patterns of segregation and convergence in the evolution of fern and

q 2005 The Paleontological Society. All rights reserved. 0094-8373/05/3101-0008/$1.00

Paleobiology, 31(1), 2005, pp. 117–140

Patterns of segregation and convergence in the evolution of fernand seed plant leaf morphologies

C. Kevin Boyce

Abstract.—Global information on Paleozoic, Mesozoic, and extant non-angiosperm leaf morphol-ogies has been gathered to investigate morphological diversity in leaves consistent with marginalgrowth and to identify likely departures from such development. Two patterns emerge from theprincipal coordinates analysis of this data set: (1) the loss of morphological diversity associatedwith marginal leaf growth among seed plants after sharing the complete Paleozoic range of suchmorphologies with ferns and (2) the repeated evolution of more complex, angiosperm-like leaftraits among both ferns and seed plants. With regard to the first pattern, morphological divergenceof fern and seed plant leaf morphologies, indirectly recognized as part of the Paleophytic-Meso-phytic transition, likely reflects reproductive and ecological divergence. The leaf-borne reproduc-tive structures that are common to the ferns and Paleozoic seed plants may promote leaf morpho-logical diversity, whereas the separation of vegetative and reproductive roles into distinct organsin later seed plant groups may have allowed greater functional specialization—and thereby mor-phological simplification—as the seed plants came to be dominated by groups originating in morearid environments. With regard to the second pattern, the environmental and ecological distri-bution of angiosperm-like leaf traits among fossil and extant plants suggests that these traits pref-erentially evolve in herbaceous to understory plants of warm, humid environments, thus sup-porting inferences concerning angiosperm origins based upon the ecophysiology of basal extanttaxa.

C. Kevin Boyce. Organismic and Evolutionary Biology, Harvard University, 26 Oxford Street, Cambridge,Massachusetts 02138

Present address: Department of the Geophysical Sciences, University of Chicago, 5734 South Ellis Avenue,Chicago, Illinois 60637. E-mail: [email protected]

Accepted: 4 June 2004

Introduction

During the Late Devonian and Early Car-boniferous, at least four vascular plant line-ages (seed plants, progymnosperms, ferns,and sphenopsids) independently evolved lam-inate leaves and followed the same early se-quence of morphological evolution. After thisinitial radiation, the ferns and seed plantsshared nearly the complete morphologicalrange found in Paleozoic leaves (Boyce andKnoll 2002). This repeated pattern of earlyevolution suggests a highly constrained radi-ation; however, this early history of morpho-logical evolution contrasts strongly with themodern world dominated by angiospermswith leaf morphologies radically differentfrom nearly all Paleozoic forms. Morphologiesreminiscent of the Paleozoic do persist, butprimarily only among ferns.

Living plants provide a developmental con-text for this evolutionary history. Marginalgrowth is found in fern laminae (Pray 1960,

1962; Zurakowski and Gifford 1988) that haveone or two orders of veins with strictly mar-ginal vein endings. A causal link betweenthese morphological and developmental traitsis consistent with current understanding ofvascular differentiation along gradients of thehormone auxin produced in growing areas(e.g., Sachs 1991; Berleth et al. 2000). Alterna-tives to strictly marginal growth, includingcell divisions dispersed throughout the leaf,are found in angiosperm leaves (Pray 1955;Poethig and Sussex 1985a,b; Hagemann andGleissberg 1996; Dolan and Poethig 1998) thathave many orders of veins and dispersed in-ternal vein endings. These correlates havebeen used to interpret the fossil record of mor-phological evolution as reflecting the indepen-dent evolution of marginal meristematicgrowth in multiple lineages in the Paleozoicand the evolution of departures from strictlymarginal leaf growth, notably in the angio-sperm lineage, which dominates modern flo-ras (Boyce and Knoll 2002).

118 C. KEVIN BOYCE

Issues regarding this transition remain un-resolved. First, the distinct morphological anddevelopmental characteristics of angiospermleaves have factored in several theories con-cerning the environmental and ecological or-igins of the group, but rarely have they beenconsidered as part of the broader history ofleaf evolution. Second, most morphologies re-lated to strictly marginal growth are now as-sociated only with ferns. This loss of seedplant morphological diversity may either sim-ply reflect the depauperate nature of the ex-tant gymnosperm flora—and thereby perhapsbe tied to the rise of an alternative form ofmorphological diversity among angio-sperms—or reflect an evolutionary trend in-dependent of the decline in gymnosperm di-versity.

