A Transitional-Combinational Theory for the Origin of Angiosperms

Author(s): Tod F. Stuessy

Reviewed work(s):
Source: Taxon, Vol. 53, No. 1 (Feb., 2004), pp. 3-16
Published by: International Association for Plant Taxonomy (IAPT)
Stable URL: http://www.jstor.org/stable/4135484 .
Accessed: 08/11/2012 07:30

Your use of the JSTOR archive indicates your acceptance of the Terms & Conditions of Use, available at .
http://www.jstor.org/page/info/about/policies/terms.jsp

.
JSTOR is a not-for-profit service that helps scholars, researchers, and students discover, use, and build upon a wide range of
content in a trusted digital archive. We use information technology and tools to increase productivity and facilitate new forms
of scholarship. For more information about JSTOR, please contact [email protected].

International Association for Plant Taxonomy (IAPT) is collaborating with JSTOR to digitize, preserve and
extend access to Taxon.

http://www.jstor.org
TAXON 53 (1) o February2004: 3-16 Stuessy o Theory for the origin of angiosperms

ANISPR EOUTO

A transitional-combinational theory for the origin of angiosperms

Tod F. Stuessy

Department of Higher Plant Systematics and Evolution, Institute of Botany and Botanical Garden, University
of Vienna, Rennweg 14, A-1030, Vienna, Austria. [email protected]

The origin of the angiospermshas long remainedproblematical.Recent new discoveries of early angiosperm
fossils and many new molecularphylogenies have helpedto identifybasal extantangiospermsandto infertheir
inter-relationships.Still unclear,however, are ties with gymnosperms,both living and fossil. In view of diffi-
culty of finding ancestorsfor angiosperms,yet also consideringtheir sudden appearanceand explosive evolu-
tionary success, a transitional-combinationaltheory is proposed. This theory suggests that angiosperms
evolved slowly from seed ferns in the Jurassicbeginning first with the carpel, followed laterby double fertil-
ization, and lastly by the appearanceof flowers. These three fundamentaltransitionsmay have takenmore than
100 million years to complete. The extant (or modem) angiospermsdid not appearuntil EarlyCretaceous,as
attested by micro- and macrofossils, when the final combinationof all three importantangiospermfeatures
occurred.This combinationprovidedthe opportunityfor explosive evolutionarydiversification,especially in
response to selection from insect pollinatorsand predators,plus attendantmodificationsin compatibilityand
breedingsystems. The transitional-combinational theory explains some of the discrepancybetween fossil and
molecular phylogenetic data, the latterof which suggests that angiospermsoriginatedpre-Cretaceous.DNA
(and protein)sequence data and oleananesreflect ancestryin an earliertransitionalcarpellatelineage (or line-
ages), much earlierthan appearanceof the final combinationthat includes floral and associated pollen modi-
fications. The theory suggests viewing all gymnosperms, other than extinct seed ferns from which carpels
arose, as having had no direct phylogenetic connections to modern angiosperms.

KEYWORDS: angiosperms, anthophytes, floral origins, phylogeny, seed plants.

ing cones, such as Coniferophytaand Gnetales. Logical

INTRODUCTION extension of this theory resulted in hypothesizing
Angiosperms, often called flowering plants, domi- Casuarinaceae as basal for angiosperms (Wettstein,
nate the earth. This large group of more than 250,000 1907, 1924) because of their cone-like female inflores-
species (Thorne, 2002) is prominentin most vegetation- cences and Equisetum-likefoliage. Morphological and
al zones ranging from deserts to tropical rain forests. anatomical (Moseley, 1948; Cronquist, 1981; Flores &
Angiosperms contribute much of earth's terrestrialbio- Moseley, 1982; Johnson& Wilson, 1989), palynological
mass, and even more importantly,they providenumerous (Doyle, 1969), and molecularstudies (e.g., Chase & al.,
foods, fibers and pharmaceuticals,upon which human 1993), however, reveal this family to be firmly embed-
survival depends. Despite this obvious importance,the ded within mainstreameudicots (EurosidsI;Angiosperm
evolutionary origins of angiosperms are still puzzling. Phylogeny Group, 1998, 2003; Judd & al., 2002) rather
Solutions to this problem remain elusive despite new than being a basal angiosperm. A modification of this
macro- and microfossils, new interpretationsof homolo- theory suggests origin of angiospermsfrom Gnetales,an
gy of morphologicaland anatomicalfeatures, more pre- extant specialized group of gymnosperms consisting of
cise morphologicalcladistic analyses of fossil and recent the generaEphedra,Gnetumand Welwitschia,or perhaps
forms, and new molecular analyses among extant repre- at minimum sharing a common ancestor with
sentatives. angiosperms (e.g., Zavada & Gabarayeva,1991; Nixon
The difficulty of understandingthe origin of the & al., 1994; Hickey & Taylor, 1996).
angiosperms has led to several different major theories. The second majortheory of angiospermorigins, the
The first theory, the pseudanthialtheory, developed by euanthial theory (also called the anthophyte,strobiloid,
Eichler (1876) and Engler (1897) and collaborators, or Ranalean theory), views much of the former
emphasized that angiosperms derived from gym- Magnoliidae(sensu Cronquist,1981), or "basalfamilies"
nosperms(i.e., otherseed plants),particularlythose bear- (sensu Judd & al., 2002) plus Magnoliales and

3
Stuessy * Theory for the origin of angiosperms TAXON 53 (1) s February2004: 3-16

Canellales, as basal among extant taxa. This theory,pro- ing many homologies with inflorescences of
posed by Delpino (1890; cited and reviewed in Endress, Chloranthaceae(his "chloranthoidorigin";the family is
1993) and Bessey (1893, 1897, 1915; supportedalso by now regarded as belonging in the basal angiosperm
Arber & Parkin, 1907), envisioned extinct Bennettitales grade, Angiosperm Phylogeny Group, 1998, 2003; Qiu
(Cycadeoidales), or their common ancestor,as progeni- & al., 1999). Melville (1962, 1975) and Retallack &
tor. Under this theory,carpels and stamens are modified Dilcher (1981) stressed a possible evolutionary origin
from leafy mega- and microsporophylls, respectively, from pteridospermousGlossopteridales (for details on
yielding a helically arrangedgynoecium and fleshy sta- fossilized pollen tubes and sperms, see Nishida & al.,
mens (for descriptionsand illustrationsof modem groups 2003). Thomas(1936), Gaussen(1946), Stebbins(1974),
with these features, see Endress & Hufford, 1989; and Doyle (1978, 1996) have emphasized another seed
Igersheim& Endress, 1997; Endress& Igersheim,2000). fern group, Caytoniales, as ancestral.
Recent molecular studies supportsome members of the
former Magnoliidae as belonging among basal
angiosperms(Qiu & al., 1993, 1999, 2000), but progeni-
tors among gymnospermous groups, either extinct or EXISTING
DATAAND ANALYSES
extant, have not been identified. Most cladistic morpho- Existing information,plus new data and new quanti-
logical analyses involving Bennettitales do not regard tative (cladistic) analyses, have providedrefined insights
them as directly sister to angiosperms (Doyle & on angiospermorigins even though the specific solution
Donoghue, 1992; Nixon & al., 1994; Doyle, 1998a). remainselusive. One of the most importanttypes of data
The third theory of angiospermorigins, what might for suggesting the age of angiosperms and their early
be called the herbaceous(or paleoherb;Taylor& Hickey, morphology has been, and continues to be, fossil pollen.
1992, 1996) hypothesis, regards the earliest representa- The excellent condition and abundanceof fossil pollen
tives of the group to have been herbaceous with small makes them extremely important(Muller, 1970, 1981;
and simple flowers possessing one carpel, two stamens, Crane & Lidgard, 1990; Doyle, 1999). Two points are
and one perianthpart[e.g., Meeuse, 1975 (he called it the clear from microfossils: (1) the age of the first recogniz-
"anthocorm"hypothesis,but this consisted of many addi- able angiospermsis c. 130 million years old (Barremian,
tional postulates); Taylor & Hickey, 1990, 1992]. Hughes & al., 1979; or at least Hauterivian or
Coalescence of these simple units could have resulted in Valanginian, Hughes & McDougall, 1987; Hughes,
many combinationsleading to modem angiospermfami- 1994; Brenner, 1996; Friis & al., 2000), i.e., early
lies (Burger,1977), includingmonocots. Thatsome basal Cretaceous (a number of studies have reported pre-
(or at least early) angiosperm families (Angiosperm Cretaceousangiospermremains, but these fossils and/or
Phylogeny Group, 1998, 2003) have small flowers of this dates have not remained convincing or at least they
nature (e.g., Ceratophyllaceae, Les, 1988, Les & al., require further scrutiny, e.g., Comet, 1989a, b, 1993);
1991; Lactoridaceae,Stuessy & al., 1998) support this and (2) the earliest pollen is monosulcate, in conformity
hypothesis as do the ovule and carpellary studies of with several hypotheses for the origin of angiosperms.
Taylor (1991). Paleoherbs,however, are a very diverse The earliest known fossil angiospermmicrofossils (com-
group, as attested by variations in gynoecium structure parable to modem taxa) to date, however, Clavatipol-
(Igersheim& Endress, 1998). lenites (Muller, 1981), are Chloranthaceous (Doyle,
The fourthalternativefor angiospermorigins is that 1969; Walker& Walker,1984; Chapman,1987; Endress,
of multiple derivations from gymnospermousancestors 1987), more supportive of the herbaceous (paleoherb)
either in morphologically similar or different lines hypothesis.
(Coulter, 1903; Gaussen, 1958; Meeuse, 1966, 1970, Recent macrofossils have also helped clarify the
1975; Krassilov, 1977, 1997; Hughes, 1994; Wu & al., early evolution of the angiosperms. Some early
2002). Arguments for this hypothesis focus on the angiosperm fossils from mid-Cretaccous (Albian-
diverse nature of basal angiosperms including woody Cenomanian) possess features of the woody magnoliid
and herbaceousassemblages.Against this position, how- complex (sensu Judd& al., 2002), such as Archaeanthus
ever, are sharedreproductivecharactersof carpel, flower, (100-95 Mya; Dilcher & Crane, 1984), Lesqueria (c. 95
and double fertilization,all of which taken together has Mya; Crane & Dilcher, 1984), and Virginianthus(c. 105
always seemed too complex to have originatedmore than Mya; Friis & al., 1994), and Laurales (c. 100 Mya;
once. Drinnan& al., 1990), but fossils of"paleoherbs"are also
Additional suggestions exist for the origin of known (e.g., Chloranthaceae,Friis & al., 1987a, 2000,
angiosperms from specific gymnospermous ancestors. possibly also Chloranthaceae[or Piperales],c. 110 Mya,
For example, Leroy (1983a, b) postulated that Taylor & Hickey, 1990; Nymphaeales, Friis & al., 2001)
angiospermsevolved from Coniferophytaancestors,see- as well as other early groups (e.g., c. 100 Mya; Platanus-

