Davis Office of the Provost, the UC Davis Library, the California State University Affordable Learning Solutions
Program, and Merlot. This material is based upon work supported by the National Science Foundation under Grant No. 1246120,
1525057, and 1413739.
Any opinions, findings, and conclusions or recommendations expressed in this material are those of the author(s) and do not
necessarily reflect the views of the National Science Foundation nor the US Department of Education.
Have questions or comments? For information about adoptions or adaptions contact [email protected]. More information on our
activities can be found via Facebook (https://facebook.com/Libretexts), Twitter (https://twitter.com/libretexts), or our blog
(http://Blog.Libretexts.org).
This text was compiled on 01/01/2024
TABLE OF CONTENTS
Licensing

1: What is Ecology?
1.1: Introduction
1.2: Definitions of Ecology
1.3: Ecology then and now
1.4: Methods of ecology

2: Ecological Theory
2.1: Levels of Ecology
2.2: Role of theory
2.3: What is a model?
2.4: Present State

3: A Basic Population Model

3.1: Characterizing populations
3.2: Bacterial Growth
3.3: A First Computer Model
3.4: Program results
3.5: Darwin's elephants

4: Modeling a Single Population

4.1: Density-independent growth
4.2: Density-enhanced growth
4.3: Density-limited growth
4.4: Parameter combinations
4.5: Generalization

5: Differential and Difference Forms

5.1: Chapter Introduction
5.2: A mathematical view
5.3: Exponential Solution
5.4: Logistic solution
5.5: Orthologistic solution

6: Human Population Growth

6.1: Chapter Introduction
6.2: Phenomological Graph
6.3: Biological-ecological graph
6.4: A global transition

7: Chaos and Randomness

7.1: Chapter Introduction
7.2: Depicting population growth

1 https://bio.libretexts.org/@go/page/25834
 7.3: Hypothetical insect data
7.4: Sensitive Dependence
7.5: Dampening of Chaos
7.6: Bifurcation diagram
7.7: Properties

8: Theory of Interactions
8.1: Dynamics of two Interacting Species
8.2: Code for two species
8.3: Summary of interactions

9: Embodied by Natural Selection

9.1: Chapter Introduction
9.2: Mutualism
9.3: Herbivory
9.4: Predation
9.5: Predatory plants
9.6: Defense
9.7: Competition

10: Phase Space

10.1: Chapter Introduction
10.2: Phase Space
10.3: Surface generating the flow
10.4: A phase space example

11: State Spaces

11.1: State Spaces

12: Predator and Prey

12.1: Further Examples of Predation
12.2: Ecological communities are complex
12.3: Predator-prey model
12.4: Phase Space
12.5: Assumption of the basic model
12.6: Independent carrying capacities
12.7: Predator satiation and starvation
12.8: Effects of Space

13: Humans as Predators

13.1: Chapter Introduction
13.2: Consciously Controlled Predation
13.3: Stochastic modeling
13.4: Present Situation

14: Humans as Prey

14.1: Chapter Introduction
14.2: Portals
14.3: Pathogen Mobility

2 https://bio.libretexts.org/@go/page/25834
 14.4: Disease more generally
14.5: The Strange Case of Polio
14.6: Modern disease trends
14.7: An ancient plague perhaps vanishing

15: Theory of Disease

15.1: New terminology
15.2: The SIR flowchart
15.3: The SIR Equations
15.4: The SI Equations
15.5: The I Equation
15.6: Analysis of the I model
15.7: 15.7 The rsN model again
15.8: Application to an actual outbreak
15.9: Conflicting ethics

16: Competition
16.1: Chapter Introduction
16.2: The niche concept
16.3: Resource competition
16.4: Competitive exclusion
16.5: Succession
16.6: Single-resource phase space
16.7: Multiple resource phase space
16.8: Lotka-Volterra formulation

17: Mutualism
17.1: Chapter Introduction
17.2: Phase Diagrams
17.3: Quasi-species
17.4: Our own species

18: Higher-order models

18.1: Chapter Introduction
18.2: One-species Example
18.3: Two-species blending
18.4: A way of organizing

Index
Glossary

Detailed Licensing

3 https://bio.libretexts.org/@go/page/25834
Licensing
A detailed breakdown of this resource's licensing can be found in Back Matter/Detailed Licensing.

1 https://bio.libretexts.org/@go/page/97893
 CHAPTER OVERVIEW

1: What is Ecology?
1.1: Introduction
1.2: Definitions of Ecology
1.3: Ecology then and now
1.4: Methods of ecology

This page titled 1: What is Ecology? is shared under a CC BY-NC 4.0 license and was authored, remixed, and/or curated by Clarence Lehman,
Shelby Loberg, & Adam Clark (University of Minnesota Libraries Publishing) via source content that was edited to the style and standards of the
LibreTexts platform; a detailed edit history is available upon request.

1
1.1: Introduction
Winston Churchill pointed out that “All the great things are simple, and many can be expressed in a single word— freedom, justice,
honor, duty, mercy, hope.” Should we try to dene these? Can we dene them?
We should at least try to dene our subject, ecology; many textbooks start with denitions. But rst, for background, consider how
we might dene life. Marvin Minsky was an articial intelligence researcher and computer scientist who thought about denitions.
When is an object alive? Think about viruses, genes, self-reproducing machines—no one has really been able to give a good
denition of “living” that satises in general. Some things are clearly living—mice—and some clearly are not—rocks. Lists of
what makes something living used to appear in textbooks:
1. Self-reproducing
2. Responds to stimuli
3. Metabolizes
4. Made of protoplasm—protein, carbohydrates, DNA.
But (1) puts out the mule, (2) and (3) put out the spore, while if those conditions are dropped, (4) will admit the frankfurter. One
can go on to extend the list with more careful qualications, but questions remain until the list grows to include special mention of
everything we can think of.

This page titled 1.1: Introduction is shared under a CC BY-NC 4.0 license and was authored, remixed, and/or curated by Clarence Lehman,
Shelby Loberg, & Adam Clark (University of Minnesota Libraries Publishing) via source content that was edited to the style and standards of the
LibreTexts platform; a detailed edit history is available upon request.

1.1.1 https://bio.libretexts.org/@go/page/25415
1.2: Definitions of Ecology

This page is a draft and is under active development.

With caveats in mind, consider denitions of ecology. In the 1860s, Ernst Haeckel, combined the term oikos—a place to live,
home, habitat—with logia—discourse, study—to coin the word “ecology.” In the 1890s Ellen Richards included humans and
harmony, quite a modern view. Variations over the years are shown in Table 1.2.1.
Table 1.2.1 Various views of ecology.
The total relations of an organism to its organic
Haeckel 1860s
and inorganic environment
Living in harmony with the environment, first
Richards 1890s
including the human species

Elton 1920s Scientic natural history

The study of structure and function of nature,

Odum 1960s
including the human species
The scientic study of the distribution and
Andrewartha 1960s
abundance of organisms
The scientific study of the interactions that
Krebs determine the distributions and abundance of
organisms
The study of relationships between organisms
Molles 1990s
and the environment

Eltis 2010s Life in context

The relationship between living organisms and

Pope Francis 2015
the environment in which they develop

Each of these denitions has merit, but the rst two and the last two are closest to the way the term is applied in this book. We
humans have become prominent in ecology, locally to globally. No modern treatment of ecology is complete without a strong dose
of anthropology.
The denition by Andrewartha has been widely quoted, but focusing merely on distribution and abundance reduces ecology to
mapping, which is why Krebs modied this definition. The Pope’s denition from his 2015 Encyclical includes the interesting idea
of development, which can be taken to mean short-term development like embryogenesis and growth, plus long-term development
like evolution. Overall, the denition by Eilts is perhaps the most general and engaging.
First and foremost, the most important concepts in ecology are about relationships, plus all of life, the whole environment, the
processes of living and development, and, above all context. And in today’s world, harmony. But also consider, “Poetry is the
subject of the poem” (Wallace Stevens, 1937) and perhaps “Ecology is what ecologists do.” With these in mind, we strive in the
remainder of this book to dene a theoretical form of ecology through examples and demonstrations, representative models and
symbols, patterns and explanations, and lessons and caveats.

1.2.1 https://bio.libretexts.org/@go/page/25577
 Figure 1.2.1 : The Rhind Papyrus, c. 1640 BC. One of the oldest known documents—and containing exercises from theoretical
ecology!

This page titled 1.2: Definitions of Ecology is shared under a CC BY-NC 4.0 license and was authored, remixed, and/or curated by Clarence
Lehman, Shelby Loberg, & Adam Clark (University of Minnesota Libraries Publishing) via source content that was edited to the style and
standards of the LibreTexts platform; a detailed edit history is available upon request.

1.2.2 https://bio.libretexts.org/@go/page/25577
1.3: Ecology then and now
Our early hominin ancestors needed aspects of ecology. To nd blueberries or other fruit, or where to dig wild onions, they had to
know where these foods grew—their distribution and abundance. These parts of ecology have thus been part of life for hundreds of
thousands of years. Ecology is connected with our species.
Some elements of the eld of ecology were formalized more than 3000 years ago. The Rhind Papyrus lists a number of ecological
exercises for students—mathematics from ancient Egypt. Among these oldest ecological problems is this:
Number 27. If a mouse eat 521 ikats of grain each year and a cat kills 96 mice a year, in each of 24 barns, how many cats are
required to control the destruction of stored grain?
This is a little problem in quantitative ecology! Even 36 centuries ago, mathematical ecology was part of life. Knowing how many
grain bins determined how many cats were to be employed.
Today, ecology has become a glamour word. A product called “Ecogate,” for example, is part of a central vacuum system that
keeps sawdust and sanding dust from being tracked around. But why the word Ecogate? Dust collection per se has nothing to do
with ecology. Advertisers, however, have found that consumers respond positively to the term.
The term “ecosystem” is frequently used in business and nance, but there it means a collection of companies, customers, and
products and their interconnections. For better or worse, ecological terminology is expanding to other domains.

This page titled 1.3: Ecology then and now is shared under a CC BY-NC 4.0 license and was authored, remixed, and/or curated by Clarence
Lehman, Shelby Loberg, & Adam Clark (University of Minnesota Libraries Publishing) via source content that was edited to the style and
standards of the LibreTexts platform; a detailed edit history is available upon request.

1.3.1 https://bio.libretexts.org/@go/page/25416
1.4: Methods of ecology
How do ecologists do ecology? Often, they start with observation, then move to theory—trying to t observations together to make
sense as a whole. Theory then leads to expectations, which in turn lead to experiments. Commonly, experiments aren’t undertaken
until there is some theory to be tested and understood.
1. Observation
2. Theory
3. Experiment
4. Serendipity
Observation, theory, and experiment, however, are not the whole story. A large part of science turns out to be serendipity—luck and
chance—capitalizing on chance and doing something with it. One example is Alexander Fleming, who discovered penicillin. Some
of the bacterial cultures in his lab became contaminated with penicillium mold and the cultures died. That ruined his experiment.
He could have written a memo to the laboratory staff ordering “Always keep mold away from our bacterial cultures. It destroys the
cultures and will ruin the hypotheses we are trying to test.” But instead he capitalized on the serendipity, wondered what was
happening, and found a substance in penicillium mold that kills bacteria. Fungi and bacteria have been archenemies for perhaps a
billion years. Fleming’s discovery has helped physicians actually cure disease, rather than being limited to diagnosing and
prognosticating.
Following up on chance is, then, a large part of science. By the way, for an interesting paper, read the original 1929 report by
Fleming about penicillium. It is so understated. He writes “the name ‘penicillin’ has been given to ltrates of broth cultures of the
mould.” No one had heard of the word before. Then he suggests that “it may be an efficient antiseptic.” One of the greatest
discoveries of all time and only, “it may be an efficient antiseptic.”
Cedar Creek is a University of Minnesota research site about thirty miles north of the University’s Saint Paul campus, and is one of
the classic ecological research sites in the world. Pictured in Figure 1.4.1 is an experiment set up by Prof. David Tilman. While
very carefully designed, it came about because of serendipity—the chance event of a deep two-year drought that altered the
abundances of species in a particular way and triggered the idea for this experiment.
Keep your eyes open for such chance events; they can crop up anywhere.

Figure 1.4.1 Observations and experiments testing theory at Cedar Creek. This entire experiment was established following up on
serendipity.

This page titled 1.4: Methods of ecology is shared under a CC BY-NC 4.0 license and was authored, remixed, and/or curated by Clarence
Lehman, Shelby Loberg, & Adam Clark (University of Minnesota Libraries Publishing) via source content that was edited to the style and
standards of the LibreTexts platform; a detailed edit history is available upon request.

1.4.1 https://bio.libretexts.org/@go/page/25417
 CHAPTER OVERVIEW

2: Ecological Theory
2.1: Levels of Ecology
2.2: Role of theory
2.3: What is a model?
2.4: Present State

This page titled 2: Ecological Theory is shared under a CC BY-NC 4.0 license and was authored, remixed, and/or curated by Clarence Lehman,
Shelby Loberg, & Adam Clark (University of Minnesota Libraries Publishing) via source content that was edited to the style and standards of the
LibreTexts platform; a detailed edit history is available upon request.

1
2.1: Levels of Ecology
Ecology covers a vast range of topics and can be viewed on multiple levels. One level is that of the individual organism— a single
bacterium, an individual wolf pup. This includes individual behavior and physiology, with behavior as part of ecology. Population
ecology covers groups of organisms of the same species—a bison herd or a grove of maples. Community ecology looks at how
different populations interact, and the communities examined can be quite large. Above this level is ecosystem ecology, which
examines how different communities interact with their environments. Finally, there is global ecology—ecology of the planetary
ecosystem.

Individual ecology Single organisms, behavior, and physiology

Population ecology Groups of organisms from a single species

Community ecology Populations of interacting species

Ecosystem ecology Multiple communities and the environment

Global ecology The planet as a biosphere

Figure 2.1.1 Levels of complexity.

This page titled 2.1: Levels of Ecology is shared under a CC BY-NC 4.0 license and was authored, remixed, and/or curated by Clarence Lehman,
Shelby Loberg, & Adam Clark (University of Minnesota Libraries Publishing) via source content that was edited to the style and standards of the
LibreTexts platform; a detailed edit history is available upon request.

2.1.1 https://bio.libretexts.org/@go/page/25422
2.2: Role of theory
From its early days, ecology has been in part a theoretical– mathematical science, and it is now also a computational science.
Mathematical theory arises where systems are relatively simple. In our modern era, computation can address somewhat more
complex systems, though creating computations on complex systems that satisfy the basic tenets of science is still problematic. For
very complex systems, narrative is all we have available.
Examine the levels in Figure 2.1.1 to think about where theory applies. Subatomic particles and atoms are the realm of quantum
mechanics, one of the most sublime and successful theories. Theory applies nicely to the hydrogen atom, a two-particle object. And
while it applies to larger atoms, the raw mathematics becomes too complex as the number of particles grows, so computation
comes into play. At higher levels like the molecular one, theory is harder to apply. Organic chemistry, for example, is not a strongly
mathematical science, and at the level of protoplasm and cells there is no comprehensive mathematical theory or computational
equivalent. This level is far too complex—with minuscule molecular machines running along tubules and carrying mitochondria on
their backs at high speed relative to their size, it is more complex than any industrial factory. At the level of tissues and organs
systems, we have only narratives to guide our understanding.
What happens, then, at the level of organisms, at the entry to ecology? Individual organisms are exceedingly complex. There is no
complete mathematical theory for the internal operation of individual organisms. But externally, organisms behave as a unit and
populations become simpler than individuals—glossing over heartbeat, neuron firing rates, white blood cell replication, and so on,
with all their enormous complexity. Details disappear. Populations can be described with basic mathematics. Communities are
more complex, but are still within the reach of mathematics and, particularly, within the reach of computation. And ecosystems are
complex, but with some unifying properties.
The whole earth thus begins to be simpler, and at the level of planets and solar systems, things once again become nicely
mathematical. This is the level where, with Newton, modern science was born. In part, this emerging simplicity is because levels of
detail again merge together. At the level of planetary orbits, it does not matter that dinosaurs once dominated the planet or that
Mozart ever wrote any concertos.
At larger scales still, solar systems are completely describable with computers, although the mathematics becomes difficult, and as
we move out into galaxies and the entire universe the descriptions become difficult again.
Changing scales thus involves the successive movement in and out of simplicity. Where is the complexity in the universe greatest?
It turns out to be at about one meter. In other words, at our scale. A great spike in complexity appears just where we and other
forms of life arose.
That is no accident. A philosophical idea called the weak anthropic principle suggests that any part of the universe that can sit
around and contemplate itself and the larger universe must itself be complex. We are constrained to live at a scale of great
complexity, or not to exist at all. That is worth some reection.
But we try to find simplicity among this complexity, to let us feel we understand, and to let us predict what can happen.

This page titled 2.2: Role of theory is shared under a CC BY-NC 4.0 license and was authored, remixed, and/or curated by Clarence Lehman,
Shelby Loberg, & Adam Clark (University of Minnesota Libraries Publishing) via source content that was edited to the style and standards of the
LibreTexts platform; a detailed edit history is available upon request.

2.2.1 https://bio.libretexts.org/@go/page/25423
2.3: What is a model?
Science strives for simplicity, and models are part of the process. What is a model? It is just a simplified view of something more
complex.
The word “model” is used here essentially as it’s used in everyday English. For example, in ordinary English, “modeling clay” can
be used to make simplified miniatures of three-dimensional images of animals, automobiles, buildings, or even full-scale three-
dimensional images of objects like the human heart. A “model airplane” can be rendered to show at a glance the physical
appearance of a large aircraft, and can even be constructed to y so as to test aerodynamics under proper rescaling. A “model
organism” is a simpler organism that may respond to medical tests or treatments in ways similar to those of a more complex
organism.
Even the fashion model on the runway meets this definition of a simplified view of something more complex. The infinite
complexity of the human spirit is not relevant on the runway; all that is relevant in this context is the person as a realistic way to
display fashions.
This book focuses on computational and mathematical models of ecological systems. What is left out of these models is as
important as what is put in. Simplification is key.

If you have a complex natural system you don’t understand, and you construct a computer
model incorporating everything you can about that natural system, you now have two
systems you don’t understand. — after Chris Payola, UMN
A designer knows he has achieved perfection not when there is nothing left to add, but
when there is nothing left to take away. — Antoine de Saint-Exupery
Two different simplifications of time are commonly used in ecological models:
Discrete time — Events happen at periodic time steps, as if time is non-existent in between.
Continuous time — Events happen smoothly and at all times.
In addition, there are two different classes of models:
Macroscale — Individual organisms are not tracked, but are measured in aggregate and represented by composite variables
such as N.
Microscale — Individual organisms are tracked separately. These are also known as agent-based or individual-based models.
Macroscale models can be handled either by computers or mathematics, but microscale models are usually restricted to computers.
Keep in mind that all four categories are only approximations of reality. Later in this book we will also explore mechanistic versus
phenomenological models.

This page titled 2.3: What is a model? is shared under a CC BY-NC 4.0 license and was authored, remixed, and/or curated by Clarence Lehman,
Shelby Loberg, & Adam Clark (University of Minnesota Libraries Publishing) via source content that was edited to the style and standards of the
LibreTexts platform; a detailed edit history is available upon request.

2.3.1 https://bio.libretexts.org/@go/page/25424
2.4: Present State
As a surprising side note, the standard models commonly taught in ecology courses are not complete, and a main purpose of this
book is to help make them more so. One aspect of theory related to simple species, for instance—called orthologistic population
growth— is rarely even studied, much less taught, yet is essential for understanding rapidly growing populations, including human
populations in millennia past. For two-species interactions, another theory concerning mutualisms and a related kind of population
growth is highly under-developed, and the theory of three-species interactions is even less complete.

Figure 2.4.1 The eternal mystery of the universe is its comprehensibility. —A. Einstein

This page titled 2.4: Present State is shared under a CC BY-NC 4.0 license and was authored, remixed, and/or curated by Clarence Lehman,
Shelby Loberg, & Adam Clark (University of Minnesota Libraries Publishing) via source content that was edited to the style and standards of the
LibreTexts platform; a detailed edit history is available upon request.

2.4.1 https://bio.libretexts.org/@go/page/25425
 CHAPTER OVERVIEW

3: A Basic Population Model

3.1: Characterizing populations
3.2: Bacterial Growth
3.3: A First Computer Model
3.4: Program results
3.5: Darwin's elephants

This page titled 3: A Basic Population Model is shared under a CC BY-NC 4.0 license and was authored, remixed, and/or curated by Clarence
Lehman, Shelby Loberg, & Adam Clark (University of Minnesota Libraries Publishing) via source content that was edited to the style and
standards of the LibreTexts platform; a detailed edit history is available upon request.

1
3.1: Characterizing populations
One way to describe populations is to look at how many individuals they contain at various times. Or, instead of individuals, it may
be more reasonable to consider total biomass— the total weight of all individuals in the population combined. For example, the
number of trees may not be as important as their total weight, or the total area of their canopy. Density may also be relevant—how
many individuals occupy a unit of area, or the percentage covered by the population in an area. All of these are gross properties of
populations that can enter models.
Additional properties that can be taken into account include the age structure—the portions of the population of various ages, and
the size structure—the portions of the population of various sizes. Such detail can be important because juveniles or older
individuals may not reproduce. Genetic structure can also be important; it is often left out of ecological models, but evolutionary
directions can affect the ecology as well.
Another important measure is the rate of change—how fast the population is changing. A population can be constant, increasing, or
decreasing, or can fluctuate in complex ways.
Remember that a model is just a simpler view of something complex, and that all scientific models are just approximations of
nature. When the full complexity cannot be understood, which is almost always, we can try to construct a simplified model in
hopes of finding the essence.

This page titled 3.1: Characterizing populations is shared under a CC BY-NC 4.0 license and was authored, remixed, and/or curated by Clarence
Lehman, Shelby Loberg, & Adam Clark (University of Minnesota Libraries Publishing) via source content that was edited to the style and
standards of the LibreTexts platform; a detailed edit history is available upon request.

3.1.1 https://bio.libretexts.org/@go/page/25429
3.2: Bacterial Growth
The place to start a discussion of basic population models is in discrete-time at the macroscale—the most basic. Consider a
hypothetical strain of bacteria reproducing every hour. Suppose a colony is started with a single bacterium when the clock reads
zero, symbolized t = 0, and one hour later that bacterium divides into two, by time t = 1. Each of those divides in two one hour after
that, by time t = 2. And so forth. Suppose you start this culture on a lab bench on Monday, let it go unchecked, and then come back
late on Friday to see the results.
If the colony grows unchecked in this way, how many bacteria will there be at the end of the work week, after five full days of
growth? You could figure that out on a calculator, or more easily you could use computer code. More easily, that is, once you know
computer coding.
Because computer coding is becoming embedded in almost every aspect of life, appreciating the basics of coding is meaningful for
each educated citizen of this century, and this book will expose you to the basics.

This page titled 3.2: Bacterial Growth is shared under a CC BY-NC 4.0 license and was authored, remixed, and/or curated by Clarence Lehman,
Shelby Loberg, & Adam Clark (University of Minnesota Libraries Publishing) via source content that was edited to the style and standards of the
LibreTexts platform; a detailed edit history is available upon request.

3.2.1 https://bio.libretexts.org/@go/page/25430
3.3: A First Computer Model
Below are two lines of computer code forming a program that models the bacterial colony as it doubles every hour. If you have not
seen computer code before, don’t be frightened by this; we will go through it carefully below.
N=1; t=0;
while(t<=5*24) { print(N); N=N*2; t=t+1; }
The above program is written in a generic programming language—it could be run as the programming language R, as the
language AWK, or, with minor adjustments as the language C, Java, or a number of others. This particular code is essentially
identical in many languages.
The first “statement” is N=1. That instructs the computer to set the number of bacteria, N, equal to 1. A semicolon (;) ends the
statement and separates it from subsequent statements. The next statement is t=0. That instructs the computer to set the time t to 0.
One bacterium at time zero forms the “initial conditions,” and once the computer has finished that line, the program is said to be
“initialized.”
The second line of the program is more involved, but can be understood in two parts. The first part on the left, while(t<=5*24),
instructs the computer to repeat a set of code for 5 simulated days of 24 hours each. The second part is the code to be repeated,
within braces on the right, {...}.
Considering the first part, while is a “keyword” that instructs the computer to repeat something until the “condition” in parentheses
is no longer true. In this case, inside the parentheses is t<=5*24, which itself consists of three parts, t, <=, and 5*24. The first part,
t, represents the time, which has just been initialized to zero in the previous line of code. The second part, <=, is the symbol for
“less than or equal to.” Six such “comparison” symbols are possible, ==, <, <=, >, >=, and !=, representing comparison for equal,
less than, less than or equal, greater than, greater than or equal, and not equal, respectively. In the third part, the asterisk (*) is a
symbol for multiplication, so 5*24 means “five times twenty-four,” a way to represent the number 120, or the number of hours
from Monday to Friday—the amount of time the hypothesized bacterial culture is to reproduce.
Computer coding is an exacting business, where tiny variations can make huge differences. The computer is the ultimate literal
interpreter. An example of this just slipped by in the previous paragraph. In coding, a single equals sign, =, means “change
something to be equal to,” whereas two consecutive equals signs, ==, means “compare to see if two things are the same.”
If you are accustomed to coding, you will already be familiar with such subtleties; if this is all new to you, it is something to get
used to. Various primers on the web can help, but don’t be discouraged if it seems difficult at first; computer coding turns out to be
one of the easiest things to jump into but one of the most difficult areas of all human endeavour to get exactly right. Time and
patience will assist.
Getting back to the code, the phrase while(t<=5*24) means, in this case, to repeat something as long as the time, t, is less than or
equal to 120 hours, 5 times 24. And that something to be repeated appears within braces to the right, {...}. (By the way, many
programming languages use three main symbols for grouping information—called braces, { }, brackets, [ ], and parentheses, ( ).
They are used for various kinds of groupings, but unfortunately their usage is not consistent across all languages.)
The first statement within braces is print(N). (Refer back to the two-line program.) “Print” is a term left from the days when
computers would communicate largely by printing on paper. Now the term just means “display.” The statement thus means
“display the number of individuals in the population, N, at this time.“ That was set to 1 in the previous line, so when the computer
runs print(N) for the first time, it will display the number 1, typically on your screen.
The next statement, N=N*2, is read “N equals N times two.” It is similar in form to the statement on the first line, N=1, which
started things off with a single bacterium. The ‘N=’ part is the same. It tells the computer that the number of bacteria, N, is about to
change. (Of course, the computer has no clue what the program is about—that you are running a program about bacteria.) What N
will change to is immediately to the right of the equal sign, N*2. The asterisk (*) means multiply. So this statement thus tells the
computer to double the value of N. That is what the hypothesized bacterial population does every hour, so this statement models
that doubling.
The third statement on the line, t=t+1, is read “t equals t plus one.” It is similar in form to the statement on the first line, t=0, which
started things off with a clock time of zero. In other words, in this example of letting bacteria grow for a five-day work week, we
are taking midnight Monday morning to be hour zero. Five days later, at midnight Friday night, that becomes hour 120 (5 days
times 24 hours per day equals 120 hours). So similarly, t= tells the computer that time t is about change. Following the equals sign

3.3.1 https://bio.libretexts.org/@go/page/25431
is what it should change to, t+1, or one more than what it is at the moment, which advances the time by one hour. This is a discrete
time model, so it approximates the real system by modeling only specific moments.
Those three statements are run in order, from left to right, first displaying the number of bacteria, then modeling the doubling of the
bacterial population, and then advancing to the next hour. By the way, it may have occurred to you that the last two statements
could be written in either order, or even run at the same time—they are independent, so the ordering would not matter.
After all three statements are run, your display will contain the number 1, N will be 2, and t will be 1. The computer next examines
the code inside the parentheses associated with the keyword while to see if the three statements inside the braces should be run
again. That condition specifies that as as long as the time t is less than or equal to 120, the three statements must be repeated. At
this point, t is equal to 1, which certainly is less than 120. Therefore the three statements will be run again.
This is called a “loop,” and now the computer will begin the second time around the loop, running the three statements again as
they ran before, but now with altered values of N and t. First it will display N, which is now equal to 2, so it will display 2. Then it
will double N again, changing it from 2 bacteria to 4, and increase the time t by 1, from hour 1 to hour 2.
Thus the process repeats. Examining the condition inside the parentheses, the computer finds that 2 is less than or equal to 120, and
so the three statements inside braces are run again. This goes on and on until t is greater than 120, at which time the loop is
finished. At the end, t will be 121 and N will be whatever number has been reached by the process of doubling.
This code illustrates two fundamental aspects of computer coding: “condition testing” and “looping.” In larger programs loops are
“nested” within other loops and condition tests are nested correspondingly. But this two-line program, with the first line initializing
and the second line running a loop, is sufficient for our first model. You will soon see that this is not a trivial model, but one that
demonstrates an inviolable law of biology, which Darwin put directly to use in creating his theory of evolution.

This page titled 3.3: A First Computer Model is shared under a CC BY-NC 4.0 license and was authored, remixed, and/or curated by Clarence
Lehman, Shelby Loberg, & Adam Clark (University of Minnesota Libraries Publishing) via source content that was edited to the style and
standards of the LibreTexts platform; a detailed edit history is available upon request.

3.3.2 https://bio.libretexts.org/@go/page/25431
3.4: Program results
Here is what the program produces, shortened to fit on a page of this book.
1
2
4
8
16
32
:
6.64614 × 1035
1.32923 × 1036
If you run this program in R or another suitable language, you should see something essentially identical to the above. Between
Monday and Friday, 120 bacterial doublings would produce over 1036 bacteria—that’s 1 followed by 36 zeros. That is the
computational result. The scientific question is how many individuals this amounts to. Worked out exactly, it is this number: 2120 =
1,329,227,995,784,915,872,903,807,060,280,344,576. To understand the size of this number, suppose the bacteria are roughly
cubical 1 µm on a side—one millionth of a meter, or about four hundred-thousandths of an inch (a suitable order-of-magnitude for
a bacterium). What volume will the colony occupy in cubic meters at the end of the work week, after five full days of growing
unchecked? You might want to speculate: will it fill the culture plate, overflow onto the lab bench, fill the lab, or what?
Work it out and you will see that the answer is 2120 bacteria times 10-18 cubic meters per bacterium equals about 1.3 × 1018 cubic
meters total. How large is that? Estimate the ocean to be a film averaging 3.7 kilometers deep and coating two-thirds of a sphere
with a 6400 kilometer radius (this approximates the amount of the earth’s surface that is covered by ocean). This is about 1.3 × 1018
cubic meters! At the end of five days, the colony unchecked would thus fill all oceans of the earth with a dense microbial mass,
from the greatest depths up to the surface!
This result has deep-reaching implications. First, even though this bacterial model can be quite accurate for a day or so, it fails
completely over the course of a week. All models are approximations to reality, at best applicable over a suitable range. Second,
there are lessons in its failure. It illustrates one of the inviolable laws of biology—that no population growth can remain unlimited
for long. And third, in a mind like Charles Darwin’s, and coupled with other biological principles, it leads to the conclusion that
organisms must evolve. That is the story of Darwin’s elephants.

This page titled 3.4: Program results is shared under a CC BY-NC 4.0 license and was authored, remixed, and/or curated by Clarence Lehman,
Shelby Loberg, & Adam Clark (University of Minnesota Libraries Publishing) via source content that was edited to the style and standards of the
LibreTexts platform; a detailed edit history is available upon request.

3.4.1 https://bio.libretexts.org/@go/page/25432
3.5: Darwin's elephants
With elephants recognized as the slowest breeders of all known animals, Darwin made a laborious calculation, similar to the
bacterial calculation above but more detailed, assuming that elephants started breeding at age 30 and continued until age 90,
producing 6 young in that time.

Figure 3.5.1 . Elephants and Kilimanjaro.

Of course he had no computers, nor calculators, and apparently kept track of 90 or more age classes and made his calculations on
paper. He calculated by hand on paper and alas those notes have never been found. But he said it cost him “some pain” to reach the
conclusion that at the end of the fifth century, fifteen million elephants would be walking the earth, descended from one original
pair. From this, he concluded that unlimited growth is impossible.

There is no exception to the rule that every organic being naturally increases at so high a
rate that, if not destroyed, the earth would soon be covered by the progeny. — Charles
Darwin, 1859
That, he explained in Chapter Three of his Origin of Species. After explaining results of selection by people in the breeding of
domestic animals, he introduced the concept of selection by natural causes in the wild, which he called “natural selection.” The
simplest model of unlimited population growth was thus useful in the extreme, leading to an inviolable law of biology and the
theory of evolution as one of its consequences. Individuals with qualities that allow them to suffer lower mortality or to reproduce
slightly faster, and who pass those qualities to their offspring, will be the ones whose qualities predominate.

Figure 3.5.2 . Charles Darwin was in his twenties when he realized that natural selection was a cause of evolution and started to
formulate his theory.

This page titled 3.5: Darwin's elephants is shared under a CC BY-NC 4.0 license and was authored, remixed, and/or curated by Clarence Lehman,
Shelby Loberg, & Adam Clark (University of Minnesota Libraries Publishing) via source content that was edited to the style and standards of the
LibreTexts platform; a detailed edit history is available upon request.

3.5.1 https://bio.libretexts.org/@go/page/25433
 CHAPTER OVERVIEW

4: Modeling a Single Population

4.1: Density-independent growth
4.2: Density-enhanced growth
4.3: Density-limited growth
4.4: Parameter combinations
4.5: Generalization

This page titled 4: Modeling a Single Population is shared under a CC BY-NC 4.0 license and was authored, remixed, and/or curated by Clarence
Lehman, Shelby Loberg, & Adam Clark (University of Minnesota Libraries Publishing) via source content that was edited to the style and
standards of the LibreTexts platform; a detailed edit history is available upon request.

1
4.1: Density-independent growth
Density, in the sense of population density, refers to how many individuals are present on average per unit area. One could say,
“The density of elk in Yellowstone National Park during the summer is about 3 to 6 per square mile.” Sometimes, however, you
will see density used as the total number in a place. You may see, “The density of elk in Yellowstone National Park during the
summer is about 10 to 20 thousand.” The symbol N is often used for population density.
In the first case above, you would write N = 4.5, the midpoint between 3 and 6. In the second case you would write N = 15000. It
should be clear from context what the area is.
With this in mind, all of the following statements are equivalent:
1. The population doubles each hour. (As in the bacterial example of the previous chapter.)
2. The population N(t) doubles each hour. Here the number of individuals is represented by the letter N, and N(t) means population
at time t. In the bacterial example, there was one bacterium at the beginning of the experiment when the clock started running,
so you would write N (0) = 1. One hour later, the hypothetical population had doubled, so you would write N (1) = 2. Doubling
successively then gives N (2) = 4, N (3) = 8, and so forth until after five days, or 120 hours, N (120) = 1036, or slightly more—
enough to fill all the oceans of the world.
3. The population N doubles each hour. Often the “(t )” is left off for simplicity, it being understood that the population N is a
function of time.
4. N doubles each hour. Since N represents a population in this case, the word “population“ will often be dropped for conciseness.
5. N (t + 1) = 2N (t ). In English this would be read “N of t plus 1 equals two times N of t.” That simply means that the population
at some time, anytime, t, when multiplied by 2, is the population in the next time step, t plus one.
6. The change in the population each hour is equal to the size of the population that hour. This may sound pretty confusing. But it
means that the amount that the population increases in the time step is equal in size to the whole population. Usually the
increase is much less than that, perhaps a few percent, but here we are dealing with a rapidly increasing bacterial population.
7. The change in the population each hour is N (t +1) minus N(t), which is to say N (t +1)− N (t ) = 2N (t )−N (t ) = N (t ). Here the
population in the next hour, N (t +1), minus the population now, N (t ), which is the change in population, is twice the current
population, 2N (t ), minus the current population, N (t ) (not less confusing, perhaps).
8. The change in the population each hour, call it “Delta N” or ∆N, is ∆N /∆t = 2N(t)−N(t) = N(t). Here the symbol delta (∆) means
change in or the difference. So ∆N means the change in N, and ∆t means the change in t. So the change in N per unit of time is
written ∆N /∆t, where delta t is the time unit being used, such as hour or day. This statement is thus the same as the previous
one, but with symbols to shorten it.
9. ∆N /∆t = N. This means the population change in each time unit is equal to the population size itself. That is just because the
population is doubling each time step.
10. 1

N
f racΔN Δt = 1 This is just dividing both sides of the previous equation by N, and perhaps looks even more confusing.

However, in what follows, it turns out to be the most useful of all.

To move forward, let’s focus on the last equation, with its parts colored in the box below.

In the first row, the “∆N = 1” refers to a change in the population of one individual, because delta (∆) means change. In the second
row, the “∆t ” in the denominator modifies this to the change in each time step—in this, case each hour. In the third row, the 1/N
modifies it drastically to mean the change in the population per individual in the population.
This could mean that one new individual is born while the parent lives on, or that two new individuals are born and the parent dies,
or that the parent divides in two, or other equivalent events. In this model, these details are abstractions that do not matter for
purposes of projecting the population. The model simply records the number of offspring produced by each member of the
population and surviving to reproduce. Multiplied by 100, this becomes the percentage growth of the population. For humans, this

4.1.1 https://bio.libretexts.org/@go/page/25436
is like the number of children per family who survive to adulthood. (Though it has to be divided by two if there are two parents per
family.) You have seen how rapidly that blows up, from the calculation in Chapter 3.

Figure 4.1.1 . Wild Rudbeckia hirta with an unusual stroke of red boldly bisecting its petals.

This page titled 4.1: Density-independent growth is shared under a CC BY-NC 4.0 license and was authored, remixed, and/or curated by Clarence
Lehman, Shelby Loberg, & Adam Clark (University of Minnesota Libraries Publishing) via source content that was edited to the style and
standards of the LibreTexts platform; a detailed edit history is available upon request.