4
TAXON 53 (1) * February2004: 3-16 Stuessy * Theory for the origin of angiosperms

like; Crane & al., 1986). In addition, the recently inves- & Olmstead, 2000; Zanis & al., 2002, 2003; see
tigated Barremian or Aptian floras of Portugal are Goremykin& al., 2003, however, for indications,based
extremely rich with many angiosperm floral structures on its entire chloroplast sequence, that Amborella may
(e.g., Friis & al., 1999, 2000). A most exciting recentdis- not be a basal group)]. What molecular data have not
covery has been Archaefructus, assigned to its own been able to do, however, is to reveal from which gym-
angiosperm family, Archaefructaceae(Sun & al., 1998, nospermous group the angiosperms have arisen
2002), and being about 125 million years old. Although (Axsmith & al., 1998). Relationshipsamong extantgym-
earlier microfossil (e.g., Hughes & al., 1979; Muller, nospermsare still not clearly resolved (e.g., Chaw & al.,
1981; Hughes & McDougall, 1987; Hughes & al., 1991) 1997, 2000; Donoghue & Doyle, 2000; Sanderson& al.,
and equally old or older macrofossil (e.g., Doyle & 2000; Magall6n & Sanderson,2002; Rydin & al., 2002;
Hickey, 1976; Hickey & Doyle, 1977; Friis & al., 2000) D. Soltis & al., 2002), and importantextinct groups can-
remains of angiosperms are known, Archaefructusis a not be analyzed molecularly (e.g., Bennettitales,
key fossil because it possesses carpels but not flowers, at Caytoniales).
least not in the usual sense. It lacks perianthparts, and Moleculardata not only provide strongevidence for
the carpels and stamens are separatedvertically on the relationships among basal angiosperms, but they also
elongated reproductiveaxis. One can debate whetherthe suggest pre-Cretaceousorigin for the group. Early stud-
entire reproductiveaxis of Archaefructusmight also be ies by Martin& al. (1989), applying a molecularclock in
interpretedas an inflorescence consisting of unisexual GAPDH sequences, suggested monocot-dicot diver-
flowers without perianthparts,and if so, the taxon might gence more than 300 million years ago, i.e., Carbonifer-
be regarded as more specialized rather than the basal ous. Troitsky& al. (1991), using rRNA sequences, esti-
family of angiosperms(Friis & al., 2003). mated gymnosperm-angiospermdivergence at c. 360
Recent morphologicalphylogenetic analyses among million years, i.e., Devonian-Carboniferousboundary.
living and extinct vascular plants have helped improve Wolfe & al. (1989) from several cpDNA gene sequences
understandingof relationships (Crane, 1985; Doyle & estimated the monocot-dicot divergence at c. 190 Mya
Donoghue, 1986, 1992; Loconte & Stevenson, 1990; (Early Jurassic;the earliest monocot fossils are known
Nixon & al., 1994; Rothwell & Serbet, 1994; Doyle, from the Upper Cretaceous,c. 90 Mya, Gandolfo & al.,
1996, 1998a; Kenrick & Crane, 1997). All studies re- 1998). Despite problemswith molecularclocks especial-
emphasize the origin of angiosperms from gym- ly involving rate constancy(e.g., Gaut, 1998; P. Soltis &
nosperms, in contrastto the creative hypothesis of Kato al., 2002), these data are certainly still provocative.
(1990) in which they are hypothesizedas having derived Sanderson& Doyle (2001) have recentlyaddressedsome
from ferns (Ophioglossales). The specific relationships of these problems in a re-evaluation of 18S and rbcL
among gymnospermsand their affinities to angiosperms DNA sequence data in embryophytes,with focus on age
vary with the particularanalysis, but there is a tendency of angiosperms. Using maximum likelihood methods,
for Gnetales to be viewed as closest living relatives of different rate assumptions,and differentsampled taxa, a
angiosperms(Doyle, 1998a, b). These studies also show range of 68-281 Mya for the origin of angiospermshas
close, but confusing, ties of "seed ferns" to the rest of been obtained, with most estimates falling between
gymnosperms and to angiosperms. A possible explana- 140-190 Mya. These valuable analyses certainly help
tion is that problemsof morphologicalhomology are too reducethe conflict between molecularestimatesand first
severe to allow convincing resolution; taxon sampling appearanceof angiospermsin the fossil record,but time
also can affect results (Rydin & KUillersji,2002). differences still remain. Further, complex oleananes,
Recent molecular sequence data and analyses have nearly restrictedto angiospermsin the modernflora, are
provided critical insights on relationshipsamong higher known in fossil deposits (althoughthey are not common)
levels of angiosperms(e.g., Chase & al., 1993; Kuzoff & in the Jurassic (Moldowan & al., 1994). Macro- and
Gasser,2000; Soltis & al., 2000). In particular,they have micro-chemical data, therefore, hint towards pre-
provided more clarity on which extant angiosperms are Cretaceousorigins for angiosperms.
basal, althoughdeterminationof which particularfamily These data and analyses suggest several points
might be most basal has varied with the genes and algo- regarding origin of angiosperms. First, micro- and
rithms used [e.g., Ceratophyllaceae,Chase & al., 1993; macro-fossils reveal origin approximately 130 million
Nandi & al., 1998 (but using combined rbcL and mor- years ago. Second, no one gymnospermous group is
phological data);Amborellaceae,or at least more broad- identified as sister to angiosperms. Each analysis gives
ly the ANITA grade (i.e., Amborellaceae,Nymphaeales, slightly differentanswers.Third,several differentlineag-
Illiciales, TrimeniaceaeandAustrobaileyaceae;Mathews es (e.g., early Cretaceousnumerousforms from Portugal
& Donoghue, 1999; Parkinson& al., 1999; Qiu & al., and Clavatipollenites, mid-Cretaceous Archaeanthus,
1999; Soltis & al., 1999; Barkman& al., 2000; Graham lower-CretaceousArchaefructus)appearearly in the fos-

5
Stuessy* Theory for the origin of angiosperms TAXON 53 (1) * February2004: 3-16

sil record. Fourth,molecularsequence data and analyses Bennettitales(perhapsalso Gnetales)certainlycould also

confirm differentassemblages as basal for angiosperms, be said to contain flowers. In the German-speaking
with Amborellaceae(or more broadly,ANITA) being the world, in fact, workers often refer to gymnosperms as
most recentcandidate.Fifth, molecularsequence data, in "Bltitenpflanzen"(Wagenitz, 1996: 346). The thirddiag-
context of a molecularclock hypothesis, suggest an older nostic feature of angiospermy is double fertilization. In
origin for angiosperms, perhaps 200 million years or addition to angiosperms, however, a type of double fer-
more. Sixth, Archaefructus demonstrates that at least tilization is known in Gnetum and Ephedra of gym-
some early angiospermshad a separationof carpels from nosperms (Friedman,1990a, b; Carmichael& Friedman,
stamens, and could, in fact, be said to lack flowers alto- 1996; Friedman& Carmichael,1996; Welwitschiais not
gether.The apparentaquatichabitof Archaefructusis not yet investigated), but whether this representsa homolo-
regardedfundamentalto understandingangiosperm ori- gous condition is controversial (Friedman & Floyd,
gins, but it is certainlyan interestingfeatureof the fossil, 2001). In my opinion, this particularcondition is a paral-
especially in light of basal Nymphaeaceae and lelism. In Ephedra and Gnetum the second sperm cell
Cabombaceae,which are also aquatic(Feild & al., 2003, fuses with a single haploid ventralcanal nucleus to form
infer a "shady stream-like" setting for the original a diploid fusion product, the function of which is yet
angiosperms). Any theory of angiosperm origins and unclear. This is very different from the triple fusion
evolution, therefore,must take all these facts and inter- nucleus and endospermformationin many angiosperms,
pretationsinto account. and still different from the formation of diploid
endosperm in some basal angiosperms (in Nuphar,
Nymphaeaceae, and related familes; Williams &
Friedman,2002).
E OF ANGIO-
CHARACTERISTICS
SPERMS
Before proceeding to the transitional-combinational
theory, it is useful to review quickly the features that
E THE TRANSITIONAL-COMBINA-
define angiosperms. Focus is on reproductivefeatures. TIONALTHEORY
Also importantare vegetative features such as net leaf In view of available new data, analyses, and inter-
venation and vessels, known in most angiosperms and pretations, it is perhaps surprising that the origin of
lacking in most gymnosperms.Vessels are also presentin angiosperms still eludes our grasp. Part of the problem
Gnetales (Muhammad& Sattler, 1982), but it is debat- perhaps can be resolved by considering that the
able whether these are homologous with those in angiosperms evolved slowly over many millions of
angiosperms (Carlquist, 1989, 1992, 1996; see also years, perhapsmore than 100 million, and that the three
Carlquist & Schneider, 2002, for comments on defini- main characteristicsof the group (carpels, double fertil-
tions and transitionsbetween vessels and tracheids);they ization, and flowers) did not all evolve at the same time,
appear to represent an evolutionary parallelism. but ratherthroughtransitions.It is proposed that carpels
Morphological cladistic analyses also support this per- evolved first, followed by double fertilization,and lastly
spective, because groups lie between Gnetales and flowers. It was only when these three featureswere com-
angiosperms (e.g., Doyle & Donoghue, 1986; Doyle, bined that modern angiosperms arose and became so
1996) and currentlybasal Amborellaceae are vesselless evolutionarily successful. This is the key feature of the
(Cronquist, 1981). Vessels have also been reported in theory.
ferns (Carlquist& Schneider, 2001), but they lack dis- Some previous authors have provided similar per-
tinctive end wall perforations that characterize such spectives, although not discussing in detail the implica-
structures in angiosperms. Of greater importance for tions. Mabberley (1984) discussed problems of finding
defining angiospermsare the three reproductivefeatures ancestors of the angiosperms from among gymnosper-
of carpels, flowers, and double fertilization.Carpels do mous groups, and he added (p. 78): "The characters
not occur in gymnosperms, by definition (Sun & al., today diagnostic of the group we call angiosperms may
1998). Flowers are also unique to angiosperms, if one have become establishedone at a time, possibly the dou-
also insists that flowers must have carpels. Withthis def- ble fertilizationfirst, ...become correlated,those without
inition, a typical hermaphroditic flower would have such successful correlations going extinct". Doyle &
carpels, anthers,and associated sterile appendages;often Donoghue (1993; Donoghue, 2002) mentioned with
all these structuresare borne closely together. If a more regard to the origin of angiosperms the distinction
general definition is used, however, such as male and between the origin of the line leading to the modern
female structureswhorled or in a close sprial on a short group (the "stem angiophytes") and the origin of the
shoot surrounded by sterile appendages, then some clade that contains all extant members(the crown-group,