4.1.2 https://bio.libretexts.org/@go/page/25436
4.2: Density-enhanced growth
Darwin made unparalleled use of a model that failed, but how can the model be improved so that it does not fail?
Think of only three Black-Eyed Susan plants (Rudbeckia hirta) becoming established in Yellowstone National Park, one near the
north-east entrance, one in the center, and a third near the south entrance—the plants thus separated by over 30 miles. How often
would the same pollinator be able to visit two of the plants so the plants could reproduce? Rarely or never, because these
pollinators travel limited distances. The plant’s growth rate will thus be 0. (In fact, it will be negative, since the three plants will
eventually die.)
Suppose instead that 1000 of these plants were scattered about the park, making them about 2 miles apart. Occasionally a pollinator
might happen by, though the chance of it visiting one of the other Black-Eyed Susans would be very low. Still, with 1000 plants in
the area, the growth rate could be slightly positive.
Now consider 1,000,000 of those plants, making them about 100 meters apart. Pollination would now become relatively frequent.
The growth rate of the population thus depends on the number of plants in the vicinity, meaning that this number must be part of
the equation used to calculate the population growth rate.
We can use the equation introduced earlier to calculate this rate. First, put a parameter in place of the 1, like this.
1

N
ΔN

Δt
= r , where r formerly was 1
Then attach a term that recognizes the density of other members of the population, N.
1

N
ΔN

Δt
= r + sN ,
Here r is related to the number of offspring each plant will produce if it is alone in the world or in the area, and s is the number of
additional offspring it will produce for each additional plant that appears in its vicinity.
Suppose r = 0 and s = 1/20, just for illustration, and start with three plants, so N (0) = 3. That is
1

N
ΔN

Δt
= 0 + 0.05 N ,
For watching the dynamics of this, multiply it out again
ΔN

Δt
= (0 + 0.05 N ) N ,
and convert the model to computer code, like this.
r=0; s=0.05; dt=1; t=0; N=3; print(N);
while(t<=14)
{ dN=(r+s*N)*N*dt; N=N+dN; t=t+dt; print(N); }
If you run this model in R (or other languages in which this code works, like C or AWK), you will see the numbers below.
3
3.45
4.045125
4.863277
6.045850
7.873465
10.97304
16.99341
31.43222
80.83145
407.5176
8,711.049

4.2.1 https://bio.libretexts.org/@go/page/25437
 3,802,830
723,079,533,905
26,142,200,000,000,000,000,000
Graph these, and you will see the numbers expand past all bounds, vertically off the page.

Figure 4.2.1 . Orthologistic growth (red) contrasted with exponential growth (blue).
The blue line shows the unlimited bacterial growth (exponential growth) that helped lead Darwin to his idea of natural selection.
The red line illustrates the new “density-enhanced growth” just being considered, where growth rate increases with density.
Because it approaches a line that is orthogonal to the line approached by the logistic model, described later, we call this an
“orthologistic model.” It runs away to infinity so quickly that it essentially gets there in a finite amount of time. In physics and
mathematics this situation is called a “singularity”—a place where the rules break down. To understand this, it is important to
remember that all models are simplifications and therefore approximations, and apply in their specific range. The orthologistic
model applies well at low densities, where greater densities mean greater growth. But a different model will take over when the
densities get too high. In fact, if a population is following an orthologistic model, the model predicts that there will be some great
change that will occur in the near future—before the time of the singularity.
In physics, models with singularities command special attention, for they can reveal previously unknown phenomena. Black holes
are one example, while a more mundane one from physics is familiar to all. Consider a spinning coin with one point touching the
table, spinning ever more rapidly as friction and gravity compel the angle between the coin and the table to shrink with time. It
turns out that the physical equations that quite accurately model this spinning coin include a singularity—a place where the
spinning of the coin becomes infinitely fast at a definite calculable time. Of course, the spinning cannot actually become infinitely
fast. As the coin gets too close to the singularity—as its angle dips too near the table—it merely switches to a different model. That
different model is a stationary coin. The exact nature of the transition between the spinning and stationary states is complex and
debated, but the inevitability of the transition is not.
It is no different in ecology. Reasonable models leading to singularities are not to be discounted, but rather considered admissible
where they apply. They arise inescapably in human population growth, considered in the next chapter.

This page titled 4.2: Density-enhanced growth is shared under a CC BY-NC 4.0 license and was authored, remixed, and/or curated by Clarence
Lehman, Shelby Loberg, & Adam Clark (University of Minnesota Libraries Publishing) via source content that was edited to the style and
standards of the LibreTexts platform; a detailed edit history is available upon request.

4.2.2 https://bio.libretexts.org/@go/page/25437
4.3: Density-limited growth
What about outside of the range of the orthologistic model? Think of the same Black-Eyed Susans, not only close enough that
pollinators can it fluently from one to another, but also crowded so that they start to shade one another, and their roots start to
compete for water and nutrients. What is a suitable model for this?
The growth rate will again depend on the number of plants, but now more plants will reduce the growth rate. That just means a
minus sign on s.
1

N
ΔN

Δt
= r + sN , s <0
Again, r is the number of offspring each will produce if it is alone in the world, but with s negative, s is the number each plant will
be unable to produce for each additional plant that appears in its vicinity.
Suppose we have r = 1 and s = −1/1000, and we start with three plants, so N (0) = 3. Here is the code, with the new negative s in
red.
r=1; s=-0.001; dt=1; t=0; N=3; print(N);
while(t<=20)
{ dN=(r+s*N)*N*dt; N=N+dN; t=t+dt; print(N); }
Now, because s is negative, the growth rate 1

N
ΔN

Δt
will drop as the population increases, so you might surmise that the rate will
eventually reach zero and the population will level off. In fact, it levels off to 1000.
Figure 4.3.1 Logistic growth (green) contrasted with orthologistic growth (red) and exponential growth (blue).
The value at which it levels off is called an “equilibrium,” a value where the dynamical system becomes quiescent and stops
changing. In the case of the logistic equation, it is also called the “carrying capacity,” a level at which the environment cannot
“carry” any larger population.
But why 1000? What value of \(\frac{1}{N}\, \frac{∆N}{∆t}\) will make the population level off? When ∆N is 0, that means “the
change in N is zero.” And that means N stops growing. And when ∆N is zero, the entire term on the left is zero and algebra
proceeds as follows.
1 ΔN
= r + sN
N Δt

0 = r + sN

−sN = r

−r
N =
s

So the carrying capacity is −r/s. In Figure 4.3, −r/s = −1/(−0.001) = 1000. Exactly where it ended up!
This is the celebrated “logistic equation,” published in 1838 by Pierre Verhulst. It is commonly written
ΔN N
= rN (1 − )
Δt K

Notice that when N is equal to K, the factor in parentheses on the right becomes 1 − N/N = 1−1 = 0, so the whole growth term ∆N /
∆t becomes zero and the population stops growing. Thus K is carrying capacity, and therefore K = −r /s.
As an exercise, you might want substitute −r /s for K in the equation above, then simplify and see if you get the r + sN formulation.

This page titled 4.3: Density-limited growth is shared under a CC BY-NC 4.0 license and was authored, remixed, and/or curated by Clarence
Lehman, Shelby Loberg, & Adam Clark (University of Minnesota Libraries Publishing) via source content that was edited to the style and
standards of the LibreTexts platform; a detailed edit history is available upon request.

4.3.1 https://bio.libretexts.org/@go/page/25438
4.4: Parameter combinations
Before moving further, consider all possible combinations of the parameters, as determined by their signs. There are six
possibilities, ignoring growth rates of exactly zero as infinitely unlikely.
1. r>0, s>0 Orthologistic growth.
2. r<0, s>0 Orthologistic growth with an Allee point.
3. r>0, s=0 Exponential growth.
4. r>0, s<0 Logistic growth with a carrying capacity.
5. r<0, s<0 Inviable population declining to extinction.
6. r<0, s=0 Same as above.
Figure 4.4.1 shows three of these possibilities pieced together to form a complete population model. On the left in the figure,
number 2 above, orthologistic growth with an Allee point, prevails at low densities, where larger numbers of other members of the
species in the vicinity enhance growth. In the middle, number 3 above, exponential growth, occurs as a transition phase. Finally on
the right, number 4 above, logistic growth with a carrying capacity, takes over when crowding and other limitations reduce growth
rates as larger numbers of other members of the species in the vicinity appear.
The vertical axis in Figure 4.4.1 shows the individual growth rate, and the horizontal axis shows the population density. On the
right, where the slope is negative, as the density approaches −r /s from the left the growth rate on the vertical axis drops to zero, so
the population stops growing.

Figure 4.4.1 . A general population growth model can be assembled from piecewise linear components.
This is the equilibrium value called the “carrying capacity.” If something pushes the population above that value— immigration of
animals from another region, for example— then the growth rate on the vertical axis drops below zero. The growth rate then is
negative, and therefore the population declines. On the other hand, if something drops the population below that value—such as
emigration of animals to another place—the growth rate on the vertical axis rises above zero. That growth rate is positive, and
therefore the population grows.
The carrying capacity is “stable.” A value is said to be stable if it tends to restore itself when it is pushed away by some outside
force.

4.4.1 https://bio.libretexts.org/@go/page/25439
 The situation is entirely different on the left in the figure, where the slope is positive. As on the right, when the density is −
r
/
s
, the growth rate on the vertical axis reaches zero, meaning the population does not change. This is an equilibrium too— not a
carrying capacity, but an Allee point. However, if the population here drifts below −
r
/
s
, the growth rate on the vertical axis becomes negative and the population declines further still. It is unstable. In this model the
population continues to decline until eventual extinction. Above the Allee point, however, the growth rate on the vertical axis is
positive, so the population increases until it reaches some other limitation.
Figure 4.4.1 ).

This page titled 4.4: Parameter combinations is shared under a CC BY-NC 4.0 license and was authored, remixed, and/or curated by Clarence
Lehman, Shelby Loberg, & Adam Clark (University of Minnesota Libraries Publishing) via source content that was edited to the style and
standards of the LibreTexts platform; a detailed edit history is available upon request.

4.4.2 https://bio.libretexts.org/@go/page/25439
4.5: Generalization
In summary, the macroscale model for population dynamics of a single species, in its simplest form, is
1 ΔN
= r + sN
N Δt

This is a straight-line form of a more general form presented by Hutchinson,

1 2 3 4
ΔN Δt = r + sN + s2 N + s3 N + s4 N +...
N

and of the most general form proposed by Kolomogorov, where f (N ) can be any function of the population density N.
1 ΔN
= f (N )
N Δt

The higher-order terms in the second equation could refine population projections if there were enough data to determine them.
They are not really needed, however, because straight-line parts can be pieced together to form a general population growth curve,
as in Figure 4.4.1. And as human population growth in Figure 6.3.1 will show, a piecewise approach can more closely approximate
the real situation.
Moreover, blending separate versions of the first equation can generalize to either the Hutchinson or Kolomorgov forms as you will
see in Chapter 18.

Figure 4.5.1 Trumpeter swans—the largest North America birds, with wingspans reaching ten feet—were nearing extinction until
deliberate protection and reintroduction programs brought their r values back to viable levels.

This page titled 4.5: Generalization is shared under a CC BY-NC 4.0 license and was authored, remixed, and/or curated by Clarence Lehman,
Shelby Loberg, & Adam Clark (University of Minnesota Libraries Publishing) via source content that was edited to the style and standards of the
LibreTexts platform; a detailed edit history is available upon request.

4.5.1 https://bio.libretexts.org/@go/page/25440
 CHAPTER OVERVIEW

5: Differential and Difference Forms

5.1: Chapter Introduction
5.2: A mathematical view
5.3: Exponential Solution
5.4: Logistic solution
5.5: Orthologistic solution

This page titled 5: Differential and Difference Forms is shared under a CC BY-NC 4.0 license and was authored, remixed, and/or curated by
Clarence Lehman, Shelby Loberg, & Adam Clark (University of Minnesota Libraries Publishing) via source content that was edited to the style
and standards of the LibreTexts platform; a detailed edit history is available upon request.

1
5.1: Chapter Introduction
1 ΔN
= r + sN ← Dif f erence equation model
N Δt

1 dN
= r + sN ← Dif f erential equation model
N dt

Recall that the delta sign (∆) means change in or the difference. Compare the difference equation with the differential form, which
uses the terminology dN and dt . These represent infinitesimally small time steps, corresponding to our common-sense perception
of time as divisible ever more finely without limit. In differential equations populations change smoothly rather than in finite steps
—growth approximating that of organisms that can reproduce at any time, such as bacterial or human populations.
It turns out that differential equations are harder for computers to solve than difference equations. Computers cannot make
infinitely fine time steps, but have to approximate by using very small time steps instead. On the other hand, difference equations
can be harder to solve mathematically.
r=1; s=-0.001; N=1; t=0;
dt=1; print(N);
while(t<=20)
{ dN=(r+s*N)*N; N=N+dN; if (N<0) N=0; t=t+dt; print(N); }
Above is computer code for a difference equation presented earlier, which leveled off at 1000, but with an addition in red. If the
population is far above its carrying capacity, the calculation could show such a strong decline that the next year’s population would
be negative—meaning that the population would die out completely. The addition in red just avoids projecting negative
populations. Below is similar code for the corresponding differential equation, with the differences again in red.
r=1; s=-0.001; N=1; t=0;
dt=1/(365*24*60*60); print(N);
while(t<=20/dt)
{ dN=(r+s*N)*N*dt; N=N+dN; if(N<0) N=0; t=t+dt; print(N); }
This intends to model infinitely small time steps. Of course it cannot do that exactly, but must settle for very small time steps.
Instead of dt = 1 , for example, representing one year, it is set here to about one second, dividing 1 year by 365 days and each day
by 24 hours, 60 minutes, and 60 seconds. This is hardly infinitely small, but for populations of bacteria and humans it is close
enough for practical purposes. Still, it is important to check for negative populations in case the time step is not small
enough.
Figure 5.1.1. Differential logistic growth (maroon) compared with discrete (green). No dots appear on the differential form, since
that represents infinitesimal time steps, whereas the difference form has a dot at each point calculated.
How small is close enough to infinitely small? is the question. To find out, you can set the time step to something small and run the
code, which will produce a set of population values through time. Then set the step smaller still and run the code again. It will run
more slowly because it is calculating more steps, but if essentially the same answer appears—if the answer “converges”—then you
can make the step larger again, speeding the calculation. With a few trials you can find a time step that is small enough to give
accurate answers but large enough to allow your code to run reasonably fast.
Figure 5.1.1 shows the results of running the differential equation version of the program (the second one above, in maroon) versus
the difference equation version (the first above, in green). The differential equation has the same parameters and general shape, but
the population approaches its carrying capacity more quickly. Because the differential time steps produce offspring earlier—not
waiting for the end of the step— offspring are available to reproduce earlier, and so forth.
This particular method for differential equations is called “Euler’s method” (pronounced “Oiler’s”), a basic approach not often used
in the twentieth century because computers not so long ago were millions of times slower than they are now. Today this method is
often fast enough, and is desirable because of its relative simplicity.
(By the way, calculating bacterial growth for five days with one-second time steps will be fast enough in any programming
language. At the time we are writing this (second decade of the twenty-first century), calculating growth for a human population,

5.1.1 https://bio.libretexts.org/@go/page/25447
second-by-second for 20 years, doing the same for 20 years second by second for humans will be too slow in R, tolerable in AWK,
and plenty fast in C, Java, or other high-speed languages.)

This page titled 5.1: Chapter Introduction is shared under a CC BY-NC 4.0 license and was authored, remixed, and/or curated by Clarence
Lehman, Shelby Loberg, & Adam Clark (University of Minnesota Libraries Publishing) via source content that was edited to the style and
standards of the LibreTexts platform; a detailed edit history is available upon request.

5.1.2 https://bio.libretexts.org/@go/page/25447
5.2: A mathematical view
Differential equations can be amenable to mathematical analysis. To repeat, here is the differential population model.
1 dN
= r + sN
N dt

It turns out there is something simple about infinity, and when time steps are infinitely small the methods of calculus developed
over the centuries can solve this differential equation exactly, mathematically. If you apply a symbolic mathematics computer
package, or the methods for integration of functions developed in calculus, you can find the population value N for any future time
t. This is called the “solution” to the differential equation.
1
N (t) = s 1 s
−rt
( + ) e −
r N r
0

Most differential equations cannot be solved this way but, fortunately, the basic equations of ecology can. This solution becomes
useful in projecting forward or otherwise understanding the behavior of a population. If you know the starting N, s, and r, you can
plug them into the formula to find the population size at every time in the future, without stepping through the differential equation.

This page titled 5.2: A mathematical view is shared under a CC BY-NC 4.0 license and was authored, remixed, and/or curated by Clarence
Lehman, Shelby Loberg, & Adam Clark (University of Minnesota Libraries Publishing) via source content that was edited to the style and
standards of the LibreTexts platform; a detailed edit history is available upon request.

5.2.1 https://bio.libretexts.org/@go/page/25443
5.3: Exponential Solution
To think about this mathematically, first set s to zero, meaning no density dependence. The differential equation then reduces to dN
/dt = r, and if you replace s in the equation above with 0 you get this:
1
N (t) = (5.3.1)
0 1 −rt 0
( + )e −
r N0 r

1
= (5.3.2)
1 −rt
e
N0

rt
= N0 e (5.3.3)

A measure often used for exponential growth, and that we will apply later in this book, is “doubling time”—the time that must
elapse for the population to double. For exponential growth, this is always the same, no matter how large or small the population.
For exponential growth, the equation above is
rt
N (t) = N0 e (5.3.4)

N0 is the starting population at time 0, N (t ) is the population at any time t, and r is the constant growth rate—the “intrinsic rate of
natural increase.” How much time, τ, will elapse before the population doubles? At some time t, the population will be N (t ), and at
the later time t+τ, the population will be N (t+τ). The question to be answered is this: for what τ will the ratio of those two
populations be 2?
N (t + τ)
= 2 (5.3.5)
N (t)

Substituting the right-hand side of the exponential growth equation gives

r (t + τ)
N0 e
= 2 (5.3.6)
rt
N0 e

The factor N0 cancels out, and taking natural logarithms of both sides gives
r(t + τ)
e
ln = ln 2 (5.3.7)
rt
e

Since the log of a ratio is the difference of the logs, this yields
r(t + τ) rt
ln e − ln e = ln 2 (5.3.8)

x
Since logarithms and exponentials are inverse processes— each one undoes the other—the natural logarithm of e is simply x. That
gives
r(t + τ) − rt = ln 2 (5.3.9)

r τ = ln 2 (5.3.10)

and finally, the doubling time τ is

ln 2
τ = (5.3.11)
r

In other words, the doubling time for exponential growth, where r is positive and s is 0, is just the natural logarithm of 2
(0.69314718...) divided by the growth rate r.

This page titled 5.3: Exponential Solution is shared under a CC BY-NC 4.0 license and was authored, remixed, and/or curated by Clarence
Lehman, Shelby Loberg, & Adam Clark (University of Minnesota Libraries Publishing) via source content that was edited to the style and
standards of the LibreTexts platform; a detailed edit history is available upon request.

5.3.1 https://bio.libretexts.org/@go/page/25444
5.4: Logistic solution
Recall that the carrying capacity is −r /s, also called K. So wherever −r /s appears, substitute K, as follows.
1
N (t) =
s 1 s
( + ) e−rt −
r r
N
0

1
=
1 1 1
( − + ) e−rt +
K N K
0

K
= K
−rt
( − 1 + ) e + 1
N
0

K
= K − N
0
−rt
( ) e + 1
N
0

This is the solution given in textbooks for logistic growth. There are slight variations in ways it is written, but they are equivalent.

This page titled 5.4: Logistic solution is shared under a CC BY-NC 4.0 license and was authored, remixed, and/or curated by Clarence Lehman,
Shelby Loberg, & Adam Clark (University of Minnesota Libraries Publishing) via source content that was edited to the style and standards of the
LibreTexts platform; a detailed edit history is available upon request.

5.4.1 https://bio.libretexts.org/@go/page/25445
5.5: Orthologistic solution
Finally, let s be positive. This creates a vertical asymptote and orthologistic growth. The position in time of the vertical asymptote
is the “singularity” mentioned earlier. The interesting question is, when s is positive, what is the time of the singularity? That is,
when will the population grow beyond all bounds in this model?
What must happen to the denominator for the population to grow to unbounded values? It has to get closer and closer to zero, for
then the N (t ) will grow closer and closer to infinity. So to find the singularity, you only have to set the denominator to zero, and
then solve for the time t. You can go through the intermediate steps in the algebra below, or use a mathematical equation solver on
your computer to do it for you.
Setting the denominator in the equation from 5.2 to zero will lead along this algebraic path:
s s 1 −rt
= ( + )e
r r N0

Multiply through by (r/s)ert, to obtain

rt r 1
e = (1 + )
s N0

Next take logarithms of both sides

r 1
rt = ln (1 + )
s N0

Finally, divide through by r to find the time of the singularity.

1 r 1
t = ln (1 + )
r s N0

In the 1960s, Heinz von Foerster wrote about this in the journal Science. Though the consequences he suggested were deadly
serious, his work was not taken very seriously at the time, perhaps in part because the time was so far away (about a human
lifetime), but perhaps also because he put the date of the singularity on Friday the 13th, 2026, his 115th birthday. In the title of his
paper he called this “doomsday”, when the human population would have demolished itself.
Von Foerster used a more complicated model than the r+sN model we are using, but it led to the same result. Some of the ideas
were picked up by Paul Ehrlich and others, and became the late-1960s concept of the “population bomb”— which was taken
seriously by many.

This page titled 5.5: Orthologistic solution is shared under a CC BY-NC 4.0 license and was authored, remixed, and/or curated by Clarence
Lehman, Shelby Loberg, & Adam Clark (University of Minnesota Libraries Publishing) via source content that was edited to the style and
standards of the LibreTexts platform; a detailed edit history is available upon request.

5.5.1 https://bio.libretexts.org/@go/page/25446
 CHAPTER OVERVIEW

6: Human Population Growth

6.1: Chapter Introduction
6.2: Phenomological Graph
6.3: Biological-ecological graph
6.4: A global transition

This page titled 6: Human Population Growth is shared under a CC BY-NC 4.0 license and was authored, remixed, and/or curated by Clarence
Lehman, Shelby Loberg, & Adam Clark (University of Minnesota Libraries Publishing) via source content that was edited to the style and
standards of the LibreTexts platform; a detailed edit history is available upon request.

1
6.1: Chapter Introduction
There is a challenge for this chapter. Coming this far in the book you have learned a little about population growth, and you have
access to computer coding, so you are ready for something big.
Imagine for fun that you have accepted a new job in Washington as a policy fellow with the United States Geological Survey, or
USGS—one of the major research branches of the federal government. This is not farfetched; many recent doctoral graduates land
such jobs at reasonably high levels. But suppose that your boss says she wants you to calculate what the world’s population will be
in 2100. Other agencies have done this, but she wants a separate USGS estimate, presented in an understandable way. She can give
you data from the 18th through the 21st centuries. She discloses that she is meeting with the Secretary General of the United
Nations tomorrow and hopes you can figure it out today. You tell her “Sure, no problem.”
Are you crazy? No! The rest of this chapter will walk you through how to it. We’ll start by piecing together the parts, as in Figure
4.4.1-the orthologistic part, if there is one, and any exponential and logistic parts as well.

This page titled 6.1: Chapter Introduction is shared under a CC BY-NC 4.0 license and was authored, remixed, and/or curated by Clarence
Lehman, Shelby Loberg, & Adam Clark (University of Minnesota Libraries Publishing) via source content that was edited to the style and
standards of the LibreTexts platform; a detailed edit history is available upon request.

6.1.1 https://bio.libretexts.org/@go/page/25455
6.2: Phenomological Graph
The excerpt of data you have been given includes the world’s population in billions, by year. That is all. Figure 6.2.1 shows the
data plotted in a phenomenological way— population size versus year, supplemented with a curve going back 2000 years to
provide perspective. The blue dots show the range of data you will be using to project the future population, and the black ‘×’
marks a great demographic transition that is not obvious in this graph, but that will become glaringly so in Figure 6.3.1.

Figure 6.2.1 . Global human population over the past 2000 years.
Can you project global population by simply extending that curve? The population is clearly rising at an enormous rate, expanding
most recently from 3 billion to 7 billion in less than half a century. Simply projecting the curve would lead to a prediction of more
than 11 billion people by the middle of the 21th century, and more than 15 billion by the century’s end.
But such an approach is too simplistic. In one sense, the data are all contained in that curve, but are obscured by the phenomena
themselves. We need to extract the biology inherent in the changing growth rate r as well as the ecology inherent in the changing
density dependence s. In other words, we want to look at data showing 1/N ∆N /∆t versus N, as in Figure 4.4.1.
Table 6.1.1 shows a subset of the original data, t and N, plus calculated values for ∆N, ∆t, and 1/N ∆N /∆t. In row 1, for example,
∆N shows the change in N between row 1 and row 2: 0.795−0.606 = 0.189 billion. Likewise, ∆t in row 1 shows how many years
elapse before the time of row 2: 1750 − 1687 = 63 years. The final column in row 1 shows the value of 1/N ∆N /∆t : 1/0.606 ×
0.189/63 = 0.004950495..., which rounds to 0.0050. Row 21 has no deltas because it is the last row in the table.
Table 6.2.1 . Human population numbers for analysis.
Point Year t N billions ∆N ∆t
1 ΔN

N Δt

1. 1687 0.606 0.189 63 0.0050

2. 1750 0.795 0.174 50 0.0044

3. 1800 0.969 0.296 50 0.0061

4. 1850 1.265 0.391 50 0.0062

5. 1900 1.656 0.204 20 0.0062

6. 1920 1.860 0.210 10 0.0113

7. 1930 2.070 0.230 10 0.0111

8. 1940 2.300 0.258 10 0.0112

9. 1950 2.558 0.224 5 0.0175

10. 1955 2.782 0.261 5 0.0188

11. 1960 3.043 0.307 5 0.0202

12. 1965 3.350 0.362 5 0.0216

13. 1970 3.712 0.377 5 0.0203

14. 1975 4.089 0.362 5 0.0177

6.2.1 https://bio.libretexts.org/@go/page/25450
 ∆N ∆t
1 ΔN
Point Year t N billions N Δt

15. 1980 4.451 0.405 5 0.0182

16. 1985 4.856 0.432 5 0.0178

17. 1990 5.288 0.412 5 0.0156

18. 1995 5.700 0.390 5 0.0137

19. 2000 6.090 0.384 5 0.0126

20. 2005 6.474 0.392 5 0.0121

21. 2010 6.866

This page titled 6.2: Phenomological Graph is shared under a CC BY-NC 4.0 license and was authored, remixed, and/or curated by Clarence
Lehman, Shelby Loberg, & Adam Clark (University of Minnesota Libraries Publishing) via source content that was edited to the style and
standards of the LibreTexts platform; a detailed edit history is available upon request.

6.2.2 https://bio.libretexts.org/@go/page/25450
6.3: Biological-ecological graph

Figure 6.3.1 . Observed human growth rate as a function of population density through the mid-1960s (blue dots), based on the
data from Table 6.1.1, with a line representing the average trend (green).
Figure 6.3.1 plots the two green columns of Table 6.1.1 through line 12—the mid-1960s—in blue dots, with a green line
representing the average trend. A line like this can be drawn through the points in various ways—the simplest with a ruler and pen
drawing what looks right. This one was done using a statistical “regression” program, with r the point at which the line intersects
the vertical axis and s the line’s slope— its ∆y /∆x. The intrinsic growth rate r for modern, global human population is apparently
negative and the slope s is unmistakably positive.
From the late 1600s to the mid 1960s, then, it’s clear that the birth rate per family was increasing as the population increased.
Greater population was enhancing the population’s growth. Such growth is orthologistic, meaning that the human population has
been heading for a singularity for many centuries. The singularity is not a modern phenomenon, and could conceivably have been
known before the 20th century.
The negative value of r, if it is real, means there is a human Allee point. If the population were to drop below the level of the
intersection with the horizontal axis—in this projection, around two hundred million people—the human growth rate would be
negative and human populations would decline. The Allee point demonstrates our reliance on a modern society; it suggests that we
couldn’t survive with our modern systems at low population levels—although perhaps if we went back to hunter–gatherer
lifestyles, this would change the growth curve. The Allee point thus indicates that there is a minimum human population we must
sustain to avoid extinction. We depend on each other.

This page titled 6.3: Biological-ecological graph is shared under a CC BY-NC 4.0 license and was authored, remixed, and/or curated by Clarence
Lehman, Shelby Loberg, & Adam Clark (University of Minnesota Libraries Publishing) via source content that was edited to the style and
standards of the LibreTexts platform; a detailed edit history is available upon request.

6.3.1 https://bio.libretexts.org/@go/page/25451
6.4: A global transition

Figure 6.2.1 to the present day.

In Figure 6.4.1 we add data from the mid-1960s to the present day. People living in the 1960s were completely unaware of the
great demographic transition that was developing. For hundreds of years prior to this time, human populations were stuck on an
orthologistic path, with a singularity ever looming and guaranteed by the positive slope. In most of the world, however, the slope
abruptly turned about and negative. Not all countries of the world turned about, but on average the world did. Humanity started
down a logistic-like path.
Where the downward-sloping line crosses the horizontal axis is where population growth would cease. From this simple r + sN
model, it appears that world’s population will stabilize between 10 and 12 billion. That is in line with other recently published
projections.
Prior to the 1960s there were dips in the increasing growth, with World Wars I and II leveling the rate of increase worldwide,
though population continued to grow rapidly. The rate also fell in 1960, corresponding to extreme social disruptions in China.
What caused this great demographic transition, averaged over the globe? The “Four Horsemen” commonly expected to check
human populations were not a primary cause. In many regions birth control, became more available. Education slowed
reproduction because people got married later. Modern medicine raised survival rates, making large families unnecessary. The
space program looked back at Earth and projected a fragile dot suspended in the black of space, viewed by billions. China’s one-
child policy had a noticeable effect. However, so did HIV, one of the few Horsemen that has made a noticeable comeback.

Figure 6.4.2 . Human fecundity as a function of national per capita income.

Plants and other animals have logistic growth forced upon them because of overcrowding. In humans, however, logistic growth has
been largely voluntary. And there could be further developments in a lifetime. In many nations, birth rates are presently below
replacement rates. In fact, in all nations with a gross national income above 16K dollars per person, the birth rate is at or below the
replacement rate of 2.1 lifetime births per female (Figure 6.4.2).
This change in demographic rates could conceivably allow present and future generations to voluntarily adjust the population to
whatever is desired. The new question just may be: what is the minimum world population we dare have?
Returning to your supervisor’s questions, you can now tell her that, in 2100, the world’s population will be between 10 and 12
billion. And you can say “The other population projections are not far off. They are slightly different from what we calculate using

6.4.1 https://bio.libretexts.org/@go/page/25452
this method. But they use very complicated methods so you have to cut them a little slack!”

Figure 6.4.3 . Earth over Moon, touching the conscience of a world.

This page titled 6.4: A global transition is shared under a CC BY-NC 4.0 license and was authored, remixed, and/or curated by Clarence Lehman,
Shelby Loberg, & Adam Clark (University of Minnesota Libraries Publishing) via source content that was edited to the style and standards of the
LibreTexts platform; a detailed edit history is available upon request.

6.4.2 https://bio.libretexts.org/@go/page/25452
 CHAPTER OVERVIEW

7: Chaos and Randomness

If the growth rate is too fast compared with that feedback, populations can overshoot their carrying capacity, which can lead to
highly complex outcomes. Delays in sensing the carrying capacity can start oscillations. For example, a modeled insect population
that grows and lays eggs one year and emerges the next year can suffer such oscillations. The insects are, in effect, “keeping their
eyes shut” about how many insects will be produced the next year. This is in contrast to species like bacteria or humans, where the
population grows more or less continuously.
7.1: Chapter Introduction
7.2: Depicting population growth
7.3: Hypothetical insect data
7.4: Sensitive Dependence
7.5: Dampening of Chaos
7.6: Bifurcation diagram
7.7: Properties

This page titled 7: Chaos and Randomness is shared under a CC BY-NC 4.0 license and was authored, remixed, and/or curated by Clarence
Lehman, Shelby Loberg, & Adam Clark (University of Minnesota Libraries Publishing) via source content that was edited to the style and
standards of the LibreTexts platform; a detailed edit history is available upon request.

1
7.1: Chapter Introduction

This page is a draft and is under active development.

In the density-limited growth examined thus far, the ecological effects of density fed back fast quickly enough that the population’s
growth could adjust and the population could reach a carrying capacity, equal to −r /s . But if the growth rate is too fast compared
with that feedback, the population can overshoot its carrying capacity, which can lead to highly complex outcomes.
Think about feedback in this way. Imagine driving down the road, keeping your eye on the road, instantly correcting any little
deviations of your car from your lane, adjusting the steering wheel without even perceiving it, and with only normal blinking of
your eyes. In this case there is very little delay in your feedback to the steering wheel, and you stay in the lane. Now suppose you
close your eyes for one second at a time, perhaps every ten seconds. (Do not run this experiment; just think about it!) You may
have drifted a bit to the left or right in that second and would have to turn the steering wheel further to get back in you lane. And
now imagine shutting your eyes for 15 seconds every minute, then opening them and correcting your path down the road. You’ll
start oscillating in your lane and precariously jerking back and forth, possibly visiting the ditch. The cause? The delay in the
feedback between stimulus and response.
So it is with populations. Delays in sensing the carrying capacity can start oscillations. For example, a modeled insect population
that grows and lays eggs one year and emerges the next year can suffer such oscillations. The insects are, in effect, “keeping their
eyes shut” about how many insects will be produced the next year. This is in contrast to species like bacteria or humans, where the
population grows more or less continuously.

This page titled 7.1: Chapter Introduction is shared under a CC BY-NC 4.0 license and was authored, remixed, and/or curated by Clarence
Lehman, Shelby Loberg, & Adam Clark (University of Minnesota Libraries Publishing) via source content that was edited to the style and
standards of the LibreTexts platform; a detailed edit history is available upon request.

7.1.1 https://bio.libretexts.org/@go/page/25648
7.2: Depicting population growth

Figure 7.2.1 . Population growth viewed four ways. (A,B) Phenomenological. (C,D) Mechanistic.
Figure 7.2.1 shows four approaches to depicting populations. While not all equally helpful, each has its use. Let’s start with
phenomenological graphs for a single species: graphs that merely depict the population phenomena observed without attempting to
describe the mechanisms causing the phenomena. Observations might come from successive bacterial plate counts or censuses of
people or, in this case, successive insect censuses. Part A in the figure represents the whole population N over time, and is a starting
place to view population change. Similarly, Part B represents the whole population’s rate of growth, dN/dt, over time, also
phenomenological.
A touch of biology is introduced in Part C by transforming the vertical axis to per-capita growth, 1/N dN/dt. This transformation
recognizes the growth rate that an individual organism achieves in a unit of time—say in a year or a week—under prevailing
conditions. There is a nominal biological limit on the number of offspring produced by an individual in each unit of time—one new
bacterium per individual bacterium in twenty minutes, say, or four goslings per family of geese in a year, or one infant per human
family each year. This subtle amount of biology can reveal patterns not evident in the phenomenological approaches of Parts A and
B—that the number of surviving offspring per individual increases with time for orthologistic growth, does not change for
exponential growth, and decreases with time for logistic growth.
Finally, in Part D, a touch of ecology is added to the biology of Part C by considering, on the horizontal axis, interactions among
organisms. This shows per-capita growth rate versus population density N, rather than versus time t. And it reveals even more
clearly the ecological mechanism behind the phenomena and the distinct nature of the three kinds of population growth—
orthologistic growth appears as a straight line slanted upward (as in Figure 6.2.1), exponential growth as a straight horizontal line,
and logistic growth as a straight line slanted downward (as in Figure 6.3.1). Population density N acts a proxy for space, food, or
other resources or limits.

This page titled 7.2: Depicting population growth is shared under a CC BY-NC 4.0 license and was authored, remixed, and/or curated by Clarence
Lehman, Shelby Loberg, & Adam Clark (University of Minnesota Libraries Publishing) via source content that was edited to the style and
standards of the LibreTexts platform; a detailed edit history is available upon request.

7.2.1 https://bio.libretexts.org/@go/page/25457
7.3: Hypothetical insect data
For a detailed illustration of the methods used in these four graphs and an illustration of population oscillations, consider the
hypothetical insect data in Table 7.3.1. Insects often have one-year reproductive cycles; these can be prone to oscillations, and also
are known for “outbreaks” (e.g. of disease or of pests). The data in Table 7.3.1 were generated by the difference equation
1 ΔN
= r + sN , (7.3.1)
N Δt

with r = 3 and s = −4 . The table shows an initial population of about 11,000 individual organisms. The next year there are about
44,000, then 168,000, then more than 500,000, then more than 900,000. But then something apparently goes wrong, and the
population drops to just over 55,000. In nature, this might be attributed to harsh environmental conditions—a drastic change in
weather or over-exploitation of the environment. But these data are simply generated from a difference equation, with oscillations
induced by overshooting the carrying capacity and getting knocked back to different places, again and again, each time the
population recovers.
Table 7.3.1 . Hypothetical insect data.
(A) (B) (C) (D)

t N ∆N ∆I
0 11,107 32,828 2.956

1 43,935 124,082 2.824

2 168,017 391,133 2.328

3 559,150 426,855 0.763

4 986,005 -930,810 -0.944

5 55,195 153,401 2.779

6 208,596 451,738 2.166

7 660,334 236,838 0.359

8 897,172 -528,155 -0.589

9 369,017 562,357 1.524

10 931,374 -675,708 -0.725

11 255,666 505,537 1.977

12 761,203 -34,111 -0.045

13 727,092 66,624 0.092

14 793,716 -138,792 -0.175

15 654,924 249,071 0.380

16 903,995 -556,842 -0.616

17 347,153 559,398 1.611

18 906,551 -567,685 -0.626

19 338,866 557,277 1.645

20 896,143

The repeated growth and setbacks are visible in the phenomenological graph of population growth (Figure 7.3.1, Part A). It’s easy
to see here that the population grows from low levels through year 4, declines drastically in year 5, then rises again and oscillates
widely in years 8 through 12. The next four years show smaller oscillations, and in years 16 through 20 there are two sets of nearly
identical oscillations.