6
TAXON 53 (1) e February2004: 3 16 Stuessy e Theory for the origin of angiosperms

the angiosperms). The stem line might be quite ancient one (Schnarf, 1931; Leon-Kloosterziel& al., 1994) or by
whereas the crown line more recent. The stem lineage integumentaryshift (Bouman & Calis, 1977). Evolution-
might have existed as fossils before the Cretaceous ary development of the second integument may have
"...though perhaps without the full complement of char- been integratedwith furtherdevelopmentof the embryo
acters found in modem angiosperms"(Donoghue, 2002: sac and eventual double fertilization. That the outer
176; see also fig. 4 in Doyle & Donoghue, 1993, for a integument may have had a separate origin from the
graphic representation).Although both of these papers innerone is suggested by expressionof the INO gene (of
touch on the main themes of the presenttheory,they did the YABBY family) only in the outer integument
not develop the ideas to the extent, nor in the way, pre- (Meister & al., 2002). The carpellarystructuresof the
sented here. Mabberley(1984) did point to the basic idea early angiospermlineage would have first evolved from
of serial acquisition of angiosperm features, but he did pteridospermsin Triassic or Early Jurassic (c. 200 mil-
not elaborate on the implications. Doyle & Donoghue lion years ago). Adaptivereasonsfor the origin of carpels
(1993) dealt with distinctions between crown- and stem- have already been emphasized by many workers in the
groups, but likewise did not elaborate on implications past (e.g., Stebbins, 1974), and these include protection
regardingincongruence with molecular data (obviously from desiccation, from predationby bacteria, fungi and
most not yet available) nor the importanceof focusing on insects (especially as the latterproliferated),both on the
seed ferns. In general,however,these comments and per- parentplant and after dispersal,establishmentof incom-
spectives are compatible with the transitional-combina- patibility systems (Zavada & Taylor, 1986; Dilcher,
tional theory presentedin detail here. 1995), and pollen competition (Mulcahy, 1979; Dilcher,
It is proposed that carpels arose first among pteri- 1996). Carpelsnot only envelop the ovules, but they also
dosperms. Defining carpels can yield unexpected com- function as transmittingtissue for pollen tube growthto
plexities (e.g., Cresens & Smets, 1989), but the term is the ovules. Overall, carpels would have provideda great
used here to mean a structure(probably of leaf origin, selective advantagein survival and in promotingspecia-
i.e., a phyllome) that nearly or completely encloses tion (e.g., Doyle & Donoghue, 1986) over the naked
ovules and that develops together with them as a unit seeds dispersedby gymnosperms.
until maturity (or near maturity). Known fossils in Evolution of double fertilization (and non-haploid
Caytoniales and Corystospermales (perhaps also in endosperm) would have taken a long time to develop.
Czekanoskiales) have already been clearly documented The change from a well-differentiatedcellular megaga-
to have structuresthat emulate carpels (e.g., see reviews metophyte with archegoniato a reducedfemale gameto-
in Retallack & Dilcher, 1988; Hughes, 1994; Taylor & phyte consisting of an embryosac with one egg, two syn-
&
al., 1994; Taylor Kirchner,1996) in the fossil record ergids, three antipodals, and two polar nuclei is not a
175-200 million years ago (Jurassic and Triassic). simple transition.Presumablythe shift involved contin-
Gaussen (1946), Stebbins (1974), and Doyle (1978, ued reduction of the female gametophyte and gradual
1994, 1996) have a
championed Caytonialeanhypothesis substitution of cells and nuclei within the embryo sac.
for the origin of angiosperms. Problematic has been Fusion of the second spermcell with the two polarnuclei
explaining either the double integument of modem created triploid endosperm surroundingthe zygote and
angiosperms in contrast to the single integument gym- of developing embryo. This may, in fact, have provideda
nosperms(including Caytonia) or the carpel itself. If the mechanismfor more efficient productionof nutritivetis-
in
cupule Caytoniales judged is to be the carpel wall, sue in the seed, only after (or concomitantwith) zygote
then the second integument is lacking. If the cupule is formation,representingan energetic investmentonly in
homologized with the second integument, then the carpel those successful ovules containing embryos (Donoghue
is lacking. Doyle (1978) attemptedto explain this latter & Scheiner, 1992; W. Burger,pers. comm.). Exactlyhow
possibility by suggesting that the number of ovules in this transitionwas accomplishedand the selective factors
each cupule might be reduced to one, that the cupule involved is unclear,but Friedman(1992) has offeredsug-
itself would become the outer integument, and that the gestions for gradual changes involving a diploid fusion
central axis would envelop the ovules, hence making product from sperm and ventral canal nuclei through
carpels.Although possible, this seems an overly complex mitoses and eventual cellular differentiationas one pos-
hypothesis. In my opinion, too much emphasis has been sible route. Friedman (1995) has also suggested more
placed on attemptingto explain the origin of the second specifically that endospermmay have developed initial-
integument.Ovules with single integuments exist within ly from a modified diploid secondaryembryo. Recently,
angiosperms (e.g., Ranunculaceae, Rosaceae, Salica- Williams & Friedman (2002, see also Friedman &
ceae; Schnarf, 1931; Shamrov, 1998; Albach & al., Williams, 2003) have shown the formation of diploid
2001), although in some cases these appearsecondarily endospermin a group of basal angiosperms(most of the
derived from lateralfusion of two integumentsinto only representativesin the ANITA clade, except Amborella),
Stuessy * Theory for the origin of angiosperms TAXON 53 (1) * February2004: 3 16

suggesting that this embryological featuremay have still dispersal, lack of attractivestructuresfor insect pollina-
been undergoing transition as the combination of tion and hence less efficient mating,and absence of com-
angiosperm features came together (also suggested by plex breedingsystems.
Friedman& Floyd, 2001). The extant patternsof double Two specific hypotheses have been advancedfor ori-
fertilization in Ephedra and Gnetum(Friedman, 1900a, gin of flowers leading to angiosperms through sexual
b; Carmichael& Friedman, 1996) suggest intermediate modifications of reproductive structures of gym-
conditions on the path to angiosperms,but this unusual nosperms. Meyen (1988) suggested that bitegmic ovules
gymnospermgroup has no carpels or flowers. originated in Bennettitales by sexual modification from
The evolution of the flower was the last stage in the microsporangia, followed by evolution of carpels
origin of angiosperms.Archaefructus(Sun & al., 1998, through enclosure of the sporophyll. Frohlich & Parker
2002) suggests thatat least some of the early angiosperm (2000) and Frohlich(2003) also follow this general line
lineages consisted not yet of typical flowers, but of of reasoning, suggesting that flowers might have come
reproductive axes with distally attached carpels and from pteridosperms,especially Corystospermales.Their
proximallyattachedpairedstamens. Reductionof intern- "mostly male" theory postulates origin of ectopic (=
odes, however, could result in carpels attached on an organs borne in the wrong place) cupules, with one or
elongated receptacle,subtendedby two stamenson short two ovules inside, among microsporangia, and subse-
filaments (e.g., as in Magnoliaceae). quentcarpel formationby enclosure by the flat and broad
The key to explosive radiationof angiosperms was microsporophyll. Whether these sexual transmutation
evolution of the flower. This combinationof ovules pro- theories for the origin of the carpel (and flower) might be
tected in carpels, with double fertilization and correct remainsto be seen, but it would not be inconsis-
endosperm, surroundedby anthers and protective (and tent with the transitional-combinationaltheory if the
showy, at least in some lines) perianthparts, was a new carpels originatedearly enough followed by lateradapta-
and highly successful reproductive combination that tions in double fertilizationand aggregations into flow-
resulted in a high rate of diversification(Dilcher, 1995, ers.
1996). One can hypothesize that the selective forces for To stress the importanceof a successful combination
the appearance of flowers came mainly from insects. of featuresfor evolutionarysuccess of angiosperms,it is
This is certainlynot a novel idea, having been remarked instructive to consider evolution of Compositae. This
upon by numerous others (e.g., Raven, 1977; Crepet, family is approximately30 million years old (Graham,
1983, 1984). Studies on diversity of insects in the fossil 1996), and it originated in southern South America
record (Grimaldi, 1990; Labandeira& Sepkoski, 1993) (DeVore& Stuessy, 1995; Stuessy & al., 1996; Bremer&
and more recent molecular phylogenetic studies (e.g., Gustafsson, 1997). After its origin, however, the family
Wiegmann& al., 2003) reveal all majortypes of insects evolved rapidly into different phylogenetic lines and
and feeding types at appearanceof modem angiosperms early dispersedwidely (Graham,1996), now found on all
c. 130 million years ago. The "constant" pollinators continents except Antarctica, and diversified to yield
(Crepet, 1983; Labandeira& Sepkoski, 1993), Hymeno- more than 23,000 species (Bremer, 1994), the largest
ptera and Lepidoptera,began to diversify in Mesozoic family of angiosperms. The closest familial relative of
followed by a stronger diversification in Cenozoic. Compositae is Calyceraceae (Olmstead & al., 1990,
Although data on diversity of fossil insects do not sup- 1992; Downie & Palmer, 1992; Michaels & al., 1993;
port that insects alone drove angiospermorigins, they do Kim & Jansen, 1995; Soltis & al., 2000), a small family
show that sufficient types of insects were available for of approximatelysix genera and 55 species (Mabberley,
co-evolutionary factors to quickly become operational 1997) still confined to southern South America, and
after the angiosperm flower appeared.The flower was which obviously has not been spectacularlyevolutionar-
the last step in a combinationthat allowed rapidand suc- ily successful. Origin of head inflorescences in Asterales
cessful evolution of defense, pollination, breeding, and and Dipsacales was a parallel trend, appearing in
dispersal systems, all in a combined system; it also Calyceraceae, Campanulaceae, Dipsacaceae, etc., per-
attracted insects by offering food resources (Pellmyr, haps due to evolution for larger pollinatortargets when
1992; Frame,2003). Shortnessof the reproductivecycle individualflorets became too small within particularlin-
in angiosperms over that of gymnosperms may also be eages. The head inflorescence was an importantstructur-
related to the appearance of flowers (Stebbins, 1974; al innovation that provided enhanced opportunities for
Verdui,2002). Cones of gymnospermsalso representevo- evolution of defense, pollination, breeding systems and
lution along similar adaptive lines, but this combination dispersal mechanisms (Stuessy & Spooner, 1988;
orientedtowardwind pollinationwas much less success- Stuessy & Garver, 1996). This alone was insufficient for
ful (or less "competitive";Midgley & Bond, 1991), per- high evolutionary success, however, as evidenced by
haps due to lack of carpels for protection of seeds after weaker success of other head-bearing families.