7.3.1 https://bio.libretexts.org/@go/page/25458
 Figure 7.3.1 . Four graphs of chaos. r = 3, s = −4. N is in millions.
The next phenomenological graph, Part B, shows not the population over time but the change in population over time. The
difference in population size from the first year to the following year is about ∆N = 33,000 (44,000 − 11,000 = 33,000). Similarly,
the difference in time between years 1 and 2 is just ∆t = 2−1 = 1. So ∆N/∆t is about 33,000/1, or in units of the graph, 0.033
million. Year 0 is therefore marked on the graph vertically at 0.033. Review Chapter 5 for why ∆ is used here rather than dN.
For the second year, the population grows from about 44,000 to about 168,000, so ∆N/∆t = (168,000−44,000)/1 = 124,000, or 0.124
million. Year 1 is therefore marked on the graph vertically at 0.124. This continues for all years, with the exact results calculated in
the ∆N column of Table 7.3.1 and plotted in Part B of Figure 7.3.1. These data are still phenomenological, and simply show the
annual changes in population levels rather than the population levels themselves.
In Part C we add a bit of biology, showing how many net offspring are produced annually by each individual in the population.
This is ∆N/∆t = 33,000/1, the number of new net offspring, divided by about 11,000 parental insects— about three net offspring per
insect (more accurately, as shown in the table, 2.956). This can mean that three new insects emerge and the parent lives on, or that
four emerge and the parent dies—the model abstracts such details away as functionally equivalent. All such per-insect (per capita)
growth rates are calculated in the ∆I column of Table 7.3.1 and plotted in Part C of Figure 7.3.1 Part C shows a little biological
information—how the net number of offspring per insect is changing through time, Over the first four years it drops from almost 3
down to almost −1. Again, this could mean that 3 new offspring emerge and survive in year 0 and that the parent survives too, and
that by year 4 almost no offspring survive and the parent dies as well. The smallest the change per insect (per capita) can ever be is
−1, because that means the individual produces no offspring and dies itself—the worst possible case. And since in this case r = 3,
the greatest the change can be per insect is 3— realized most closely when N is very close to 0. In the end, however, even with this
touch of biology added to the graph, Part C still oscillates wildly.
The order underlying the chaos finally is revealed in Part D by retaining the biology with per capita growth on the vertical axis, but
adding ecology with density N on the horizontal axis. Successive years are numbered in red above the corresponding dot. Suddenly,
all points fall on a straight line!
This line reveals the underlying growth equation. Remember that the growth rate is represented as r+sN, which is a straight line. It
is equivalent to the algebraic form y = mx +b, only rewritten with s in place of m, N in place of x, and r in place of b. Remember
also that it is a “first-order approximation” to the general form proposed by G. Evelyn Hutchenson,
2 3
r + sN + s N2 + s N3 +. . . , (7.3.2)

usable when the parameters s2, s3, and so on are small, so that a straight line is a good approximation. And finally, remember that in
terms of human population growth, for which we have reasonably good data, a straight line is indeed a good approximation (Figure
6.3.1).

7.3.2 https://bio.libretexts.org/@go/page/25458
Part D of Figure 7.3.1 thus exposes these population dynamics as density-limited growth, because the individual growth rate on
the vertical axis, 1/N dN/dt, gets smaller as density on the horizontal axis, N, gets larger. And because it is a straight line, it is
logistic growth. But it is different in that the finite time steps allow the population to go above its carrying capacity, forcing its
growth rate negative and pulling the population back down in the next time step—whereupon the growth rate becomes positive
again and is pushed up again in a confusing cascade of chaos.

This page titled 7.3: Hypothetical insect data is shared under a CC BY-NC 4.0 license and was authored, remixed, and/or curated by Clarence
Lehman, Shelby Loberg, & Adam Clark (University of Minnesota Libraries Publishing) via source content that was edited to the style and
standards of the LibreTexts platform; a detailed edit history is available upon request.

7.3.3 https://bio.libretexts.org/@go/page/25458
7.4: Sensitive Dependence

alt

Figure 7.4.1. Sensitive dependence on initial conditions. Both parts have r = 3 and s = −(r+1), but Part A starts at 0.011107
million and Part B starts at 0.012107 million.
Below is the computer code that produced the hypothetical data for the graphs of Figure 7.2.1, quite similar to other code you have
seen before.
r=3; s=-4; N=0.011107; t=0; print(N);
while(t<=20)
{ dN=(r+s*N)*N; N=N+dN; t=t+1; print(N); }
The initial condition is 11,107 insects—0.011107 million in this representation— which produces the time-pattern of Figure 7.2.1,
Part A. But change that initial condition just a little, to 0.012107 million, and the time-pattern changes considerably. (Compare
Parts A and B of Figure 7.4.1) Part A is identical to Part A of Figure 7.2.1, but Part B of the new figure has quite a different
pattern, of repeated population outbreaks—not unlike those seen in some insect populations. The emergence of very different
patterns from slightly different starting points is called “sensitive dependence on initial conditions,” and is one of the characteristics
of chaos.
The carrying capacity in these graphs is
r −3
K =− = = 0.75 (7.4.1)
s −4

which represents 750,000 insects and marked with a horizontal gray dashed line. It is clear that the population is fluctuating about
that equilibrium value.
Added to the right of the time line is a distribution of the points, showing the proportion of the time that the population occurs at
the corresponding place on the vertical axis, with values to the right in the distribution representing higher proportions of time. In
this case, the population spends much of the time at very low or very high values. These distributions can be determined by letting
the program run for a hundred million steps, more or less, and keeping track of how many times the population occurred at specific
levels. In some cases, however, as in this particular case with r = 3, the distribution can be determined algebraically. Here it is
called the arcsine distribution and is equal to
1
x = − −−−−−−. (7.4.2)
π √ y(1 − y)

Though it is not particularly important to population ecology, isn’t it curious to see the value π = 3.14159... emerge from a
difference equation developed to understand population growth!

This page titled 7.4: Sensitive Dependence is shared under a CC BY-NC 4.0 license and was authored, remixed, and/or curated by Clarence
Lehman, Shelby Loberg, & Adam Clark (University of Minnesota Libraries Publishing) via source content that was edited to the style and
standards of the LibreTexts platform; a detailed edit history is available upon request.

7.4.1 https://bio.libretexts.org/@go/page/25459
7.5: Dampening of Chaos

Figure 7.5.1 . Bifurcations with changing parameters. In all parts s = −r−1. (A) r = 2.84, period 3, (B) r = 2.575, near period 4, (C) r
= 2.48, period 4, (D) r = 2.26, period 2.
alt

Figure 7.5.2 . Changing nature of equilibrium, period 1. In all parts s = −r−1. (A) r = 2.00, (B) r = 1.86, (C) r = 1.24, (D) r = 0.60.
If the growth rate r diminishes, the amount that the population can overshoot its carrying capacity also diminishes, meaning that the
size and severity of the uctuations should diminish as well. Figures 7.5.1 and 7.5.2 show this happening.
When r diminishes from 3 to 2.84, for example, as in Part A of Figure 7.5.1, chaos vanishes and the oscillations become regular,
jumping from a specific low value to a specific medium value to a specific high value, then dropping back to repeat the cycle. This
is called “period three.” Sensitive dependence on initial conditions has also vanished; slight changes in the starting population will
not produce different patterns, as in Figure 7.3.1, but will end up approaching exactly the same three levels as before. The pattern is
stable. Moreover, changing parameter r slightly will not change the period-three pattern to something else. The exact values of the
three levels will shift slightly, but the period-three pattern will remain.
But when r is changed more than slightly—to 2.575, for example, as in Part B— the period-three pattern vanishes and, in this case,
a chaos-like pattern appears. The population fluctuates among four distinct bands, with a complex distribution within each, as

7.5.1 https://bio.libretexts.org/@go/page/25460
shown on the right in Part B. With r somewhat lower—at 2.48, for example, as in Part C— the bands coalesce into a period-four
pattern, which is stable like the period-three pattern in Part A. With further reductions in r, the period-four pattern is cut in half to a
period-two pattern, as in Part D, and finally to a period-one, an equilibrium pattern.
Figure 7.5.2 shows the progression from r = 3 downward, as it changes from an oscillation toward the equilibrium, as in Parts A
and B, and to a smooth approach, as in Parts C and D. This smooth approach begins when the growth rate is small enough that the
population does not overshoot its carrying capacity.

This page titled 7.5: Dampening of Chaos is shared under a CC BY-NC 4.0 license and was authored, remixed, and/or curated by Clarence
Lehman, Shelby Loberg, & Adam Clark (University of Minnesota Libraries Publishing) via source content that was edited to the style and
standards of the LibreTexts platform; a detailed edit history is available upon request.

7.5.2 https://bio.libretexts.org/@go/page/25460
7.6: Bifurcation diagram

Figure 7.6.1 . Bifurcation diagram summarizing the discrete logistic equation.

The dynamics of all possible values of r can summarized in a “bifurcation diagram” (Figure 7.6.1). In mathematical terminology, a
bifurcation is a place where a tiny change in a parameter causes an extensive and discontinuous change in the behavior of the
system. Figure \(\PageIndex{1}\) shows this by amalgamating the distributions on the right in Figures 7.3.1 through 7.4.2, plus
distributions for all other possible values of r. Shading shows where the population spends most of its time. Starting at the right of
this figure, fully in the domain of chaos, and moving to the left by reducing r, the behavior moves in and out of chaos-like patterns
that never repeat and thus have no period, and also hits stable patterns of every possible period from one up toward infinity.

This page titled 7.6: Bifurcation diagram is shared under a CC BY-NC 4.0 license and was authored, remixed, and/or curated by Clarence
Lehman, Shelby Loberg, & Adam Clark (University of Minnesota Libraries Publishing) via source content that was edited to the style and
standards of the LibreTexts platform; a detailed edit history is available upon request.

7.6.1 https://bio.libretexts.org/@go/page/25461
7.7: Properties
Chaos is not precisely defined in mathematics, but it occurs where:
1. Population dynamics appear to oscillate erratically, without outside influence.
2. The population has an unlimited set of different patterns of oscillation, all for the same parameter values.
3. The slightest change in the number of individuals can change the population from one pattern of oscillations to any other
pattern.
It is not important that you learn all the details of chaos. The important scientific point here is that complexity can arise from
simplicity. Complex behavior of something in nature does not imply complex causes of that behavior. As you have seen, as few as
two lines of computer code modeling such systems can generate extremely complex dynamics. The important biological point is
that populations can oscillate chaotically on their own, with no outside influences disturbing them, and that their precise future
course can be unpredictable.
Chaos and randomness in deterministic systems were discovered by mathematician Henri Poincaré late in the 19th century, but the
knowledge scarcely escaped the domain of mathematics. In the 1960s, meteorologist Edward Lorenz discovered their effects in
models of weather, and in the 1970s theoretical ecologist Robert May made further discoveries, publishing a high-profile paper that
landed in scientific fields like a bombshell. The details of chaos were then worked out by a large assortment of mathematicians and
scientists during the last quarter of the twentieth century. The discrete-time logistic equation examined in this chapter is now
designated by Ian Steward as number sixteen of seventeen equations that changed the world.

Figure 7.7.1. The final six of seventeen equations that changed the world, as designated by Ian Steward.

This page titled 7.7: Properties is shared under a CC BY-NC 4.0 license and was authored, remixed, and/or curated by Clarence Lehman, Shelby
Loberg, & Adam Clark (University of Minnesota Libraries Publishing) via source content that was edited to the style and standards of the
LibreTexts platform; a detailed edit history is available upon request.

7.7.1 https://bio.libretexts.org/@go/page/25462
 CHAPTER OVERVIEW

8: Theory of Interactions
8.1: Dynamics of two Interacting Species
8.2: Code for two species
8.3: Summary of interactions

This page titled 8: Theory of Interactions is shared under a CC BY-NC 4.0 license and was authored, remixed, and/or curated by Clarence
Lehman, Shelby Loberg, & Adam Clark (University of Minnesota Libraries Publishing) via source content that was edited to the style and
standards of the LibreTexts platform; a detailed edit history is available upon request.

1
8.1: Dynamics of two Interacting Species
In the first part of this book you’ve seen the two main categories of single-species dynamics—logistic and orthologistic, with
exponential growth being an infinitely fine dividing line between the two. And you’ve seen how population dynamics can be
simple or chaotically complex.
Moving forward you will see three kinds of two-species dynamics—mutualism, competition, and predation—and exactly forty
kinds of three-species dynamics, deriving from the parameters of the population equations and their various combinations.
To review, the population dynamics of a single species are summarized in the following equation.
1 dN
= r + sN (8.1.1)
N dt

Here parameter r is the “intrinsic growth rate” of the species— the net rate at which new individuals are introduced to the
population when the population is vanishingly sparse, and s is a “density dependence” parameter that reflects how the size of the
population affects the overall rate. Parameter s is key. If s is negative, the population grows “logistically,” increasing to a “carrying
capacity” of −r /s, or decreasing to that carrying capacity if the population starts above it. If s is positive, then the population grows
“orthologistically,” increasing ever faster until it encounters some unspecified limit not addressed in the equation. Exponential
growth is the dividing line between these two outcomes, but this would only occur if s remained precisely equal to zero.
How should this single-species equation be extended to two species? First, instead of a number N for the population size of one
species, we need an N for each species. Call these N1 for species 1 and N2 for species 2. Then, if the two species do not interact at
all, the equations could be
1 dN1
= r1 + s1,1 N1 (8.1.2)
N1 dt

1 dN2
= r2 + s2,2 N2 (8.1.3)
N2 dt

Here r1 and r2 are the intrinsic growth rates for N1 and N2, respectively, and s1,1 and s2,2 are the density dependence parameters for
the two species. (The paired subscripts in the two-species equations help us address all interactions.)

There are thus four possible si,j parameters here:

s1,1 : How density of species 1 affects its own growth.
s1,2 : How density of species 2 affects the growth of species 1.
s2,1 : How density of species 1 affects the growth of species 2.
s2,2 : How density of species 2 affects its own growth.
With these parameters in mind, here are the two-species equations. The new interaction terms are in blue on the right.
1 dN1
= r1 + s1,1 N1 + s1,2 N2 (8.1.4)
N1 dt

1 dN2
= r2 + s2,2 N2 + s2,1 N1 (8.1.5)
N2 dt

In the single-species equations, the sign of the s term separates the two main kinds of population dynamics—positive for
orthologistic, negative for logistic. Similarly, in the two-species equations, the signs of the interaction parameters s1,2 and s2,1
determine the population dynamics.
Two parameters allow three main possibilities—(1) both parameters can be negative, (2) both can be positive, or (3) one can be
positive and the other negative. These are the main possibilities that natural selection has to work with.

8.1.1 https://bio.libretexts.org/@go/page/25464
 Figure 8.1.1. Both interaction parameters negative, competition.

Competition. First consider the case where s1,2 and s2,1 are both negative, as in Figure 8.1.1.

Figure 8.1.2 . Time course of competition, one example.

For a single species, parameter s being negative causes the population to approach a carrying capacity. The same could be expected
when parameters s1,2 and s2,1 are both negative—one or both species approach a carrying capacity at which the population remains
constant, or as constant as external environmental conditions allow.
One example is shown in Figure 8.1.2, where the population of each species is plotted on the vertical axis and time on the
horizontal axis. Here Species 2, in red, grows faster, gains the advantage early, and rises to a high level. Species 1, in blue, grows
more slowly but eventually rises and, because of the mutual inhibition between species in competition, drives back the population
of Species 2. The two species eventually approach a joint carrying capacity.
In other cases of competition, a “superior competitor” can drive the other competitor to extinction—an outcome called
“competitive exclusion.” Or, either species can drive the other to extinction, depending on which gains the advantage first. These
and other cases are covered in later chapters.
In any case, when both interaction terms s1,2 and s2,1 are negative, in minus–minus interaction, each species inhibits the other’s
growth, which ecologists call the “interaction competition”.
Mutualism. The opposite of competition is mutualism, where each species enhances rather than inhibits the growth of the other.
Both s1,2 and s2,1 are positive.

Figure 8.1.3 : Both interaction parameters positive, mutualism.

Depicted in Figure 8.1.3 is a form of “obligate mutualism,” where both species decline to extinction if either is not present. This is
analogous to a joint Allee point, where the growth curves cross the horizontal axis and become negative below certain critical
population levels. If this is not the case and the growth curves cross the vertical axis, each species can survive alone; this is called
“facultative mutualism,” and we’ll learn more about it in later chapters.
For now, the important point is how mutualistic populations grow or decline over time. A single species whose density somehow
enhances its own rate of growth becomes orthologistic, increasing ever more rapidly toward a singularity, before which it will grow
so numerous that it will be checked by some other inevitable limit, such as space, predation, or disease.
It turns out that the dynamics of two species enhancing each other’s growth are similar to those of a single species enhancing its
own growth. Both move to a singularity at ever increasing rates, as illustrated earlier in Figure 4.2.1 and below in Figure 8.1.4. Of
course, such growth cannot continue forever. It will eventually be checked by some force beyond the scope of the equations, just as
human population growth was abruptly checked in the mid-twentieth century— so clearly visible earlier in Figure 6.3.1.

8.1.2 https://bio.libretexts.org/@go/page/25464
 Figure 8.1.4 : Time course of unchecked mutualism
Predation. The remaining possibility for these two-species equations is when one interaction parameter si,j is positive and the
other is negative. In other words, when the first species enhances the growth of the second while the second species inhibits the
growth of the first. Or vice versa. This is “predation,” also manifested as parasitism, disease, and other forms.
Think about a predator and its prey. The more prey, the easier it is for predators to catch them, hence the easier it is for predators to
feed their young and the greater the predator’s population growth. This is the right part of Figure 8.1.5. The more predators there
are, however, the more prey are captured; hence the lower the growth rate of the prey, as shown on the left of the figure. N1 here,
then, represents the prey, and N2 represents the predator.

Figure 8.1.5 . Interaction terms of opposite signs, predation.

Prey can survive on their own, without predators, as reflected on the left in positive growth for N1 when N2 is 0. Predators,
however, cannot survive without prey, as reflected on the right in the negative growth for N2 when N1 is 0. This is like an Allee
point for predators, which will start to die out if the prey population falls below this point.
The question here is this: what will be the population dynamics of predator and prey through time? Will the populations grow
logistically and level off at a steady state, as suggested by the negative parameter s1,2, or increase orthologistically, as suggested by
the positive parameter s2,1?
Actually, they do both. Sometimes they increase faster than exponentially, when predator populations are low and growing prey
populations provide ever increasing per capita growth rates for the predator, according to the right part of Figure 8.1.5. In due time,
however, predators become abundant and depress prey populations, in turn reducing growth of the predator populations. As shown
in Figure 8.1.6, the populations oscillate in ongoing tensions between predator (red line) and prey (blue line).

Figure 8.1.6 . Time course of predation.

Examine this figure in detail. At the start, labeled A, the prey population is low and predators are declining for lack of food. A
steady decline in the number of predators creates better and better conditions for prey, whose populations then increase
orthologistically at ever accelerating per capita rates as predators die out and conditions for prey improve accordingly.
But then the situation turns. Prey grow abundant, with the population rising above the Allee point of the predator, at B. The number
of predators thus start to increase. While predator populations are low and the number of prey is increasing, conditions continually
improve for predators, and their populations grows approximately orthologistically for a time.
Then predators become abundant and drive the growth rate of the prey negative. The situation turns again, at C. Prey start to
decline and predator growth becomes approximately logistic, leveling off and starting to decline at D. By E it has come full circle
and the process repeats, ad infinitum.
While Figure 8.1.6 illustrates the classical form for predator-prey interactions, other forms are possible. When conditions are right,
the oscillations can dampen out and both predator and prey populations can reach steady states. Or the oscillations can become so

8.1.3 https://bio.libretexts.org/@go/page/25464
wild that predators kill all the prey and then vanish themselves. This assumes some effectively-zero value for N1 and N2, below
which they “snap” to zero. Or prey populations can become so low that predators all die out, leaving the prey in peace. Or both can
go extinct. Or, in the case of human predators, the prey can be domesticated and transformed into mutualists. More on all such
dynamics in later chapters.

This page titled 8.1: Dynamics of two Interacting Species is shared under a CC BY-NC 4.0 license and was authored, remixed, and/or curated by
Clarence Lehman, Shelby Loberg, & Adam Clark (University of Minnesota Libraries Publishing) via source content that was edited to the style
and standards of the LibreTexts platform; a detailed edit history is available upon request.

8.1.4 https://bio.libretexts.org/@go/page/25464
8.2: Code for two species
Below is computer code for two-species dynamics—a logical expansion of the code for one-species dynamics you saw earlier. The
code here produced the graph in Figure 8.1.2 and, with other values for Ni, ri, si,j, also produced the graphs in Figures 8.1.4 and
8.1.6.
N1=.01; N2=.01;
r1=0.5; r2=0.8;
s11=-0.08; s12=-0.03; s21=-0.09; s22=-0.06;
t=0; dt=.0001; tmax=75; step=0;
print(c(t, N1, N2));
while(t<tmax)
{ dN1=(r1+s11*N1+s12*N2)*N1*dt;
dN2=(r2+s21*N1+s22*N2)*N2*dt;
N1=N1+dN1; if(N1<0) N1=0;
N2=N2+dN2; if(N2<0) N2=0;
t=t+dt; step=step+1;
if(step==1000)
{ print(c(t, N1, N2)); step=0; }
This code uses a time step dt of 1/10,000 but displays only every 1000th time step, using the variable named step. This is a reliable
way to display output periodically, because step takes on only integer values 0, 1, 2.... Watching variable t for the same purpose
may not be reliable; t has rounding errors in its fractional portion, which typically is represented only approximately by the
computer.
By the way, here is a crucial requirement for coding multi-species dynamics: you must update all of the Ni together. You might be
tempted to shorten the code by writing the following, getting rid of the dN1 and dN2 variables.
N1=N1+(r1+s11*N1+s12*N2)*N1*dt; if(N1<0) N1=0;
N2=N2+(r2+s21*N1+s22*N2)*N2*dt; if(N2<0) N2=0;
This shortened code, however, contains a serious bug that could be difficult to detect. (A “bug” is the computerese term for any
mistake in computer code, typically one that goes undetected.) The calculation of N2 on the second line uses the new value of N1,
not the present value. This will generate subtly incorrect results that could be costly—if, for example, you were using the code to
project the course of an epidemic.
Careful code reviews with experienced colleagues are one way to help avoid such bugs. If the problem is important enough, having
the code written independently by two or more people not communicating with each other, except by accepting the same
specifications and comparing the final results, is a way to approach correctness. This is like independent replication of scientific
experiments. Other methods of ensuring that code is correct are addressed later.

This page titled 8.2: Code for two species is shared under a CC BY-NC 4.0 license and was authored, remixed, and/or curated by Clarence
Lehman, Shelby Loberg, & Adam Clark (University of Minnesota Libraries Publishing) via source content that was edited to the style and
standards of the LibreTexts platform; a detailed edit history is available upon request.

8.2.1 https://bio.libretexts.org/@go/page/25465
8.3: Summary of interactions
In summary, based on the effects of each population on the other, two species can interact mainly in three different ways, as shown
in Figure 8.3.1. Competition is a ‘−−’ combination, mutualism is ‘++’, and predation is ‘+−’, in either order.

Figure 8.3.1 . Summary of species interactions based on signs of their interaction parameters.
Sandwiched between the boxes above are special cases where one of the interaction terms is zero, or very close to zero. These are
called “commensalism,” when one parameter is positive and the other is zero, or “amensalism,” when one parameter is negative
and the other is zero. We won’t focus further on these special cases.

This page titled 8.3: Summary of interactions is shared under a CC BY-NC 4.0 license and was authored, remixed, and/or curated by Clarence
Lehman, Shelby Loberg, & Adam Clark (University of Minnesota Libraries Publishing) via source content that was edited to the style and
standards of the LibreTexts platform; a detailed edit history is available upon request.

8.3.1 https://bio.libretexts.org/@go/page/25466
 CHAPTER OVERVIEW

9: Embodied by Natural Selection

9.1: Chapter Introduction
9.2: Mutualism
9.3: Herbivory
9.4: Predation
9.5: Predatory plants
9.6: Defense
9.7: Competition

This page titled 9: Embodied by Natural Selection is shared under a CC BY-NC 4.0 license and was authored, remixed, and/or curated by
Clarence Lehman, Shelby Loberg, & Adam Clark (University of Minnesota Libraries Publishing) via source content that was edited to the style
and standards of the LibreTexts platform; a detailed edit history is available upon request.

1
9.1: Chapter Introduction

This page is a draft and is under active development.

The absolute simplicity of the three plus–minus combinations for two-species dynamics, explored in the previous chapter, is
converted by natural selection into physical forms of splendid complexity—unexpected and subtle, spectacular and sublime,
stunning and beautiful.
From our human perspective, some details appear to embody timeless altruism and kindness, others unbounded horror and cruelty.
But nature seems to implement material possibilities without regard to human values, and we are left merely to observe and
wonder, and strive to understand.
At this point, to gain perspective before proceeding further, it is well to examine at least a tiny subset of ecological reality. The
range of examples of multi-species interactions looks endless, and even within individual organisms are interactions among
multiple species.

This page titled 9.1: Chapter Introduction is shared under a CC BY-NC 4.0 license and was authored, remixed, and/or curated by Clarence
Lehman, Shelby Loberg, & Adam Clark (University of Minnesota Libraries Publishing) via source content that was edited to the style and
standards of the LibreTexts platform; a detailed edit history is available upon request.

9.1.1 https://bio.libretexts.org/@go/page/25664
9.2: Mutualism
At the deepest level, the eukaryotic cell seems to have been created of interactions among separate prokaryotic species one or two
billion years ago. Mitochondrial genomes are separate genomes within the cell with parallels to bacterial genomes, but
mitochondria are no longer able to live on their own except under very special circumstances. Chloroplasts are similar. These are
mutualisms at the very basis of complex life. Other mutualisms, like those with gut bacteria, form higher levels.
Mutualisms among pollinators and flowers are an ingenious arrangement for the fertilization of immobile organisms. A bee, for
example, collects pollen as a food source and in the process spreads pollen from plant to plant. Plants advertise the pollen with
bright flowers, and provide sweet nectar as an additional attraction (Figure 9.2.1, left).

Figure 9.2.1 . Pollination mutualism.

A small mammal can act similarly, drinking nectar from one plant, getting pollen on its whiskers, and transferring the pollen when
it drinks from another plant, in turn pollinating that plant (Figure 9.2.1, middle).
A fascinating pollination mutualism extends across the wetlands of the North American Upper Midwest, in marsh milkweed,
Asclepias incarnata (Figure 9.2.1, right). Mutualisms are not necessarily perfect, and each member can be exploited in some small
way. In this case, pollinators land on the milkweed flower and stand on “landing platforms” while taking nectar. But the platforms
are not secure; they are curved and slightly difficult footholds. A pollinator’s legs slip into cracks between the platforms and, when
the pollinator pulls its leg out, there are “saddle bags” of pollen (pollinia) wrapped around its leg, which can’t be dislodged. As it
flies to a another flower, the saddle bags rotate forward so they easily dislodge, and they’re pulled off when the pollinator slips on
another insecure landing platform of the corresponding species.
In another example, warthogs (Figure 9.2.2 left) attract abundant ticks. What better place, then, for Oxpeckers to find morsels of
food than in the bristly fur of a water buffalo? A little blood in the tick is probably a bonus. This is good for the water buffalo, good
for the oxpecker, and not good for the tick. This three-species interaction is (1) predation on the water buffalo by the tick, (2)
predation on the tick by the oxpecker, and thus (3) mutualism between water buffalo and oxpecker. It is an “enemy of my enemy is
my friend” interaction, one of the forty kinds of three-species interactions you will see in upcoming chapters.

Figure 9.2.2 Mutualisms through a third species.

Likewise, ants ward off potential predators from aphids (Figure 9.2.2 middle), and “cleaner fish” swim freely within the mouth of a
large fish (Figure 9.2.2 right) while they remove the fish’s parasites.

Figure 9.2.3 . Mutualism with a dangerous partner.

Sea anemones have stinging tentacles that other fish must avoid, but clown fish can resist the sting (Figure 9.2.3 left). This
mutualism is more complex. The clown fish protect themselves from predators by living among the anemones, and their bright
colors may attract predators who become prey for the anemone. Clown fish eat scraps of food missed by the anemone, plus their

9.2.1 https://bio.libretexts.org/@go/page/25471
own food from the water column, and provide nitrogen to the anemone from their digested waste. These are intricate mutualisms,
in which specifically matched species of clown fish and anemone have become permanent partners.
Pom-pom crabs employ sea anemones as weapons (Figure 9.2.3, right), carrying two anemones and waving them in a dance to
dissuade approaching predators. Crabs are sloppy eaters, so the sea anemones get the benefit of the mess as payment.

Figure 9.2.4 . Mutualism with people.

Early on, our hunter–gatherer predecessors were predators, but later they domesticated some of their prey, changing some predator–
prey interactions to mutualisms. From the point of view of domesticated sheep (Figure 9.2.4, left), humans may not be ideal
mutualists. We protect them from wolves, harbor them from disease, and shield them from the worst vagaries of weather. But we
also confine them to pens, shear off their wool, and kill and eat their lambs. Yet as agriculture advanced, the more people there
were on the planet, and the more sheep there were, and vice versa. This is the ecological making of a mutualism.
It is similar with crops. Instead of gathering grain and fruit from forest and field, we cleared areas specifically for domesticated
plants. We protect the crops from competition with other plants, work to keep them free of disease, and add water and nutrients to
the soil to help them grow. For thousands of years we used beasts of burden to accelerate cultivation, also in mutualisms with those
beasts. Now we are entering a wholly synthetic phase, in which teams of satellite-guided tractors navigate fields with the power of
a thousand horses (Figure 9.2.4 right).
For a thought exercise, you may want to ponder our relationship with machines. If we were mutualists with draft horses, are we
now mutualists with the machines that have replaced them? As you proceed through the next chapters, consider whether our
relationships with our machines meet the ecological requirements of mutualisms.

This page titled 9.2: Mutualism is shared under a CC BY-NC 4.0 license and was authored, remixed, and/or curated by Clarence Lehman, Shelby
Loberg, & Adam Clark (University of Minnesota Libraries Publishing) via source content that was edited to the style and standards of the
LibreTexts platform; a detailed edit history is available upon request.

9.2.2 https://bio.libretexts.org/@go/page/25471
9.3: Herbivory

Figure 9.3.1 . Herbivory.

Herbivory is a kind of predation in which the prey is a plant. Upon detecting herbivory, the plant may pump toxins into the leaf to
dissuade the herbivore. In response, the herbivore may quickly chew out a pattern to isolate part of the leaf from the toxins and then
dine in relative salubrity (Figure 9.3.1, left).
Multitudes of leaf cutter ants can be formidable herbivores (Figure 9.3.1, middle), cutting luggable-sized pieces of leaves to take
back to their nests. The ants do not eat the leaves, but chew them further, feed them to fungus, and then eat the fungus—creating an
ant–fungus mutualism, with ants being predators on the trees.
At a larger scale, multitudes of bison were formidable herbivores of the prairie (Figure 9.3.1, right). The native tallgrass prairies
were wildflower gardens, with a few dozen species of grasses but hundreds of species of owers—said to bloom a different color
each week.

Figure 9.3.2 . Beaver–aspen herbivory. Or is it mutualism?

Beavers fell whole aspen trees (Figure 9.3.2, left) to make their dams and lodges (middle and right), and also for food. This sounds
like herbivory. At the same time, however, they girdle and kill other species of trees that they do not use as food, clearing the way
for new aspen. This is mutualism.
The photo on the right in Figure 9.3.2 shows a pond colonized by beaver after an absence of more than a century. Though it is
midsummer, a few of the trees are defoliated— including one giant bur oak—because beaver have chewed off the cambium layer
all around the trees, girdling and killing them. Judiciously followed, this practice would keep the forest in an early successional
stage, a condition which favors staple aspen.

This page titled 9.3: Herbivory is shared under a CC BY-NC 4.0 license and was authored, remixed, and/or curated by Clarence Lehman, Shelby
Loberg, & Adam Clark (University of Minnesota Libraries Publishing) via source content that was edited to the style and standards of the
LibreTexts platform; a detailed edit history is available upon request.

9.3.1 https://bio.libretexts.org/@go/page/25472
9.4: Predation
Bears catching fish and a kookaburra ambushing a frog (Figure 9.4.1, left and middle) are simple kinds of predation. Other ambush
strategies are also common. The stonefish (right) is disguised to match its background and waves a lure to attract other fish, who
are then instantly swallowed whole. Despite hundreds of millions of years of evolution, the trick still works.

Figure 9.4.1 . Predation.

Cats typically pursue ambush strategies, lying in wait (Figure 9.4.2, left), whereas dogs typically pursue sustained chasing
strategies. This means that cats must be relatively odorless to avoid detection, but dogs need not be.

Figure 9.4.2 . Ambush predation.

In a curious ambush strategy only recently achieved, herons wave gathered feathers to attract fish, then drop the feathers and grab
the fish (Figure 9.4.2, right). Such use of feathers is apparently an animal “meme” that has spread rapidly through heron
populations after being discovered by some Einstein-heron.

Figure 9.4.3 . Predation and parasitoidism.

Species interactions can lead to remarkable evolutionary adaptations, including the non-messy way of eating an egg (Figure 9.4.3

left).
Nature acknowledges neither kindness nor cruelty, but parasitoidism seems one of the cruelest strategies (Figure 9.4.3, right). Here
an Ammophilia wasp is carrying a caterpillar not to kill and eat, but as a living hatchery for her eggs. Young wasps developing
from these eggs consume the caterpillar from within as the caterpillar remains alive, transforming it into wasp larvae. Parasitoidism
is widespread.

9.4.1 https://bio.libretexts.org/@go/page/25473
 Figure 9.4.4 . Parasitism.

When the predator is small relative to the prey, predator– prey interactions are called “parasitism.” At left in Figure 9.4.4 is a
blood-sucking mosquito attached to an unfortunate human. At right is a beetle seemingly overwhelmed with mites. When the
predator is much smaller still, it is called a “pathogen” and the interactions are called “infection” and “disease.”

This page titled 9.4: Predation is shared under a CC BY-NC 4.0 license and was authored, remixed, and/or curated by Clarence Lehman, Shelby
Loberg, & Adam Clark (University of Minnesota Libraries Publishing) via source content that was edited to the style and standards of the
LibreTexts platform; a detailed edit history is available upon request.

9.4.2 https://bio.libretexts.org/@go/page/25473
9.5: Predatory plants

Figure 9.5.1 . Plants as predators.

Peatlands can form over vast areas where the habitat is isolated from a normal supply of nutrients (Figure 9.5.1, left) and life there
must endure low levels of nitrogen. In this situation, some plants become predators.
The pitcher plant (Figure 9.5.1, middle), for example, attracts insects with nectar and then eats them to obtain nitrogen, trapping
them in a pool of digestive fluid at the bottom of a tall green vase with slippery, unidirectional sides of downward-pointing hairs.
The sundew (Figure 9.5.1, right) seems simpler, capturing unwary insects in sticky droplets and then consuming them.

This page titled 9.5: Predatory plants is shared under a CC BY-NC 4.0 license and was authored, remixed, and/or curated by Clarence Lehman,
Shelby Loberg, & Adam Clark (University of Minnesota Libraries Publishing) via source content that was edited to the style and standards of the
LibreTexts platform; a detailed edit history is available upon request.

9.5.1 https://bio.libretexts.org/@go/page/25474
9.6: Defense

Figure 9.6.1 . Predation, warnings, and deception.

Prey develop remarkable defenses against predators, following the processes of evolution, and provide warnings of the existence of
their defenses. Some such defenses and advertisements are real, like the fetid fluid sprayed by skunks (Figure 9.6.1 left) that can
deter even large bears from attacking. The fire-bellied toad (next to left) is filled with toxins and its bright color advertises, “Do not
eat me!”
Others species benefit from complex deception to keep predators away, such as the harmless clear-winged moth colored to look
anything but harmless (next to right), and an edible caterpillar in disguise with a viper-like tail (right).

Figure
9.6.2

.
Predation, camouflage, and temptation.

Hiding is a simple, common strategy for escaping predators. On the left in Figure 9.6.2 is a young grasshopper on pebbles. Can you
see it? Zoom in and look just a little below and to the left of center. The grasshopper’s head is downward and its tail is upward and
slightly to the left, with one antenna and one leg clearly visible once you see them.
The killdeer (Figure 9.6.2, right) tempts predators who get too close by stumbling away from the nest, feigning an injured wing and
making itself look like an easy catch. Once it has lured the predator far from the nest, it lifts quite competently into the air and flies
off.

This page titled 9.6: Defense is shared under a CC BY-NC 4.0 license and was authored, remixed, and/or curated by Clarence Lehman, Shelby
Loberg, & Adam Clark (University of Minnesota Libraries Publishing) via source content that was edited to the style and standards of the
LibreTexts platform; a detailed edit history is available upon request.

9.6.1 https://bio.libretexts.org/@go/page/25475
9.7: Competition
Examples of competition seem both more subtle and more ordinary than examples of mutualism and predation. But competition is
pervasive.

Figure 9.7.1 . Competition at the North American triple ecotone.

North America has many ecosystems—the tundra of the Arctic, the deserts of the Southwest, the giant conifers of the Pacific
Northwest—but the three largest ecosystems merge in a triple ecotone in the Upper Midwest, an area which exemplifies
competition among plants. Here, the needle-leaf forests stretch north to the arctic, the broad-leaf forests extend east to the Atlantic,
and the prairies’ amber waves of grain ow west to the Rockies. Figure 9.7.1 shows a whirlpool of competition at this broad triple
ecotone. White pines stand tall above the deciduous trees in the background, with Big Bluestem and other native prairie grasses
setting seed in the foreground. Staghorn Sumac in red fall colors tries to hold its own in the middle, with pines invading behind it.
Leaves of Bur Oak are browning for the oncoming winter.
While it may seem a peaceful scene, for the plants it is a scene of intense competition for their very existence. Fire is a foe of trees
(Figure 9.7.1 right), killing many instantly, and thus favoring grasses and prairie flowers. Times of moister conditions allow trees
to reenter the grasslands, eventually shading the grassland vegetation to death if the moisture persists. But complexities of weather
and climate have kept this competitive tension zone intact for thousands of years, with no group permanently gaining the
competitive advantage.

This page titled 9.7: Competition is shared under a CC BY-NC 4.0 license and was authored, remixed, and/or curated by Clarence Lehman,
Shelby Loberg, & Adam Clark (University of Minnesota Libraries Publishing) via source content that was edited to the style and standards of the
LibreTexts platform; a detailed edit history is available upon request.

9.7.1 https://bio.libretexts.org/@go/page/25476
 CHAPTER OVERVIEW

10: Phase Space

10.1: Chapter Introduction
10.2: Phase Space
10.3: Surface generating the flow
10.4: A phase space example

This page titled 10: Phase Space is shared under a CC BY-NC 4.0 license and was authored, remixed, and/or curated by Clarence Lehman, Shelby
Loberg, & Adam Clark (University of Minnesota Libraries Publishing) via source content that was edited to the style and standards of the
LibreTexts platform; a detailed edit history is available upon request.

1
10.1: Chapter Introduction

This page is a draft and is under active development.

In differential equation models, the basic population dynamics among species become visible at a glance in “phase space.” The
concepts and applications of phase spaces were originally worked out late in the nineteenth century by
Henri Poincaré and others for the dynamical systems of physics, but the mathematical foundations also
apply to the theories of ecology. (At left is Poincaré seated with Marie Curie at the initial Solvay
Conference in 1911. As a point of interest, Marie Curie is the only person to have twice won the Nobel
Prize for science—and she was nominated the first time before her doctoral defense!)
In a phase space with two species interacting—in competition, predation, or mutualism—the abundance of one species occupies the
horizontal axis and that of the other occupies the vertical axis. This makes each possible pair of population values (N1,N2) into a
point in the phase space.

Figure 10.1.1 . Phase space with two species abundances and directional derivatives also marked.
For example, a measured average abundance of 1.55 individuals per square meter for Species 1 and of 1.1 individuals per square
meter for Species 2 corresponds to the point marked with an ‘×’ in Figure 10.1.1 — 1.55 units to the right on the horizontal axis
and 1.1 units up the vertical axis. If Species 1 is rare, at 0.05 individuals per square meter, and Species 2 is at 0.85 individuals per
square meter, the point is that marked with a ‘+’, near the left in Figure 10.1.1. A phase space, however, is not about the size of
populations, but rather about how the populations are changing over time. That change
dN1 dN2

dt
= f1 (N1 , N2 ),
dt
= f2 (N1 , N2 )) is made visible as arrows emerging from each point.

Figure 10.1.2 . Phase space with two species abundances and trajectories to equilibria.
Suppose that, at the time of the measurement of the populations marked by ×, Species 1 is decreasing slightly and Species 2 is
increasing relatively strongly. Decreasing for Species 1 means moving to the left in the phase space, while increasing for Species 2
means moving up, as shown in the inset of Figure 10.1. The net direction of change would thus be north-northwest. In the opposite
direction, for the populations marked by +, if Species 1 is increasing slightly and Species 2 is decreasing relatively strongly, the
direction of change would be south-southeast.
Arrows in the phase space point in the direction of immediate population change. But as the populations change, ecological
conditions also change and the paths curve. Figure 10.1.2 shows in green how the populations change in this example as time
passes. The pair of abundances starting from × moves up, with Species 1 decreasing at first and then both species increasing and
finally coming to rest at the green dot, which marks a joint carrying capacity.

10.1.1 https://bio.libretexts.org/@go/page/25711
From the +, on the other hand, Species 2 decreases uniformly but Species 1 increases at first and then reverses direction. In this
case both species go extinct by arriving at the origin (0,0). Something significant separates the + from the ×.

Figure 10.1.3 . Phase space with multiple equilibria.

What separates them can be judged by calculating an arrow at many points throughout the phase space (Figure 10.1.3). Following
the arrows from any pair of abundances (N1,N2) traces the future abundances that will arise as time progresses, and following the
arrows backwards shows how the populations could have developed in the past. Note that the arrows seem to be avoiding the open
circle near the lower left (at about 0.5, 0.3). That is an Allee point.
Some points in the phase space are exceptional: along certain special curves, the arrows aim exactly horizontally or exactly
vertically. This means that one of the two populations is not changing—Species 1 is not changing along the vertical arrows, and
Species 2 is unchanged along the horizontal arrows. These special curves are the isoclines—from the roots ‘iso-,’ meaning ‘same’
or ‘equal,’ and ‘-cline,’ meaning ‘slope’ or ‘direction.’)

Figure 10.1.4 . Phase space with multiple equilibria and isoclines.

The two isoclines of Species 2 are shown in red in Figure 10.1.4, one along the horizontal axis and the other rising and curving to
the right. On the horizontal axis, the abundance of Species 2 is zero. Therefore it will always stay zero, meaning it will not change
and making that entire axis an isocline. Along the other red isocline, the arrows emerging exactly from the isocline point exactly
right or left, because the system is exactly balanced such that the abundance of Species 2 does not change—it has no vertical
movement.
The situation is similar for the two isoclines of Species 1, shown in blue in Figures 10.1.4 and 10.1.5 —one along the vertical axis
and the other rising and curving upward. Along the blue curves, the arrows emerging exactly from the isocline point exactly up or
down. Again, along the blue isocline the system is exactly balanced such that the abundance of Species 1 does not change—it has
no horizontal movement.

10.1.2 https://bio.libretexts.org/@go/page/25711
 Figure 10.1.5 . Phase space with multiple equilibria, isoclines, and trajectories.
Understanding the isoclines of a system goes a long ways toward understanding the dynamics of the system. Where an isocline of
one species meets an isocline of the other, the population of neither species changes and therefore an equilibrium forms. These are
marked with circles. Notice that the arrows converge on the filled circles (stable equilibria) and judiciously avoid the open circle
(unstable equilibrium). And notice that wherever a population (N1,N2) starts, the arrows carry it to one of two outcomes (except,
technically, starting on the Allee point itself, where it would delicately remain until perturbed).
For further illustration, four population growth curves are traced in green in Figure 10.1.5 and marked as A, B, C, and D. All start
with one of the populations at 2.0 and the other at a low or moderate level. And they head to one of the two stable equilibria,
avoiding the unstable equilibrium in between. You can view these four growth curves plotted in the usual way, as species
abundances versus time, in Figure 10.1.6. Blue indicates Species 1, while red indicates Species 2.

Figure 10.1.5 through time.

This page titled 10.1: Chapter Introduction is shared under a CC BY-NC 4.0 license and was authored, remixed, and/or curated by Clarence
Lehman, Shelby Loberg, & Adam Clark (University of Minnesota Libraries Publishing) via source content that was edited to the style and
standards of the LibreTexts platform; a detailed edit history is available upon request.

10.1.3 https://bio.libretexts.org/@go/page/25711
10.2: Phase Space
A good way to understand the arrows of phase spaces is to imagine raindrops falling on a curvilinear rooftop and owing across its
surface. Figure 10.2.1 shows such a surface.

Figures 10.0.3 to 10.0.5.

Why should thinking of raindrops on rooftops help us understand phase spaces? It is because the differential equations themselves
are situated on mathematically surfaces—albeit sometimes higher-dimensional surfaces—with points owing dynamically across
the surfaces, just like raindrops owing across a roof. It is not completely the same, of course, but is a useful aid to thought.
Instead of raindrops, it can also be useful to think of a marble rolling on the surface. At the bottom of the basin at
Point C in Figure 10.2.1, a marble is trapped. The surface goes up in every direction from this point, so after any
small disturbance the marble will roll to the bottom again.
Point B corresponds to the equilibrium at the origin, stable in this case, where both species are extinct. A marble
resting on this surface and experiencing a small positive disturbance away from the origin must roll uphill in every direction, so it
will return to that equilibrium as well. It is below the two-species Allee point.
For example, Point A divides rain owing to the left and rain owing to the right. The basin at Point C corresponds to the carrying
capacity, Point B corresponds to extinction at the origin, and Point A corresponds to the unstable Allee point.
Point A, on the other hand, corresponds to the Allee point. A marble could be balanced precariously at that place, with the slightest
breath of air sending it to extinction at B or carrying capacity in the basin at C, depending on miniscule variations in the breath.
Marbles starting close to either of the axes roll to the origin, equilibrium B. Marbles starting farther from the axes are on the other
side of a long ridge and roll to the carrying capacity at C.

This page titled 10.2: Phase Space is shared under a CC BY-NC 4.0 license and was authored, remixed, and/or curated by Clarence Lehman,
Shelby Loberg, & Adam Clark (University of Minnesota Libraries Publishing) via source content that was edited to the style and standards of the
LibreTexts platform; a detailed edit history is available upon request.

10.2.1 https://bio.libretexts.org/@go/page/25478
10.3: Surface generating the flow
Think about shapes on which a marble could remain stationary, possibly balanced precariously, on a complicated two-dimensional
surface in three-dimensional space, with peaks and valleys in their structure. Figure 10.3.1 shows seven possible configurations for
remaining stationary.

Figure 10.3.1 . Seven kinds of equilibria possible as part of surfaces defining two-species phase spaces.
Configuration A is a summit, a high point with respect to its surroundings. It curves downward in every direction. It is therefore
unstable—a marble perched on top can roll away in any direction.
Configuration C is the opposite, a basin. The surface curves upward in every direction. It is stable because a marble resting at the
bottom rolls back from a disturbance in any direction.
Configuration B is like a combination of A and C. It is called a “saddle” for its once-ubiquitous shape (right). It
curves upward in some directions and downward in others. A marble resting at its very center is unstable because
it can roll away in many directions.
Configurations D, E, and F are related to A, B, and C, but are level in at least one direction. A “ridge,”
Configuration D, has equilibria all along its very top. A marble balanced precariously there and nudged could
conceivably move to a new equilibrium along the ridge, if the nudge were aligned with infinite exactitude;
almost certainly, however, the marble would roll off. This configuration has an infinite number of equilibria—all
along the ridge—but none of them are stable.
A “trough,” Configuration F, is the opposite of a ridge, with equilibria all along its lowest levels. A marble resting there and nudged
will move to a new equilibrium position at the base of the trough. Again there are an infinite number of equilibria—all along the
base—but none are stable because, pushed along the trough, the marble does not return to its previous location. The equilibria are
neutrally stable, however, in the ecological view.
An “inection,” Configuration E, is like a combination of D and F, changing slope and becoming level, but then resuming the same
direction as before and not changing to the opposite slope. It too has a level line with an infinite number of equilibria, all unstable,
with marbles rolling away from the slightest nudge in half of the possible directions.
Configuration G, a perfectly flat “plain,” is perhaps easiest to understand. A marble can rest anywhere, so every point on the flat
surface is an equilibrium. But a marble will not return to its former position if nudged, so no equilibrium on the flat surface is
stable. In ecology this situation is sometimes called “neutrally stable;” in mathematics it is called “unstable.”
With three-dimensional images and human cognitive power, it is possible to visualize a surface at a glance, such as in Figure
10.1.1, and judge the implications for populations growing according to equations that correspond to that surface. You can see at a
glance if it is curving up everywhere or down everywhere, if it combines upward and downward directions, or if it has level spots.
You can classify each equilibrium into the configurations of Figure 10.3.1. But how can that judgment be quantified, made
automatic?
The method of eigenvectors and eigenvalues accomplishes this. Think of the prefix eigen- as meaning “proper,” as in “proper
vector” or “proper value.” The idea will become clear shortly. The method of eigenvectors and eigenvalues was developed in stages
over the course of more than two centuries by some of the best mathematical minds, and now we can apply it intact to ecology.

10.3.1 https://bio.libretexts.org/@go/page/25479
 p
p
aa
g
g
ee
1
1
3
3
8
8
ii
m
m
aa
g
g
ee
3
3
7
7
4
4
4
4
3
3
5
5
8
8
4
4

Figure 10.1.1, with the slope of the surface (its first derivative, dy/dx, middle) and changes in that slope (its second derivative, d
2y/dx2, bottom)

Think of a one-dimensional slice through the surface of Figure 10.1.1, successively passing through points B, A, C, and beyond. It
would look like the top of Figure 10.3.2 As before, a marble balanced precisely at A would be unstable, ready to roll either toward
B or C. The equilibrium points are level points where the slope is zero, as they are on the surface of Figure 10.1.1 From calculus,
this is where the derivative is zero, where dy /dx = 0. Those slopes of zero are marked by green horizontal lines in the figure.
The middle graph of Figure 10.3.2 shows sign of the derivative, dy/dx, plus or minus. The sign function on the vertical axis, sgn(u
), is equal to zero if u is zero but is equal to plus or minus one if u is positive or negative, respectively. Whether an equilibrium is in
a trough or at a summit is determined by how the slope is changing exactly at the equilibrium point. From calculus, that is the
2
d y
second derivative, dx2
, recording changes in the first derivative, dy/dx —just as the first derivative records changes in the surface
itself.
The sign of the second derivative is shown in the bottom part of Figure 10.3.2. Wherever the slope is increasing at an equilibrium
point—that is, changing from sloping down on the left to sloping up on the right—that is a basin. Wherever it is decreasing at an
equilibrium point—changing from sloping up at the left to sloping down at the right—that is a summit. Whether an equilibrium
point is stable or not can thus be determined mathematically merely from the sign of the second derivative of surface at that point!
This is easy if there is only one species, as in the models of earlier chapters, with only one direction to consider. But it becomes
tricky when two or more species are interacting, for an infinite number of directions become available.
alt

Figure 10.3.3 Partial derivatives, eigenvectors, and stability.

It might seem that a configuration will be a basin if the surface curves upward in both the x and y directions, as in Configuration C
of Figure 10.3.1. But have a look at the three parts of Figure 10.3.3. Part A is a surface with a trough aligned with the axes.
Looking along the x-axis—which would be the N1 axis showing the abundance of Species 1—the surface is curving up on both
sides of its minimum (white curve). However, looking along the y -axis—the N2 axis showing the abundance of Species 2—reveals
that it is exactly level in that direction (green line), meaning the equilibrium is not stable.
But suppose the same surface is rotated 45 degrees, as in part B of the figure. The surface curves upward not only along the x -axis
(white curve) but also along the y -axis (green curve). Yet the surface is the same. Contrary to what might have been expected,
curving upward in both the x and y directions does not mean the configuration is a basin! Understanding the structure means
looking in the proper directions along the surface, not simply along the axes.
This is what eigenvalues and eigenvectors do. They align with the “proper” axes for the surface, as illustrated in part C. No matter
how twisted, skewed, or rescaled the surface is with respect to the axes, the eigenvectors line up with the “proper” axes of the
surface, and the eigenvalues measure whether the slope is increasing or decreasing along those axes at an equilibrium.

10.3.2 https://bio.libretexts.org/@go/page/25479
 Box 10.3.1 rules of eigenvalues for hill-climbing systems

1. If all eigenvalues are negative, the equilibrium is stable.

2. If any eigenvalue is positive, the equilibrium is unstable.
3. If some or all of the eigenvalues are zero and any remaining eigenvalues are negative, there is not enough information in
the eigenvalues to know whether the equilibrium is stable or not. A deeper look at the system is needed.

In short, if all the eigenvalues are positive, the equilibrium is a basin, as in Figures 10.1.1C and 10.3.1C. If all the eigenvalues are
negative, the equilibrium is a summit, as in Figures 10.1.1A and 10.3.1A. And if the eigenvalues are of mixed signs, or if some are
zero, then we get one of the other configurations. (See Box 10.3.1)

he influence of gravity. Dynamical systems can do the opposite— they can climb to the highest level in the locality. For example,
natural selection is commonly described as climbing “fitness peaks” on abstract “adaptive landscapes.” Of course, for mathematical
surfaces rather than real mountain ranges, this is just a point of view. Whether a system climbs up or slides down depends
arbitrarily on whether the mathematical surface is affixed with a plus or a minus sign in the equations.
Figure 10.3.2 , an equilibrium in a hill-climbing system is stable if the slope is decreasing (negative sign) rather than increasing
(positive sign). Because human intuition applies so well to marbles rolling off heights and settling into shallows, we have chosen to
explain it intuitively that way first, but it is important to understand both ways.
It turns out that the proper axes at each equilibrium point—the eigenvectors—can be determined exactly
from only four numbers, and how much the slope is increasing or decreasing at each equilibrium point—
the eigenvalues—can be determined at the same time from the same four numbers. These are the four
partial derivatives in what is called the “Hessian matrix” of the surface, or, equivalently in the “Jacobian
matrix” of the population growth equations. An understanding of these matrices and their applications has developed in
mathematics over the past two centuries.
By expending some effort and attention you can work the eigenvalues out mathematically with pencil and paper. However, you will
likely employ computers to evaluate the eigenvalues of ecological systems. This can be done with abstract symbols in computer
packages such as Mathematica or Maxima, or numerically in programming languages such as R. For standard two-species systems,
we have worked out all equilibria and their corresponding eigenvalues. These are recorded in Table 10.3.1 in mathematical
notation and in Program 10.3.1 as code, and identify the equilibria and stability for all predation, mutualism, and competition
systems represented by the two-species formulae, which is copied into the table for reference.
Table 10.3.1. Two-species formulae

Location Equilibrium Eigenvalues

Origin
(0,0) ( r1 , r2 )
(Both species extinct)

Horizontal axis r1 q
− 0 −r1 , )
(Species 1 at K1) s1, 1 s1, 1

Vertical axis r2 p
0, − −r2
(Species 2 at K2) s2, 2 s2, 2

Interior p q −b±√b −4ac

2

,
(Coexistence) a a
2a

with
a = s1,2 s2,1 − s1,1s2,2
b = r1 s2,2 (s2,1 −s1,1) + r2 s1,1 (s1,2 −s2,2)
c = −pq

10.3.3 https://bio.libretexts.org/@go/page/25479
 p = r1 s2,2 −r2 s1,2
q = r2 s1,1 −r1 s2,1
in the ecological equations for two interacting species,
1 dN1
= r1 + s1,1 N1 + s1,2 N2
N1 dt

1 dN2
= r2 + s2,2 N1 + s2,1 N1
N2 dt

where
N1, N2 are population abundances of Species 1 and 2
r1, r2 are intrinsic growth rates
s1,1, s2,2 measure species effects on themselves
s1,2, s2,1 measure effects between species
p = r1*s22 -r2*s12; # Compute useful sub
q = r2*s11 -r1*s21; # formulae.
a = s12*s21 -s11*s22;
b = r1*s22*(s21-s11) +r2*s11*(s12-s22);
c = -p*q; # Compute the equilibria.
x00=0; y00=0; # (at the origin)
x10=-r1/s11; y10=0; # (on the x-axis)
x01=0; y01=-r2/s22; # (on the y-axis)
x11=p/a; y11=q/a; # (at the interior)
v00= r1; w00=r2; # Compute the corresponding
v10=-r1; w10=q/s11; # four pairs of eigenvalues
v01=-r2; w01=p/s22; # (real part only).
v11=(-b-Sqrt(b^2-4*a*c))/(2*a);
w11=(-b+Sqrt(b^2-4*a*c))/(2*a);
Program 10.3.1. The code equivalent to Table 10.3.1, for use in computer programs. Sqrt(w) is a specially written function that
returns 0 if w is negative (returns the real part of the complex number 0 +√−
−
w i ).

The formulae in Table work for any two-species RSN model—that is, any model of the form
10.3.1
dNi
1

Ni dt
with constant coefficients—but formulae for other models must be derived separately, from a
= r1 + si,i Ni + si,j Nj

software package, or following methods for Jacobian matrices.

box 10.3.2 parameters for a sample Competitive system

r1 = 1.2 r2 = 0.8 Intrinisic growth rate

s1,1 = −1 s2,2 = −1 Self-limiting terms

s1,2 = −1.2 s2,1 = −0.5 Cross-limiting terms

This page titled 10.3: Surface generating the flow is shared under a CC BY-NC 4.0 license and was authored, remixed, and/or curated by Clarence
Lehman, Shelby Loberg, & Adam Clark (University of Minnesota Libraries Publishing) via source content that was edited to the style and
standards of the LibreTexts platform; a detailed edit history is available upon request.

10.3.4 https://bio.libretexts.org/@go/page/25479
10.4: A phase space example
For an example of finding equilibria and stability, consider two competing species with intrinsic growth rates r1 = 1.2 and r2 = 0.8.
Let each species inhibit itself in such a way that s1,1 =−1 and s2,2 =−1, let Species 2 inhibit Species 1 more strongly than it inhibits
itself, with s1,2 =−1.2, and let Species 1 inhibit Species 2 less strongly than it inhibits itself, with s2,1 =−0.5. These conditions are
summarized in Box 10.2.2 for reference. The question is, what are the equilibria in this particular competitive system, and what
will their stability be?
First, there is an equilibrium at the origin (0,0) in these systems, where both species are extinct. This is sometimes called the
“trivial equilibrium,” and it may or may not be stable. From Table 10.2.1, the eigenvalues of the equilibrium at the origin are r1 and
r2 —in this case 1.2 and 0.8. These are both positive, so from the rules for eigenvalues in Box 10.2.1, the equilibrium at the origin
in this case is unstable. If no individuals of either species exist in an area, none will arise. But if any individuals of either species
somehow arrive in the area, or if both species arrive, the population will increase. This equilibrium is thus unstable. It is shown in
the phase space diagram of Figure 10.4.1, along with the other equilibria in the system.
box 10.4.1 calculated results for the sample competitive system

Equilibrium Coordinates Eigenvalues Condition

Origin (0,0) (1.2,0.8) Unstable

Horizontal axis (1.2,0) (-1.2,0.2) Unstable

Vertical axis (0,0.8) (-0.8,0.24) Unstable

Interior (0.6,0.5) (-0.123,-0.977) Stable

On the horizontal axis, where Species 2 is not present, the equilibrium of Species 1 is N'1 = −r1/s1,1 = 1.2. That is as expected—it is
just like the equilibrium of N' = −r /s for a single species—because it indeed is a single species when Species 2 is not present. As to
the stability, one eigenvalue is −r1, which is −1.2, which is negative, so it will not cause instability. For the other eigenvalue at this
equilibrium, you need to calculate q from Table 10.2.1. You should get q =−0.2, and if you divide that by s1,1, you should get 0.2.
This is positive, so by the rules of eigenvalues in Box 10.2.1, the equilibrium on the horizontal axis is unstable. Thus, if Species 1
is at its equilibrium and an increment of Species 2 arrives, Species 2 will increase and the equilibrium will be abandoned.
Likewise, on the vertical axis, where Species 1 is not present, the equilibrium of Species 2 is N2 = −r2/s2,2 = 0.8. Calculate the
eigenvalues at this equilibrium from Table 10.2.1 and you should get p=−0.24, and dividing by s2,2 give eigenvalues of −0.8 and
0.24. With one negative and the other positive, by the rules of eigenvalues in Box 10.2.1 the equilibrium on the vertical axis is also
unstable.
Finally, for the fourth equilibrium—the interior equilibirum where both species are present—calculate a, b, and c from the table.
You should get a =−0.4, b =−0.44, and c=−0.048. Now the interior equilibrium is N'1 = p/a = 0.6 and N'2 = q/a = 0.5.

10.4.1 https://bio.libretexts.org/@go/page/25480
 Figure 10.4.1 . Phase space for the example of competition with stable coexistence summarized in Boxes 10.2.2 and 10.4.1 . The
arrows are calculated by code like that of Program 10.4.1 , included for reference.
But is it stable? Notice the formula for the eigenvalues of the interior equilibrium in Table 10.2.1, in terms of a, b, and c. It is
simply the quadratic formula! This is a clue that the eigenvalues are embedded in a quadratic equation, ax2 + bx + c = 0. And if you
start a project to derive the formula for eigenvalues with pencil and paper, you will see that indeed they are. In any case, working it
out more simply from the formula in the table, you should get −0.123 and −0.977. Both are negative, so by the rules of Box 10.2.1
the interior equilibrium for this set of parameters is stable.
As a final note, the presence of the square root in the formula suggests that eigenvalues can have imaginary parts, if the square root
covers a negative number. The rules of eigenvalues in Box 10.2.1 still apply in this case, but only to the real part of the eigenvalues.
−1±√−5
Suppose, for example, that the eigenvalues are 2
= −0..5 ± 1.118i. These would be stable because the real part, −0.5, is

negative. But it turns out that because the imaginary part, ± 1.118i, is not zero, the system would cycle around the equilibrium
point, as predator–prey systems do.
In closing this part of the discussion, we should point out that eigenvectors and eigenvalues have broad applications.
They reveal, for instance, electron orbitals inside atoms (right), alignment of multiple variables in statistics, vibrational
modes of piano strings, and rates of the spread of disease, and are used for a bounty of other applications. Asking how
eigenvalues can be used is a bit like asking how the number seven can be used. Here, however, we simply employ them
to evaluate the stability of equilibria.

10.4.2 https://bio.libretexts.org/@go/page/25480
Program 10.4.1 Sample program in R to generate a phase space of arrows, displaying the locations of the beginning and ends of
the arrows, which are passed through a graphics program for display. The ‘while(1)’ statement means “while forever”, and is just
an easy way to keep looping until conditions at the bottom of the loop detect the end and break out.

This page titled 10.4: A phase space example is shared under a CC BY-NC 4.0 license and was authored, remixed, and/or curated by Clarence
Lehman, Shelby Loberg, & Adam Clark (University of Minnesota Libraries Publishing) via source content that was edited to the style and
standards of the LibreTexts platform; a detailed edit history is available upon request.

10.4.3 https://bio.libretexts.org/@go/page/25480
 CHAPTER OVERVIEW

11: State Spaces

11.1: State Spaces

This page titled 11: State Spaces is shared under a CC BY-NC 4.0 license and was authored, remixed, and/or curated by Clarence Lehman, Shelby
Loberg, & Adam Clark (University of Minnesota Libraries Publishing) via source content that was edited to the style and standards of the
LibreTexts platform; a detailed edit history is available upon request.

11.1 https://bio.libretexts.org/@go/page/25491
11.1: State Spaces
Closely related to “phase spaces” are “state spaces”. While phase spaces are typically used with continuous systems, described by
differential equations, state spaces are used with discrete-time systems, described by difference equations. Here the natural system
is approximated by jumps from one state to the next, as described in Chapter 7, rather than by smooth transitions. While the two
kinds of spaces are similar, they differ in important ways.

Figure 11.1.1 . A prairie ecosystem setting seed in fall colors— Bluebird Prairie, U.S. Upper Midwest. If not burned, grazed, or
mowed, grasslands can leave a thick layer of litter that inhibits the next year’s growth.
Inspired by the complexities of ecology, and triggered in part by Robert May’s bombshell paper of 1976, an army of
mathematicians worked during the last quarter of the twentieth century to understand these complexities, focusing on discrete-time
systems and state spaces. One endlessly interesting state space is the delayed logistic equation (Aronson et al. 1982), an outgrowth
of the discrete-time logistic equation described in Chapter 7.
For a biological interpretation of the delayed logistic equation, let’s examine the example of live grassland biomass coupled with
last year’s leaf litter. Biomass next year (Nt+1) is positively related to biomass this year (Nt), but negatively related to biomass from
the previous year (Nt−1). The more biomass in the previous year, the more litter this year and the greater the inhibitory shading of
next year’s growth. The simplest approximation here is that all biomass is converted to litter, a fixed portion of the litter decays
each year, and inhibition from litter is linear. This is not perfectly realistic, but it has the essential properties for an example. Field
data and models have recorded this kind of inhibition (Tilman and Wedin, Nature 1991).

Program 11.1.1. A program to compute successive points in the state space of the delayed logistic equation.
The basic equation has N1 as live biomass and N2 as accumulated leaf litter. N1 and N2 are thus not two different species, but two
different age classes of a single species.
N1 (t + 1) = rN1 (t) (1 − N2 (t))

N2 (t + 1) = N1 (t) + p N2 (t)

The above is a common way to write difference equations, but subtracting Ni from each side, dividing by Ni , and making p = 0 for
simplicity gives the standard form we have been using.
1 ΔN1
= (r − 1) − rN2 = r1 + s1,2 N2
N1 Δt

1 ΔN2 1
= −1 + N1 = r2 + s2,1 N1
N2 Δt N2

Notice something new. One of the coefficients, s2,1, is not a constant at all, but is the reciprocal of a dynamical variable. You will
see this kind of thing again at the end of the predator–prey chapter, and in fact it is quite a normal result when blending functions
(Chapter 18) to achieve a general Kolomogorov form. So the delayed logistic equation is as follows:
1 ΔN1
= r1 + s1,2 N2
N1 Δt

11.1.1 https://bio.libretexts.org/@go/page/25485
 1 ΔN2
= r2 + s2,1 N1
N2 Δt

where r1 = r−1, r2 = −1, s1,2 = −r, and s2,1 = 1/N1. Notice also that ri with a subscript is different from r without a subscript.

Figure 11.1.2 . Delayed logistic phase space with r =1.9, spiraling inward toward an equilibrium.
For small values of r, biomass and litter head to an equilibrium, as in the spiraling path of Figure 11.1.2. Here the system starts at
the plus sign, at time t = 0, with living biomass N1 =0.5 and litter biomass N2 =0.1. The next year, at time t =1, living biomass
increases to N1 =0.85 and litter to N2 = 0.5. The third year, t = 2, living biomass is inhibited slightly to N1 = 0.81 and litter builds up
to N2 = 0.85. Next, under a heavy litter layer, biomass drops sharply to N1 =0.22, and so forth about the cycle. The equilibrium is
called an “attractor” because populations are pulled into it.

Figure 11.1.3 . Delayed logistic phase space with r=2.03, spiraling toward a stable limit cycle, with the equilibrium point for r=1.9
shown as a dot.
For larger values of r, the equilibrium loses its stability and the two biomass values, new growth and old litter, permanently
oscillate around the state space, as in the spiraling path of Figure 11.1.3. The innermost path is an attractor called a “limit cycle.”
Populations starting outside of it spiral inward, and populations starting inside of it spiral outward— except for populations
balanced precariously exactly at the unstable equilibrium point itself.

Figure 11.1.4 . Delayed logistic phase space with a stable equilibrium and three stable limit cycles for r = 1.9 to r = 2.27, as
marked.

11.1.2 https://bio.libretexts.org/@go/page/25485
For still larger values of r, the system moves in and out of chaos in a way that itself seems chaotic. By r = 2.15 in Figure 11.1.4,
the limit cycle is becoming slightly misshapen in its lower left. By r = 2.27 it has become wholly so, and something very strange
has happened. A bulge has appeared between 0 and about 0.5 on the vertical axis, and that bulge has become entangled with the
entire limit cycle, folded back on itself over and over again. What happens is shown by magnifying Region 1, inside the red square.
Figure 11.1.5 shows the red square of Figure 11.1.4 magnified 50 diameters. The tilted U-shaped curve is the first entanglement of
the bulge, and the main part of the limit cycle is revealed to be not a curve, but two or perhaps more parallel curves. Successive
images of that bulge, progressively elongated in one direction and compressed in the other, show this limit cycle to be infinitely
complex. It is, in fact, not even a one-dimensional curve, but a “fractal,” this one being greater than one-dimensional but less than
two-dimensional!

Figure 11.1.4 magnified 50 diameters.

Figure 11.1.6 magnifies the red square of Figure 11.1.5, an additional 40 diameters, for a total of 2000 diameters. The upper line
looks single, but the lower fatter line from Figure 11.1.5 is resolved into two lines, or maybe more. In fact, every one of these lines,
magnified sufficiently, becomes multiple lines, revealing finer detail all the way to infinity! From place to place, pairs of lines fold
together in U-shapes, forming endlessly deeper images of the original bulge. In the mathematical literature, this strange kind of
attractor is, in fact, called a “strange attractor.”

Figure 11.1.5 magnified an additional 40 diameters.

Such strange population dynamics that occur in nature, with infinitely complex patterns, cannot arise in phase spaces of dynamical
systems for one or two species owing in continuous time, but can arise for three or more species in continuous time. And as
covered in Chapter 7, they can arise for even a single species approximated in discrete time.
What we have illustrated in this chapter is perhaps the simplest ecological system with a strange attractor that can be visualized in a
two-dimensional state space.

11.1.3 https://bio.libretexts.org/@go/page/25485
 Figure 11.1.7 . Landscape of the Mandelbrot set, which has been called the most complex object envisioned by the human mind. It
is the division between two different behaviors in a particular two-dimensional dynamical system—an infinitely complex object
that is less than 2-dimensional but greater than 1.9-dimensional. We display it here for your enjoyment—for wonder and for art.

This page titled 11.1: State Spaces is shared under a CC BY-NC 4.0 license and was authored, remixed, and/or curated by Clarence Lehman,
Shelby Loberg, & Adam Clark (University of Minnesota Libraries Publishing) via source content that was edited to the style and standards of the
LibreTexts platform; a detailed edit history is available upon request.

11.1.4 https://bio.libretexts.org/@go/page/25485
 CHAPTER OVERVIEW

12: Predator and Prey

Competition and mutualism can be understood without much attention to the sizes of the species involved. But predation is quite
different. Think of a producer, the prey, and a consumer, the predator. When the consumer is extremely large and the producer very
small, as with a whale and krill, the relationship is called filter feeding. There the predator kills the prey. When the producer and
consumer are roughly matched in size, they are called predators and prey. When the producer is much smaller than the consumer,
but still independently mobile, the relationship is called parasitism. And when the consumer is much smaller still, and has
difficulty moving around on its own, the relationship is called infection and disease. Symptoms of disease are part of the ecology
of disease. In parasitism and disease, the consumer often does not kill the producer.
12.1: Further Examples of Predation
12.2: Ecological communities are complex
12.3: Predator-prey model
12.4: Phase Space
12.5: Assumption of the basic model
12.6: Independent carrying capacities
12.7: Predator satiation and starvation
12.8: Effects of Space

This page titled 12: Predator and Prey is shared under a CC BY-NC 4.0 license and was authored, remixed, and/or curated by Clarence Lehman,
Shelby Loberg, & Adam Clark (University of Minnesota Libraries Publishing) via source content that was edited to the style and standards of the
LibreTexts platform; a detailed edit history is available upon request.

1
12.1: Further Examples of Predation
Amazing mechanisms for both capturing prey and avoiding predators have been discovered through evolution. Fulmar chicks, for
example, can direct “projectile vomit” at predators approaching the nest too closely (Figure 12.1.1). This is not just icky for the
predator. By damaging the waterproofing of an avian predator’s feathers, this ultimately can kill the predator. The chick’s projectile
vomit is thus a lethal weapon.

Figure 12.1.1 . Fulmar chick with projectile vomit as predator defense.

One of the most remarkable predatory weapons is that of pistol shrimp. These shrimp have one special claw adapted to cavitation,
and are capable of shooting bullets at their prey; colonies of these shrimp are loud from the sound of these bullets. But where does
an underwater crustacean get bullets? Actually, it creates them from nothing at all—from cavitation. If you’ve ever piloted a
powerful motorboat and pushed the throttle too hard, or watched a pilot do so, you’ve seen the propellers start kicking out bubbles,
which look like air bubbles. But the propellers are well below the water line, where there is no air. The propellers are in fact
creating bubbles of vacuum—separating the water so instantly that there is nothing left in between, except perhaps very low density
water vapor. Such bubbles collapse back with numerous blasts, each so powerful that it rips off pieces of bronze off the propeller
itself, leaving a rough surface that is the telltale sign of cavitation.

Figure 12.1.2. Pistol-packing shrimp shoot cavitation bullets.

A pistol crab snap its pistol claw together so quickly that it creates a vacuum where water used to be. With the right circulation of
water around the vacuum bubble, the bubble can move, and a crab can actually project its bullet of vacuum toward its prey. When
the bubble collapses, the effect is like thunder attending a lightning bolt, when air snaps together after the lightning has created a
column of near-vacuum. But the consequences are quite different in water. While a loud sound might hurt the ears of a terrestrial
animal, the sound does not rip apart the fabric of the animal’s body. This is, however, what intense sounds in water can do,
traveling through water and through the water-filled bodies of animals. In effect, pistol shrimp shoot bullets that explode near their
prey and numb them into immobility. Somehow evolution discovered and perfected this amazing mechanism!

Figure 12.1.3 . Rie-carrying humans shoot lead bullets—bison bones and remnant herd.

12.1.1 https://bio.libretexts.org/@go/page/25492
The ultimate weapons of predation, however, are those of our own species. Figure 12.1.3 (right) shows a remnant bison herd, a few
hundred of the hundreds of millions of bison that migrated the plains not many generations ago. No matter how vast their numbers,
they were no match for gunpowder and lead bullets, and they dropped to near extinction by the beginning of the twentieth century.
The image at the left illustrates the epic efficiency of lead bullets by showing a nineteenth-century pile of bison bones, with
members of the predator species positioned atop and aside.

This page titled 12.1: Further Examples of Predation is shared under a CC BY-NC 4.0 license and was authored, remixed, and/or curated by
Clarence Lehman, Shelby Loberg, & Adam Clark (University of Minnesota Libraries Publishing) via source content that was edited to the style
and standards of the LibreTexts platform; a detailed edit history is available upon request.

12.1.2 https://bio.libretexts.org/@go/page/25492
12.2: Ecological communities are complex
Before proceeding with simplified models of predation, we want to stress that ecological communities are complex (Figure 12.2.1).
Fortunately, progress in understanding them comes piece by piece. Complex food webs, like the one illustrated in the figure, can be
examined in simpler “motifs.”
You have seen in earlier chapters that there are two motifs for a single Species: logistic and orthologistic, with exponential growth
forming a fine dividing line between them. And in the prior chapter, you saw three motifs for two species: predation,
competition, and mutualism.
Later you will see that there are exactly forty distinct three-species motifs, one of which is two prey pursued by one predator. This
is called “apparent competition” because it has properties of competition.

Figure 12.2.1 . Cretaceous terrestrial food web, Mitchell et al. 2012 PNAS.

This page titled 12.2: Ecological communities are complex is shared under a CC BY-NC 4.0 license and was authored, remixed, and/or curated by
Clarence Lehman, Shelby Loberg, & Adam Clark (University of Minnesota Libraries Publishing) via source content that was edited to the style
and standards of the LibreTexts platform; a detailed edit history is available upon request.

12.2.1 https://bio.libretexts.org/@go/page/25493
12.3: Predator-prey model
For the next several chapters we will consider two species, starting with one predator and one prey. Figure 12.3.1 depicts this
situation, with one line sloping down and the other up.

Figure 12.3.1 . Predator–prey interactions with corresponding equations.