8
TAXON 53 (1) e February2004: 3--16 Stuessy * Theory for the origin of angiosperms

Additional specific adaptationsaccrued in Compositae features(i.e., more compact arrangementwith surround-

that apparentlyprovided needed evolutionary stimulus. ing perianth).Analogous might be Homo erectus coex-
The calyx became modified for more efficient defense isting with H. sapiens (Swisher & al., 1996), the former
and dispersal (Stuessy & Garver, 1996). Also important going eventually extinct, and the latterbecoming incred-
was fusion of epidermis of anthers laterally forming a ibly successful.
tube (syngenesious condition), which allowed develop- Second, discrepancy between molecular data that
ment of alternative pollen presentation mechanisms suggest angiospermsbeing at least 200 million years old
(either pumpor brush;Leins & Erbar,2003) plus numer- and appearanceof the flower (or pollen) at c. 130 million
ous variations in pollination ecology (Lane, 1996). years ago, can now be explained. Angiospermsbegan to
Structuraland reproductivemodifications of individual evolve (carpellaryinnovations)early,and the extant lines
florets also occurred, some being hermaphroditic,some still reflect this ancestry in their genes. Troitsky & al.
only functionally female and others only male, which (1991) are in general agreement with this perspective,
allowed development of different breeding systems and stressing some early angiosperm innovations but final
gender strategies in heads, e.g., simple monoecy, combination of traits (with all diagnostic features) not
gynomonoecy, etc. (Mani & Saravanan, 1999). The being achieved until Cretaceous. Flowers in seed plants
important point is that there is no single feature that (but not in angiosperms) occur also relatively early, c.
defines Compositae(Crepet& Stuessy, 1978), but rather 175 million years ago in Bennettitales (Crepet, 1974),
a combination of features that is unique. By analogy, but without carpelsand perhapswithoutdouble fertiliza-
therefore,the innovationof flowers in angiosperms,plus tion (still unknown, however). The "Jurassic"appear-
carpels and double fertilization,and with associated bio- ance of nectar-eatingBrachyceraflies has been used to
logical correlations of compatibility and mating strate- arguefor pre-Cretaceousangiospermfloral origins (Ren,
gies (e.g., Barrett, 1998), provided the combinationthat 1998). The Yixian Formation,from which these fossils
allowed explosive success of angiosperms, in a fashion were obtained, however, is the same as that in which
similar to how the combinationof featuresin the head of Archaefructusoccurs, now dated c. 125 Mya (Swisher &
Compositae led to their rapidevolutionarysuccess. al., 1999; Sun & al., 2002; the Yixian and the Jiufotang
Formations are yielding many valuable fossils, both
plants and animals, Zhou & al., 2003). In any event,
sources of nectar must have been present,perhapseven
I OF THE TRANSI-
IMPLICATIONS in a carpellaryline with nectaraccumulatedin cupules or
TIONAL-COMBINATIONALTHEO- some other associated structures, or in extinct gym-
RY
nosperm lines perhaps with flowers with nectaries but
Acceptance of the transitional-combinationaltheory withoutcarpels(e.g., Bennettitales).Nectardoes occur in
of the origin of angiosperms allows a numberof prob- some extant gymnosperms (e.g., Gnetophyta;Endress,
lems to be resolved or at least better understood.First, 1996).
because the three majorfeaturesof angiospermsevolved Third, the transitional-combinationaltheory allows
at differenttimes, there is no single progenitorthat is the us to focus more clearly on the questionof monophylyof
missing link; rather,many transitionalforms should be angiosperms.In a narrowsense, it may be thatthe extant
found, some with different stages of carpel closure and angiosperms are polyphyletic (as often emphasized by
differentiation, others with the slow development of Meeuse, 1975; Hughes, 1976, 1994; Krassilov, 1997),
endosperm,and still others with early flowers. This point with the combination of critical features having been
has alreadybeen stressedby Mabberley(1984). As a case achieved through several transitions. The Laurales,
in point, althoughArchacfructushas been describedas a Magnoliales, and paleoherbs,etc., may all in fact repre-
basal angiospermfamily (Archaefructaceae),it certainly sent different lineages from carpellary ancestors. It is
would be unusualwith carpels and stamens widely sepa- certainlythe case that several well-defined evolutionary
rated on the axis. It could also be interpretedto possess lines are identifiable in these basal angiosperms (e.g.,
an inflorescencewith specialized unisexual flowers, sub- Doyle & Donoghue, 1993) and might well reflect differ-
tending bracts having fallen off or reduced (Friis & al., ent carpellary ancestors; the carpels, in fact, are quite
2003). Findingthis fossil at 125 million years, laterthan diverse in this basal group (Endress& Igersheim,2000)
the first microfossil records of angiosperms (Hughes & as is endospermdevelopment(Floyd & Friedman,2000).
McDougall, 1987; Hughes & al., 1979, 1991; Hughes, In any molecular phylogenetic analysis, however, these
1994), suggests to me that Archaefructusmay be a sur- lineages will appear monophyletic because they came
viving experimenttowardsangiospermsin a lineage that fromthe same ancientgeneral group of carpellaryances-
subsequently died out, unable to compete effectively tors (in pteridosperms)far removed from modern gym-
with angiospermswith a more successful combinationof nosperms. Thus, in a widened and more usual sense of

9
Stuessy * Theory for the origin of angiosperms TAXON 53 (1) s February2004: 3-16

monophyly (with regardto origin of higher groups), the nosperms (see Doyle, 2001, for a similar point of view).
carpellary transition to angiosperms would be mono- Troitsky& al. (1991: 259) are in agreement:"...the par-
phyletic because it arose in pteridospermous gym- ticular molecular data at our disposal show that none of
nosperms. the gymnosperm lineages could have been an ancestral
Fourth,the question of homology of the three diag- one for angiosperms".Cones may, in fact, representan
nostic featuresof angiospermsalso can be more fruitful- attemptto combine structuresinto more efficient flower-
ly discussed. Carpels would be the best candidateto be like aggregations,but this combinationwas less success-
homologous, having come, at least, from a pteridosperm ful due to lack of carpels. Some evidence exists that
monophyletic lineage in a broadsense (Sun & al., 1998, Bennettitalesmight have been visited (if not pollinated)
2002). Examinationof the many pteridospermousgroups by beetles (Crepet, 1974, 1983), but this evolutionary
that show ovules enclosed in different sterile experiment also failed. The successful combinationwas
appendages, e.g., in Caytoniales, Corystospermales,or simply not yet available. An importantnote, however, is
Glossopteridales, suggests that evolution of the carpel that searching for relationships in the fossil record
(derived from an ovuliferous phyllome; Cresens & involving gradual transitions from pteridosperms and
Smets, 1989) might have been occurring frequently in final floral punctuationto angiosperms occurring over
this general group. Even so, in a narrow sense, "...it is more than 100 million years is not an easy task.
likely that carpel closure is polyphyletic, because both One pertinentquestion might be: Why did it take so
incompletely closed carpels and completely closed long for the critical combinationof angiosperm features
carpels are found among the earliest diverging families to evolve? If carpels originated 100 million years earlier
of flowering plants" (Raven & Weyers, 2001: 112; than flowers, and if this by itself was a major morpho-
Endress& Igersheim, 2000, point out, however, that all logical innovation, why didn't double fertilization and
carpels are really closed, but four different patterns of the flower follow soon thereafter? Changes in the
closure exist). Stebbins (1974: 244) also remarkedthat embryo sac leading to double fertilization are complex
carpels might have originated independentlymore than (Friedman, 1994; Floyd & Friedman, 2000), and the
once. As for stamens, numerousdifferent male clusters adaptive explanationsfor such changes are not yet clear.
exist in pteridosperms (Dilcher, 1979; Retallack & The collection of structuresinto a flower, however, must
Dilcher, 1988), and hence the likelihood seems high of surely relate in large measure to insect pollination
having had different experiments of aggregations of (Crepet, 1983). Once insects reached sufficient diversity
microsporangiacombined into flowers. The sterile asso- to provide strong selection for floral aggregations for
ciated appendages,i.e., petals and sepals, might be least pollination, flowers (and inflorescences) began to occur.
likely to be homologous across angiosperms (they are This final transitionbroughttogetherthe favorablecom-
extremely diverse in the basal and magnoliidcomplexes, bination that allowed further evolution into defense,
e.g., Endress & Hufford, 1989), although obviously breeding, and dispersal systems that have much to do
homologous in selected lineages. These features are so with angiosperm success as well as co-evolution with
intimatelyassociated with pollination and defense func- insects (both as predatorsand pollinators; see Pellmyr,
tions that the specific selective pressures would have 1992).
been different in various early lineages. Another diag- Another question is: Why are fossil pollen of
nostic feature of angiosperms,double fertilization, may angiosperms not detected priorto 130 million years ago
be homologous within this groupbut non-homologousin even though carpellaryangiospermline(s) are postulated
seed plants, having originatedalso in Gnetales. Gnetum, to have existed much earlier?This may be due to corre-
for example, has its own set of apomorphic,perhapshet- lations with development of the flower. Prior to the last
erochronic, female gametophytic features (Friedman& transition to flowers, the pollen of the angiophyte line
Carmichael, 1998). Recent molecular data suggesting would have been gymnospermous and most probably
thatGnetalesare more closely allied to otherextantgym- wind-pollinated. It may be that only after the flower
nospermsthan to angiosperms(Goremykin& al., 1996; evolved in combinationwith otherfeaturesthatnew biol-
Hansen& al., 1999; Samigullin& al., 1999; Bowe & al., ogy of pollination and breeding systems was established
2000; Schmidt & Schneider-Poetsch,2002; D. Soltis & that favored outcrossing (Dilcher, 2000). Pollen grains
al., 2002), supportthis position. then began to change rapidly,developing the typical wall
Fifth, meaningful morphological cladistic analyses patterninto a non-laminarendexine and an ectexine dif-
of basal angiosperms and extinct and extant gym- ferentiatedinto a foot-layer,columellae and tectum (Friis
nosperms should now focus on ties between the pteri- & al., 1987b), perhaps in response to development of
dosperms and angiospermsdirectly (see, e.g., the many compatibility systems regulated by pollen wall proteins
suggested alternatives in Dilcher, 1979; Hughes, 1994; and pollen competition (Mulcahy & Mulcahy, 1987) as
and Krassilov, 1997) and not include the rest of the gym- well as complex interactionswith insect visitors. Some

10
TAXON 53 (1) * February2004: 3-16 Stuessy * Theory for the origin of angiosperms

Bremer, K. 1994. Asteraceae: Cladistics & Classification.