The graph on the left describes the prey, because its numbers N1 are reduced when the numbers of predator, N2, increase. Likewise,
the graph on the right describes the predator, because its numbers, N2, increase with the density of its prey, N1. The equations of
growth are revealed by the slopes and intercepts of the two lines.
Since these are both straight lines, y = mx + b , the equations can be written down simply from the geometry. The intercept on
the left is +1 and the slope is −1. The intercept on the right is −1/2 and its slope is +1/2. The equivalent equations of the two lines
appear below the graphs.
These specific equations can be generalized using symbols in place of actual numbers, writing r1, s1,2, r2, and s2,1 for the intercept
+1.0 and slope −1.0 on the left and the intercept −0.5 and slope +0.5 on the right, as follows.
1 dN1
= r1 + s1,2 N2 with r1 = + 1.0, s1,2 = − 1.0
N1 dt

1 dN2
= r2 + s2,1 N1 with r2 = − 0.5, s2,1 = + 0.5
N2 dt

Merely by writing down the form of these geometric graphs, the classic Lotka–Volterra predator–prey equations have appeared:
1 dN1
= r1 + s1,2 N2 with r1 > 0, s1,2 < 0
N1 dt

1 dN2
= r2 + s2,1 N1 with r2 < 0, s2,1 > 0
N2 dt

Here is how the equations look in many textbooks, with V for prey density and P for predator density:
dV
= rV −α V P
dt

dP
= β V P − qP
dt

Volterra arrived at the equation rather differently than we did, with a growth rate r for the prey, reduced by a rate α for each
encounter between predator and prey, V ⋅ P , and with a natural death rate q for predators and compensatory growth rate β for each
encounter,V ⋅ P , between predator and prey.
To see the equivalence, divide the first equation through by V and the second by P, then set
V = N1 , P = N2 , r = r1 , q = −r2 , α = −s1,2 , β = s2,1 . The Lotka–Volterra formulation will be revealed to be just
the r + sN equations in disguise.
Figure 12.3.1 exposes the basic predator–prey equations from geometry, which reveal the unity of the equations of ecology, as you
saw in Chapter 5. That analysis revealed a form of one-dimensional equation not considered in ecological textbooks—the
orthologistic equation—and which is needed for understanding human and other rapidly growing populations.
Now analyze these equations a bit. Suppose predator and prey densities are both 1, say 1 individual per hectare (N1 = N2 = 1).
Substitute 1 for both N1 and N2. What are the growth rates?
1 dN1
= 1.0 − 1.0 × 1 = 0
1 dt

1 dN2
= −0.5 + 0.5 × 1 = 0
1 dt

12.3.1 https://bio.libretexts.org/@go/page/25494
The population growth is zero for both species, so the populations do not change. This is an equilibrium.
dN1 dN2
This can be seen in the graphs below. The fact that both growth rates, 1

N1 dt
and 1

N1 dt
, cross the horizontal axis at N1 = N2 = 1
(position of the dots) means that growth stops for both. This is called an equilibrium, a steady state, or, sometimes, a fixed point.

But what will happen if both populations are 2, say 2 individuals per hectare?

dN1
The prey growth rate, 1

N1 dt
, is negative at N2 = 2 (the line is below the horizontal axis) and the predator growth rate, \(\frac{1}
{N_1}\frac{dN_2}{dt}, is positive at N1 = 2 (the line is above the horizontal axis). So the prey population will decrease and the
predator population will increase. Exactly how the populations will develop over time can be worked out by putting these
parameters into the program in Chapter 8. Here it what it shows.

Figure 12.5 as calculated by Program 95.

For comparison, here is what early experimenters such as Gause and Huffaker showed for populations of protozoa, mites, and other
small systems in the middle of the twentieth century:

Figure 12.3.3 . Dynamics in an experimental predator–prey system conducted by C. B. Huffaker in the 1950s with two species of
mite.
The dynamics here are much the same as those shown in the calculated version of Figure 12.3.2 and the experimental version of
Figure 12.3.3, but with stochasticity overlayed on the experimental system. Experimenters, however, had difficulty achieving
continual cycling. In simple conditions, the predators would find the prey and eat every last one, and then the predators themselves
would all die. Continual cycling could be achieved by providing the prey with places to escape, or making it difficult for the
predators to move around the environment.

12.3.2 https://bio.libretexts.org/@go/page/25494
This page titled 12.3: Predator-prey model is shared under a CC BY-NC 4.0 license and was authored, remixed, and/or curated by Clarence
Lehman, Shelby Loberg, & Adam Clark (University of Minnesota Libraries Publishing) via source content that was edited to the style and
standards of the LibreTexts platform; a detailed edit history is available upon request.

12.3.3 https://bio.libretexts.org/@go/page/25494
12.4: Phase Space
The cycling can be understood better in phase space, where the densities of the two species are represented as two-dimensional
points.

Figure 12.4.1 . Predator–prey phase space. The circle marks an equilibrium where growth of both predator and prey stops. The plus
sign marks a population value of N1 = 1.5, N2 = 0.5.
For example, as explained in Chapter 10, if the prey population is 1.5 and the predator population is 0.5, the population will be 1.5
units to the right on the horizontal axis and 0.5 units up on the vertical axis, at the location of the blue plus sign in the graph.
Where does the predator population cease to grow? In the equation
1 dN2
= r2 + s2,1 N1 (12.4.1)
N2 dt

−r2
it ceases to grow where 0 = r2 + s2,1 N1 , or N1 =
s2,1
. With r2 = −0.5 and s2,1 = 0.5, this is a vertical line—the predator
isocline—at N1 = 1, as in Figure 12.4.2.

Figure 12.4.2 . The predator isocline, at which the predator population ceases to grow.
To the left of the isocline, prey are sparse and predators decline, as indicated by the downward arrows. To the right of the isocline,
in contrast, prey are abundant and predators can increase, as indicated by the upward arrows.
Likewise, where does the prey population cease to grow? In the equation
1 dN1
= r1 + s1,2 N2 (12.4.2)
N1 dt

−r1
it ceases to grow where 0 = r 1 + s1,2 N2 , which means N2 =
s1,2
. With r = 1 and
1 s1,2 = −1, this is a horizontal line—the prey
isocline—at N2 = 1.

12.4.1 https://bio.libretexts.org/@go/page/25495
 Figure 12.4.3 . The prey isocline, at which the prey population ceases to grow.
Below the isocline, predators are sparse so prey can increase, as indicated by the arrows pointing right. Above the isocline, in
contrast, predators are abundant and prey decreases, as indicated by the arrows pointing left. Putting Figures 12.4.2 and 12.4.3
together gives Figure 12.4.4, which shows rotation in the combined arrows.

Figure 12.4.4 . Both isoclines, with both sets of arrows combined, showing the cycling.
Here the rotation can be deduced by thinking about the dynamics of predator and prey. The rotation is corroborated by using Table
10.2.1 to calculate the eigenvalues. The eigenvalues of the interior equilibrium turn out to be 0±0.707i, a number with both real and
imaginary parts. The existence of an imaginary part, ±0.707i, implies cycling. The real part, 0, means that eigenvalues alone cannot
determine the stability—it could be stable, unstable, or neutral. In fact, for this particular case with no self-limitation, deeper
mathematical examination shows that the stability is neutral. The dynamics will rotate indefinitely, maintaining whatever cycle it
started on.

Figure 12.3.2.
Taking all the data from Figure 12.3.2 and plotting N1 versus N2 gives Figure 12.4.5. The process starts at day 0 with N1 = N2 = 2.
One day later, prey have dropped to N1 ≈ 0.5 and predators have increased to N2 ≈ 2.2, marked by the red numeral 1 on the cycle.
(By the symbol ‘≈’, we mean “approximately equal to.”) Two days later, prey have dropped to N1 ≈ 0.2 and predators have dropped
to N2 ≈ 1.0, marked by the numeral 3. With predators at relatively low levels, prey then start to increase and, four days later, have
reached N1 ≈ 1.0, while predators have dropped further to N2 ≈ 0.3, marked by the numeral 7. Two days later, prey have increased
to N1 ≈ 3.0 and predators have increased to N2 ≈ 1.0, marked by the numeral 9. Finally, one day later the cycle begins to repeat, as
marked with the numeral 10. This is another way of showing the cycling of Figure 12.3.2.

12.4.2 https://bio.libretexts.org/@go/page/25495
 Figure 12.4.6 . The ow across the phase space.
In Figure 12.4.6—a ow diagram, the entire phase space can be filled with arrows to show how cycling proceeds everywhere. The
path of Figure 12.3.2, displayed in Figure 12.4.5, is overlayed in blue.

This page titled 12.4: Phase Space is shared under a CC BY-NC 4.0 license and was authored, remixed, and/or curated by Clarence Lehman,
Shelby Loberg, & Adam Clark (University of Minnesota Libraries Publishing) via source content that was edited to the style and standards of the
LibreTexts platform; a detailed edit history is available upon request.

12.4.3 https://bio.libretexts.org/@go/page/25495
12.5: Assumption of the basic model
This basic model of interacting predators and prey reveals the tension between population growth and decline, and shows the kind
of cycling that characterizes predator–prey systems. It has, however, a number of simplifying assumptions that can be relaxed in
more detailed studies, including the following:
1. The predator lives only on this prey, and perishes without it.
2. The prey has no carrying capacity of its own—only that imposed by the predator.
3. The environment is homogeneous and prey have no hiding places.
4. Growth is continuous, with no age structure, maturation periods, and so forth.
5. The number of prey taken is proportional to the number of prey present. In other words, predators are never satiated.
6. Genetics are uniform within each species, and there is no evolution.
Relaxing all of these assumptions is a book in itself, but we will relax some of them in sections ahead.

This page titled 12.5: Assumption of the basic model is shared under a CC BY-NC 4.0 license and was authored, remixed, and/or curated by
Clarence Lehman, Shelby Loberg, & Adam Clark (University of Minnesota Libraries Publishing) via source content that was edited to the style
and standards of the LibreTexts platform; a detailed edit history is available upon request.

12.5.1 https://bio.libretexts.org/@go/page/25496
12.6: Independent carrying capacities
Self-limitation or self-enhancement of population growth are within the r + sN framework. Below these terms are in red.
1 dN1
= r1 s1,2 N2 + s1,1 N1
N1 dt

1 dN2
= r2 s2,1 N1 + s2,2 N2
N2 dt

The self-feedback term for the prey, s , is typically negative, reflecting a carrying capacity for the prey in the absence of
1,1

predators, K = −r / s . This tends to stabilize the system, dampening oscillations and leading to a joint equilibrium of
1 1 1,1

predator and prey.

On the other hand, the self-feedback term for the predator, s , is typically zero, meaning the predators vanish in the absence of
2,2

prey. But it could be positive, indicating benefits from group hunting and the like. A positive value for s tends to destabilize the
2,2

system, leading to enlarging oscillations.

This page titled 12.6: Independent carrying capacities is shared under a CC BY-NC 4.0 license and was authored, remixed, and/or curated by
Clarence Lehman, Shelby Loberg, & Adam Clark (University of Minnesota Libraries Publishing) via source content that was edited to the style
and standards of the LibreTexts platform; a detailed edit history is available upon request.

12.6.1 https://bio.libretexts.org/@go/page/25497
12.7: Predator satiation and starvation

This page is a draft and is under active development.

Figure 12.7.1 . Predators satiated versus sick and starving.

In the original Lotka–Volterra formulation, doubling the number of prey in the environment doubles the number of prey taken. The
same is true in the equivalent r + sN formulation explored above. While this may be reasonable at low prey densities, eventually
the predators become satiated and stop hunting, as in the image at left in Figure 12.7.1. Satiation will therefor truncate the predator
growth curve at some maximum rate, as at right in Figure 12.7.2.

Figure 12.7.2 . Predator growth truncated at a maximum number of prey taken per predator per unit of time.
In the opposite direction, if prey are not available, the predator population starves, as in the sad image at right in Figure 12.7.1. As
indicated by a negative vertical intercept in Figures 12.7.2 and elsewhere, the population does not reach a maximum rate of
decline. In the complete absence of prey, vertebrate predators decline more and more rapidly, reaching extinction at a definite time
in the future, as induced by the increasingly large rates of decline shown in the right part of Figure 12.7.3. This is a different kind
of singularity that can actually occur in a finite time.

Figure 12.7.3 . Vertebrate predators with asymptotic rates of consumption.

As an exercise and illustration, let us create a predator– prey system in which the predators become satiated and reach a maximum
growth rate, but for which there is no maximum death rate in the absence of prey, and see where it leads.
For predators, we want to mimic the shape at right in Figure 12.7.3. This has the shape of a hyperbola, y = 1/x, but reflected
about the horizontal axis and shifted upwards. The equation would be y = a − b/x , where a and b are positive constants. When
x approaches infinity, the term b/x goes to zero and y therefore approaches a . It crosses the horizontal axis where x = b/a , then

heads downward toward minus infinity as x declines to zero. Such a curve has the right general properties.
The predator equation can therefore be the following, with r for a , s1 1,2 for −b , and N for x.
1

1 dN2 1
= r2 + s2,1
N2 dt N1

When there are ample prey N will be large, so the term s /N will be small and the predator growth rate will be near r . As
1 2,1 1 2

prey decline, the term s /N will grow larger and larger without limit and, since s is less than zero, the predator growth rate
2,1 1 2,1

will get more and more negative, also without limit.

12.7.1 https://bio.libretexts.org/@go/page/25782
What about the prey equation? The important point here is that predators become satiated, so the chance of an individual prey being
caught goes down as the number of prey in the environment goes up. So instead of a term like s N for the chance that an 1,2 2

individual prey will be taken, it would be more like s N /N . 1,2 2 1

1 dN1 N2
= r1 + s1,2 + s1,1 N1
N1 dt N1

In other words, the rate of prey being taken increases with the number of predators in the environment, but is diluted as there are
more and more prey and predators become satiated. Eventually, with an extremely large number of prey in the area relative to the
number of predators, the effect of predators on each individual prey becomes negligible. This creates the following predator–prey
system, which takes satiation and starvation into account:
1 dN1 N2
= r1 + s1,2 + s1,1 N1
N1 dt N1

1 dN2 1
= r2 + s2,1
N2 dt N1

This system could be criticized because it is not “mass balanced.” In other words, one unit of mass of prey does not turn directly
into a specific amount of mass of predators. But this is not a simple molecular system, and it at least fits more closely the realities
of predator and prey behavior.
In any case, keep in mind that s is less than 0, to reflect limitation of prey due to crowding and other effects; s is equal to 0,
1,1 2,2

assuming the predator is limited only by the abundance of prey; s is less than 0 because the abundance of predators decreases the
1,2

growth of prey; and s is also less than 0 because as the number of prey decreases there is an increasingly negative effect on the
2,1

growth of the predator.

The next step is to examine the isoclines for this new set of equations, making a phase-space graph with N on the horizontal axis 1

dN1
versus N on the vertical. Where does the prey growth,
2
1

N1 dt
, cease? Working through some algebra, it is as follows:
1 dN1 N2
= 0 = r1 − s1,2 − s1,1 N1
N1 dt N1

N2
⇒ s1,2 = r1 − s1,1 N1
N1

r1 s1,1
2
⇒ N2 = N1 − N
s1,2 s1,2 1

dN2
Similarly, where does the predator growth, 1

N2 dt
, cease for the same phase-plane? Let's follow similar algebra:
1 dN2 1
= 0 r2 + s2,1
N2 dt N1

⇒ −r2 = s2,1
1

N1
\
s2,1
⇒ N1 = −
r2

This predator isocline is simply a vertical line, as before in Figures 12.4.2 and following figures. But note that the prey curve has
the form of an inverted parabola—a hump, as graphed for two cases in Figure 12.7.4.
Remarkably, this formulation of predator–prey equations closely matches what earlier researchers deduced logically and
graphically, when computers were slow or not yet available. If you want to better understand the shape of the prey curve, read
Rosenzweig’s 1969 paper entitled “Why the prey curve has a hump.” For interest, his hand-drawn published figure with
experimental data points is reproduced in Figure 12.7.5.
Rosenzweig pointed out a paradoxical effect, which he called “the paradox of enrichment.” At left in Figure 12.7.5, the prey have a
relatively low carrying capacity, with K = −r / s about halfway along the horizontal axis. If you analyze the ow around the
1 1,1

red dot that marks the equilibrium point to the right of the hump, or run a program to simulate the equations we just derived, you
will find that the populations spiral inward. The equilibrium is stable.

12.7.2 https://bio.libretexts.org/@go/page/25782
 Figure 12.7.4 . The paradox of enrichment, Rosenzweig, AmNat 1969.
The paradox is this: if you try to improve conditions for the prey by increasing their carrying capacity—by artificially providing
additional food, for example— you can drive the equilibrium to the left of the hump, as in the right part of Figure 12.7.4. Around
the equilibrium marked by the red circle, the populations spiral outward. The system has become unstable.
This is a warning from ecological theory. In conservation efforts where predators are present, trying to enhance a prey population
by increasing its carrying capacity could have the opposite effect. This is not to say that efforts to enhance prey populations should
not be undertaken, only that they should proceed with appropriate caution and study.

Figure 12.7.5 . Rosenzweig’s original 1969 graph with phase-space arrows, interpreting Huffaker’s 1958 experiments with mites.

This page titled 12.7: Predator satiation and starvation is shared under a CC BY-NC 4.0 license and was authored, remixed, and/or curated by
Clarence Lehman, Shelby Loberg, & Adam Clark (University of Minnesota Libraries Publishing) via source content that was edited to the style
and standards of the LibreTexts platform; a detailed edit history is available upon request.

12.7.3 https://bio.libretexts.org/@go/page/25782
12.8: Effects of Space

This page is a draft and is under active development.

In predator–prey systems, especially in confined areas, the predator tends to capture all the prey and then starve, so the systems
“crash.” But over large areas it is conceivable that a predator can completely wipe out its prey in one area and not go extinct,
because it can simply move to another area where the prey still exist. Prey can then repopulate the area from which they had been
depleted.
Imagine a series of interconnected cells where, with some restrictions, predator and prey can migrate between adjacent cells. Now,
even though the system may be locally unstable and crash in individual cells, the entire system across all cells could be stable and
persist indefinitely.
In the 1930s the Russian ecologist Gause conducted a very famous set of early experiments on competition among protozoa, but he
also studied predation of Didinium on Paramecium. The populations he set up would commonly crash and go extinct, with the
Didinium eating all the Paramecia and then finding themselves without food. If he made places for the Paramecium prey to hide,
however, the systems could persist for many cycles.
In the 1960s Krebs noticed that populations of fenced mice, even those with a full half-acre within the fence, would crash and
disappear after grossly overgrazing their habitat. But in areas where they were allowed to disperse, the populations would persist.
Huffaker also ran extensive experiments, again in the 1960s, with mites and oranges. A single population of mites on a single
orange would crash and the whole population would disappear. Using multiple oranges with limited migration paths between them,
however, allowed the system to persist for many generations.
And in the 1970s Lukinbill did similar work with protozoa in aquatic tubs— larger and larger tubs holding miniature predator–prey
systems. He found that the larger the tub, the longer the system persisted.
The point to remember here is that the mere presence of spatial structure, in one form or another, can allow a let predator–prey
system to persist. The basic reason is simply that species can go extinct in some areas while continually recolonizing other areas,
always maintaining a population that blinks in and out locally, but persists globally.

This page titled 12.8: Effects of Space is shared under a CC BY-NC 4.0 license and was authored, remixed, and/or curated by Clarence Lehman,
Shelby Loberg, & Adam Clark (University of Minnesota Libraries Publishing) via source content that was edited to the style and standards of the
LibreTexts platform; a detailed edit history is available upon request.

12.8.1 https://bio.libretexts.org/@go/page/25824
 CHAPTER OVERVIEW

13: Humans as Predators

13.1: Chapter Introduction
13.2: Consciously Controlled Predation
13.3: Stochastic modeling
13.4: Present Situation

This page titled 13: Humans as Predators is shared under a CC BY-NC 4.0 license and was authored, remixed, and/or curated by Clarence
Lehman, Shelby Loberg, & Adam Clark (University of Minnesota Libraries Publishing) via source content that was edited to the style and
standards of the LibreTexts platform; a detailed edit history is available upon request.

1
13.1: Chapter Introduction
Early humans suffered from predators just as other primates suffer still. Eventually, though, they developed spears longer than the
longest teeth and became one of the top predators on land. Even if they did not eat sabre-tooth tigers, they were able to kill them.
We know from vivid paintings on rock walls in the protected shelter of caves that our ancestors at least fancied themselves as
hunters (Figure 13.1.1).

Figure 13.1.1 Hunter and hunted.

Humans are now the dominant large vertebrate on the planet. But on all the continents beyond Africa, large vertebrates were
prominent in the ecosystems before our ancestors arrived. In Africa many remain, coevolved with humans and perhaps wiser to our
ways. Elsewhere, however, they had little fear when our ancestors arrived (Figure 13.1.2).

Figure 13.1.2 Artist’s conception. The Moa had no clue that the diminutive primate was aiming a barb that would shortly pierce its
heart. Unafraid and then gone.
The reasons behind the extinction of so many megafauna are controversial. Archeologist Haynes and others believe humans are
responsible. Archeologist Grayson and others blame climate change, though animals had been through many glaciations and
deglaciations before. Mammalogist MacPhee and virologist Marx postulate a virulent “hyper- disease” brought by humans. And
geologist Kennett and colleagues assign a comet impact as the cause—an interesting theory, as some major human hunters
disappeared at about the same time as the larger mammals.

Figure 13.1.3 The remaining mass of Earth’s megafauna lives in the sea.
In any case, when something extreme happens, several causes may be working in concert. We do know from historical records that
only about one human lifetime ago our predecessors in North America hunted bison almost to extinction (Figure 13.1.3). Now the
large remaining populations of megafauna are in the seas. What is their fate?

This page titled 13.1: Chapter Introduction is shared under a CC BY-NC 4.0 license and was authored, remixed, and/or curated by Clarence
Lehman, Shelby Loberg, & Adam Clark (University of Minnesota Libraries Publishing) via source content that was edited to the style and
standards of the LibreTexts platform; a detailed edit history is available upon request.

13.1.1 https://bio.libretexts.org/@go/page/25499
13.2: Consciously Controlled Predation
Consider predation by humans that is not subject to the cyclical dynamics of natural predator–prey interactions, but is instead
consciously controlled to provide steady, reliable returns—from the world’s fisheries, for example. How is this attempted?
Recall logistic population growth, where the ecological term s is negative. Figure 13.2.1 shows the individual growth rate on the
vertical axis and the population density on the horizontal axis, as you have seen before. Individual growth means the same thing as
growth per capita, or relative growth, or percentage growth if multiplied by 100. The intention is to apply this to logistically
growing populations of prey.

Figure 13.2.1 : Logistic population growth, individual rate.

It is the growth rate of the whole population, however, that is of interest in controlled predation—not the per capita growth rate—
since it is a fraction of the whole population that is to be taken. So the vertical axis should show dN /dt rather than 1/N dN /dt.
Start with the individual growth rate, which reaches its maximum r as N approaches 0. Here the population produces very few
individuals because the population itself is almost nonexistent.
1 dN
= r + sN (13.2.1)
N dt

The goal is to maximize the population growth rate so that the greatest number of prey can be taken. To find that number, multiply
both sides of the equation above by the number of individuals, N , to get the growth rate of the entire population—in other words,
to determine how many individuals are added to the population in a unit of time. The result is
dN 2
= rN + sN (13.2.2)
dt

The growth of the entire population, dN /dt , has the shape of an inverted parabola, shown in Figure 13.2.2, since s is negative.
Population growth is lowest when the population is very small, near 0, or when it is high, near its carrying capacity,−r/s. It
reaches its maximum growth rate midway, at half the carrying capacity, (−r/s) / 2. So if the population is kept at half its carrying
capacity, it will be growing its fastest and the greatest amount can be “harvested” each year.

Figure 13.2.2 : Logistic population growth, population rate.

What is that maximum rate? To find it, substitute half the carrying capacity, −r/(2s), for N in the equation above, giving

13.2.1 https://bio.libretexts.org/@go/page/25500
 dN ∣ r r 2

∣ = r (− ) + s(− )
dt ∣ 2s 2s
max

2 2
r r
=− + s
2s 4s2

2
r
= −
4s

So in this theory the population grows most rapidly at rate −r /(4s), producing the greatest number of new individuals if drawn
2

down to half its carrying capacity. This has been called the “maximum sustainable yield.”

Figure 13.2.3 Fishing then and now.

Let us introduce a harvesting intensity, H . When H is zero, there is no harvesting, and when H is 1, harvesting is at the maximum
sustainable rate. In between it is proportional.
2
dN r
2
= (rN + sN ) + H (13.2.3)
dt 4s

dN

dt
⇒ The net rate of population growth: the number of individuals per time unit, considering births, deaths, and hunting
(rN + sN
2
) ⇒ The rate of addition: the number of individuals born per time unit minus those dying from causes other than
hunting
2

H
r

4s
⇒ The rate of removal: the number of individuals caught per time unit
If individual fishermen predominate (Figure 13.2.3, left), H will be small. This pulls the curve down, as in Figure 13.2.4, lowering
the carrying capacity slightly and leaving somewhat fewer fish in the sea. It also introduces an Allee point, though that point is far
below the equilibrium and therefore not a significant danger.

Figure 13.2.4: Population growth with light harvesting (H small).

But with increasingly focused and mechanized fishing (Figure 13.2.3, right), H approaches 1 and the curve is pulled farther down
(Figure 13.2.5). The carrying capacity is markedly reduced and the population produces new individuals at a large rate. And the
Allee point is pulled close to the carrying capacity, introducing a danger that unforeseen fluctuations in the population could push
the population below the Allee point and collapse the fishery.

13.2.2 https://bio.libretexts.org/@go/page/25500
 Figure 13.2.5: Population growth with heavy harvesting (H near 1).

Finally, with hunting or fishing at the maximum sustainable yield, the Allee point coincides with the carrying capacity and in effect
annihilates it (Figure 13.2.6). This introduces a dynamical conflict because there is a stable situation to the right but an unstable
one to the left, making it inevitable that the population will fall below the Allee point and collapse. The maximum yield is not
sustainable!

Figure 13.2.6. Population growth with maximal harvesting (H=1).

This page titled 13.2: Consciously Controlled Predation is shared under a CC BY-NC 4.0 license and was authored, remixed, and/or curated by
Clarence Lehman, Shelby Loberg, & Adam Clark (University of Minnesota Libraries Publishing) via source content that was edited to the style
and standards of the LibreTexts platform; a detailed edit history is available upon request.

13.2.3 https://bio.libretexts.org/@go/page/25500
13.3: Stochastic modeling
Program 13.3.1 simulates harvesting at the so-called maximal sustainable yield. It introduces small random fluctuations in the
population—so small that they cannot be discerned in a graph. The slight stochasticity makes the program take a different
trajectory each time it runs, with widely different time courses. Inevitably, however, the populations drift below the Allee point and
rapidly collapse, as in the sample run of the program shown in Figure 13.3.1.
In the age of sailing, at the arrow marked "A", fishing was high-effort but low-impact and fisheries stayed approximately at their
carrying capacity, K . “Optimal harvesting” was introduced once mathematical ecology combined with diesel technology, and
fisheries helped feed the growing human and domestic animal populations, with fish populations near “maximum sustainable
yield,” as expected. But throughout the 20th century, as shown on either side of the arrow marked "B", fish populations continued to
decline, and before 2015—at the arrow marked "C"—it becomes clear that something is seriously amiss.

# SIMULATE ONE YEAR

#
# This routine simulates a differential equation for optimal harvesting
# through one time unit, such as one year, taking very small time steps
# along the way.
#
# The ’runif’ function applies random noise to the population. Therefore it
# runs differently each time and the collapse can be rapid or delayed.
#
# ENTRY: ’N’ is the starting population for the species being simulated.
# ’H’ is the harvesting intensity, 0 to 1.
# ’K’ is the carrying capacity of the species in question.
# ’r’ is the intrinsic growth rate.
# ’dt’ is the duration of each small time step to be taken throughout
# the year or other time unit.
#
# EXIT: ’N’ is the estimated population at the end of the time unit.

SimulateOneYear = function(dt)
{ for(v in 1:(1/dt)) # Advance the time step.
{ dN = (r+s*N)*N - H*r^2/(4*s)*dt; # Compute the change.
N=N+dN; } # Update the population value.
if(N<=0) stop("Extinction"); # Make sure it is not extinct.
assign("N",N, envir=.GlobalEnv); } # Export the results.

r=1.75; s=-0.00175; N=1000; H=0; # Establish parameters.

for(t in 1850:2100) # Advance to the next year.

{ if(t>=1900) H=1; # Harvesting lightly until 1990.
print(c(t,N)); # Display intermediate results.
N = (runif(1)*2-1)*10 + N; # Apply stochasticity.
SimulateOneYear(1/(365*24)); } # Advance the year and repeat.

Program 13.3.1. This program simulates maximal harvesting with small fluctuations in the populations.

13.3.1 https://bio.libretexts.org/@go/page/25501
What happened? A collapse is part of the dynamics of this kind of harvesting. Inevitable stochasticity in harvest combines
unfavorably with an unstable equilibrium in the prey population. In some runs it collapses in 80 years, in others it may take 300.
The timing is not predictable; the main predictable property of the simulation is that ultimately the system will collapse.

Figure 13.3.1. One sample run of Program 13.4, showing the collapse typical of such runs.

This page titled 13.3: Stochastic modeling is shared under a CC BY-NC 4.0 license and was authored, remixed, and/or curated by Clarence
Lehman, Shelby Loberg, & Adam Clark (University of Minnesota Libraries Publishing) via source content that was edited to the style and
standards of the LibreTexts platform; a detailed edit history is available upon request.

13.3.2 https://bio.libretexts.org/@go/page/25501
13.4: Present Situation
Many of the world’s fisheries are under collapse, and in the oceans we seem to be on a path like the one our predecessor predators
took on land.
There are better ecological approaches—a “constant effort” approach, for example, rather than the “constant harvest” examined
here—but economic, social, and political pressures have kept them from extensive use.
We hope this chapter has shown you that insufficient examination of ecological equations applied on a large scale can generate
disasters, that equilibria must not be considered apart from their stability, and that management of real ecological systems requires
attention to natural history and social conditions.

Figure 13.4.1. Excruciatingly efficient modern methods of fishing.

This page titled 13.4: Present Situation is shared under a CC BY-NC 4.0 license and was authored, remixed, and/or curated by Clarence Lehman,
Shelby Loberg, & Adam Clark (University of Minnesota Libraries Publishing) via source content that was edited to the style and standards of the
LibreTexts platform; a detailed edit history is available upon request.

13.4.1 https://bio.libretexts.org/@go/page/25502
 CHAPTER OVERVIEW

14: Humans as Prey

14.1: Chapter Introduction
14.2: Portals
14.3: Pathogen Mobility
14.4: Disease more generally
14.5: The Strange Case of Polio
14.6: Modern disease trends
14.7: An ancient plague perhaps vanishing

This page titled 14: Humans as Prey is shared under a CC BY-NC 4.0 license and was authored, remixed, and/or curated by Clarence Lehman,
Shelby Loberg, & Adam Clark (University of Minnesota Libraries Publishing) via source content that was edited to the style and standards of the
LibreTexts platform; a detailed edit history is available upon request.

1
14.1: Chapter Introduction
Our species no longer suffers large predators as so many species do (Figure 13.0.1). But dangerous predators remain part of the
human condition. We call them diseases.
By analogy, think of visible organisms living in a pond— from birds, fish, frogs, insects, and plankton to aquatic plants (Figure
14.1.1, left). To be successful across a region, the organisms must gain resources in the pond while competing with other species,

survive predators, and enable their offspring to disperse to another pond. To be successful they must not destroy the pond in which
they live, at least until they or their offspring have dispersed to another pond.

Figure 14.1.1. Macroscopic organisms living in a pond are analogous to microscopic pathogens living in a body.

What is the body of a plant or animal to an infectious virus or bacterium? It is like a pond. There may be other pathogens in the
body competing for its metabolic resources. There are predators in the body in the form of an immune system. And to be successful
the pathogen must disperse to another body without destroying the body in which it lives, at least until it or its offspring has
orchestrated a way to disperse into and colonize another body.

This page titled 14.1: Chapter Introduction is shared under a CC BY-NC 4.0 license and was authored, remixed, and/or curated by Clarence
Lehman, Shelby Loberg, & Adam Clark (University of Minnesota Libraries Publishing) via source content that was edited to the style and
standards of the LibreTexts platform; a detailed edit history is available upon request.

14.1.1 https://bio.libretexts.org/@go/page/25511
14.2: Portals
A pond has entry and exit portals. If an organism is small enough, it can exit a pond on wind and spray. Some organisms can oat
downstream to another pond, while birds and insects can simply y. Larger organisms such as amphibians can hop or walk from
pond to pond to lay eggs. And a new exit portal has recently appeared in the form of boats and trailers that carry invasive weeds
and animals from one pond and deposit them in another.
Think analogously of the entry and exit portals of an animal. How can pathogens leave one body and enter another? While skin
covers at most two square meters in humans, mucus membranes of the respiratory, digestive, and reproductive tracts cover more
than 400 square meters. So, mucus membranes become good portals.
Successful diseases can exploit obligate behaviors. Animals, for example, must breathe continually, so exploiting the respiratory
pathway—being breathed out into the air by one animal and breathed in by another— is a reliable and ever-present method of
transmission.
How infections can leave and enter the body (mix and match):
Portals of exit:
Breath
Droplets (sneezing, coughing)
Saliva
Sweat
Tears
Feces
Urine
Seminal uids
Vaginal uids
Blisters, boils, zits (via scratching, breaking the skin)
Blood (mosquitoes, ticks, hemorrhage)
Portals of entry:
Blood (cuts, wounds, insects, needles)
Lungs
Nose
Eyes (conjunctivia)
Mouth (mucosa)
Mouth (tooth/gingivia junction)
Gut
Skin (scabes, triconosis, warts, etc.)
Urinary tract
Rectum
Vagina (mucosa)
Penis (mucosa)
Animals must also eat frequently and periodically, so exploiting the oral pathway—leaving through urine and feces and getting
back through the alimentary canal—is another reliable path, at least under conditions in the wild or with animals that live on and
eat grass (Figure 14.0.1, right).
Animals must reproduce, so exploiting the genital pathway—where many animals come into direct bodily contact— is a third
reliable path. This can be especially productive, as mucosal tissues of high surface area are touching and infected fluids can be
transferred from one sex to another. A pathogen that can get into seminal fluid of a mammal, for example, has a direct path for
transmission.
Finally, many animals care for young, so exploiting parental care can be a fourth reliable path. A pathogen that can enter a mother’s
milk has a direct path for infecting her offspring. Such transmission from parent to offspring is called “vertical transmission,” with
other forms called “horizontal transmission.”

14.2.1 https://bio.libretexts.org/@go/page/25506
This page titled 14.2: Portals is shared under a CC BY-NC 4.0 license and was authored, remixed, and/or curated by Clarence Lehman, Shelby
Loberg, & Adam Clark (University of Minnesota Libraries Publishing) via source content that was edited to the style and standards of the
LibreTexts platform; a detailed edit history is available upon request.

14.2.2 https://bio.libretexts.org/@go/page/25506
14.3: Pathogen Mobility
While microscopic pathogens are not independently mobile, their hosts are, and pathogens have evolved ingenious ways of
modifying the behavior of a host to enable their transferal to another host.

Figure 14.3.1 . Tiny predators getting from one host to another.

Sneezing and coughing reflexes, for instance, are ancient responses for clearing obstructions from the nose and throat, and some
pathogens deceptively induce those responses for a pathway out (Figure 14.3.1). What we call “symptoms of disease” are not, then,
random effects of a disease, but can often be a pathogen’s way of getting out of one pond, so to speak, and into another. Any of the
pathways listed in Chapter 14.1 can be exploited by a pathogen, which in doing so may upset these pathways and cause the host
great distress.

Figure 14.3.2. Pink Eye and Cold Sores.

Some pathogens, for example, get into your eyes and tears and deceptively cause itching and soreness (Figure 14.3.2, left),
inducing you to rub your eyes and transfer the pathogens to your fingers. This in turn can successfully move them to other locations
like food, from which they can enter another host by the oral pathway.
Other pathogens are able to break the skin and get out of the body on their own. Cold sores (Figure 14.3.2, right), a form of oral
herpes, form around the mouth and nose, transmitting to what touches the sore. A related genital herpes is transmitted sexually,
though evolution has been proceeding and the genital form is now able to infect orally and the oral form genitally. This and other
sexually transmitted diseases have been increasing since about the middle of the twentieth century.
One of the most successful pathogens to use the lungs and skin as pathways out of the body is smallpox (Figure 14.3.3). It leaves
its host with permanent scars, often over the entire body. Smallpox is an ancient disease, dating from the time before the
pyramids, that can kill a majority of those it infects and at times can infect a majority of the population.

Figure 14.3.3. Smallpox, the first disease consciously driven extinct from the natural world.

14.3.1 https://bio.libretexts.org/@go/page/25507
The very success and horror of smallpox was part of its ultimate destruction—its eradication was the first complete victory in the
conquest of disease. Thanks to prolonged diligent attention throughout the world, and of course to the invention of vaccine,
smallpox has been made extinct in the natural world. (We say “natural world” because laboratory samples are being retained.)
William Foege, a key player in orchestrating the extinction of smallpox, said that we can conquer disease because we evolve so
much more rapidly than the disease. This may be startling to hear, given that our physiological evolution is much slower than that
of viruses or bacteria. But, he explained, because we evolve socially much more rapidly than a disease can evolve biologically, we
are able to “outsmart” the disease.
Rinderpest, a viral disease causing high rates of mortality in cattle and wild mammals, was the second disease declared extinct in
the natural world. Others—such as polio and Guinea-worm disease— may soon follow, though the latter may simply be eradicated
from human populations by our continual isolation from sources of infection.
Diseases can exploit blood-sucking parasites to move directly from the blood stream of one host to that of another. Lyme disease
(Figure 14.3.4, left), for example, is spread by ticks, which puncture the skin to obtain a blood meal for themselves but in the
process can transfer pathogens. And Ebola (Figure 14.3.4, right) leaves by almost every exit portal listed, destroying those portals
by carrying not just the pathogen but chunks of lung, intestine, or skin in the process.

Figure 14.3.4. Lyme disease (left) and one form of Ebola (right).

Human populations are so large and dense that even relatively inefficient pathogens can be successful. And diseases, of course, also
affect wild and domestic animals as well as crops and other plants.
As the next chapter illustrates, many diseases can evolve to be relatively harmless to their hosts, promoting transmission and
allowing the disease to become widespread. Rust fungus infections, for example, are common in many plant species, but seldom
lead to the death of the host. Powder mildew on the prairie plant Monarda fistulosa is so widespread that it is used in plant
identification books as a way to identify the species (Figure 14.3.5, right).
Pathogens can dramatically alter animal behavior. Rabies first gets through the salivary glands and into the saliva of an infected
host. Physiological changes then make the host animal salivate profusely—foaming at the mouth—while psychological changes
make it appear crazy and angry. The animal then bites through the skin of another animals, transferring the pathogen to that
animal’s bloodstream, and the cycle continues. Both the physiological and psychological changes are caused by the
pathogen and allow it to spread, even after the death of the initial host.

Figure 14.3.5. Prarie Bush Clover and Monarda fistulosa.

“Mad dog behavior” is thus not an accidental consequence of the disease, but precisely the means the pathogen has developed for
getting, so to speak, from pond to pond.
It is useful to try to think of all the ways a pathogen might alter the behavior of its host to force the host to transfer the pathogen.
This is not just an intellectual exercise, but could help identify potential for new emerging diseases. For example, what should a
sexually transmitted disease do to its host in order to spread faster? It should render its host more active sexually! And indeed this
happens. Female chimpanzees would normally mate only every two years or so, after having given birth and nursed their young to
the point of weaning. But female chimps infected with SIV (simian immunodeficiency virus) reach estrus every month or so, and

14.3.2 https://bio.libretexts.org/@go/page/25507
do not conceive. The pathogen changes their mating behavior to spread itself more than an order of magnitude faster than it would
otherwise spread.
Inspired to think in this way, one student came up with a novel idea: Imagine a disease that can escape through the sweat glands
without harming its host. As behavioral modification, it makes infected hosts want to undergo strenuous exercise in groups, such as
in gymnasiums—thus explaining the entire modern exercise phenomenon as a disease! (Gerbils may also harbor this disease.)

This page titled 14.3: Pathogen Mobility is shared under a CC BY-NC 4.0 license and was authored, remixed, and/or curated by Clarence
Lehman, Shelby Loberg, & Adam Clark (University of Minnesota Libraries Publishing) via source content that was edited to the style and
standards of the LibreTexts platform; a detailed edit history is available upon request.

14.3.3 https://bio.libretexts.org/@go/page/25507
14.4: Disease more generally
The tiniest predators—infection and disease—differ from what are usually thought of as predators in a number of ways, some of
which you have seen. First, the disease organism is much smaller than its victim and not independently mobile. It must be carried
by wind or water, or induce its host to transfer it in one of numerous ways.
Second, disease does not necessarily kill its victims. Many diseases, in fact, leave their victims largely intact, the better to transmit
the pathogen to another host.
And third, after infection, the prey may become forever immune to future infections, both by that pathogen and related ones. This
immunity is created by the enormously elaborate “immune system” of vertebrates and other animals— a system recognizing and
killing incoming pathogens before they can incubate and do much harm, and as elaborate and complex as the brain and central
nervous system.
One of the great discoveries of the last millennium was that the immune system could be primed to recognize a pathogen before it
invaded, though what was happening inside the body was not understood until the twentieth century. Vaccination played a central
role in the eradication of smallpox. The twentieth century also saw the discovery of antibiotics such as penicillin, which allow
doctors to cure disease after an infection has progressed.
These discoveries show us that humans must be considered separately from plants and other animals, for we have developed
special powers against disease. We are not passive prey, and do not simply suffer a disease or make behavioral modifications to
avoid it. Instead we actively and globally strive to destroy disease, or subdue it. And we extend these efforts to diseases affecting
the animals and plants we depend on.
With respect to disease, plants have distinct properties that are in direct contrast with those of animals. Animals, in general, are
high-energy organisms— metabolizing rapidly, moving about, and perpetually pumping oxygen throughout the body. Plants are
nothing like this. Rather than hearts and rapid fluid ow to distribute food and oxygen and to cleanse waste materials, plants use the
passive effects of capillary action and evaporation. This requires the tiniest of veins, or capillary action will fail. And these veins
are too small to transport plant cells, or to allow larger pathogens like protozoa and many bacteria to gain access to the entire
organism. This also means that plants cannot have the same kind of immune system as animals, with their own cells traveling
through their tissues on patrol.
In addition, plants are typically modular. An infected part—leaf, flower, or whole limb—can be discarded and grown again. The
apical meristem cells at the tips of branches and roots are capable of developing entirely new plants. While cancer cells can spread
through the body of an animal and kill it, such cells would simply plug the veins of plants. So while animals get cancer, plants get
cankers. Plants have longevity, while animals have mortality—the cost of being a high-energy organism.

This page titled 14.4: Disease more generally is shared under a CC BY-NC 4.0 license and was authored, remixed, and/or curated by Clarence
Lehman, Shelby Loberg, & Adam Clark (University of Minnesota Libraries Publishing) via source content that was edited to the style and
standards of the LibreTexts platform; a detailed edit history is available upon request.

14.4.1 https://bio.libretexts.org/@go/page/25508
14.5: The Strange Case of Polio
Polio had long been a relatively rare disease of infants, called “infantile paralysis.” In the middle of the twentieth century, however,
it became more common and started affecting older children and adults. A new form of the disease seemed to be emerging.

Human Health and Hygiene

Which of these people, do you think, performed the greatest service to human health and hygiene in the twentieth
century, but inadvertently triggered this mid-century polio epidemic?
a. Louis Pasteur, discoverer of pasteurization
b. Alexander Fleming, discoverer of penicillin
c. Jonas Salk, creator of the polio vaccine
d. Franklin Delano Roosevelt, President of the United States and polio victim
e. Henry Ford, creator of the production line.

Answer
This seems a strange question, with industrialist Henry Ford under consideration. But indeed, the answer is
Henry Ford!

At the beginning of the twentieth century, most local transportation was by horse and powered, of course, almost entirely by the
biofuel hay. While it has now largely left social memory, in the early decades of the twentieth century the streets were a slurry of
gravel and horse manure. Flies were everywhere, and caused little concern paid, for this was the norm. People’s outhouses were
ventilated to the open air, and flies laid eggs there and in the streets, then freely entered houses and landed on food. A number of
diseases take advantage of the fecal–oral pathway, and polio is one of them.
But automobiles and tractors intervened. As the horse-drawn era closed, manure generally vanished, running water arrived and
flush toilets arrived, hygiene improved, sealed screen doors became common, and flies died in vast numbers. Without intending it,
Henry Ford became the greatest y killer of all time. The availability of the fecal–oral pathway diminished, and the infectivity of
related diseases fell.

Figure 14.5.1. Horses and estimated cars in the United States, the source of Henry Ford's public health legacy.

Figure P ageI ndex1 shows the horse population declining slowly until World War I, then falling rather steadily as the number of
cars increased in stages. The first increase in the number of cars ended around 1930 with the Great Depression, when many people
could not afford cars. The end of World War II in 1945 brought another boom in car purchases, and by 1975 society had replaced
almost every horse per capita with a car.
As the chance of catching polio fell, the average age of catching it increased. To understand this, consider residents of the northern
hemisphere living at various latitudes. Because residents of the High Arctic have a chance to see the northern lights—the aurora
borealis—every week, children living there will likely see the aurora before their first birthday. Farther south, at 50 degrees north
latitude, the aurora may appear only once every few years, especially near the lights of cities, so a child could be 5 or 10 years old
before ever seeing it. And finally, say at 35 degrees north latitude, the aurora may appear but once or twice in a lifetime, so many

14.5.1 https://bio.libretexts.org/@go/page/25509
people could be in middle age before viewing them, and others might go an entire lifetime without being touched by their hypnotic
display.
So it is with disease. The number of opportunities for catching a highly infectious disease, naturally, is high. If the quantity of
pathogens in the environment is such that all individuals encounter them on average once a year, only about one-third of infants
will avoid infection in their first year. (Actually the number is 1/e = 0.367..., if the chance of infection is completely random.) The
same fraction of the remaining infants will catch the disease during their first year, and the rest will be age two or older when they
catch the disease. Therefore, as the pathways for transmitting polio diminished during the twentieth century, the chances of
catching it in any year decreased and the age of onset correspondingly increased.
Polio is like some other diseases that are not usually virulent in infants and young children. A baby infected with polio might have
a cold and a runny nose, and the infection might go without particular notice. In an older child, however, it can stop bone growth
and muscle development, crippling the child. The polio epidemic of mid-century America was thus not a new disease emerging, but
an ancient disease dying out.
Albert Sabin, of polio vaccine fame, suspected a connection with flies. In 1941 he and his colleagues reported in Science on a study
they performed in areas of the United States where polio had struck. They captured flies, pureed them in sterile fluid, and gave
them to monkeys in feedings, nosedrops, or injections. As they put it, “Down came the monkeys with polio.”

Figure 14.5.2. The recent course of polio in the United States.

With further improvements in hygiene and broad use of vaccines, rates of polio have dropped to nearly zero. Figure 14.8 shows a
moderate number of cases of polio before the late 1940s, an outbreak lasting until the early 1960s, nearly nothing in the years
following.

Figure 14.5.3. Tillie and Oscar Lehman's adz-hewn log cabin with frame addition, circa 1930s—the end of the horse-drawn era.
They raised ten children in this cabin, all of whom likely had polio unnoticed as infants.

This page titled 14.5: The Strange Case of Polio is shared under a CC BY-NC 4.0 license and was authored, remixed, and/or curated by Clarence
Lehman, Shelby Loberg, & Adam Clark (University of Minnesota Libraries Publishing) via source content that was edited to the style and
standards of the LibreTexts platform; a detailed edit history is available upon request.

14.5.2 https://bio.libretexts.org/@go/page/25509
14.6: Modern disease trends

Figure 14.6.1. Paths of four diseases. Vertical axis shows cases in thousands.

Many ordinary diseases have been subdued since the last half of the twentieth century, some to the point of extinction from the
natural world. Smallpox and rinderpest are gone, and polio nearly so—as we are writing this (2016–17), polio workers are
anticipating its extinction in the foreseeable future. Diptheria is on a similar path (Figure 14.6.1), with no cases at all in the United
States during the twentieth century.
Diseases such as whooping cough and measles (Figure 14.6.1) have been subdued but remain with us, with some cycling through
periodic outbreaks. The rates of many ordinary diseases are being reduced, and infectious disease is no longer the major cause of
deaths in human populations.
Rates of various sexually transmitted diseases, however, are on a different course (Figure 14.6.2). Gonorrhea rates have declined
but remain considerably above zero, and it is a commonly reported disease in the United States. The appearance of syphilis appears
to be cyclic, as rates had declined but are now rising again. Rates of chlamydia—which can lead to serious outcomes, including
infertility in women— have been increasing steadily, without an end in sight, and rates of genital herpes and other sexually
transmitted diseases are rising similarly. Sexually transmitted diseases are a prominent problem to be solved in the twentieth
century.

Figure 14.6.2. Sexually transmitted diseases in the U.S. The vertical axis shows the rate per 100,000 population.

This page titled 14.6: Modern disease trends is shared under a CC BY-NC 4.0 license and was authored, remixed, and/or curated by Clarence
Lehman, Shelby Loberg, & Adam Clark (University of Minnesota Libraries Publishing) via source content that was edited to the style and
standards of the LibreTexts platform; a detailed edit history is available upon request.

14.6.1 https://bio.libretexts.org/@go/page/25510
14.7: An ancient plague perhaps vanishing

This page is a draft and is under active development.

We close this chapter with a graph to ponder. Examine Figure 14.7.1, an epidemiological view of annual deaths per 1000
population during the twentieth century, from a widespread and ancient cause. Imagine what it represents. Is it a sexually
transmitted disease, an ordinary disease with an effective treatment introduced around 1946, or something else entirely?

Figure 14.7.1. Worldwide deaths per year from an ancient human malady.

The solid blue regression line goes through the average number of deaths over time from 1946 forward, projected back on the
dashed line through the outbreaks earlier in the century.
Figure 14.7.1 is actually an epidemiological view of deaths from warfare, which have been declining per capita over the past 70
years. The vertical line marks the beginning of the atomic era, and the numbers in parenthesis indicate (1) World War I, (2) World
War II, (3) Korea, (4) Vietnam, (5) Cambodia and Ethiopia, (6) USSR–Afganistan and Iran–Iraq, and (7) Rwanda.
Why include a chart of war deaths in a book on ecology, and in a chapter on the ecology of disease? First, war is directly connected
with ecology and the environment. Throughout human history, warfare has been caused by environmental change, as existing
territories became unproductive and new territories were sought, and in turn it has caused environmental change through habitat
alteration and other forces. Second, humans are a dominant ecological force, whose impact we examine in this book, and warfare
has been a prominent theme in the human condition. And third, warfare has some of the properties of a disease. It can spread from
places of origin like a disease, and has analogs of competitors and mutualists in addition to obvious roles of predators and prey.
Moreover, it involves an infectious agent—replicating not as biological agents spread between bodies of their hosts, but abstractly,
like ideas—as seen by Richard Dawkins,⋆ — replicate as memes between minds of their hosts.
Warfare has enough abstract similarities to biological agents, and enough tangible effects on ecology of the planet, that we want to
offer these ideas for your future consideration, and with the hope that some progress can be made by enough minds examining
them.
From discussions with students and colleagues thus far, and from parts of the literature, here are some thoughts for your
consideration.
Nuclear weapons. It seems indisputable that these caused an initial collapse in warfare, but they may also have affected the number
of war deaths during the rest of the century. Of course, they could have led to unprecedented numbers of deaths had political
arrangements worked out differently.
Immediate journalism. Photographic news coverage becoming ever more immediate gave the world a different view of war. Cell-
phone cameras, social networks, and the internet expand that indefinitely today.
International law. Most international law may not yet be written, but we have seen its beginnings. How much has the encoding of
war crimes since World War II contributed to the decline?
Self-government. The rapid expansion of self-government since the middle of the twentieth century may have contributed to the
decline in war-related deaths, as self-governing nations tend to avoid war with other self-governing nations.
International trade. In the same way, nations that trade mutualistically may also tend to avoid war so as to avoid destroying trading
partnerships.

14.7.1 https://bio.libretexts.org/@go/page/25840
Expanding ethics. At the end of 1957 the Soviets launched the space dog “Little Curly” into orbit, intending him to die while
orbiting our planet. Though this went largely unchallenged in the twentieth century, would any nation be able to do something like
this in the twenty-first? Do expanding ethics in other realms contribute at all to the decline in war deaths?
Women in power. It is worth considering whether the increasing proportion of women in government has an effect on the number of
war-related deaths. Among primates such as chimpanzees and baboons, males are the more aggressive sex. If this is true in humans,
might it have continued effects in the future?
Improved medicine. Serious wounds once meant infection and death, but now victims can recover. And mortality from diseases
which can spread rapidly in wartime—like inuenza—has been reduced. The same level of warfare now manifests fewer deaths,
making part of the decline an artifact.
Reduced overkill. The percentage of a population killed during a war decreased from nearly 100% in some ancient times to “what is
necessary” in more recent cases. Has the development of precision weaponry contributed to continuing decrease in war deaths?
This would also make part of the decline an artifact.
Humanity has already unexpectedly broken the millennia-long rush of ever-accelerating population levels. Could something similar
be happening with the millennia-long scourge of war? According to projections along the regression line, and for whatever the
reasons, if the trends of Figure 14.7.1 are real and can be understood and continued, humanity may be on a path toward the
elimination of background warfare, even before the end of this century.
This material is a partial encapsulation of Steven Pinker’s 800-page book, “The Better Angels of our Nature” (2011). That
downward slope indicates a plausible goal to understand and a plausible hope to maintain. It is plausible, but we cannot know if it
is practical without dedicating ourselves to it.
There is a level of self-fulfillment in such things, for if we collectively do not believe a goal like this can be achieved, it likely will
not, but if we believe in it and work toward it, we might succeed. Along present trends, you can work with reasonable, rational,
data-based hope to make background warfare vanish in your lifetime. And regardless of the outcome, all will be ennobled by the
effort.

Figure 14.7.2. Arlington National Cemetery, Virginia.

This page titled 14.7: An ancient plague perhaps vanishing is shared under a CC BY-NC 4.0 license and was authored, remixed, and/or curated by
Clarence Lehman, Shelby Loberg, & Adam Clark (University of Minnesota Libraries Publishing) via source content that was edited to the style
and standards of the LibreTexts platform; a detailed edit history is available upon request.

14.7.2 https://bio.libretexts.org/@go/page/25840
 CHAPTER OVERVIEW

15: Theory of Disease

15.1: New terminology
15.2: The SIR flowchart
15.3: The SIR Equations
15.4: The SI Equations
15.5: The I Equation
15.6: Analysis of the I model
15.7: 15.7 The rsN model again
15.8: Application to an actual outbreak
15.9: Conflicting ethics

This page titled 15: Theory of Disease is shared under a CC BY-NC 4.0 license and was authored, remixed, and/or curated by Clarence Lehman,
Shelby Loberg, & Adam Clark (University of Minnesota Libraries Publishing) via source content that was edited to the style and standards of the
LibreTexts platform; a detailed edit history is available upon request.

1
15.1: New terminology

This page is a draft and is under active development.

Terminology of the theory of disease is not completely consistent in the epidemiological literature, but we will use it consistently as
follows.
Virulence: How much or how quickly a pathogen harms its host. Often symbolized as alpha, α , in disease equations. Example: if
one-tenth of infected organisms die in a particular time period, and everything is random, α = 1/10.
Infectivity: How readily a pathogen arrives at and invades a new host. Often symbolized as beta, β, in disease equations. Example:
in an otherwise uninfected population, if each infected host is expected to infect three others in a particular time period, β = 3.
Basic reproductive number: In an otherwise uninfected population, how many new infections an infected individual is expected to
produce during the duration of the infection. Often symbolized as R in disease equations, and pronounced “are not.” This is a
0

crucial number; if R is greater than 1, the disease will spread through the population, while if R is less than 1, the disease will
0 0

die out.

This page titled 15.1: New terminology is shared under a CC BY-NC 4.0 license and was authored, remixed, and/or curated by Clarence Lehman,
Shelby Loberg, & Adam Clark (University of Minnesota Libraries Publishing) via source content that was edited to the style and standards of the
LibreTexts platform; a detailed edit history is available upon request.

15.1.1 https://bio.libretexts.org/@go/page/25841
15.2: The SIR flowchart
A standard starting point for examining the theory of disease is the “SIR model” (Figure 15.2.1). In this model individuals are born
“susceptible,” into the box marked S at the left. They may remain there all their lives, leaving the box only upon their ultimate
death—marked by the red arrow pointing downward from the box. The label δ s on this arrow represents the rate of ow from the
box—the rate of death of individuals who have never had the disease. The model assumes a per capita death rate of δ deaths per
individual per time unit. If δ = 1/50, then one-fiftieth of the population will die each year. Multiplying by the number of individuals
in the box, S , gives the ow out of the box, δ S individuals per year.

Figure 15.2.1. Flow through an SIR sysytem, a prototypical model of epidemiology.

The only other way out of the S box is along the red arrow pointing right, indicating susceptible individuals who become infected
and move from the left box to the middle box. (The blue arrow pointing up indicates new individuals created by births, not existing
individuals moving to a different box.) This rate of ow to the right is more complicated, depending not just on the number of
susceptible individuals in the left-hand box but on the number of infected individuals I in the middle box. In the label on the right-
pointing arrow out of the S box is the infectivity coefficient β, the number of susceptible individuals converted by each infected
individual per time unit if all individuals in the whole population are susceptible. This is multiplied by the number of individuals
who can do the infecting, I , then by the probability that an “infection propagule” will reach and infect someone who is susceptible,
S / (S + I + R) . This is just the ratio of the number in the S box to the number in all boxes combined, and in effect “discounts”

the maximum rate β. The entire term indicates the number of individuals per time unit leaving the S box at left and entering the I
box in the middle.
All other ows in Figure 15.2.1 are similar. The virulence symbolized with α is the rate of death from infected individuals—those
in the I box. This results in α I deaths per year among infected individuals, transferring from the blue I box to the gray box below
it. Note that if infected individuals can also die from other causes, the actual virulence might be more like α − δ , though the
situation is complicated by details of the disease. If a disease renders its victims bedridden, for example, their death rate from other
causes such as accidents, such as being hit by a train, may be reduced. Such refinements can be addressed in detailed models of
specific diseases, but are best not considered in an introductory model like this.
The other way out of the blue I box in the middle of Figure P ageI ndex1 is by recovery, along the red arrow leading to the blue R
box on the right. In this introductory model, recovered individuals are permanently immune to the disease, so the only exit from the
R box is by death—the downward red arrow—with δ R recovered individuals dying per year. Note that recovered individuals are

assumed to be completely recovered, and not suffering any greater rate of death than susceptible individuals in the S box (both
have the same death rate δ .) Again, refinements on this assumption can be addressed in more detailed models of specific diseases.
The blue arrows represent offspring born and surviving, not individuals leaving one box for another. In this introductory model, all
individuals have the same birth rate b, so that being infected or recovering does not affect the rate. The total number of offspring
born and surviving is therefore b (S + I + R) . This is the final red arrow in Figure 15.2.1, placing newborns immediately into
the box of susceptible individuals.

This page titled 15.2: The SIR flowchart is shared under a CC BY-NC 4.0 license and was authored, remixed, and/or curated by Clarence
Lehman, Shelby Loberg, & Adam Clark (University of Minnesota Libraries Publishing) via source content that was edited to the style and
standards of the LibreTexts platform; a detailed edit history is available upon request.

15.2.1 https://bio.libretexts.org/@go/page/25513
15.3: The SIR Equations
The entire process depicted in Figure 15.2.1 is equivalent to the following set of equations.
dS S
= b(S + I + R) − β I −δ S
dt S+I+R

dI S
=β I −γ I −α I
dt S+I+R

dR
= γ I −δ R
dt

At the left in each equation is the net rate of change of each box, accounting for all arrows transferring individuals out of one box
and into another. Again, S is the density of susceptible individuals, I the density of infected individuals, and R the density of
recovered individuals. Note that the terms are balanced— the term γ I , for example, representing individuals entering the recovered
box in the last equation, is balanced by the complementary term −γ I , leaving the infected box in the middle equation.
The SIR model is another “macroscale model.” With recent changes in computation, “microscale models,” processing tens or
hundreds of millions of individual hosts, are becoming more widely used. They can reliably take you beyond what purely
mathematical formulations can do. More about them in later chapters.

This page titled 15.3: The SIR Equations is shared under a CC BY-NC 4.0 license and was authored, remixed, and/or curated by Clarence
Lehman, Shelby Loberg, & Adam Clark (University of Minnesota Libraries Publishing) via source content that was edited to the style and
standards of the LibreTexts platform; a detailed edit history is available upon request.

15.3.1 https://bio.libretexts.org/@go/page/25514
15.4: The SI Equations
SIR is also a “frequency-dependent” model, at one end of a spectrum which has “density-dependent” models at the other end.
Frequency dependence approximates situations in which infection propagules are limited, while density dependence approximates
situations in which potential victims are limited.
But do not be concerned with the full SIR model for now. We shall simplify it here to reveal its basic properties. First, suppose
there is no recovery— this is an incurable disease that, once contracted, stays with its victim forever. Many viral diseases
approximate this situation—herpes and HIV, for example. In gray below are all the terms that will drop out if there is no recovery.
dS S
= b(S + I +R ) −β I −δ S (15.4.1)
dt S + I +R

dI S
=β I −γ I −α I (15.4.2)
dt S + I +R

dR
= γ I −δ R (15.4.3)
dt

Removing those terms gives an “SI” model.

dS S
= b(S + I ) − β I −δ S (15.4.4)
dt S +I

dI S
=β I −α I (15.4.5)
dt S +I

But we won’t be concerned with this model just now.

This page titled 15.4: The SI Equations is shared under a CC BY-NC 4.0 license and was authored, remixed, and/or curated by Clarence Lehman,
Shelby Loberg, & Adam Clark (University of Minnesota Libraries Publishing) via source content that was edited to the style and standards of the
LibreTexts platform; a detailed edit history is available upon request.

15.4.1 https://bio.libretexts.org/@go/page/25515
15.5: The I Equation
We shall simplify the SI model further to an “I” model, the most basic epidemiological formulation. This can be done by
considering a constant population, with births always matching deaths. To accomplish this, the birth term, b(S + I ) , can be made
equal to the death term, δ (S + I ) , giving
dS S
= δ (S + I ) − β I −δ S
dt S +I

dI S
=β I −α I
dt S +I

However, since the total population—call it N —is constant, S is not needed in the equations at all. It is always equal to the total
population N minus the number infected: S = N − I . You can forget about the S equation and substitute S = N − I into the
I equation. This gives

dI N −1
=β I −α I (15.5.1)
dt (N − 1) + I

dI N −I
=β I −α I (15.5.2)
dt N

dI I
= β I (1 − ) −α I (15.5.3)
dt N

1 dI I
= β (1 − ) −α (15.5.4)
I dt N

This is the beginning of the I model. It needs a little more work before proceeding with its analysis, but first some additional
terminology:
Infection: A term applied as soon as a pathogen has taken hold in a host.
Disease: A term often applied when an infection starts producing symptoms in a host.
Incidence: The number of new infections or cases of disease appearing in a population per time unit. Often expressed as a
fraction of the total population.
Prevalence: The total number of infections or cases of disease existing in a population. Often expressed as a fraction of the total
population.
With this terminology in mind, let us put the I model into the form of prevalence, as a fraction of the entire population. In this case,
prevalence is just p = I/N. Start where we left off, with
1 dI I
= β (1 − ) −α (15.5.5)
I dt N

multiply through by I ,
dI I
= β I (1 − ) −α I (15.5.6)
dt N

divide through by N ,
I
d I
N
= βI (1 − ) − αI (15.5.7)
dt N

substitute prevalence
dp
= β p(1 − p) − α p (15.5.8)
dt

and, finally, divide through by p,

1 dp
= β p(1 − p) − α (15.5.9)
p dt

15.5.1 https://bio.libretexts.org/@go/page/25516
Now think about the (1 − p) term. This “one minus the prevalence” represents the fraction of the population that is susceptible to
the disease. But some fraction of the population may have natural immunity to the disease, and another may have been successfully
vaccinated against the disease. Let us call this fraction v and subtract it too from the fraction that is susceptible, finally giving
1 dp
= β(1 − v − p) − α (15.5.10)
p dt

Whew, that is the final I model—the starting place for analysis!

You will see in an upcoming chapter that Equation ??? is identical to a model of habitat destruction in which plants are isomorphic
to “infections” of the landscape, and that habitat destruction, which “protects” the landscape from infection by plants, is isomorphic
to vaccination.

Here is the equation with term-by-term explanations:

1 dp
= β(1 − v − p) − α (15.5.11)
p dt

dp

dt
⇒ Relative growth in prevalence under conditions prevailing at time t
β ⇒ Number of new infections induced by each infected individual in each time unit in a completely uninfected population.
v⇒ Probability of encountering an individual who cannot be infected.
p ⇒ Probability of encountering an infected individual.
α ⇒ Fraction of infected individuals lost per time unit.
Keep in mind that this is a constant-population approximation. Whenever an individual dies of the disease, a new susceptible
individual enters the population.

This page titled 15.5: The I Equation is shared under a CC BY-NC 4.0 license and was authored, remixed, and/or curated by Clarence Lehman,
Shelby Loberg, & Adam Clark (University of Minnesota Libraries Publishing) via source content that was edited to the style and standards of the
LibreTexts platform; a detailed edit history is available upon request.

15.5.2 https://bio.libretexts.org/@go/page/25516
15.6: Analysis of the I model
With v as the prevalence of vaccination in a population, what will the I model reveal about vaccination? In an uninfected
population the prevalence of infection will be zero (p = 0) , so the number of new infections produced per infected individual per
time unit will be β (1 − v) .
On average, a fraction α will die per time unit, so the average duration of infection will be 1/α , assuming complete randomness. If
1/10 die per year, for example, the average duration of infection will be 10 years.
This makes R (v) = β (1 − v) × (1/α) = (β/α) (1 − v) . And the disease will decline to extinction if R
0 0 (v) < 1 —that is,
if R (v) = (β/α) (1 − v) < 1 , which you can work out algebraically in a few steps to v > 1 − α/β .
0

Look what this means. A disease that infects 4 individuals per year in a totally uninfected
population
β = 4), andwhichremainsinf ectiousf oroneyear\(1/α = 1), willdeclinetoextinctionif \(v > 1 − 1/4 = 3/4 . If only
slightly more than 3/4 of the population is vaccinated, that disease will eventually vanish. Remarkably, a disease can be eradicated
even if the whole population cannot be vaccinated! Largely because of this, society can develop programs striving toward the
conquest of disease.
What does the I model reveal about the evolution of infectious disease? The pathogen has many more generations and can
therefore evolve biologically more rapidly than the host, and α and β can evolve to benefit the pathogen.
Because β enters the equation with a plus sign and α enters with a minus sign, the disease will spread more rapidly—
(1/p) (dp/dt) will be larger—if β increases and alpha decreases.
This means that, genetics permitting, a successful disease operating according to this or any similar model will tend to become
more infectious (higher β) and less virulent (lower alpha) over time. At the limit, nearly everybody will be infected but the effect
on anybody will be minimal. Polio in humans before the 20th century, and SIV in monkeys, are examples.
In the ultimate limit, a disease could evolve to negative virulence—that is, to be a mutualism with the host. Rhizobial bacterial in
legumes may be an example.
As usual, there are refinements to this idea, in part because infectivity and virulence are not independent. Diseases that evolve to be
more infectious may have to use more of their victims’ metabolic resources, and consequently may become more virulent in the
process. Again, such refinements can be addressed in more specific models.

This page titled 15.6: Analysis of the I model is shared under a CC BY-NC 4.0 license and was authored, remixed, and/or curated by Clarence
Lehman, Shelby Loberg, & Adam Clark (University of Minnesota Libraries Publishing) via source content that was edited to the style and
standards of the LibreTexts platform; a detailed edit history is available upon request.

15.6.1 https://bio.libretexts.org/@go/page/25517
15.7: 15.7 The rsN model again
Let’s return to shorter time frames without evolution. Notice that prevalence p changes with time as the epidemic spreads, but β, α ,
and v remain constant. So starting with
1 dp
= β(1 − v − p) − α (15.7.1)
p dt

rearranging terms,
1 dp
= (β(1 − v) − α) − β p (15.7.2)
p dt

and substituting r = β(1 − v) − α, s = −β , and N = p ,

1 dN
= r + sN (15.7.3)
N dt

Voilà, the epidemiological I model is revealed to be just the standard model of ecology in another disguise! But now, with
mechanisms included (infectivity, virulence, vaccination), deeper conclusions can be reached.
In getting to this standard model, for example, we set s equal to −β. Because β is positive, the density dependence term s is
negative, Negative s implies logistic growth (positive s is orthologistic), meaning that N will reach a carrying capacity—an
equilibrium, a steady state. But to arrive at the standard model, we set N equal to p, so since N reaches an equilibrium in the
standard model, the prevalence p will reach an equilibrium in the I model.
Without further analysis you can therefore conclude that a disease will not necessarily infect an entire population, but that its
prevalence will level out when it reaches a carrying capacity, at the equivalent of −r/s. Substitute backwards
(r = β (1 − v) − α and s = −β ) and you will find the carrying capacity of the disease:
α
^ =1−
p −v (15.7.4)
b

The small hat atop the p is just a reminder that this is not the variable prevalence p, but rather the fixed value of the equilibrium
prevalence, p^ .
Think about the approach in 15.6, which used R . Here is another way to get the result. Since v , the proportion of the population
0

vaccinated, appears in the equation with a minus sign, that means that the greater the proportion vaccinated, the lower the
equilibrium prevalence p^ . In fact, setting p^ = 0 and solving for v , when v = 1 − α /β , the prevalence p will be zero and the
disease will be eradicated. (Actually, for a margin of error, when v ≥ 1 − α /β .

This page titled 15.7: 15.7 The rsN model again is shared under a CC BY-NC 4.0 license and was authored, remixed, and/or curated by Clarence
Lehman, Shelby Loberg, & Adam Clark (University of Minnesota Libraries Publishing) via source content that was edited to the style and
standards of the LibreTexts platform; a detailed edit history is available upon request.

15.7.1 https://bio.libretexts.org/@go/page/25518
15.8: Application to an actual outbreak

This page is a draft and is under active development.

An ominous outbreak of Ebola in West Africa became widely known in 2014, with the rate of deaths repeatedly doubling and
redoubling. By the fall of that year cases started appearing on other continents.
Ebola apparently enters human populations from wild animals such as bats, in whom it is not particularly virulent. It has not,
however, adapted to the human body. It becomes such a dread disease in humans because it exploits almost all of the body’s exit
portals— not simply by finding limited passages through them, but by completely destroying them. Some diseases may induce
vomiting or diarrhea, for example, as a means of for the pathogen to exit the alimentary canal, but with Ebola, entire tissue systems
are destroyed and chunks of intestine accompany the exit.
At the time of the outbreak, two of us (CL, SW) were jointly teaching courses in the United States in quantitative ecology and in
the ecology of disease, using tools illustrated thus far in this book. A level of fear prevailed in the country because the disease had
just reached the U.S., with a few deaths in U.S. hospitals. Along with our students, we decided to make Ebola a case study for
application of the disease equations, applying the principles week-by-week as the outbreak was advancing. Hundreds of thousands
of deaths had been predicted by health organizations. This section describes what we did and what we discovered.
Real-time data. Data from the World Health Organization (WHO) and other official sources had been tabulated on a website about
Ebola in West Africa, and you can select a date there to see exactly which data were available when we started tracking the
outbreak, or at any subsequent time. The site listed both the number of individuals infected with Ebola and the number who had
died, but in the early days of the outbreak we guessed that the number of deaths would be more reliable. Moreover, the world was
paying greatest attention to deaths, so we hypothesized that these numbers would most strongly influence social efforts to combat
the disease.

Figure 15.8.1 . Ebola deaths reported at the time Ebola began to appear outside of Africa.

We thus started with total deaths, plotting them as in Figure 15.8.1. The number of deaths was not only increasing but accelerating,
seemingly on an exponential course. The calculated doubling time was 31.5 days, not quite what U.S. Centers for Disease Control
(CDC) had found earlier, but within reasonable correspondence. They estimated 15–20 days to double in one country and 30–40 in
another (Meltzer et al., 2014).

Figure 15.8.2 . Ebola deaths extended according to doubling times seen early in the outbreak.

15.8.1 https://bio.libretexts.org/@go/page/26178
Calculating from these doubling times and extending for several more months results in the number of deaths shown in Figure
15.8.2. Before five months with the estimates we made during class (blue curve in the figure) and before three months with the

earlier estimates (red curve), the number of deaths was predicted to exceed 100,000.
But there is a flaw in this approach. As shown by the calculations on bacteria and Darwin’s calculations on elephants (Chapter 3),
exponential growth models cannot be extended very far. They can be quite accurate a limited number of time units in the future.
But when applied to biological populations both exponential and orthologistic models inevitably fail when extended indefinitely.
Actually, one view is that they do not really fail—they just warn that some other growth model will supplant them before
populations grow too large.
Table 15.8.1 . Consequences of unlimited doubling in Ebola deaths.
Days Years Total Deaths

0 0.00 892

100 0.27 8,412

200 0.55 79,335

300 0.82 74,8267

400 1.10 7,057,440

500 1.37 66,563,800

600 1.64 627,811,000

700 1.92 5,921,330,000

710 1.94 7,411,030,000

In fact, assuming the unrestrained doublings of exponential growth is tantamount to assuming a disease will kill the entire world,
with the only question being when. Table 15.8.1 shows the results of doubling at the rate illustrated by the blue curve in Figure
15.8.2, extended farther. At this rate the entire human population would be extinguished in less than two years!

Expected moderation. Of course, the entire human population will not be extinguished by a manageable disease. Draconian
measures would be put in place long before— isolating those infected, closing borders, and more. In effect, the growth rate r will
be moderated by strong negative social pressure, term s .
When would such negative pressure appear in the data? Could it be seen early in the Ebola outbreak, when we and the students
began watching? Because the basic disease equation is equivalent to the rsN model, we thought early trends might appear if we
examined the data in terms of r and s . The individual growth rate in deaths, (1/N )dN /dt, could be examined and plotted against
the total number of deaths, N . This is Figure 15.8.3, with the same data as Figure 15.8.1, just reformulated.

Figure 15.8.3 . The data of Figure 15.8.1 converted to individual change in death rate as a function of number of deaths.

The data show quite a bit of noise, but with a clear downward trend, with the rate in the total number of new deaths decreasing as
deaths increase. Both the upper and lower ranges of points are decreasing (gray lines). The green line through the averages (least-
square regression line, solid, with r and s as shown in the figure) projects forward (dashed) to about 12,000 deaths before the
outbreak would be over—an extensive human tragedy, but far below the straightforward projections of Figure 15.8.2.

15.8.2 https://bio.libretexts.org/@go/page/26178
If this decline in death rate was real, it was likely developing from more and more attention to moderating the disease—medical
workers expanding hospitals, populations practicing more careful burials, governments cautioning only justified travel, and the
like. When we made the projection of 12,000 in the fall of 2014, we had no certainty about what would happen; we were simply
looking at the data, which showed not unrestrained exponential growth, but moderating growth instead.
The next step was to plug the r and s derived from the fitted curve and project forward six months or a year. This projection is
shown in Figure 15.8.4. The green curve is the projection from r = 0.028 and s = −0.0024. It is markedly different than the
other two curves, leveling off early and reaching about 12,000 deaths. Also in the figure are three additional points of actual data,
in yellow, not sufficiently advanced to tell which of the three curves—red, green, or blue—will be the real one.

Figure 15.8.4 . Ebola deaths projected by Equation 5.2 using parameters of Figure 15.8.3 .

In only a few weeks, however, it became apparent that the r + sN curve was the most accurate. By the end of the winter semester
(about day 140 in the graphs) it was clear that the outbreak was coming under control, and that the number of deaths was fairly
close to our initial projection. After tracking the outbreak with our students through the winter and spring semesters, we could see
how remarkable that early projection was (Figure 15.8.5). Reading the data carefully early in the crisis gave an accurate projection
of the outcome, from a simple model indeed!
Doubling times. This example provides a good place for us to reconsider doubling times, introduced with exponential growth.
Recall that exponential growth is the infinitely thin dividing line between logistic and orthologistic growth, and has the property of
a fixed “doubling time.” In other words, there is a specific time interval—call it tau (τ )—during which the population exactly
doubles. In logistic growth the doubling time constantly decreases, while in orthologistic growth it constantly increases. Earlier the
doubling time is shown to be the natural logarithm of 2 divided by r— (ln2)/r, or approximately 0.693/r. This is in years if r is
measured per year, days if per day, an so forth.
Thus the doubling time for the exponential curve of Figure 15.8.1, with r = 0.022, is 0.693/0.022 = 31.5days. The logistic
growth of Figure 15.8.3 has no fixed doubling time. However, at all times there will be an “instantaneous doubling time,” which
will hold approximately for a short time. In particular, near the very beginning of the outbreak—when N is close to 0—the growth
rate is r + sN = r + s ⋅ 0 = r . For the Ebola data early in the outbreak, we found r = 0.028 and s = 0.0024. The Ebola
doubling time, averaged over the countries in which it spread, therefore started out at 0.693/r = 0.693/0.028 = 25days. This is
in accord with early estimates made by health organizations.
When we began following the outbreak, about 4.5 thousand deaths had occurred, meaning that the growth rate was
r + sN = 0.028 − 0.0024 ⋅ 4.5 = 0.0172 , and the doubling time was 0.693/0.0172 = 40days.