early angiosperms, however, may have been columella- Timber Press, Portland.
less (Walker & Skvarla, 1975; Sampson, 2000; Hesse, Bremer, K. & Gustafsson, M. H. G. 1997. East Gondwana
2001). ancestry of the sunflower alliance of families. Proc. Natl.
Acad. Sci. U.S.A. 94: 9188-9190.
Brenner, G J. 1996. Evidence for the earliest stage of
angiosperm pollen evolution: a paleoequatorial section
E ACKNOWLEDGEMENTS from Israel. Pp. 91-115 in: Taylor, D. W. & Hickey, L. J.
(eds.), Flowering Plant Origin, Evolutionand Phylogeny.
Thanks go to: the Austrian National Science Foundation
Chapman& Hall, New York.
(FWF) for general supportunder grant No. P15225; David L. Burger, W. C. 1977. The Piperalesand the monocots.Alternate
Dilcher for stimulatingdiscussionson angiospermoriginsduring hypotheses for the origin of the monocotyledonousflow-
the banquet of the American Botanical Society in Madison, ers. Bot. Rev. 43: 345-393.
Wisconsin,7 August2002, which led to this paperbeing written; Carlquist, S. 1989. Woodand barkanatomyof the New World
William C. Burger, David L. Dilcher, James A. Doyle, Peter species of Ephedra.Aliso 12: 441-483.
Endress, Else Marie Friis, Klaus Kubizki, Gar Rothwell, and Carlquist, S. 1992. Wood, bark, and pith anatomy of Old
World species of Ephedra and summary for the genus.
RobertThornefor criticalandvery helpfulcommentson the man- Aliso 13: 255-295.
uscript;and CharlesS. Gasserfor excellent and specific sugges- Carlquist, S. 1996. Wood, bark, and stem anatomy of
tions andreferencesdealingwithgeneticcontrolof integumentsof Gnetales: a summary. Int. J. P1. Sci. 157 (6, suppl.):
ovules. S58-S76.
Carlquist, S. & Schneider, E. L. 2001. Vessels in ferns: struc-
tural, ecological, and evolutionary significance. Amer.J.
Bot. 88: 1-13.
I LITERATURE
CITED Carlquist, S. & Schneider, E. L. 2002. The tracheid-vessel
element transitionin angiospermsinvolves multiple inde-
Albach, D. C., Soltis, P. S. & Soltis, D. E. 2001. Patternsof
pendentfeatures:cladistic consequences.Amer.J. Bot. 88:
embryological and biochemical evolution in the Asterids. 1-13.
Syst. Bot. 26: 242-262. Carmichael, J. S. & Friedman, W. E. 1996. Double fertiliza-
Angiosperm Phylogeny Group. 1998. An ordinal classifica- tion in Gnetumgnemon (Gnetaceae): its bearing on the
tion for the families of flowering plants. Ann. Missouri evolution of sexual reproductionwithin the Gnetales and
Bot. Gard. 85: 531-553. the anthophyteclade. Amer J. Bot. 83: 767-780.
Angiosperm Phylogeny Group. 2003. An update of the Chapman, J. L. 1987. Comparisonof Chloranthaceaepollen
Angiosperm Phylogeny Groupclassification for the orders with the Cretaceous Clavatipollenitescomplex. Pollen et
and families of flowering plants:APG II. Bot. J Linn.Soc.
Spores 29: 249-272.
141: 399-436.
Chase, M. W., Soltis, D. E., Olmstead, R. G., Morgan, D.,
Arber, E. A. N. & Parkin, J. 1907. On the origin of Les, D. H., Mishler, B. D., Duvall, M. R., Price, R. A.,
angiosperms.J. Linn. Soc. London38: 29-80. Hills, H. G., Qiu, Y.-L., Kron, K. A., Rettig, J. H.,
Axsmith, B. J., Taylor, E. L. & Taylor, T. N. 1998. The limi- Conti, E., Palmer, J. D., Manhart, J. R., Sytsma, K. J.,
tations of molecular systematics: a palcobotanical per-
Michaels, H. J., Kress, W. J., Karol, K. G., Clark, W. D.,
spective. Taxon47: 105-108. Hedren, M., Gaut, B. S., Jansen, R. K., Kim, K.-J.,
Barkman, T. J., Chenery, G., McNeal, J. R., Lyons-Weiler,
Wimpee, C. F., Smith, J. F., Furnier, G. R., Strauss, S.
J. & DePamphilis, C. W. 2000. Independentand com-
H., Xiang, Q.-Y., Plunkett, G. M., Soltis, P. S., Swensen,
bined analyses of sequences from all three genomic com- S. M., Williams, S. E., Gadek, P. A., Quinn, C. J.,
partments converge on the root of flowering plant phy- Eguiarte, L. E., Golenberg, E., Learn, G. H., Jr.,
logeny. Proc. Natl. Acad. Sci. U.S.A. 97: 13166--13171. Graham, S. W., Barrett, S. C. H., Dayanandan, S. &
Barrett, S. C. H. 1998. The evolution of mating strategies in Albert, V.A. 1993. Phylogeneticsof seed plants:an analy-
flowering plants. TrendsP1. Sci. 3: 335-341. sis of nucleotide sequences from the plastid gene rhcL.
Bessey, C. E. 1893. Evolution and classification. Bot. Gaz. 18: Ann. Missouri Bot. Gard. 80: 528-580.
329-332.
Chaw, S.-M., Parkinson, C. L., Cheng, Y., Vincent, T. M.,
Bessey, C. E. 1897. Phylogeny and taxonomy of the Palmer, J. D. 2000. Seed plant phylogeny inferredfrom
angiosperms.Bot. Gaz. 24: 145-178. all three plant genomes: monophyly of extant gym-
Bessey, C. E. 1915. The phylogenetic taxonomy of flowering nospermsand origin of Gnetales fromconifers. Proc. Natl.
plants. Ann. Missouri Bot. Gard. 2: 109-164. Acad Sci. U.S.A. 97: 4086-4091.
Bouman, F. & Calls, J. I. M. 1977. Integumentaryshifting-- Chaw, S.-M., Zharkikh, A., Sung, H. M., Lau, T. C. & Li,
a third way to unitegmy. Be,: Deutsch. Bot. Ges. 90: W. H. 1997. Molecularphylogeny of extantgymnosperms
15-28.
and seed plant evolution-analysis of nuclear 18S riboso-
Bowe, L. M., Coat, G. & DePamphilis, C. W. 2000. mal-RNA. Molec. Biol. Evol. 14: 56-68.
Phylogeny of seed plants based on all three plantgenomic Cornet, B. 1989a. The reproductivemorphology and biology
compartments:extant gymnospermsare monophyleticand of Sanmiguelialewisii, and its bearingon angiospermevo-
Gnetales are derived conifers. Proc. Natl. Acad. Sci. lution in the late Triassic.Evol. TrendsPlants 3: 25-51.
U.S.A. 97: 4092-4097. Cornet, B. 1989b. Late Triassic angiosperm-likepollen from

11
Stuessy * Theory for the origin of angiosperms TAXON 53 (1) * February2004: 3-16

the Richmond rift basin of Virginia, U.S.A. Palaeonto- Donoghue, M. J. 2002. An overview of green plant phylogeny.
graphica B, 213: 37-87. Pp. 153-183 in: Judd,W. S., Campbell, C. S., Kellogg, E.
Cornet, B. 1993. Dicot-like leaf and flowers from the Late A., Stevens, P. F. & Donoghue, M. J. (eds.), Plant
Triassic tropical Newark Supergroup rift zone, U.S.A. Systematics: a Phylogenetic Approach, ed. 2. Sinauer
Modern Geol. 19: 81-99. Associates, Sunderland,Massachusetts.
Coulter, J. M. 1903. The phylogeny of angiosperms. The Donoghue, M. J. & Doyle, J. A. 2000. Seed plant phylogeny:
Decennial Publications 10: 191-194. demise of the anthophyte hypothesis'?CurrentBiol. 10:
Crane, P. R. 1985. Phylogeneticanalysis of seed plantsand the RI06-R109.
origin of angiosperms. Ann. Missouri Bot. Gard. 72: Donoghue, M. J. & Scheiner, S. M. 1992. The evolution of
716-793. endosperm: a phylogenetic account. Pp. 356-389 in:
Crane, P. R. & Dilcher, D. L. 1984. Lesqueria: an early Wyatt, R. E. (ed.), Ecology and Evolution of Plant
angiosperm fruiting axis from the mid-Cretaceous.Ann. Reproduction. New Approaches. Chapman & Hall, New
Missouri Bot. Gard. 71: 384-402. York.
Crane, P. R., Friis, E. M. & Pedersen, K. R. 1986. Lower Downie, S. R. & Palmer, J. D. 1992. Restrictionsite mapping
Cretaceous angiosperm flowers: fossil evidence on early of the chloroplast DNA inverted repeat:a molecular phy-
radiationof dicotyledons. Science 232: 852-854. logeny of the Asteridae. Ann. Missouri Bot. Gard. 79:
Crane, P. R. & Lidgard, S. 1990. Angiosperm radiationand 266-283.
patterns of Cretaceous palynological diversity. Pp. Doyle, J. A. 1969. Cretaceous angiosperm pollen of the
377-407 in: Taylor,P. D. & Larwood, G. P. (eds.), Major Atlantic Coastal Plain and its evolutionarysignificance. J.
EvolutionaryRadiations. ClarendonPress, Oxford. ArnoldArbor 50: 1-35.
Crepet, W. L. 1974. Investigations of North American Doyle, J. A. 1978. Origin of angiosperms. Annual Rev. Ecol.
cycadeoids: the reproductive biology of Cycadeoidea. Syst. 9: 365-392.
Science 148: 144-169. Doyle, J. A. 1994. Origin of the angiosperm flower: a phylo-
Crepet, W. L. 1983. The role of insect pollination in the evo- genetic perspective. Pl. Syst. Evol., Suppl. 8: 7-29.
lution of angiosperms. Pp. 31-50 in: Real, L. (ed.), Doyle, J. A. 1996. Seed plant phylogeny and the relationships
Pollination Biology. Academic Press, Orlando. of Gnetales. Int. J. Pl. Sci. 157 (6, suppl.): S3-S39.
Crepet, W. L. 1984. Advanced (constant) insect pollination Doyle, J. A. 1998a. Phylogeny of the vascular plants. Annual
mechanisms: patternof evolution and implications vis-a- Rev. Ecol. Syst. 29: 567-599.
vis angiosperm diversity. Ann. Missouri Bot. Gard. 71: Doyle, J. A. 1998b. Molecules, morphology, fossils, and the
607-630. relationship of angiosperms and Gnetales. Molec.
Crepet, W. L. & Stuessy, T. F. 1978. A reinvestigationof the Phylogen. Evol. 9: 448-462.
fossil Viguiera cronquistii (Compositae). Brittonia 30: Doyle, J. A. 1999. The rise of angiosperms as seen in the
483-491. African Cretaceous pollen record. Pp. 3-29 in: Scott, L.,
Cresens, E. M. & Smets, E. F. 1989. The carpel-a problem Cadman, A. & Verhoeven, R. (eds.), Palaeoecology of
child of floral morphology and evolution. Bull. Jard. Bot. Africa and the Surrounding Islands. A. A. Balkema,
Nat. Belg. 59: 377-409. Rotterdam.
Cronquist, A. 1981. An IntegratedSystem of Classification of* Doyle, J. A. 2001. Significance of molecular phylogenetic
Flowering Plants. ColumbiaUniv. Press, New York. analyses for paleobotanicalinvestigations on the origin of
DeVore, M. L. & Stuessy, T. F. 1995. The place and time of angiosperms.Palaeobotanist 50: 167-188.
origin of the Asteraceae,with additionalcomments on the Doyle, J. A. & Donoghue, M. J. 1986. Seed plant phylogeny
Calyceraceaeand Goodeniaceae. Pp. 23-40 in: Hind, D. J. and the origin of angiosperms: an experimental cladistic
N., Jeffrey,C. & Pope, G. (eds.), Advances in Compositae approach.Bot. Rev. 52: 321-402
Systematics. Royal Botanic Gardens,Kew. Doyle, J. A. & Donoghue, M. J. 1992. Fossils and seed plant
Dilcher, D. L. 1979. Early angiospermreproduction:an intro- phylogenies reanalyzed.Brittonia 44: 89-106.
ductory report.Rev. Palaeobot. Palynol. 27: 291-328. Doyle, J. A. & Donoghue, M. J. 1993. Phylogenies and
Dilcher, D. L. 1995. Plant reproductivestrategies: using the angiosperm diversification.Paleobiology 19: 141-167.
fossil record to unravel current issues in plant reproduc- Doyle, J. A. & Hickey, L. J. 1976. Pollen and leaves from the
tion. Pp. 187-198 in: Hoch, P. C. & Stephenson, A. G. mid-CretaceousPotomacGroupand theirbearingon early
(eds.), Experimentaland Molecular Approaches to Plant angiosperm evolution. Pp. 139-206 in: Beck, C. B. (ed.),
Biosystematics. Missouri Bot. Garden, St. Louis. Origin and Early Evolution of Angiosperms. Columbia
[Monogr.Syst. Bot. 53.] Univ. Press, New York.
Dilcher, D. L. 1996. La importancia del origen de las Drinnan, A. N., Crane, P. R., Friis, E. M. & Pedersen, K. R.
angiospermas y como formaron el mundo alrededor de 1990. Lauraceousflowers from the Potomac Group (mid-
ellas. Pp. 29-48 in: Fortunato,R. H. (ed.), VI Congreso Cretaceous) of eastern North America. Bot. Gaz. 151:
Latinoamericanode Botinica [Mar del Plata,Argentina]: 370-384.
Conferencias.Royal Botanic Gardens,Kew. Eichler, A. W. 1876. Syllabus der Vorlesungeniiber spezielle
Dilcher, D. L. 2000. Towarda new synthesis: majorevolution- und medizinisch-pharmazeutische Botanik, ed. 1.
ary trends in the angiosperm fossil record. Proc. Natl. Borntraeger,Berlin.
Acad. Sci. U.S.A. 97: 7030-7036. Endress, P. K. 1987. The early evolution of the angiosperm
Dilcher, D. L. & Crane, P. R. 1984. Archaeanthus:an early flower. TrendsEcol. Evol. 2: 300-304.
angiosperm from the Cenomanianof the western interior Endress, P. K. 1993. Federico Delpino and early views on
of NorthAmerica.Ann. Missouri Bot. Gard. 71: 351-383. angiosperm origin and macroevolution. Diss. Bot. 196:

12
TAXON 53 (1) * February2004: 3-16 Stuessy e Theory for the origin of angiosperms

77-83. Friis, E. M., Crane, P. R. & Pedersen, K. R. 1987a. Floral

Endress, P. K. 1996. Structureand function of female and evidence for Cretaceous chloranthoid angiosperms.
bisexual organ complexes in Gnetales. Int. J. Pl. Sci. 157 Nature 320: 163-164.
(6, suppl.): S I13-S125. Friis, E. M., Doyle, J. A., Endress, P. K. & Leng, Q. 2003.
Endress, P. K. & Hufford, L. D. 1989. The diversity of sta- Archaefructus-angiosperm precursoror specialized early
men structures and dehiscence patterns among angiosperm?TrendsPI. Sci. 8: 369-373.
Magnoliidae.Bot. J. Linn. Soc. 100: 45-85. Friis, E. M., Eklund, H., Pedersen, K. R. & Crane, P. R.
Endress, P. K. & Igersheim, A. 2000. Gynoecium structure 1994. Virginianthuscalycanthoides gen. et. sp. nov.--a
and evolution in basal angiosperms.Int. J. PI. Sci. 161 (6, calycanthaceous flower from the Potomac Group (Early
suppl.): S211-S223. Cretaceous)of easternNorthAmerica. Int. J. PI. Sci. 155:
Engler, A. 1897. Nachtrige. Pp. 364-369 in: Engler, A. & 772-785.
Prantl, K. (eds.), Die natiirlichen Pflanzenfamilien. Friis, E. M., Pedersen, K. R. & Crane, P. R. 1999. Early
Engelmann,Leipzig. angiospermdiversification:the diversity of pollen associ-
Feild, T. S., Arens, N. C. & Dawson, T. E. 2003. The ances- ated with angiosperm reproductive structures in Early
tral ecology of angiosperms:emerging perspectives from Cretaceousfloras from Portugal.Ann. MissouriBot. Gard.
extant basal lineages. Int. J. PI. Sci. 164 (3, suppl.): 86: 259-296.
S129-S142. Friis, E. M. Pedersen, K. R. & Crane, P. R. 2000.
Flores, E. M. & Moseley, M. F., Jr. 1982. The anatomyof the Reproductive structure and organization of basal
pistillate inflorescence and flower of Casuarina verticilla- angiosperms from the early Cretaceous (Barremian or
ta Lamarck(Casuarinaceae).Amer.J. Bot. 69: 1673-1684. Aptian)of western Portugal.Int. J. PI. Sci. 161 (6, suppl.):
Floyd, S. K. & Friedman, W. E. 2000. Evolution of S169-S182.
endospermdevelopmentalpatternsamong basal flowering Friis, E. M., Pedersen, K. R. & Crane, P. R. 2001. Fossil evi-
plants. Int. J. PI. Sci. 161 (6, suppl.): S57-S81. dence of water lilies (Nymphaeales) in the Early
Frame, D. 2003. Generalistflowers, biodiversityand florivory. Cretaceous.Nature410: 357-360.
Taxon52: 681-685. Frohlich, M. W. 2003. An evolutionaryscenario for the origin
Friedman, W. E. 1990a. Double fertilization in Ephedra, a of flowers. Nature Rev. Gen. 4: 559-566.
nonflowering seed plant: its bearing on the origin of Frohlich, M. W. & Parker, D. S. 2000. The mostly male the-
angiosperms.Science 247: 951-954. ory of flower evolutionaryorigins: from genes to fossils.
Friedman, W. E. 1990b. Sexual reproduction in Ephedra Syst. Bot. 25: 155-170.
nevadensis (Ephedraceae):furtherevidence of double fer- Gandolfo, M. A., Nixon, K. C., Crepet, W. L., Stevenson, D.
tilization in a nonflowering seed plant. Amer J. Bot. 77: W. & Friis, E. M. 1998. Oldest known fossils of mono-
1582-1598. cotyledons. Nature 394: 532-533.
Friedman, W. E. 1992. Evidence of a pre-angiospermorigin of Gaussen, H. 1946. Les GymnospermesActuelles et Fossiles.
endosperm: implications for the evolution of flowering Travauxdu laboratoireforestierde Toulouse, tome 2, sec-
plants. Science 255: 336-339. tion 1, chapitre 5 (Les autres Cycadophytes; L'origine
Friedman, W. E. 1994. The evolution of embryogeny in seed cycadophytiquedes Angiospermes). Faculte du Sciences,
plants and the developmental origin and early history of Toulouse.
endosperm.Amer J. Bot. 81: 1468-1486. Gaussen, H. 1958. L'originedes Angiospermes.Compt.Rend.
Friedman, W. E. 1995. Organismalduplication, inclusive fit- Acad. Sci. Fr 247: 169-172.
ness theory and altruism:understandingthe evolution of Gaut, B. S. 1998. Molecularclocks and nucleotidesubstitution
endosperm and the angiosperm reproductive syndrome. rate in higher plants. Evol. Biol. 30: 93-120.
Proc. Natl. Acad. Sci. U.S.A. 92: 3913-3917. Goremykin, V., Bobrova, V., Pahnke, J., Troitsky, A.,
Friedman, W. E. & Carmichael, J. S. 1996. Double fertiliza- Antonov, A. & Martin, W. 1996. Noncoding sequences
tion in Gnetales:implications for understandingreproduc- from the slowly evolving chloroplast inverted repeat in
tive diversificationamong seed plants. Int. J. PI. Sci. 157 additionto rbcLdatado not supportGnetaleanaffinities of
(6, suppl.): S77-S94. angiosperms.Molec. Biol. Evol. 13: 383-396.
Friedman, W. E. & Carmichael, J. S. 1998. Heterochrony Goremykin, V. V., Hirsch-Ernst, K. I., Wijlfl, S. & Hellwig,
and developmentalinnovation:evolution of female game- F. 2003. Analysis of the Amborellatrichopdachloroplast
tophyte ontogeny in Gnetum, a highly apomorphic seed genome sequence suggests that Amborella is not a basal
plant. Evolution 52: 1016-1030. angiosperm.Molec. Biol. Evol. 20: 1499-1505.
Friedman, W. E. & Floyd, S. K. 2001. The origin of flower- Graham, A. 1996. A contributionto the geologic historyof the
ing plants and their reproductivebiology-a tale of two Compositae.Pp. 123-140 in: Hind, D. J. N. & Beentje, H.
phylogenies. Evolution 55: 217-231. J. (eds.), Compositae:. Systematics. Proceeding of the
Friedman, W. E. & Williams, J. H. 2003. Modularityof the InternationalCompositae Conference, Kew, 1994. Royal
angiosperm female gametophyte and its bearing on the Botanic Gardens,Kew.
early evolution of endosperm in flowering plants. Graham, S. W. & Olmstead, R. G 2000. Utility of 17 chloro-
Evolution 57: 216-230. plast genes for inferring the phylogeny of the basal
Friis, E. M., Chaloner, W. G & Crane, P. R. 1987b. angiosperms.Amer J. Bot. 87: 1712-1730.
Introductionto angiosperms. Pp. 1-16 in: Friis, E. M., Grimaldi, D. A. (ed.). 1990. Insects from the Santana
Chaloner, W. G. & Crane, P. R. (eds.), The Origins of Formation,Lower Cretaceous,of Brazil. Bull. Amer Mus.
Angiosperms and Their Biological Consequences. Nat. Hist. 195: 1-191.
CambridgeUniv. Press, Cambridge. Hansen, A., Hansmann, S., Samigullin, T., Antonov, A. &