The doubling time continued to decrease until the outbreak was conquered and all deaths from Ebola ceased.

15.8.3 https://bio.libretexts.org/@go/page/26178
 Figure 15.8.5 . Actual development of the outbreak (yellow dots), falling along the projection from October, 2014 (green curve).

Caveats and considerations. There was some fortuity in the timing of our initial projection. Shortly after we started the death rate
dropped, possibly from increased efforts following intense world-wide attention. Had we made the projection a few weeks later we
would have seen two slopes and would have had to guess which would prevail. It turns out that the first slope prevailed, returning
about 40 days after our initial projection. But we had no way to know this from the data at the time.
Our approach could also be criticized because the r + sN equation we used is a hybrid—a single equation trying to represent two
different things, in this case infections and deaths. It is reasonable here to base the amelioration parameter s on the total number of
deaths, since deaths were the parameter of world concern. Deaths influence social attention to the disease and efforts to control it.
However, what sense is there in saying that deaths grow at rate r, based on the total number of deaths thus far? Ebola can be
transmitted to others shortly after death, but in general deaths do not cause new deaths. Infections cause new infections, which in
turn cause new deaths. If N represents total deaths, it seems the equations should also have an I to represent infections, and some
fraction of infections should result in deaths—as in this two-dimensional system of equations.
dI
= β I − γ I − α I − sN I (15.8.1)
dt

dN
= α I (15.8.2)
dt

The two dimensions are I , the number of existing infections, and N , the cumulative number of deaths. Infectivity is β, the rate of
recovery from infection is γ, and the rate of death from the disease is α . These correspond to the notation in Figure 15.2.1. In
addition, the term sN moderates the growth of the infection by representing all the cautions and infrastructure put in place against
the disease as the number of deaths increased.
No factor like 1 − v − p , representing the fraction of susceptible individuals, is needed to multiply the β term here, because both in
the early stage and the outbreak the prevalence was low and there was no vaccine, so almost everyone was susceptible. And with
this rapidly progressing disease the population remained almost constant, so births could be considered negligible. Equation is thus
still a simplified equation.
As simplified as it is, however, Equation carries more parameters than can be perceived in the raw data. The death rate from the
disease α , the recovery rate γ, the infectivity β, and even the number of infections I , can only be discovered through the use of
special research programs. But even at the beginning of an outbreak, and even for a poorly understood disease, there may be
enough data to determine the doubling time for deaths from the disease, and how the doubling time is changing—enough data to
determine r and s , even if little more.
This is fortunate, because Equation can be reduced in dimension and approximated by the r + sN form. Without the amelioration
term sN I , the two dimensions are independent and growth of each is exponential, with N being the integral of I . Because the
integral of an exponential is still exponential, the two can be approximated by a single equation of one fewer dimension, with the
amelioration term sN I reinstated. In this way, total deaths become a legitimate surrogate for infections in low-prevalence
outbreaks, as in the present example of Ebola.
And as we noted, this gave an accurate projection on the course of a dread disease, from a very simple model indeed!

This page titled 15.8: Application to an actual outbreak is shared under a CC BY-NC 4.0 license and was authored, remixed, and/or curated by
Clarence Lehman, Shelby Loberg, & Adam Clark (University of Minnesota Libraries Publishing) via source content that was edited to the style
and standards of the LibreTexts platform; a detailed edit history is available upon request.

15.8.4 https://bio.libretexts.org/@go/page/26178
15.9: Conflicting ethics

This page is a draft and is under active development.

Fundamental conflicts in ethical behavior are hidden just below the surface of the cold mathematics in these equations. Virulence α
appearing in the numerator of a term with a minus sign means that the greater the virulence, the lower the equilibrium level of
disease, p^ . In the absence of vaccination (v = 0) , setting p^ = 0 and solving for α yields α = β . Artificially increasing the
virulence α of a disease to equal or exceed the infectivity β will therefore drive the disease extinct.
The enduring ethical conflict here is between the individual and the population. With all other things being equal, working to
reduce virulence benefits the individual but may cause more individuals in the population to become infected. Working to increase
virulence, in contrast, harms individuals but may reduce the number who become infected.
The ethics of modern medicine emphasizes individuals— working to cure disease and reduce virulence, ameliorating symptoms,
reducing discomfort, and recognizing patient needs. Increasing the virulence of a disease in a human patient to reduce its spread is
unthinkable, both in medicine and public health. The ethics of modern agriculture, however, are the diametric opposite. If a crop is
infected with a destructive communicable disease, entire fields of the crop may be mowed, burned, or otherwise disposed of.
Infected populations of poultry and livestock are treated similarly, killed en masse and buried or burned to contain the disease.
Artificially altering the infectivity β is also a possibility. In Equation 15.4, β appears in the denominator of a term having a minus
sign—meaning that decreasing β will decrease the equilibrium level of the disease, p^ . Ethical conflicts also arise here, though they
are not as stark as the conflicts connected with α . During the influenza epidemic of 1918–19, San Francisco leaders required
citizens to wear breathing masks—for “conscience, patriotism, and self-protection,” wrote the mayor.

Figure 15.9.1 . Funeral pyre of infected cows. Exterminating a living organism when it is infected artificially increases the
virulence of the disease to something like α → ∞ .

This lowered β by containing respiratory droplets from infected individuals, and lowered the chance of infected droplets entering
the respiratory systems of susceptible individuals, in turn reducing the infectivity β. Some citizens, however, refused to wear the
masks.
During the Ebola outbreak of 2014–15, amidst fears and warnings of the disease becoming established around the world, some U.S.
governors ordered temporary quarantine of returning medical workers who had been in direct proximity with Ebola, until it was
clear that they were not infected. At least one refused the quarantine based on individual rights, and the courts upheld the refusal.
These ethical conflicts surrounding β are not as grim as those surrounding α , with options currently practiced for domestic plants
and animals but so extreme that they are never proposed for human populations.
It is curious that these high-level social dilemmas are perceptible within the most basic equations of ecology. Science can inform
such ethical issues, but society must decide them.

15.9.1 https://bio.libretexts.org/@go/page/26183
 Figure 15.9.2 . The slogan “Obey the laws and wear the gauze” accompanied regulations intended to protect public and individual
health in 1918–19 San Francisco. Thousands of city residents died of influenza, but at a rate lower than that of many other cities.
Most of the public complied with the regulations; some who refused were quarantined in jail. Pictured above are police officers in
Seattle in 1918.

This page titled 15.9: Conflicting ethics is shared under a CC BY-NC 4.0 license and was authored, remixed, and/or curated by Clarence Lehman,
Shelby Loberg, & Adam Clark (University of Minnesota Libraries Publishing) via source content that was edited to the style and standards of the
LibreTexts platform; a detailed edit history is available upon request.

15.9.2 https://bio.libretexts.org/@go/page/26183
 CHAPTER OVERVIEW

16: Competition
16.1: Chapter Introduction
16.2: The niche concept
16.3: Resource competition
16.4: Competitive exclusion
16.5: Succession
16.6: Single-resource phase space
16.7: Multiple resource phase space
16.8: Lotka-Volterra formulation

This page titled 16: Competition is shared under a CC BY-NC 4.0 license and was authored, remixed, and/or curated by Clarence Lehman, Shelby
Loberg, & Adam Clark (University of Minnesota Libraries Publishing) via source content that was edited to the style and standards of the
LibreTexts platform; a detailed edit history is available upon request.

1
16.1: Chapter Introduction

This page is a draft and is under active development.

Ecological competition is a potent force driving organic evolution. When Charles Darwin reached the Galapagos Islands at age 26,
he studied an assortment of fifteen similar species that are now called “Darwin’s Finches” (Figure 16.1.2). His observation that
various species had different beak structures, specialized for different foods, led him to question the stability of species. Indeed,
such “character displacement” is one of several consequences of ecological competition.

Figure 16.1.1 . The Galapagos Archipelago, on the equator more 500 miles off the west coast of South America.

Figure 16.1.2 . Three species of Darwin’s finches: Sharp-beaked ground finch, Geospiza difficilis (left); cactus ground finch, G.
scandens (middle); and large ground finch, G. magnirostris (right).

Seeing the gradation and diversity of structure in one small, intimately related group of
birds, one might really fancy that from an original paucity of birds in this archipelago,
one species had been taken and modified for different ends. —Charles Darwin

This page titled 16.1: Chapter Introduction is shared under a CC BY-NC 4.0 license and was authored, remixed, and/or curated by Clarence
Lehman, Shelby Loberg, & Adam Clark (University of Minnesota Libraries Publishing) via source content that was edited to the style and
standards of the LibreTexts platform; a detailed edit history is available upon request.

16.1.1 https://bio.libretexts.org/@go/page/26184
16.2: The niche concept
A species’ niche is the range of environmental factors that allow that species to survive and reproduce. A particular tree species, for
example, may be able to live where temperatures do not drop below −40 ◦ , and where yearly precipitation is at least 750 mm.
Perhaps it also needs open sunlight and an appropriate collection of root fungi. Such are the parameters of a niche.
G. Evelyn Hutchinson, one of the great ecologists of the twentieth century, envisioned the parameters that form a niche as an “n-
dimensional hyperspace.” The “fundamental niche” is the set of conditions allowing the species to survive if there are no other
species interfering. Physical conditions are chief among those. The “realized niche” is the real life niche—where species are
restricted by interactions with other species.
Consider latitude on the earth’s surface, which is connected to several parameters such as sunlight and temperature. And consider
two species that can thrive anywhere between 40◦ and 60◦ latitude, and whose density drops slowly with increasing latitude (Figure
16.2.1).

Figure 16.2.1 : Species living apart along a spatial gradient (above) and living together along the same gradient (below).
At the top of the figure are two nearly horizontal lines representing the abundance you might observe of the two species as you
travel north. If free of Species 2 (its competitor), Species 1 (blue line) declines slowly in abundance in more northerly climates.
Species 2 similarly declines in abundance (red line), but compared with Species 1 fares a little better in the north and a little worse
in the south.
When these two species are together they compete with each other—each suppressing the other. Using competition equations like
those presented in this chapter, we see that in the south, where Species 1 fares better, it takes over and dominates. In the north, in
contrast, where Species 2 fares better, it dominates instead (Figure 16.2.1, bottom).
You see that there can be a sharp change in abundance even with only very slight changes in species characteristics. A range of one
species can end and that of a new species can begin, even though you may not be able to discover anything from either species
alone as to why they switch their dominance. And the switch-over point need not correspond to the exact place in which their
dominance switches. Here the actual switch-over point is a few degrees to the north because of the migration simulated in the
model. This phenomenon is called “competitive exclusion” and, when it occurs over space like this, “zonation.”

16.2.1 https://bio.libretexts.org/@go/page/25520
 Figure 16.2.2 : Zonation on a three-year-old fen, Bluebird Prairie, Upper Midwest.
Any environmental gradient can induce zonation. Figure 16.2.2 illustrates this on a restored prairie in the North American Upper
Midwest. Though less than ten meters, the variation in elevation is enough to induce a mild moisture gradient. The entire area had
been converted from a corn field to a restored prairie and seeded uniformly with a mixture of prairie grasses and prairie flowers,
but distinct boundaries arose only three years after restoration.
All the upland areas, labeled 1, contained standard restored prairie flora such as Andropogon gerardi (Big Bluestem grass) and
rapidly emerging flowers such as Rudbekia hirta (Black-eyed Susan). But in successive zones surrounding mild depressions in the
landscape, labeled 2 to 4, there were sharp transitions to moisture-loving genera such as Typha latifolia (cattail) and Stachys
palustris (Smartweed).

Definitions

Fundamental niche: The conditions under which a species can live, absent interference from other species.
Realized niche: The conditions to which a species is restricted by interactions with other species.

This page titled 16.2: The niche concept is shared under a CC BY-NC 4.0 license and was authored, remixed, and/or curated by Clarence Lehman,
Shelby Loberg, & Adam Clark (University of Minnesota Libraries Publishing) via source content that was edited to the style and standards of the
LibreTexts platform; a detailed edit history is available upon request.

16.2.2 https://bio.libretexts.org/@go/page/25520
16.3: Resource competition
Competition among two species occurs when the interaction terms s and s in Equation 8.1 are both negative. This rather
1,2 2,1

abstract approach does lead to broad insights, but for other kinds of insights let us proceed to a more mechanistic view. Instead of
abstract coefficients representing inhibition among species, let us consider resources which species need to thrive and survive. The
species will not interact directly— they never even need to come into contact—but will influence each other through their use of a
common resource, which they both need for maintenance and growth of their populations.
Resource competition is one of the oldest parts of ecological theory, introduced in the late 1920s by mathematician Vivo Volterra.
We will start where he started, considering what have been called “abiotic resources”.
Species require sunlight, space, nitrogen, phosphorous, and other resources in various amounts. If a resource is too rare,
populations cannot grow, and in fact will decline. In Figure 16.3.1 this is shown in the region to the left of the arrow marked A, in
which the individual growth rate 1/N dN /dt is negative.
At higher resource levels the growth rate increases and, at point A, the population can just barely maintain itself. Here the
individual growth rate 1/N dN /dt is zero. This level of resources is called R , pronounced “are star”. At higher levels of
∗

resource, above R , the population grows because 1/N dN /dt becomes positive.
∗

Figure 16.3.1 Growth curve generalized for any resource, with a piece-wise linear approximation sketched as dashed lines.

Once resources are abundant—approximately above B in the figure—needs are satiated and the addition of more resources does not
make any large difference. Population growth stays approximately the same between marks B and C.
At very high levels, too much resource can actually harm the population. Too much sunlight can burn leaves, for example, while
too much nitrogen can damage roots. At this point, above C in Figure 16.3.1, the growth rate starts falling. By D the species can
again just barely hold its own, and to the right of D the species is killed by an overabundance of resources.

Figure 16.3.2 Growth curve with only the left linear piece of Figure 16.3.1 , usable because species tend to keep resources low.

Such high resource levels, however, are not usually observed, because species draw resources concentrations down by using them
up. Unless extreme environments are being modeled, only the left dashed linear piece of Figure 16.3.1 needs to be modeled, as
shown in Figure 16.3.2.
At this point, it is helpful to review various forms for the equation of a straight line. The usual slope–intercept form, y = mx + b ,
which is a y-intercept form—is not as useful here. It’s the x-intercept form, y = m(x − a) , that comes into play for writing a
mechanistic resource model for single species population growth.

16.3.1 https://bio.libretexts.org/@go/page/25521
 1. Slope–intercept form: y = mx + b
(slope m, y-intercept b)
2. Intercept–intercept form: y = b(1 − x/a)
(x-intercept a, y-intercept b)
(m = −b/a)
3. Slope–x-intercept form: y = m (x − a)
(slope m, x-intercept a)
(b = −m/a)

The zero-growth point, R , is important in the theory of resource competition. It is the amount of resource that just barely sustains
∗

the species. If the resource level is less than R , the species dies out; if it is greater, the species grows and expands. The resource
∗

level in the environment therefore is expected to be at or near the R value of the dominant species. If it is above that level, the
∗

population grows, new individuals use more resource, and the resource level is consequently reduced until growth stops.
R
∗
can be measured in the greenhouse or the field. In the greenhouse, for example, you might arrange plants in 20 pots and give
them different amounts of nitrogen fertilizer in sterile, nutrient-free soil. In the complete absence of fertilizer, the plants will die.
With larger amounts of fertilizer the plants will be luxuriant, and when there is too much fertilizer, the resource will become toxic,
again leading to dieback in the plants. You can thus measure the curve of Figure 16.3.1 fairly easily and find the point on the left
where the plants just survive. This is their R . ∗

You can also measure this value for different species independently, and from the results estimate how plants will fare living
together. To start, suppose that one resource is the most limiting. Represent the amount of that limiting resource available in the
environment at time t by the symbol R(t) or, for shorthand, simply R . The amount in excess of minimal needs is R − R , and that ∗

amount of excess will determine the rate of growth. A tiny excess will mean slow growth, but a larger excess can support faster
growth. So the equation in Figure 16.3.2 shows the individual growth rate, 1/N dN /dt, being proportional to how much resource
exists in excess.
As before, N measures the size of the population at time t , in number of individuals, total biomass, or whatever units are relevant
to the species being studied. R is the smallest amount of resource that can support a viable population, and m tells how the
∗

individual growth rate, (1/N )dN /dt, depends on the amount of resource available in excess of minimal needs.
Now let us say that R is the maximum amount of resource available in the environment, in absence of any organisms, and u is
max

the amount of resource used by each living organism in the population. Then uN is the amount of resource tied up in the
population at time t . R − uN = R is the amount of resource not used by the population. This is the basis of a resource
max

theory that assumes resources are released immediately upon death of an organism, and it has many of the important properties of
more complex resource models.
Start with the above statements in algebraic form,
1 dN
∗
= m(R − R ) and R = Rmax − uN (16.3.1)
N dt

then substitute the equation on the right above into the one on the left. This gives
1 dN ∗
= m (Rmax − uN − R )
N dt

Multiplying through on the right by m and rearranging terms gives

1 dN
∗
= m(Rmax − R ) − umN (16.3.2)
N dt

Notice that the first term on the right is a constant and the second term is a constant times N. Does this look familiar? This is just
density-regulated population growth in disguise— the r + sN model again!
1 dN /ast
= r + sN , r = m(Rmax − R ), s = −um
N dt

Recall that this also happened for the epidemiological I model. And it will arise occur again in future models.

This page titled 16.3: Resource competition is shared under a CC BY-NC 4.0 license and was authored, remixed, and/or curated by Clarence
Lehman, Shelby Loberg, & Adam Clark (University of Minnesota Libraries Publishing) via source content that was edited to the style and
standards of the LibreTexts platform; a detailed edit history is available upon request.

16.3.2 https://bio.libretexts.org/@go/page/25521
16.4: Competitive exclusion
Consider what will happen with two species using the same resource, such as light or space or nitrogen fertilizer. The amount of
resource available will be the maximum amount, R , minus what is tied up in all individuals of all species. With u being the
max i

amount of resource tied up in each individual of species i, the resource remaining at any time will be
R = Rmax − u1 N1 − u2 N2

Or for many species

R = Rmax − u1 N1 − u2 N2 − u3 N3 − ⋯ − uh Nh

R = Rmax − ∑ ui Ni (16.4.1)

i=1

Each species has its own growth equation, identical in form for all species, but different in the critical level of resource, R , and the ∗
i

growth coefficient, m : i

1 dNi
∗
= mi (R − R ) (16.4.2)
i
Ni dt

What remains is to consider how the growth coefficient m relates to the minimal level of resource tolerated, R . It turns out to be
i
∗
i

a tradeoff between the two. Consider, for example, a plant species that is limited by the amount of nitrogen available, as plants are.
And to have a large growth coefficient m the plant must produce abundant seed. To have superior nitrogen use, measured by a low
i

value of R , it needs abundant roots. But it cannot do both. There is a limited amount of solar energy to exploit, so if the plant
∗

allocates more to roots there is less to allocate to seeds, and vice versa.

Figure 16.4.1 . Tradeoff measured between colonization and ability to exploit nitrogen (Tilman 1994, Ecology 75:2–16).

It therefore turns out that species which are good colonizers, producing abundant seed, are poorer competitors for resources, having
a higher value of R . This idea is illustrated by measurements reported in Figure 16.4.1.
∗

Figure 16.4.2 . Two species individual growth rate versus resource level, whose time-course is illustrated in Figure 16.4.1

In Figure 16.4.2, tradeoffs are formulated for modeling. Species 2 grows more rapidly when resources are abundant. This is the
case at time 0, marked with t on the top axis.
0

16.4.1 https://bio.libretexts.org/@go/page/25522
As the populations grow they reduce the amount of resource available in the environment. At time 1, marked with t on the upper 1

axis, Species 2 can continue to grow faster than Species 1, though the margin is deteriorating. But there comes a point at which the
resources become depleted enough that the characteristics of Species 2 do not let it gather enough resources to maintain its
advantage. This is the crossing point of the blue and red lines in the figure. At time 2, both species are still growing, but Species 1
is growing faster. At time 3, with still lower resource levels—drawn down by Species 1—the resource falls below the minimal level
for Species 2, R . The growth rate of Species 2 falls negative and Species 2 starts to die out.
∗
2

Finally, at time 4, Species 1 depletes the resource to the level that it can just barely survive, and it stands alone, having wiped out
its competitor. This process is called “competitive exclusion.”
How this plays out over time is illustrated in Figure 16.4.3. At the top, Species 2 alone does just fine, rapidly rising to its carrying
capacity of 50 and pulling the resource down to its \(R^{\ast}\
) of 2. In the middle, Species 1 alone also does just fine, rapidly rising to its carrying capacity of 60 and pulling the resource down
to its R of 1.
∗

But grown together, Species 2 makes an initial splash and then declines. This is due to the incessant growth of Species 1, which
outcompetes it. Species 1 simply draws the resource down below the level at which Species 2 can survive.
Competitive exclusion, which assumed that no more species could exist than there were resources, was treated as an inviolable law
of ecology for over fifty years. In the 1970s, however, this was shown not to be the case (Armstrong and McGehee 1980). More
about that later in the chapter.

Figure 16.4.3 . Competitive exclusion based on the tradeoffs of Figure 16.4.2 .

This page titled 16.4: Competitive exclusion is shared under a CC BY-NC 4.0 license and was authored, remixed, and/or curated by Clarence
Lehman, Shelby Loberg, & Adam Clark (University of Minnesota Libraries Publishing) via source content that was edited to the style and
standards of the LibreTexts platform; a detailed edit history is available upon request.

16.4.2 https://bio.libretexts.org/@go/page/25522
16.5: Succession
A similar process for more than two species results in a succession of species taking over, one after the other, in an ecological
process known as “succession.”

Figure 16.5.1 . Multi-species individual growth rate versus resource level, whose time-course is illustrated in Figure 16.5.2

In natural systems many species compete, with tradeoffs between their R values and their growth rates, as in Figure 16.5.1. The
∗

following is a program to simulate the differential equations for five species competing for the same resource and producing the
curves of Figure 16.5.2. With only two species, this same program can produce the curves of Figure 16.3.3.

# SIMULATE ONE YEAR

# This routine simulates competition differential equations through one time
# unit, such as one year, taking very small time steps along the way.
# Accuracy should be checked by reducing the size of small time steps until
# the results do not significantly change.
# This routine implements <Q>Euler’s Method</Q> for solving differential
# equations, which always works if the time step is small enough.
#
# ENTRY: ’N1’ to ’N5’ are the starting populations for species 1-5.
# ’m1’ to ’m5’ specify the sensitivity of the corresponding species
# to the available amount of resource.
# ’u1’ to ’u5’ specify the resource tied up in each species.
# ’R1star’ to ’R5star’ are the minimum resource levels.
# ’Rmax’ is the greatest amount of resource possible.
# ’dt’ is the duration of each small time step to be taken throughout
# the year or other time unit.
#
# EXIT: ’N1’ to ’N5’ are the estimated populations of species 1-5 at the
# end of the time unit.
# ’R’ is the estimated resource level at the end of the time step.

Rmax=R=7;
R1star=1.0; R2star=2.0; R3star=3.0; R4star=4.0; R5star=5.0;
N1=0.000001; N2=0.000010; N3=0.000100; N4=0.001000; N5=0.010000;
m1=0.171468; m2=0.308642; m3=0.555556; m4=1.000000; m5=1.800000;
u1=0.001000; u2=0.001000; u3=0.001000; u4=0.001000; u5=0.001000;

# SIMULATE ONE YEAR

SimulateOneYear = function(dt)

16.5.1 https://bio.libretexts.org/@go/page/25523
 { for(v in 1:(1/dt)) # Advance a small time step.
{ R=Rmax-u1*N1-u2*N2-u3*N3- # Compute resource remaining.
u4*N4-u5*N5;
dN1=m1*(R-R1star)*N1*dt; # Estimate the change in the
dN2=m2*(R-R2star)*N2*dt; # population of each species.
dN3=m3*(R-R3star)*N3*dt;
dN4=m4*(R-R4star)*N4*dt;
dN5=m5*(R-R5star)*N5*dt;

N1=N1+dN1; N2=N2+dN2; # Add the estimated change to

N3=N3+dN3; N4=N4+dN4; N5=N5+dN5; } # each population and repeat.

assign("N1",N1, envir=.GlobalEnv); # At the end, export the results and return

assign("N2",N2, envir=.GlobalEnv);
assign("N3",N3, envir=.GlobalEnv);
assign("N4",N4, envir=.GlobalEnv);
assign("N5",N5, envir=.GlobalEnv); }

# SIMULATE ALL YEARS

for(t in 0:100) # Advance one year.
{ print(round(c(t,N1,N2,N3,N4,N5))); # Display results.
SimulateOneYear(1/(365*24)); } # Repeat.

An environment can change because species living in it have effects that can “feed back” and change the environment itself. In this
case the feedback is change in the resource level, which each successive species changes in a way that is compatible with its own
existence. There is nothing teleological in this; any species that change the environment in ways not compatible with their own
existence simply do not persist, and hence are not observed. When the program runs, it produces a file excerpted below, which is
graphed in Figure 16.5.2.

t N1 N2 N3 N4 N5

1 0 0 0 0 0

2 0 0 0 0 13

3 0 0 0 7 391

4 0 0 0 45 1764

5 0 0 1 127 1894

6 0 0 3 329 1743

7 0 0 9 790 1392

: : : : : :

60 1649 3536 0 0 0

61 1891 3324 0 0 0

62 2157 3094 0 0 0

63 2445 2864 0 0 0

64 2751 2584 0 0 0

65 3070 2313 0 0 0

66 3397 2039 0 0 0

16.5.2 https://bio.libretexts.org/@go/page/25523
 67 3725 1767 0 0 0

: : : : : :

96 5999 1 0 0 0

97 5999 0 0 0 0

98 6000 0 0 0 0

99 6000 0 0 0 0

100 6000 0 0 0 0

Figure 16.5.2 . Succession based on the tradeoffs of Figure 16.5.1 .

At the beginning in Figure 16.5.2, from time 0 to about time 3, the resource is at its maximum level, R , and the abundances of
max

all species are at very low levels. Between times 3 and 5—Species 5, the one with the highest growth rate when resources are
abundant—increases rapidly while resources drop accordingly. But near the end of that time the next in the series, Species 4, starts
to increase, pulling the resource down below the level that allows Species 5 to survive. Species 5 therefore declines while Species 4
increases.
This process continues in succession, with Species 3 replacing 4, 2 replacing 3, and, finally, Species 1 replacing 2. The resource
falls in stages as each successive species gains dominance. Finally, when no more superior species exist, the system reaches what is
called its “climax condition” at about time 90, with resources at a low level.
There is nothing especially remarkable about Species 1. It is simply (1) the best competitor living in the region, meaning better
competitors cannot readily arrive on the scene, or (2) the best competitor that evolutionary processes have yet produced. In either
case, it is subject to replacement by another—for example, by an “invasive species” arriving by extraordinary means.
Of course, succession in complex natural systems may not be as clear-cut as in our simple models. Multiple resources are involved,
species may be very close to each other in their ecological parameters, and stochastic events may intervene to add confusion.

Figure 16.5.3 . Schizacharium at Cedar Creek with oaks invading.

This page titled 16.5: Succession is shared under a CC BY-NC 4.0 license and was authored, remixed, and/or curated by Clarence Lehman,
Shelby Loberg, & Adam Clark (University of Minnesota Libraries Publishing) via source content that was edited to the style and standards of the
LibreTexts platform; a detailed edit history is available upon request.

16.5.3 https://bio.libretexts.org/@go/page/25523
16.6: Single-resource phase space
Some aspects of competition for a resource are clarified by looking at the phase space, as introduced in Chapter 10. Combining
Equations 16.3.1 and 16.3.2 gives the following as a starting point:
1 dN1
∗
= m1 (Rmax − R ) − u1 m1 N1 − u2 m2 N2 (16.6.1)
1
N1 dt

1 dN2
∗
= m2 (Rmax − R ) − u2 m2 N2 − u1 m1 N1 (16.6.2)
2
N2 dt

As before, m is the rate of growth of Species i for each level of resource above its minimum resource requirement R , and u is
i
∗
i i

the amount of resource tied up in each individual of Species i. For reference, here is the assignment of parameters in terms of r i

and s .
i,j

∗
r1 = m1 (Rmax − R ), s1,1 = −u1 m1 , s1,2 = −u2 m2 (16.6.3)
1

∗
r2 = m2 (Rmax − R ), s2,1 = −u1 m1 , s2,2 = −u2 m2 (16.6.4)
2

Where in the phase space will the growth rate be 0 for each species? For Species 1 it will be where
1 dN1
= 0 = r1 + s1,1 N1 + s1,2 N2
N1 dt

Solving for N gives

2

r1 s1,1
N2 = − − N1 ← Species 1 isocline (16.6.5)
s1,2 s1,2

Figure 16.6.1. Species 1 increases below its isocline, shaded with gray copies of the numeral 1.

Anywhere along that line, the population of Species 1 will not change, but on either side of the line it will (Figure 16.6.1).
Formulae for the four possible equilibria and their stability are in Table 10.1. The vertical intercept of the isocline, where N = 0 , 1

is −r /s , and the horizontal intercept, where N = 0 , is −r /s . The slope is −s /s ) = (u m )/(u m ) .

1 1,2 2 1 1,1 1,1 1,2 1 1 2 2

Similarly, growth of Species 2 will be 0 where

1 dN2
= 0 = r2 + s2,2 N2 + s2,1 N1
N2 dt

16.6.1 https://bio.libretexts.org/@go/page/25524
 Figure 16.6.2 . Species 2 increases below its isocline, shaded with gray copies of the numeral 2.

Solving for N gives

2

r2 s2,1
N2 = − − N1 ← Species 2 isocline (16.6.6)
s2,2 s2,2

Again, anywhere along that line the population of Species 2 will not change, but on either side of the line it will (Figure 16.6.2).
The vertical intercept of that line, where N = 0 , is −r /s , the horizontal intercept, where N = 0 , is -r /s , and the slope
1 2 2,2 2 2 2,1

is −s /s = (u m )/(u m ) .
2,1 2,2 1 1 2 2

Figure 16.6.3 . Single-species parallel isoclines. Each species increases only below its respective isocline, shaded with gray with the
species number, 1 or 2.

Notice this: In terms of the resource, the slope of the isocline for Species 2 is identical to the slope for Species 1—both are equal to
(u m )/(u m ). What does this mean? It means that the two isoclines are parallel. And that, in turn, means that the two species
1 1 2 2

cannot permanently coexist.

The populations can fall into only one of the three regions of Figure 16.6.3. If they start in the upper region, they decrease until
they enter the middle region. If they start in the lower region, they increase until they also enter the middle region. Once in the
middle region, only Species 2 increases. That means the population of Species 1 is driven leftward, toward lower values of N , 1

while the population of Species 2 is driven upward, toward higher values of N . 2

16.6.2 https://bio.libretexts.org/@go/page/25524
 Figure 16.6.4 . Flow across the phase space, as explained in Chapter 10, converging on a stable equilibrium where Species 2
excludes Species 1. (r = 0.75, r = 0.52, s = −1.875, s = −0.533, s = s = −1) .
1 2 12 21 11 22

These dynamics show up in the ow diagram of Figure 16.6.4. The origin (0,0) is an unstable equilibrium. In this single-resource
system, any populations near extinction, but not completely extinct, increase until they hit the middle region. The horizontal axis
has another unstable equilibrium, where Species 1 is at its carrying capacity and Species 2 is extinct (−r /s ,0). Any populations
1 1,1

near that unstable equilibrium soon arrive in the middle region. All populations not precisely on one of those two unstable
equilibria converge on the red disc on the vertical axis, where Species 2 is at its carrying capacity, K = −r /s , and Species 1
2 2 2,2

is extinct (0,−r /s ). This equilibrium is called a “global attractor.”

2 2,2

Phase spaces thus provide another view of competitive exclusion, the theory of which applies at least to two species competing for
a single resource at equilibrium.

This page titled 16.6: Single-resource phase space is shared under a CC BY-NC 4.0 license and was authored, remixed, and/or curated by
Clarence Lehman, Shelby Loberg, & Adam Clark (University of Minnesota Libraries Publishing) via source content that was edited to the style
and standards of the LibreTexts platform; a detailed edit history is available upon request.

16.6.3 https://bio.libretexts.org/@go/page/25524
16.7: Multiple resource phase space
As the next step, consider the case of two essential resources. This can be done mathematically following the approach we used
earlier for a single resource.
Call the two resources R and R . Suppose that of the total of these two resources used by Species 1, a proportion p is Resource
A B 1

A and therefore a proportion q = 1 − p is Resource B. Likewise, for Species 2 a proportion p is Resource A and q = 1 − p
1 1 2 2 2

is Resource B.
This is confusing, so for a clarifying example, suppose R is phosphate, PO4, and R is silicate, SiO2, both essential to two
A B

species of algae in a waterway. Take Species 1 to be an Asterionella species and Species 2 to be a Cyclotella species, as in a
pioneering study by David Tilman (1977). In this case, Species 1 needs phosphate and silicate resources in about a 1:99 ratio, while
Species 2 needs them in about a 6:94 ratio. If silicate is low, Species 1 will thus suffer first, since it needs a larger proportion of it,
while Species 2 will suffer first if phosphate is low, for the related reason. Here it would be p = 0.01, q = 1 − p = 0.99 for 1 1 1

the use of Resources A and B by Species 1, and p = 0.06, q = 1 − p = 0.94 for use by Species 2.
2 1 1

With u being the total amount of resource tied up by each individual of Species 1, p u will be the amount of Resource A tied up
1 1 1

by Species 1 and q u the amount of Resource B tied up the same way. Similarly, p u will be the amount of Resource A tied up
1 1 2 2

by each individual of Species 2 and q u the amount of Resource B. Also, assume as before that the resources under consideration
2 2

disappear from the environment when they are taken up by individuals newly born, are released immediately when individuals die.
With this in mind, the resources remaining at any time, as functions of the maximum resource and the abundance of each species,
will be
RA = RAmax − p1 u1 N1 − p2 u2 N2 (16.7.1)

RA = RBmax − q1 u1 N1 − q2 u2 N2 (16.7.2)

Suppose that populations of Species 1 and 2 grow based on which resource is closest to the R experienced by that species for the ∗

resource. This can be represented by the “min” function, min(a, b), which selects the smaller of two values. For example,
min(200,10) = 10, min(−200,10) = −200. Now the two-species, two-resource growth equations, generalizing the single-species,
single-resource growth in Equation 16.1.1, are
1 dN1 ∗ ∗
= m1 min(RA − R , RB − R )
N1 dt 1A 1B

1 dN2 ∗ ∗
= m2 min(RA − R , RB − R )
N2 dt 2A 2B

This could be refined, so that the growth rates m and m would depend on which resource was limiting, but this does not matter
1 2

in the present analysis. If the species are similar enough and the resource level is such that they are limited by the same resource,
one will tend to be competitively excluded, as in the previous section. But if the two species are quite different, they can be limited
by different resources and the equations can be simplified.
1 dN1
∗
= m1 (RA − R ) (16.7.3)
1A
N1 dt

1 dN2
∗
= m2 (RB − R ) (16.7.4)
2B
N2 dt

Some algebra will reveal the basic properties. If you substitute the expressions for R1 and R2 from Equations 16.6.1 and 16.6.2
into Equations 16.6.3 and 16.6.4 you will get
1 dN1
∗
= m1 (RAmax − p1 u1 N1 − p2 U2 N2 − R ) (16.7.5)
1A
N1 dt

1 dN2
∗
= m2 (RBmax − q1 u1 N1 − q2 u2 N2 − R ) (16.7.6)
2B
N2 dt

Now if you expand, collect, and rearrange terms, you get this equivalent form:
1 dN1
∗
= m1 (RAmax − R ) − m1 p1 u1 N1 − m1 p2 u2 N2 (16.7.7)
1A
N1 dt

16.7.1 https://bio.libretexts.org/@go/page/25525
 1 dN2
∗
= m2 (RBmax − R ) − m2 q2 u2 N2 − m2 q1 u1 N1 (16.7.8)
2B
N2 dt

Notice that, again, a mechanistic model with measurable parameters is just the general ecological RSN model in disguise.
The RSN formulation can expose the possibilities of two-resource situations in phase space. The isoclines are Equations 16.5.5 and
16.5.6 of this chapter, with slopes −s /s and −s /s for Species 1 and 2, respectively.
1,1 1,2 2,1 2,2

These two slopes can be written in terms of the resource. With the values for s from Equations 16.6.7 and 16.6.8 i,j

(s1,1= −m p u , s1 = −m p u , s
1 1 1,2 = −m q u , and s
1 2 2 = −m q u ), the slopes of the isoclines become
2,1 2 1 1 2,2 2 2 2

s1,1 −m1 p1 u1 u1 p1
= =
s1,2 −m1 p2 u2 u2 p2

s1,1 −m1 p1 u1 u1 1−p1

= =
s1,2 −m1 p2 u2 u2 1−p2

First notice that if the two species use the two resources in equal proportions (if p = p ), both slopes become u /u . The slopes 1 2 1 2

are parallel, as in the single-resource case in Figures 16.5.1 through 16.5.4. Therefore, if two species use two different resources
identically—that is, in equal proportions— they do not coexist. Coexistence requires some difference in how they use resources.
However, using the resources differently does not guarantee coexistence. Depending on their p’s and q’s, one isocline could still
enclose the other completely. Figure 16.7.1 has the same properties as Figure 16.5.4. Everywhere below the red isocline, Species 2
will increase, including the broad band between the red and blue isoclines where Species 1 will decrease.
If the blue and red isoclines are reversed, the result is similar, but with Species 1 excluding Species 2. Figure 16.7.2 shows this,
with the arrows reversed as Species 1 increases everywhere below the blue isocline, including the broad band between the isoclines
where Species 2 decreases.