13
Stuessy * Theory for the origin of angiosperms TAXON 53 (1) February2004: 3-16

Martin, W. 1999. Gnetumand the angiosperms:molecu- International Compositae Conference,Kew, 1994, vol. 2.
lar evidence thattheirsharedmorphologicalcharactersare Royal Botanic Gardens,Kew.
convergent, rather than homologous. Molec. Biol. Evol. Leins, P. & Erbar, C. 2003. Floraldevelopmentalfeaturesand
16: 1006-1009. molecular data in plant systematics. Pp. 81-105 in:
Hesse, M. 2001. Pollen characters of Amborella trichopoda Stuessy, T. F., Mayer, V. & H6randl, E. (eds.), Deep
(Amborellaceae): a reinvestigation. Int. J. Pl. Sci. 162: Morphology: Toward a Renaissance of Morphology in
201-208. Plant Svstematics.A. R. G. Gantner,Ruggell.
Hickey, L. J. & Doyle, J. A. 1977. EarlyCretaceousfossil evi- Leon-Kloosterziel, K. M., Keijzer, C. J. & Koornneef, M.
dence for angiospermevolution. Bot. 43: 3-104. 1994. A seed shape mutantof Arabidopsis that is affected
Hickey, L. J. & Taylor, D. W. 1996. Rev.
Origin the angiosperm
of in integumentdevelopment. PI. Cell 6: 385-392.
flower. Pp. 176-231 in: Taylor, D. W. & Hickey, L. J. Leroy, J.-F. 1983a. The origin of angiosperms: an unrecog-
(eds.), Flowering Plant Origin, Evolution & Phylogeny. nized ancestral dicotyledon, Hedyosum (Chloranthales),
Chapman& Hall, New York. with a strobiloid flower is living today. Taxon 32:
Hughes, N. F. 1976. Paleobiology of Angiosperm Origins. 169-175.
CambridgeUniv. Press, Cambridge. Leroy, J.-F. 1983b. Interpretationnouvelle des appareils sex-
Hughes, N. F. 1994. The Enigma of Angiosperm Origins. uels des Chloranthacees(Chloranthales,Magnoliidees). C.
CambridgeUniv. Press, Cambridge. R. Acad. Sc. Paris, ser. 3, 296: 747-752.
Hughes, N. F., Drewry, G E. & Liang, J. F. 1979. Barremian Les, D. H. 1988. The origin and affinities of the Cerato-
earliest angiospermpollen. Palaeontolog, 22: 513-535. phyllaceae. Taxon37: 326-345.
Hughes, N. F. & McDougall, A. B. 1987. Records of Les, D. H., Garvin, D. K. & Wimpee, C. F. 1991. Molecular
angiospermid pollen entry into the English Early evolutionaryhistoryof ancient aquaticangiosperms.Proc.
Cretaceous succession. Rev. Paleobot. Palvnol. 50: Natl. Acad Sci. U.S.A. 88: 10119-10123.
255-272. Loconte, H. & Stevenson, D. W. 1990. Cladistics of the
Hughes, N. F., McDougall, A. B. & Chapman, J. L. 1991. Spermatophyta.Brittonia42: 197-211.
Exceptional new record of Cretaceous Hauterivian Mabberley, D. J. 1984. The optimistic in pursuitof the unrec-
angiospermid pollen from Southern England. J. ognizable: a note on the origin of angiosperms. Taxon33:
Micropalaeontol. 10: 75-82. 77-84.
Igersheim, A. & Endress, P. K. 1997. Gynoecium diversity Mabberley, D. J. 1997. The Plant-Book. a Portable
and systematics of the Magnoliales and winteroids.Bot. J. Dictionary of the VascularPlants, ed. 2. CambridgeUniv.
Linn. Soc. 124: 213-271. Press, Cambridge.
Igersheim, A. & Endress, P. K. 1998. Gynoecium diversity Magall6n, S. & Sanderson, M. J. 2002. Relationshipsamong
and systematics of the paleoherbs.Bot. J. Linn. Soc. 127: seed plants inferredfrom highly conserved genes: sorting
289-370. conflicting phylogenetic signals among ancient lineages.
Johnson, L. A. S. & Wilson, K. L. 1989. Casuarinaceae;a Amer J. Bot. 89: 1991-2006.
synopsis. Pp. 167-188 in: Crane, P. R. & Blackmore, S. Mani, M. S. & Saravanan, J. M. 1999. Pollination Ecology
(eds.), Evolution, Systematics and Fossil History of the and Evolution in Compositae (Asteraceae). Science
Hamamelidae,vol. 2, "Higher"Hamamelidae.Clarendon Publishers,Enfield, New Hampshire.
Press, Oxford. Martin, W., Gierl, A. & Saedler, H. 1989. Molecularevidence
Judd, W. S., Campbell, C. S., Kellogg, E. A., Stevens, P. F. & for pre-Cretaceous angiosperm origins. Nature 339:
Donoghue, M. J. 2002. Plant Systematics:a Phylogenetic 46-48.
Approach, ed. 2. Sinauer Associates, Sunderland, Mathews, S. & Donoghue, M. J. 1999. The root of
Massachusetts. angiosperm phylogeny inferred from duplicate phy-
Kato, M. 1990. Ophioglossaceae:a hypotheticalarchetypefor tochrome genes. Science 286: 947-950.
the angiospermcarpel. Bot. J. Linn. Soc. 102: 303-311. Meeuse, A. D. J. 1966. Fundamentals of Phytomorphology.
Kenrick, P. & Crane, P. R. 1997. The Origin and Early The Ronald Press Co., New York.
Diversification of Land Plants.:a Cladistic Study. Smith- Meeuse, A. D. J. 1970. The descent of the flowering plants in
sonian InstitutionPress, Washington,D.C. the light of new evidence from phytochemistryand from
Kim, K.-J. & Jansen, R. K. 1995. ndhF sequence evolution other sources. Acta. Bot. Neerl. 19: 61-72.
and the major clades in the sunflower family. Proc. Natl. Meeuse, A. D. J. 1975. Floral evolution as the key to
Acad. Sci. U.S.A. 92: 10379-10383. angiospermsdescent. Acta Bot. Indica 3: 1-18.
Krassilov, V. A. 1977. The origin of angiosperms.Bot. Rev.43: Meister, R. J., Kotow, L. M. & Gasser, C. S. 2002. SUPER-
143-176. MAN attenuatespositive INNERNO OUTER autoregula-
Krassilov, V. A. 1997. Angiosperm Origins: Morphological tion to maintainpolar development of Arabidopsis ovule
and Ecological Aspects. Pensoft, Sofia. outer integuments.Development 129: 4281-4289.
Kuzoff, R. K. & Gasser, C. S. 2000. Recent progressin recon- Melville, R. 1962. A new theory of the angiosperm flower. I.
structing angiosperm phylogeny. Trends Pl. Sci. 5: The gynoecium. Kew Bull. 16: 1-50.
330-336. Melville, R. 1975. The distributionof Australian relict plants
Labandeira, C. C. & Sepkoski, J. J., Jr. 1993. Insect diversi- and its bearingon angiospermevolution. Bot. J. Linn.Soc.
ty in the fossil record.Science 261: 310-315. 71: 67-88.
Lane, M. A. 1996. Pollination biology of Compositae. Pp. Meyen, S. V. 1988. Origin of the angiosperm gynoecium by
61-80 in: Caligari, P. D. S. & Hind, D. J. N. (eds.), gamoheterotopy.Bot. J. Linn. Soc. 97: 171-178.
Compositae: Biology & Utilization. Proceedings of the Michaels, H. J., Scott, K. M., Olmstead, R. G, Szero, T.,

14
TAXON 53 (1) * February2004: 3-16 Stuessy * Theory for the origin of angiosperms

Jansen, R. K. & Palmer, J. D. 1993. Interfamilialrela- mal fusion and intercalarygrowth for angiosperm evolu-
tionships of the Asteraceae: insights from rbcL sequence tion. TrendsPl. Sci. 6: 111-113.
variation.Ann. Missouri Bot. Gard. 80: 742-751. Raven, P. H. 1977. A suggestion concerning the Cretaceous
Midgley, J. J. & Bond, W. J. 1991. Ecological aspects of the rise to dominance of the angiosperms. Evolution 31:
rise of angiosperms:a challenge to the reproductivesupe- 451-452.
riorityhypothesis. Biol. J. Linn. Soc. 44: 81-92. Ren, D. 1998. Flower-associatedBrachyceraflies as fossil evi-
Moldowan, J. M., Dahl, J., Huizinga, B. J., Fago, F. J., dence for Jurassic angiosperm origins. Science 280:
Hickey, L. J., Peakman, T. M. & Taylor, D. W. 1994. 85-88.
The molecularfossil recordof oleanane and its relationto Retallak, G & Dilcher, D. L. 1981. Arguments for a glos-
angiosperms.Science 265: 768-771. sopteridancestryof angiosperms.Paleobiology 7: 54-67.
Moseley, M. F. 1948. Comparativeanatomyand phylogeny of Retallak, G & Dilcher, D. L. 1988. Reconstructionsof select-
the Casuarinaceae.Bot. Gaz. 110: 231-280. ed seed ferns. Ann. Missouri Bot. Gard. 75: 1010-1057.
Muhammad, A. F. & Sattler, R. 1982. Vessel structureof Rothwell, G. W. & Serbet, R. 1994. Lignophyte phylogeny
Gnetumand the origin of angiosperms.Amer J. Bot. 69: and the evolution of spermatophytes:a numericalcladistic
1004-1021. analysis. Syst. Bot. 19: 443-482.
Mulcahy, D. L. 1979. The rise of the angiosperms:a geneco- Rydin, C. & Kaillersjo, M. 2002. Taxon sampling and seed
logical factor.Science 206: 20-23. plant phylogeny. Cladistics 18: 485-513.
Mulcahy, D. L. & Mulcahy, G B. 1987. The effects of pollen Rydin, C., Kaillersjo,M. & Friis, E. M. 2002. Seed plantrela-
competition.Amer Sci. 75: 44-50. tionships and the systematic position of Gnetales based on
Muller, J. 1970. Palynological evidence on early differentia- nuclear and chloroplast DNA: conflicting data, rooting
tion of angiosperms.Biol. Rev. 45: 417-450. problems, and the monophyly of conifers. Int. J. Pl. Sci.
Muller, J. 1981. Fossil pollen records of extant angiosperms. 163: 197-214.
Bot. Rev. 47: 1-146. Samigullin, T. K., Martin, W. F., Troitsky,A. V. & Antonov,
Nandi, O. I., Chase, M. W. & Endress, P. K. 1998. A com- A. S. 1999. Molecular data from the chloroplast rpoCI
bined cladistic analysis of angiosperms using rbcL and gene suggest a deep and distinct dichotomy of contempo-
non-molecular data sets. Ann. Missouri Bot. Gard. 85: rary spermatophytesinto two monophyla: gymnosperms
137-212. (including Gnetales) and angiosperms.J. Molec. Evol. 49:
Nishida, H., Pigg, K. B. & Rigby, J. F. 2003. Swimming 310-315.
sperm in an extinct Gondwanan plant. Nature 422: Sampson, F. B. 2000. Pollen diversity in some modern mag-
396-397. noliids. Int. J. Pl. Sci. 161 (6, suppl.): S193-S210.
Nixon, K. C., Crepet, W. L., Stevenson, D. & Friis, E. M. Sanderson, M. J. & Doyle, J. A. 2001. Sources of errorand
1994. A reevaluation of seed plant phylogeny. Ann. confidence intervalsin estimating the age of angiosperms
Missouri Bot. Gard. 81: 484-533. from rbcL and 18S rDNA data. Amer J. Bot. 88:
Olmstead, R. G, Jansen, R. K., Michaels, H. J., Downie, S. 1499-1516.
R. & Palmer, J. D. 1990. ChloroplastDNA and phyloge- Sanderson, M. J., Wojciechowski, M. F., Hu, J.-M., Sher
netic studies in the Asteridae. Pp. 119-134 in: Kawano,S. Khan, T. & Brady, S. G. 2000. Error,bias, and long-
(ed.), Biological Approaches and Evolutionary Trendsin branchattractionin data for two chloroplastphotosystem
Plants. Academic Press, London. genes in seed plants. Molec. Biol. Evol. 17: 782-797.
Olmstead, R. G., Michaels, H. J., Scott, K. M. & Palmer, J. Schmidt, M. & Schneider-Poetsch, H. A. W. 2002. The evo-
D. 1992. Monophyly of the Asteridae and identificationof lution of gymnosperms redrawn by phytochromegenes:
their major lineages inferred from DNA sequences of the Gnetatae appear at the base of the gymnosperms.J.
rbcL. Ann. Missouri Bot. Gard. 79: 249-265. Molec. Evol. 54: 715-724.
Parkinson, C. L., Adams, K. L. & Palmer, J. D. 1999. Schnarf, K. 1931. VergleichendeEmbrvologieder Angiosper-
Multigene analyses identify the three earliest lineages of men. Borntraeger,Berlin.
extant flowering plants. Curr Biol. 9: 1485-1488. Shamrov, 1. I. 1998. Ovule classification in flowering plants-
Pellmyr, O. 1992. Evolution of insect pollination and new approaches and concepts. Bot. Jahrb. Syst. 120:
angiospermdiversification. TrendsEcol. Evol. 7: 46-49. 377-407.
Qiu, Y.-L., Chase, M. W., Les, D. H. & Parks, C. R. 1993. Soltis, D. E., Soltis, P. S., Chase, M. W., Mort, M. E.,
Molecular phylogenetics of the Magnoliidae: cladistic Albach, D. C., Zanis, M., Savolainen, V., Hahn, W. H.,
analysis of nucleotide sequences of the plastid gene rbcL. Hoot, S. B., Fay, M. F., Axtell, M., Swensen, S. M.,
Ann. Missouri Bot. Gard. 80: 587-606. Prince, L. M., Kress, W. J., Nixon, K. C. & Farris, J. S.
Qiu, Y.-L., Lee, J., Bernasconi-Quadroni, F., Soltis, D. E., 2000. Angiosperm phylogeny inferred from 18S rDNA,
Soltis, P. S., Zanis, M., Zimmer, E. A., Chen, Z., rbcL, and atpB sequences. Bot. J. Linn. Soc. 133:
Savolainen, V. & Chase, M. W. 1999. The earliest 381-461.
angiosperms: evidence from mitochondrial, plastid, and Soltis, P. S., Soltis, D. E. & Chase, M. W. 1999. Angiosperm
nucleargenomes. Nature 402: 404-407. phylogeny inferredfrom multiple genes as a researchtool
Qiu, Y.-L., Lee, J., Bernasconi-Quadroni, F., Soltis, D. E., for comparativebiology. Nature402: 402-404.
Soltis, P. S., Zanis, M., Zimmer, E. A., Chen, Z., Soltis, P. S., Soltis, D. E., Savolainen, V., Crane, P. R. &
Savolainen, V. & Chase, M. W. 2000. Phylogeny of basal Barraclough, T. G. 2002. Rate heterogeneityamong line-
angiosperms:analyses of five genes from three genomes. ages of tracheophytes:integrationof molecularand fossil
Int. J P1. Sci. 161 (6, suppl.): S3-S27. data and evidence for molecular living fossils. Proc. Natl.
Raven, J. A. & Weyers, J. D. B. 2001. Significance of epider- Acad. Sci. U.S.A. 99: 4430-4435.

15
Stuessy * Theory for the origin of angiosperms TAXON 53 (1) * February2004: 3- 16

Soltis, D. E., Soltis, P. S. & Zanis, M. J. 2002. Phylogeny of Thorne, R. F. 2002. How many species of seed plants are
seed plants based on evidence from eight genes. Amer.J. there'?Taxon51: 511-512.
Bot. 89: 1670-1681. Troitsky, A. V., Melekhovets, Y. F., Rakhimova, G. M.,
Stebbins, G. I., Jr. 1974. Flowering Plants: EvolutionAbove Bobrova, V. K., Valiejo-Roman, K. M. & Antonov, A. S.
the Species Level. HarvardUniv. Press, Cambridge. 1991. Angiosperm origin and early stages of seed plant
Stuessy, T. F., Crawford, D. J., Anderson, G. J. & Jensen, R. evolution deduced from rRNA sequence comparisons. J.
J. 1998. Systematics, biogeography and conservation of Molec. Evol. 32: 253-261.
Lactoridaceae.Pers. Pl. Ecol. Evol. Syst. 1: 267-290. Verdui,M. 2002. Age at maturityand diversificationin woody
Stuessy, T. F. & Garver, D. 1996. The defensive role of pap- angiosperms.Evolution 56: 1352-1361.
pus in heads of Compositae. Pp. 81-91 in: Caligari, P. D. Wagenitz, G 1996. Wiirterbuchder Botanik.:die Terminiin
S. & Hind, D. J. N. (eds.), Compositae:. Biology & ihremhistorischenZussammenhang.Gustav Fischer,Jena.
Utilization. Proceedings of the International Compositae Walker, J. W. & Skvarla, J. J. 1975. Primitively columella-
Confierence,Ken: 1994, vol. 2. Royal Botanic Gardens, less pollen: a new concept in the evolutionarymorphology
Kew. of angiosperms.Science 187: 445-447.
Stuessy, T. F., Sang, T. & DeVore, M. L. 1996. Phylogeny and Walker, J. W. & Walker, A. G. 1984. Ultrastructureof Lower
biogeography of the subfamily Barnadesioideae with Cretaceous angiosperm pollen and the origin and early
implications for early evolution of Compositae. Pp. evolution of flowering plants. Ann. Missouri Bot. Gard.
463-490 in: Hind, D. J. N. & Beentje, H. (eds.), 71: 464-521.
Compositae: Systematics. Proceedings of the Inter- Wettstein, R. 1907. Handbuch der svstematischen Botanik.
national CompositaeConference,Ken: 1994, vol. 1. Royal Deuticke, Leipzig.
Botanic Gardens,Kew. Wettstein, R. 1924. Handbuchder systematischenBotanik, ed.
Stuessy, T. F. & Spooner, D. M. 1988. The adaptive and phy- 3. Deuticke, Leipzig.
logenetic significance of receptacular bracts in the Wiegmann, B. M., Yeates, D. K., Thorne, J. L. & Kishino,
Compositae. Taxon37: 114-126. H. 2003. Time flies, a new molecular time-scale for
Sun, G., Dilcher, D. L., Zheng, S. & Zhou, Z. 1998. In search Brachyceranfly evolution without a clock. Syst. Biol. 52:
of the first flower: a Jurassic angiosperm:Archaeftructus 745-756.
from NortheastChina. Science 282: 1692-1695. Williams, J. H. & Friedman, W. E. 2002. Identification of
Sun, G., Ji, Q., Dilcher, D. L., Zheng, S., Nixon, K. C. & diploid endosperm in an early angiospermlineage. Nature
Wang, X. 2002. Archaefructaceae,a new basal angio- 415: 522-526.
sperm family. Science 296: 899-904. Wolfe, K. H., Gouy, M., Yang, Y.-W., Sharp, P. M. & Li, W.-
Swisher, C. C., III, Rink, W. J., Ant6n, S. C., Schwarcz, H. H. 1989. Date of the monocot-dicot divergence estimated
P., Curtis, G. H. & Suprijo, A. & Widiasmoro. 1996. from chloroplast DNA sequence data. Proc. Natl. Acad.
Latest Homo erectus of Java: potential contemporaneity Sci. U.S.A. 86: 6201-6205.
with Homo sapiens in southeast Asia. Science 274: Wu, Z.-Y., Lu, A.-M., Tang, Y.-C., Chen, Z.-D. & Li, D.-Z.
1870-1874. 2002. Synopsis of a new "polyphyletic-polychromic-poly-
Swisher, C. C., III, Wang, Y.-Q., Wang, X.-L., Xu, X. & typic" system of the angiosperms.Acta Phytotaxon.Sinica
Wang, Y. 1999. Cretaceous age for the feathered 40: 289-322.
dinosaursof Liaoning, China. Nature 400: 58-61. Zanis, M. J., Soltis, D. E., Soltis, P. S., Mathews, S. &
Taylor, D. W. 1991. Angiospermovules and carpels:theirchar- Donoghue, M. J. 2002. The root of the angiosperms
acters and polarities, distribution in basal clades, and revisted. Proc. Natl. Acad. Sci. U.S.A. 99: 6848-6853.
structuralevolution. Postilla 208: 1-40. Zanis, M. J., Soltis, P. S., Qiu, Y. L., Zimmer, E. & Soltis, D.
Taylor, D. W. & Hickey, L. J. 1990. An Aptian plant with E. 2003. Phylogenetic analyses and perianthevolution in
attached leaves and flowers: implications tor angiosperm basal angiosperms.Ann. MissouriBot. Gard.90: 129-150.
origin. Science 247: 701-704. Zavada, M. S. & Gabarayeva, N. 1991. Comparativepollen
Taylor, D. W. & Hickey, L. J. 1992. Phylogeneticevidence for wall development of Welwitschiamirabilis and selected
the herbaceousorigin of angiosperms.Pl. Syst. Evol. 180: primitive angiosperms. Bull. Torr. Bot. Club 118:
137-156. 292-302.
Taylor, D. W. & Hickey, L. J. 1996. Evidence for and impli- Zavada, M. S. & Taylor, T. N. 1986. The role of self-compat-
cations of an herbaceous origin for angiosperms. Pp. ibility and sexual selection in the gymnosperm-angio-
232-266 in: Taylor, D. W. & Hickey, L. J. (eds.), sperm transition:a hypothesis. Amer Nat. 128: 538-550.
Flowering Plant Origin, Evolution & Phylogeny. Zhou, Z., Barrett P. M. & Hilton, J. 2003. An exceptionally
Chapman& Hall, New York. preserved Lower Cretaceous ecosystem. Nature 421:
Taylor, D. W. & Kirchner, G. 1996. The origin and evolution 807-814.
of the angiospermcarpel. Pp. 116-140 in: Taylor,D. W. &
Hickey, L. J. (eds.), Flowering Plant Origin, Evolution &
Phylogeny.Chapman& Hall, New York.
Taylor, T. N., Del Fueyo, G. M. & Taylor, E. L. 1994.
Permineralized seed fern cupules from the Triassic of
Antarctica:implications for cupule and carpel evolution.
Amer.J. Bot. 81: 666-677.
Thomas, H. H. 1936. Paleobotany and the origin of the
angiosperms.Bot. Rev. 2: 397-418.

16