Figure 16.7.1 . Two species using resources differently, with Species 2 excluding Species 1. (
r1 = 0.75, r2 = 0.65, s12 = −1.875, s21 = −0.666, s11 = s22 = −1 )

In all three cases, from Figures 16.5.4 to 16.7.2, the system has three equilibria—at the origin (0,0), where both species are absent,
at the carrying capacity K for Species 1 alone (−r /s ,0), and at the carrying capacity K for Species 2 alone (0,−r /s ). The
1 1 1,1 2 2 2,2

origin is unstable and only one of the other two equilibria is stable, depending on which isocline encloses the other.

Figure 16.7.2 . The opposite of Figure 16.7.1 , with Species 1 excluding Species 2. (
r1 = 0.975, r2 = 0.4, s12 = −1.5, s21 = −0.533, s11 = s22 = −1 ).

16.7.2 https://bio.libretexts.org/@go/page/25525
Figures 16.7.1 and 16.7.2 can be combined to give each species a chance to exclude the other, depending on circumstances. This
means not allowing one isocline to completely enclose the other, as in Figure 16.7.3. Intersecting isoclines introduce a fourth
equilibrium at the interior of the phase space. This equilibrium is unstable, marked with an open circle, and the two equilibria for
the individual species, on the axes, are stable. They are no longer “globally stable,” however, since only local regions of the phase
space lead to either of them.

Figure 16.7.3 . Isoclines intersecting in a way that allows each species to exclude the other, depending on starting conditions.

Depending on where the populations start, one of the two species will exclude the other. A curve called a “separatrix,” dividing
these starting points according to which species will exclude the other, is shown with the dashed black line in Figure 16.7.4. That
separatrix corresponds to a long curved ridge in any surface above the phase space, as described in Chapter 10. It necessarily passes
through the unstable interior equilibrium. Here it is a simple curve, though in some cases (such as the Mandelbrot system, not
representing competition), related curves can be infinitely complex.

Copy and paste image here

Figure 16.7.4 . Flow in the bistable system of Figure 16.7.3 . The black dashed line is a separatrix, dividing the phase space into
regions that reach one stable equilibrium or the other. (r = 0.975, r = 0.65, s = −2.437, s21 = −0.866, s = s = −1 ).
1 2 12 11 22

Finally, the isoclines can intersect in the opposite way, as in Figure 16.7.5. In this case, neither species has any region where its
isocline encloses the other, as seen in each of the phase spaces of Figures 16.5.4 through 16.7.4. What will happen when neither
species can exclude the other in any part of the phase space? They are forced to coexist. The individual equilibria on the axes
become unstable and the interior equilibrium becomes stable—indeed, globally stable.

16.7.3 https://bio.libretexts.org/@go/page/25525
 alt

Figure 16.7.5 . Isoclines intersecting at a globally stable coexistence. (

r1 = 0.75, r2 = 0.4, s12 = −1.153, s21 = −0.410, s11 = s22 = −1 ).

In summary, the isoclines in competitive systems have four different configurations, as in Figure 16.7.6. Cases 1 and 2 can
represent competition for a single resource or, equivalently, competition for two different resources that the two species handle
identically. One of the two is a superior competitor that excludes the other. Case 3 is “bistable,” where either of the species can
exclude the other, depending on how the system starts out. Finally, Case 4 is globally stable, where neither species can exclude the
other, and stable coexistence prevails. Cases 3 and 4 can represent competition for two different resources.

Figure 16.7.6 . Case 1: Species 2 excludes Species 1. Case 2: Species 1 excludes Species 2. Case 3: one excludes the other,
depending on starting conditions. Case 4: both species coexist.

This page titled 16.7: Multiple resource phase space is shared under a CC BY-NC 4.0 license and was authored, remixed, and/or curated by
Clarence Lehman, Shelby Loberg, & Adam Clark (University of Minnesota Libraries Publishing) via source content that was edited to the style
and standards of the LibreTexts platform; a detailed edit history is available upon request.

16.7.4 https://bio.libretexts.org/@go/page/25525
16.8: Lotka-Volterra formulation
Competition equations are usually presented in textbooks as the Lotka–Volterra competition model. This first appeared in the
ecological literature in the 1920s and is defined not just in terms of the interactions among species, but also in terms of the species’
carrying capacities, as follows.

Figure 16.8.1. Case 1: Species 2 excludes Species 1. Case 2: Species 1 excludes Species 2. Case 3: one excludes the other,
depending on starting conditions. Case 4: both species coexist.

1 dN1 N1 + a1,2 N2
= r1 (1 − ) (16.8.1)
N1 dt K1

1 dN2 N2 + a2,1 N1
= r2 (1 − ) (16.8.2)
N2 dt K2

K1 and K are the carrying capacities for Species 1 and 2, respectively. Parameters a and a represent the interference by each
2 1,2 2,1

species on the other. If a = 2 , for example, each individual of Species 2 interferes with the growth of Species 1 as if it were two
1,2

individuals of Species 1. If, on the other hand, a = 1/2 , it takes two individuals of Species 2 to have the same negative effect on
1,2

the growth of Species 1 as one individual of Species 1 itself.

To compare this with the RSN formulation, represent Equations 16.7.1 and 16.7.2 with i and j subscripts
dNi Ni +ai,j Nj
1
= ri (1 − )
Ni dt Ki

and multiply the right hand side through by r , i

dNi ri ai,j ri
1
= ri − Ni − Nj
Ni dt Ki Ki

This shows that the Lotka–Volterra formulation is isomorphic to the RSN formulation. All conclusions about competitive systems
examined thus far apply to the Lotka–Volterra formulation as well, with the appropriate translation of parameters. Parameter r is i

the same in both in the Lotka–Volterra formulation and the RSN formulation, but s = −r /K and s = −a r /K .
i,i i i i,j i,j i i

Beware, however, of a widely quoted statement derived from this formulation that appears throughout the ecological literature and
textbooks. Statements like “Coexistence requires that each species inhibit itself more than it inhibits the other species” are
abundant in textbooks and in the ecological literature. Alas, those statements are not correct.

box 16.8.2

Coexistence in the Lotka–Volterra model requires that each species be able to increase from low densities when
the other species is at its single-species equilibrium.

16.8.1 https://bio.libretexts.org/@go/page/26231
To see this, examine Figure 16.21. Here Species 1 inhibits itself with s11 =−1, while it inhibits the other species more strongly with
s12 =−1.153. Yet there is global coexistence. Evidently, coexistence does not require that each species inhibit itself more than it
inhibits the other, as in the conventional wisdom. The confusion in the literature has apparently emerged from the presence of the
carrying capacity terms, K1 and K2, in the Lotka–Volterra formulation. These terms obscure the effects of the interaction terms, a1,2
and a2,1, when carrying capacities differ between the species.
What, then, is a correct statement about coexistence? It can be put in terms of increasing from low densities, as shown in Box
16.8.2. The qualifying phrase “from low densities” is required, because the species not present can increase from high densities in a

bistable system, as in Figure16.20, and ip it to the other state, even though coexistence cannot occur.
Another way in which coexistence has been explained is by Vandermeer (1981 Bioscience), connecting coexistence with a certain
kind of “overyielding,” wherein two crops require less land for the same annual productivity when growing together than when
growing apart.

Figure 16.8.3. Coexistence in terms of a kind of overyielding (Vandermeer 1981, Bioscience).

The test is whether the joint equilibrium is above a line connecting the single-species equilibria (dashed gray in Figure 16.8.3A ) or
below the line (Figure 16.8.3B).
This view is correct for models we have been considering with straight-line isoclines, but incorrect for more general models with
curved isoclines (Figure 16.8.3C,D ). The statement in Box 16.8.2, however, holds true in each of these cases.

All things considered, rather than relying on rules of thumb, it can be better to evaluate a system directly, for example with the
methods of eigenvectors and eigenvalues described in Chapter 10.

This page titled 16.8: Lotka-Volterra formulation is shared under a CC BY-NC 4.0 license and was authored, remixed, and/or curated by Clarence
Lehman, Shelby Loberg, & Adam Clark (University of Minnesota Libraries Publishing) via source content that was edited to the style and
standards of the LibreTexts platform; a detailed edit history is available upon request.

16.8.2 https://bio.libretexts.org/@go/page/26231
 CHAPTER OVERVIEW

17: Mutualism
17.1: Chapter Introduction
17.2: Phase Diagrams
17.3: Quasi-species
17.4: Our own species

This page titled 17: Mutualism is shared under a CC BY-NC 4.0 license and was authored, remixed, and/or curated by Clarence Lehman, Shelby
Loberg, & Adam Clark (University of Minnesota Libraries Publishing) via source content that was edited to the style and standards of the
LibreTexts platform; a detailed edit history is available upon request.

1
17.1: Chapter Introduction
In Equation 8.1, mutualism occurs when the interaction terms, s and s , are both positive. This is the defining feature of
1,2 2,1

mutualism. The self-engagement terms, s and s , may be positive or negative, and the intrinsic growth terms, r and r , may
1,1 2,2 1 2

also be positive or negative.

If both growth terms r are positive, then either species can survive on its own, without the presence of the other. This is called
i

“facultative mutualism.” An example is the mutualism between oxpecker and warthog (Figure 9.2 left). Oxpeckers have other
sources for food, and warthogs may suffer but likely not die from larger parasite loads.
If the r are both negative, then neither species can survive on its own without the presence of the other. This is called “obligate
i

mutualism.” An example is lichen, composed of an algae and a fungus. If the fungus alone is placed on a rock, it will die. Although
it can eat rock—or, more accurately, can dissolve nutrients from rock—it cannot produce carbon compounds from rock. And if the
algae alone is placed on a rock, it will die. Although it can produce carbon compounds from the air through photosynthesis, it
cannot eat rock. But the two together become a perfect partnership, each covering the other’s weaknesses.

Figure 17.1.1. Photosynthesis is the basis of many mutualisms. Euglena is a protist–chloroplast combination that can fix carbon
dioxide from the water, changing the nature of the protist, and can move distances following nutrients, supporting the chloroplast.
Chloroplasts appear to be much-modified cyanobacteria, rendered wholly dependent on their partners by eons of evolution (left,
through a microscope). Lichens are fungal–algal combinations in which the algal component can fix carbon and the fungal
component can dissolve nutrients from rock (middle, shoulder height). Coral are anthozoa—marine animals with stinging tentacles
to capture food—that are often reef-building and associated with marine algae. A significant fraction of the world’s carbon is
sequestered in coral reefs (right, from space).

When r is positive and r is negative, or vice versa, it is a partially obligate mutualism. One species depends on a second, but the
1 2

second can get along without the first. This is a mathematical possibility, and some cases such as chloroplast– euglena may be
examples (Figure 17.1.1 left), where the chloroplast cannot live without the protist, but the protist can.
Besides obligate versus facultative mutualism, there is another major distinction. This is between “restrained” and “unrestrained”
mutualism. This distinction is not typically made in textbooks because it is related to the difference between logistic and
orthologistic population growth, which typically is not covered.
When self-limiting effects of terms s and s are stronger than the interspecies enhancement terms s and s , the mutualistic
1,1 2,2 1,2 2,1

pair has a carrying capacity that can be computed from the properties of the individual species. They reach their joint carrying
capacity along a path that looks like logistic growth, and is identical in form to that of two competing species that coexist and reach
a joint carrying capacity. However, when the self-limiting effects of s and s are weaker than the inter-species enhancement
1,1 2,2

terms s and s , the carrying capacity of the pair of species cannot be computed from the value of those parameters, and they
1,2 2,1

increase ever more rapidly until some other limitation is hit.

Think about lichen again. The carrying capacity of each species is essentially zero when they are living separately on the surface of
a rock. Together, however, they can cover the rock, and their joint carrying capacity on Earth is related to the total area of rocks on
Earth—plus other suitable habitat such as the bark of trees, the exterior surface of an automobile junked and neglected for decades,
and myriad other surfaces otherwise uninhabitable. The area of rocks in the world, and other suitable surfaces, certainly is not part
of the ri and si,j parameters! This is quite unlike the situation for competition and predation, in which the joint carrying capacity is
encoded in the parameters. Why this should be is revealed by some diagrams and corresponding mathematics.

17.1.1 https://bio.libretexts.org/@go/page/25527
Figure 17.1.2. Phase space for competition weaker than that of Figure 10.11. The joint equilibrium (black-filled circle) is closer
along the axes to the individual carrying capacities (open circles).

Figure 17.1.3 . Phase space for mild mutualism, causing the opposite effect of that shown in Figure 17.1.2 . The joint equilibrium is
larger than the individual carrying capacities.

17.1.2 https://bio.libretexts.org/@go/page/25527
Figure 17.1.4 Phase space for stronger mutualism, but still restrained, because a joint equilibrium still exists (though far off the
chart at the intersection of the red and blue isoclines).

Figure 17.1.5. With slightly stronger mutualism still, the joint equilibrium vanishes—cannot be determined from the parameters—
and the mutualistic relationship becomes “unrestrained.”

17.1.3 https://bio.libretexts.org/@go/page/25527
Figure 17.1.6. As mutualistic species become more dependent upon each other, their individual carrying capacities become smaller
(open circles here are closer to the origin).

Figure 17.1.7 . Finally, in obligate mutualisms, individual carrying capacities vanish and an Allee point arises (open circle).
Extinction becomes a present danger (closed circle).

This page titled 17.1: Chapter Introduction is shared under a CC BY-NC 4.0 license and was authored, remixed, and/or curated by Clarence
Lehman, Shelby Loberg, & Adam Clark (University of Minnesota Libraries Publishing) via source content that was edited to the style and
standards of the LibreTexts platform; a detailed edit history is available upon request.

17.1.4 https://bio.libretexts.org/@go/page/25527
17.2: Phase Diagrams
The principles can be visualized in phase diagrams, with arrows showing how populations change. Earlier, Figure 10.11 showed
strong competition with coexistence at equilibrium, as could be the case if the species were competing for two different resources.
Nonetheless, each species restricts the other to much lower levels than it could maintain on its own. Weaker competition means that
each species is restricted less so each can maintain higher levels, as in Figure 17.1.2. You can tell this is competition because the
two parameters s and s are both negative. This is shown in the negative slopes of the two diagrams in the upper left. The
1,2 2,1

carrying capacity of each species together is only slightly reduced from the individual carrying capacities, which would be about
1.2 for Species 1 living alone and about 0.8 for Species 2 living alone. (for example, an average of 1.2 individuals per square meter,
or 1,200,000 individuals per square mile if measured in millions). However, for species living together, the carrying capacity of
each is slightly reduced, perhaps 10 to 20 percent.
In this case, by the way, the two species together have a higher total population than would be the case if either was living alone.
This is called “over yielding,” and is a recurrent theme in studies of plant communities.
Figure 17.1.3 shows a similar situation, but now with the inter-species interaction terms s and s both positive, shown by the
1,2 2,1

positive slopes in the two upper-left diagrams of the figure. It looks quite similar to Figure 17.1.2, but the two together are each
more abundant than they would be apart— the joint equilibrium is larger than the individual carrying capacities.
This joint equilibrium can be computed from the r and s parameters. It will occur where the growth of each species
i i,j

simultaneously reaches 0. You can find the numerical value for this equilibrium with pencil and paper by setting the first species
growth rate to 0, solving for the populations of Species 1, substituting that into the equation for Species 2, and solving for when the
growth of that species reaches 0. Alternatively, you can pose the problem to a symbolic mathematics program and ask it to solve
the two equations simultaneously. In any case, you would start with both growth rates set to zero at equilibrium,
1 dN1
= r1 + s1,1 N1 + s1,2 N2 = 1.2 − 1 N1 + 0.1 N2 = 0 (17.2.1)
N1 dt

1 dN2
= r2 + s2,2 N2 + s2,1 N1 = 0.8 − 1 N2 + 0.1 N1 = 0 (17.2.2)
N2 dt

and end up with N = 1.2929 and N = 0.929.

1 2

As mutualisms become stronger—meaning that the interspecific interactions become more positive—the equilibrium point moves
further out. It can be very large, as in in Figure 17.1.4, but, in restrained mutualism, the equilibrium is finite and computable from
the parameters of the individual species.
On the other hand, when the inter-species enhancement terms are stronger still, a bifurcation occurs and the joint equilibrium
ceases to exist at all. (Figure 17.1.5). The calculated equilibrium point has, in effect, moved to infinity, or in a sense beyond,
meaning that the carrying capacity cannot be computed from the parameters of the species and their interactions. Some further
information is needed about the system.
Beyond this, the mutualists can become more dependent on each other, so that the r terms become smaller, as in Figure 17.1.6, or
i

negative, as in Figure 17.1.7. The mutualism can be unrestrained even if the intrinsic growth rates r are negative. What arises is a
i

kind of Allee point, where the populations run away if they start above that point, but decline to extinction if they start below.

This page titled 17.2: Phase Diagrams is shared under a CC BY-NC 4.0 license and was authored, remixed, and/or curated by Clarence Lehman,
Shelby Loberg, & Adam Clark (University of Minnesota Libraries Publishing) via source content that was edited to the style and standards of the
LibreTexts platform; a detailed edit history is available upon request.

17.2.1 https://bio.libretexts.org/@go/page/25528
17.3: Quasi-species
A new phenomenon arises in equations for unrestrained mutualism, one not possible for competition or predation. It can be shown
that unrestrained mutualists are pulled into fixed ratios which are based on how their interaction terms differ from their self-
limiting terms. As populations grow and move toward these fixed ratios, the two equations governing the individual species
collapse, in effect, into an equation of a single-species. This single species grows orthologistically, a form of growth you saw
earlier in Figure 4.3. The system behaves mathematically as if only one species were participating.
What does the mathematical collapse to a single equation mean biologically? It suggests that the biological world could respond
analogously—that two actual species growing as unrestrained mutualists could intermingle into a single quasi-species—at least into
the best approximation of a single species that biology could accomplish using genetically different entities. Lichen and the
eukaryotic cell are examples, genetically separate but biologically merged. Indeed, lichens were thought to be individual species
until the nineteenth century, and the eukaryotic cell was only accepted as the result of mutualistic combinations late in the
twentieth.
It therefore appears that natural selection has not overlooked this possibility. In a community model with mutualistic species locked
in approach to a singularity, the two mutualists may be replaced in the model by a new quasi-species, representing the two species
jointly but ultimately having a non-singular form of population growth. The resulting quasi-species may grow without obvious
inhibition toward a singularity, then switch to a different model, as you have seen for single-species models in Figure 4.4 and for
our own species in Figure 6.3.

This page titled 17.3: Quasi-species is shared under a CC BY-NC 4.0 license and was authored, remixed, and/or curated by Clarence Lehman,
Shelby Loberg, & Adam Clark (University of Minnesota Libraries Publishing) via source content that was edited to the style and standards of the
LibreTexts platform; a detailed edit history is available upon request.

17.3.1 https://bio.libretexts.org/@go/page/25529
17.4: Our own species

Figure 17.4.1. One of the most conspicuous mutualisms on the planet involves our species. Our mutualisms are biological, but have
become technological as well, with teams of satellite-guided tractors navigating our fields with the power of a thousand horses.

We reached our present numbers through a complex of mutualisms. A variety of increasingly refined crops and increasing
domesticated animals acted as mutualists, switching our population dynamics from boom-and-bust predator–prey cycling (Figure
17.4.2A) to prolonged runaway growth (17.4.2B). Through cultivation of plants and animals that were formerly prey, agriculture

increased the carrying capacity of the cultivated species, which in turn increased human carrying capacity, forming a positive
feedback loop. It may seem dubious to consider domesticated plants and animals as our mutualists, but their parameters meet the
mathematical requirements of mutualism. And both populations can expand so rapidly that they would soon exceed all former
bounds.

Figure 17.4.2 . Stages in human population dynamics. Representative predator abundance is shown in red, prey abundance in
green, and abundance and our ancestors in blue.

Mutualisms in nature typically do not long follow the unrestrained growth toward astronomical levels shown in Figure 17.4.2B.
Instead, natural mutualisms increase only until they are checked by some other force, such as predators or disease. Our ancestors,
however, kept their mutualisms pure by eliminating third-species interferences. They hunted and killed predators of domesticated
animals, weeded crops to eliminate plant competitors, and fenced crops to exclude herbivores. Indeed, the human species does not
now dominate the earth, as often suggested. The mutualisms dominate. The joint biomass of cows, horses, pigs, chickens, and dogs
exceeds the biomass of humans, to say nothing of the biomass of crops. We did not get here alone.
We have recently outgrown the need for animal mutualists to power plows and support locomotion, substituting engines consuming
fossil fuels instead. These are like synthetic mutualists. We also outgrew the need for animals to fertilize crops, substituting
artificial nitrogen fertilizer created in Haber–Bosch process furnaces from the vast supply of nitrogen in the air. The need for
animals and plants for clothing diminished with synthetic cloth and furs. All of these brought us to a partially non-biological world,
still developing, where not even the carrying capacities of our living mutualists need be limiting.

17.4.1 https://bio.libretexts.org/@go/page/25530
However, advancing medicine and public health eventually decreased child mortality. Death rates fell and, later, birth rates began to
fall even faster. The mutualisms remain, but—unexpectedly and abruptly— twelve millennia of rapidly accelerating population
growth ended (Figures 6.3, 17.4.2C).
As a closing note, in the twenty-first century, after only about fifty years of decelerating growth, the global population continues to
increase. But, births in many societies have dropped below replacement levels in a final stage characterized by declines in the
resident populations, compensated in many countries by immigration. The relative rapidity of this change may spill over into social
disjunction between conditions of the present and memories of times past. It may also foretell a future time when our species may
voluntarily reduce populations to sustainable values yet unknown (dotted lines in Figure 17.4.2C).

Figure 17.4.3 . A complex of mutualisms are represented here. What are they? Dog and human together are one pairwise
mutualism, among the oldest for our species. Sheep and humans are another. What else? Grass with sheep keeping the forest from
invading? Humans protecting forests for firewood? Anything else?

This page titled 17.4: Our own species is shared under a CC BY-NC 4.0 license and was authored, remixed, and/or curated by Clarence Lehman,
Shelby Loberg, & Adam Clark (University of Minnesota Libraries Publishing) via source content that was edited to the style and standards of the
LibreTexts platform; a detailed edit history is available upon request.

17.4.2 https://bio.libretexts.org/@go/page/25530
 CHAPTER OVERVIEW

18: Higher-order models

18.1: Chapter Introduction
18.2: One-species Example
18.3: Two-species blending
18.4: A way of organizing

This page titled 18: Higher-order models is shared under a CC BY-NC 4.0 license and was authored, remixed, and/or curated by Clarence
Lehman, Shelby Loberg, & Adam Clark (University of Minnesota Libraries Publishing) via source content that was edited to the style and
standards of the LibreTexts platform; a detailed edit history is available upon request.

1
18.1: Chapter Introduction
The RSN formulation of Equations 4.1 and 8.1 incorporate the classical competition, predation, and logistic equations of ecology,
plus mutualistic and orthologistic equations that are not part of the classical set. For this, the parameters can remain constant.
But they are not constant in all cases. Volterra (1928) considered parameters that vary through time, but more generally the
parameters can be functions of population density as well. In this way the RSN formulation can be extended to cover not just the
classical equations of ecology and related equations, but any equations of ecology whatever, as with Equation 4.3, proposed by
Kolomogorov.

This page titled 18.1: Chapter Introduction is shared under a CC BY-NC 4.0 license and was authored, remixed, and/or curated by Clarence
Lehman, Shelby Loberg, & Adam Clark (University of Minnesota Libraries Publishing) via source content that was edited to the style and
standards of the LibreTexts platform; a detailed edit history is available upon request.

18.1.1 https://bio.libretexts.org/@go/page/25534
18.2: One-species Example
To see the basic idea, start with a single species. With variable parameters, population growth for a single species can be written as
follows, with the growth rate r and the density dependence term s depending on the population level N and optionally on the time
t.

1 dN
= r(N , t) + s(N , t)N (18.2.1)
N dt

Because anything can be embedded in the term r(N , t), or in the corresponding s(N , t) , the equation is perfectly general and can
cover any ecological situation for a single species modeled by a differential equation. A basic example, which you have already
seen for human population growth, is where the parameters are approximately constant for long periods, but change at certain
“bifurcation” events. The parameters for human population growth changed abruptly at the beginning of the modern era, resulting
in overall population dynamics that were neither orthologistic nor logistic, but a piecewise combination of the two. In that case the
two were blended by varying the parameters as follows.

1 dN − 0.001185 + 0.00684N , when N ≤ 3.28 billion

={
N dt
0.03077 − 0.00289N , when N > 3.28 billion

This led to the population growth curve of Figure 6.3, which nicely modeled human population growth over the centuries.
Human population growth called for a piecewise blending of the parameters, because the parameters changed rather abruptly from
one constant set to another. The parameters can also be blended continuously for parameters that change gradually.
For example, take an orthologistic equation, 1/N dN /dt = −2 + 2N , and a logistic equation, 1/N dN /dt = 4 − 2N , and
consider N as it ranges from 0 to 1. Have the orthologistic equation apply exactly as N approaches 0, and the logistic equation
apply exactly when N reaches 1. Then let r change uniformly from −2 to +4 and s change uniformly from +2 to −2 as N goes from
0 to 1, as follows.
r(N , t) = 6N − 2

s(N , t) = −4N + 2

Plugging that into Equation 18.1 gives

1 dN
= (6N − 2) + (−4N + 2)N (18.2.2)
N dt
2
= −2 + 8N − 4N (18.2.3)

2
= r + sN = s2 N (18.2.4)

This kind of blending between orthologistic and logistic has simply added one more term to the population growth equation, an N 2

term—one of the terms proposed by Hutchinson (Equation 4.2). The result is graphed in Figure 18.2.1.

Figure 18.2.1 . Orthologistic and logistic equations smoothly blended.

This smoothly blends the orthologistic, which has an Allee point but no carrying capacity, with the logistic, which has a carrying
capacity with no Allee point, providing both in the blended curve of Figure 18.2.1. The curve has an Allee point at about N = 0.3
and a carrying capacity at about N = 1.7 . Compare this with the piecewise blending depicted earlier in Figure 4.4.

This page titled 18.2: One-species Example is shared under a CC BY-NC 4.0 license and was authored, remixed, and/or curated by Clarence
Lehman, Shelby Loberg, & Adam Clark (University of Minnesota Libraries Publishing) via source content that was edited to the style and
standards of the LibreTexts platform; a detailed edit history is available upon request.

18.2.1 https://bio.libretexts.org/@go/page/25535
18.3: Two-species blending
Blending two-species systems is a similar process, but has more options in the parameters. Equation 18.3.1 is an example with
limited options that produced the phase spaces in Figures 10.1.3 through 10.1.5.
1 dN1
= r1 (N1 ) + s1,1 N1 + s1,2 (N1 )N2 (18.3.1)
N1 dt

1 dN2
= r2 (N2 ) + s2,2 N2 + s2,1 (N2 )N1
N2 dt

Changing the parameters uniformly from a b value when the corresponding N value is 0 to an a + b value when the
corresponding N value is 1 is analogous to the blending that produced Figure 4.4.1. The parameters would vary as follows, using
four distinct a values (a , a , a , a ), plus four distinct b values with matching subscripts (b , b , b , b ).
1 2 1,2 2,1 1 2 1,2 2,1

r1 (N1 ) = a1 N1 + b1 , s1,2 (N1 ) = a1,2 N1 + b1,2

r2 (N2 ) = a2 N2 + b2 , s2,1 (N2 ) = a2,1 N2 + b2,1

Substituting the above into Equation 18.3.1 and collecting terms gives an equation having all the RSN terms present, but now with
a cross-product in terms of N N added at the end:
1 2

1 dN1
= b1 + (a1 + s1,1 )N1 + b1,2 N2 + a1,2 N1 N2 (18.3.2)
N1 dt

1 dN2
= b2 + (a2 + s2,2 )N2 + b2,1 N1 + a2,1 N1 N2
N2 dt

In the specific case of Figures 10.1.3 through 10.1.5, we used s 1,1 = s2,2 = −0.98 and
r1 (N1 ) = 0.75 N1 − 0.5 s1,2 (N1 ) = −1.15 N1 + 2.5

r2 (N2 ) = 0.75 N2 − 0.5 s2,1 (N2 ) = −0.45 N2 + 1.3

which gave
1 dN1
= −0.5 − 0.23 N1 + 2.50 N2 − 1.15 N1 N2
N1 dt

1 dN2
= −0.5 − 0.98 N2 + 2.50 N1 − 0.45 N1 N2
N2 dt

for the ow in the figures.

This page titled 18.3: Two-species blending is shared under a CC BY-NC 4.0 license and was authored, remixed, and/or curated by Clarence
Lehman, Shelby Loberg, & Adam Clark (University of Minnesota Libraries Publishing) via source content that was edited to the style and
standards of the LibreTexts platform; a detailed edit history is available upon request.

18.3.1 https://bio.libretexts.org/@go/page/25536
18.4: A way of organizing
The methods we have described in this chapter provide an organized way of designing higher-order
models of ecological interactions. Characteristics of the individual models purely in RSN form can
be mixed and matched as needed to produce the higher-order model, then the resulting parameters
can be fit to observations or modeling needs.
For example, the bistable phase space in Figures 10.1.3 through 10.1.5, with an unstable Allee point
combined with a stable carrying capacity, was easy to achieve by combining sub-models, each with
one of those characteristics. The result was simply an RSN system with an N N cross-product term added, but how easy would it
1 2

have been to conceive that term a priori and perceive its sufficiency?
Moreover, a blended model in the form of Equation 18.3.1 can be visualized in terms of (1) the intrinsic growth rate of each
species, (2) how each species affects the other, (3) how each species affects itself, and (4) how those effects change with population
densities. In a way, any higher-order system constructed in this way can be viewed as an RSN system at every point, yet as a
higher-order system overall.
In summary, we cautiously recommend this method of creating and organizing ecological models, and are eager to hear of
successes and failures you might have with it.
The characters acting in this book—the equations of population growth, the parameters and their relationships, the graphs and
phase planes—live in variations and disguises within an enormous variety of actual ecological applications, detailed in thousands
of scientific papers you will be able to sample as your journey continues. We hope you have enjoyed this introductory book and
that it has been useful. We welcome your comments on any aspects at all. An ebook is never finished but can constantly be
improved.
—C.L., S.L., A.C.

This page titled 18.4: A way of organizing is shared under a CC BY-NC 4.0 license and was authored, remixed, and/or curated by Clarence
Lehman, Shelby Loberg, & Adam Clark (University of Minnesota Libraries Publishing) via source content that was edited to the style and
standards of the LibreTexts platform; a detailed edit history is available upon request.

18.4.1 https://bio.libretexts.org/@go/page/25537
Index
A L R
Allee point Lotka–Volterra model realized niche
10.2: Phase Space 12.3: Predator-prey model 16.2: The niche concept
16.8: Lotka-Volterra formulation
C S
competition M SIR model
9.7: Competition mutualism 15.2: The SIR flowchart
competitive exclusion principle 9.2: Mutualism state spaces
16.4: Competitive exclusion 11.1: State Spaces
N
F niche T
fundamental niche 16.2: The niche concept The I Equation
16.2: The niche concept 15.5: The I Equation
P
H phase space Z
Herbivory 10: Phase Space zonation
10.1: Chapter Introduction
9.3: Herbivory 16.2: The niche concept
10.2: Phase Space
12.4: Phase Space
Polio
14.5: The Strange Case of Polio

1 https://bio.libretexts.org/@go/page/27887
Glossary
Sample Word 1 | Sample Definition 1

1 https://bio.libretexts.org/@go/page/39883
Detailed Licensing
Overview
Title: Quantitative Ecology - A New Unified Approach (Lehman, Loberg, and Clark)
Webpages: 121
Applicable Restrictions: Noncommercial
All licenses found:
CC BY-NC 4.0: 97.5% (118 pages)
Undeclared: 2.5% (3 pages)

By Page
Quantitative Ecology - A New Unified Approach (Lehman, 6.1: Chapter Introduction - CC BY-NC 4.0
Loberg, and Clark) - CC BY-NC 4.0 6.2: Phenomological Graph - CC BY-NC 4.0
Front Matter - CC BY-NC 4.0 6.3: Biological-ecological graph - CC BY-NC 4.0
TitlePage - CC BY-NC 4.0 6.4: A global transition - CC BY-NC 4.0
InfoPage - CC BY-NC 4.0 7: Chaos and Randomness - CC BY-NC 4.0
Table of Contents - Undeclared 7.1: Chapter Introduction - CC BY-NC 4.0
Licensing - Undeclared 7.2: Depicting population growth - CC BY-NC 4.0
1: What is Ecology? - CC BY-NC 4.0 7.3: Hypothetical insect data - CC BY-NC 4.0
7.4: Sensitive Dependence - CC BY-NC 4.0
1.1: Introduction - CC BY-NC 4.0
7.5: Dampening of Chaos - CC BY-NC 4.0
1.2: Definitions of Ecology - CC BY-NC 4.0
7.6: Bifurcation diagram - CC BY-NC 4.0
1.3: Ecology then and now - CC BY-NC 4.0
7.7: Properties - CC BY-NC 4.0
1.4: Methods of ecology - CC BY-NC 4.0
2: Ecological Theory - CC BY-NC 4.0 8: Theory of Interactions - CC BY-NC 4.0

2.1: Levels of Ecology - CC BY-NC 4.0 8.1: Dynamics of two Interacting Species - CC BY-
2.2: Role of theory - CC BY-NC 4.0 NC 4.0
2.3: What is a model? - CC BY-NC 4.0 8.2: Code for two species - CC BY-NC 4.0
2.4: Present State - CC BY-NC 4.0 8.3: Summary of interactions - CC BY-NC 4.0

3: A Basic Population Model - CC BY-NC 4.0 9: Embodied by Natural Selection - CC BY-NC 4.0

3.1: Characterizing populations - CC BY-NC 4.0 9.1: Chapter Introduction - CC BY-NC 4.0
3.2: Bacterial Growth - CC BY-NC 4.0 9.2: Mutualism - CC BY-NC 4.0
3.3: A First Computer Model - CC BY-NC 4.0 9.3: Herbivory - CC BY-NC 4.0
3.4: Program results - CC BY-NC 4.0 9.4: Predation - CC BY-NC 4.0
3.5: Darwin's elephants - CC BY-NC 4.0 9.5: Predatory plants - CC BY-NC 4.0
9.6: Defense - CC BY-NC 4.0
4: Modeling a Single Population - CC BY-NC 4.0
9.7: Competition - CC BY-NC 4.0
4.1: Density-independent growth - CC BY-NC 4.0
10: Phase Space - CC BY-NC 4.0
4.2: Density-enhanced growth - CC BY-NC 4.0
10.1: Chapter Introduction - CC BY-NC 4.0
4.3: Density-limited growth - CC BY-NC 4.0
10.2: Phase Space - CC BY-NC 4.0
4.4: Parameter combinations - CC BY-NC 4.0
10.3: Surface generating the flow - CC BY-NC 4.0
4.5: Generalization - CC BY-NC 4.0
10.4: A phase space example - CC BY-NC 4.0
5: Differential and Difference Forms - CC BY-NC 4.0
11: State Spaces - CC BY-NC 4.0
5.1: Chapter Introduction - CC BY-NC 4.0
5.2: A mathematical view - CC BY-NC 4.0 11.1: State Spaces - CC BY-NC 4.0
5.3: Exponential Solution - CC BY-NC 4.0 12: Predator and Prey - CC BY-NC 4.0
5.4: Logistic solution - CC BY-NC 4.0 12.1: Further Examples of Predation - CC BY-NC 4.0
5.5: Orthologistic solution - CC BY-NC 4.0 12.2: Ecological communities are complex - CC BY-
6: Human Population Growth - CC BY-NC 4.0 NC 4.0

1 https://bio.libretexts.org/@go/page/97894
 12.3: Predator-prey model - CC BY-NC 4.0 15.7: 15.7 The rsN model again - CC BY-NC 4.0
12.4: Phase Space - CC BY-NC 4.0 15.8: Application to an actual outbreak - CC BY-NC
12.5: Assumption of the basic model - CC BY-NC 4.0 4.0
12.6: Independent carrying capacities - CC BY-NC 4.0 15.9: Conflicting ethics - CC BY-NC 4.0
12.7: Predator satiation and starvation - CC BY-NC 16: Competition - CC BY-NC 4.0
4.0 16.1: Chapter Introduction - CC BY-NC 4.0
12.8: Effects of Space - CC BY-NC 4.0 16.2: The niche concept - CC BY-NC 4.0
13: Humans as Predators - CC BY-NC 4.0 16.3: Resource competition - CC BY-NC 4.0
13.1: Chapter Introduction - CC BY-NC 4.0 16.4: Competitive exclusion - CC BY-NC 4.0
13.2: Consciously Controlled Predation - CC BY-NC 16.5: Succession - CC BY-NC 4.0
4.0 16.6: Single-resource phase space - CC BY-NC 4.0
13.3: Stochastic modeling - CC BY-NC 4.0 16.7: Multiple resource phase space - CC BY-NC 4.0
13.4: Present Situation - CC BY-NC 4.0 16.8: Lotka-Volterra formulation - CC BY-NC 4.0
14: Humans as Prey - CC BY-NC 4.0 17: Mutualism - CC BY-NC 4.0
14.1: Chapter Introduction - CC BY-NC 4.0 17.1: Chapter Introduction - CC BY-NC 4.0
14.2: Portals - CC BY-NC 4.0 17.2: Phase Diagrams - CC BY-NC 4.0
14.3: Pathogen Mobility - CC BY-NC 4.0 17.3: Quasi-species - CC BY-NC 4.0
14.4: Disease more generally - CC BY-NC 4.0 17.4: Our own species - CC BY-NC 4.0
14.5: The Strange Case of Polio - CC BY-NC 4.0 18: Higher-order models - CC BY-NC 4.0
14.6: Modern disease trends - CC BY-NC 4.0 18.1: Chapter Introduction - CC BY-NC 4.0
14.7: An ancient plague perhaps vanishing - CC BY- 18.2: One-species Example - CC BY-NC 4.0
NC 4.0 18.3: Two-species blending - CC BY-NC 4.0
15: Theory of Disease - CC BY-NC 4.0 18.4: A way of organizing - CC BY-NC 4.0
15.1: New terminology - CC BY-NC 4.0 Back Matter - CC BY-NC 4.0
15.2: The SIR flowchart - CC BY-NC 4.0
Index - CC BY-NC 4.0
15.3: The SIR Equations - CC BY-NC 4.0
Glossary - CC BY-NC 4.0
15.4: The SI Equations - CC BY-NC 4.0
Detailed Licensing - Undeclared
15.5: The I Equation - CC BY-NC 4.0
15.6: Analysis of the I model - CC BY-NC 4.0

2 https://bio.libretexts.org/@go/page/97894