Analysis of Morphological Diversity in theLeaves of Fossil and Extant Plants

Leaf morphologies of extant plants and Pa-leozoic and Mesozoic fossils were surveyed atthe generic level for 19 discrete characters de-scribing the lamina and venation (see Appen-dices 1, 2, and 3 for character list, references,and morphological data and ranges). Thisdata set consists of 281 fossil and 185 extantgenera. Of the fossil genera, 107 are seedplants, 60 are ferns, seven are progymnos-perms, four are sphenopsids, and 103 are ofother or unknown affinities. Of the extant gen-era, 168 are ferns and 17 are seed plants. Oc-currence data were assigned to geologic epochor period, on the basis of durations stated intaxonomic descriptions and expanded by oc-currence information reported from individ-ual localities.

The data were summarized with a principalcoordinates analysis (PCO). The pairwisecomparison of all taxa was used to create adissimilarity matrix, the eigenvectors ofwhich form the axes of the PCO after Gowertransformation of the matrix (Gower 1966)and multiplication of each eigenvector by itscorresponding eigenvalue (for greater detailsee Foote 1995; Lupia 1999; Boyce and Knoll2002). PCO provides a method for visualizinglarge quantities of morphological informationby geometrically summarizing as much of thevariability between taxa as possible on a few

axes in the form of a morphospace. The firsttwo PCO axes (Fig. 1) contain 55.7% of the in-formation of the original data matrix, and thefirst three axes (Fig. 2) contain 74.1% as esti-mated by the sum of their eigenvalues dividedby the sum of all eigenvalues (Foote 1995).PCO was supplemented by plotting of the av-erage pairwise dissimilarity for all taxa andfor several groups analyzed individually, aswell as by plotting of the partitioned contri-butions of individual groups to the overallvariance (Fig. 3). Partitioned variance is basedon the squared Euclidean distances betweenthe members of a group and the overall cen-troid for the 123 principal coordinate axeswith positive eigenvalues (Foote 1993; Lupia1999).

The character list was designed to describethe morphological diversity within leaves con-sistent with marginal growth and to identifyleaves that suggest the evolution of departuresfrom marginal growth. Hence, the large mor-phological diversity found within angio-sperms would not be circumscribed with thecurrent character list except as the evolution ofan alternative to marginal growth. Proper in-vestigation of morphological patterns withinangiosperm-like leaves would require a largenumber of characters (e.g., Hickey 1974; LeafArchitecture Working Group 1999) that wouldnot apply to and would obscure patternswithin leaf forms consistent with marginalgrowth. Beyond an inadequate description ofangiosperm morphological diversity with thecurrent character list, including a large num-ber of nearly identically coded angiospermleaf genera to the current data set would over-emphasize the narrow range of characterstates found within the angiosperms duringthe PCO analysis and hinder investigation ofpatterns within marginally organized leaves.(For further discussion of the sensitivities ofsuch analyses, see McGhee 1999; Boyce andKnoll 2002.) The angiosperms have thereforebeen excluded. Similarly, linear leaves, such asthose of the lycopods and many conifers andsphenopsids, would all be coded identicallyand were not included. The morphologicalcodings that angiosperms and linear leaveswould produce with this character set are ap-proximated respectively by the Gnetales and

119FERN AND SEED PLANT LEAF MORPHOLOGIES

by Czekanowskia and several progymnos-perms. All other leaf taxa with adequate de-scription and preservational detail were in-cluded.

Results

This global analysis demonstrates that thelate Paleozoic maximum of morphological di-versity documented previously (Boyce andKnoll 2002) among leaves consistent withmarginal growth largely represents the mor-phological range of such leaves in later time aswell (Figs. 1, 3). This morphological range isshared by the ferns and seed plants early intheir history but is subsequently partitionedbetween the two groups, with the seed plantsprogressively losing much of this range aftera Permian maximum (Figs. 1, 3). When not re-duced to linear leaves, as in most conifers,leaves are less compound (a characteristic notincluded in the principal coordinates analy-sis); lamina attachment is often broad withmultiple equivalent veins entering the leaf; ve-nation is typically simplified to a single orderwithout a midvein; and venation is parallel,typically open, and runs a straight course end-ing at the distal margin. These characteristicsare found in Cordaitales, Ginkgoales, Cyca-dales, Bennettitales, and conifers includingthose with nonlinear leaves, as well as morepoorly understood groups such as the Czek-anowskiales and Vojnovskyales. Ferns rarelyhave exhibited the morphologies typical ofpost-Paleozoic seed plants, but otherwise theyhad occupied the morphological range of Pa-leozoic seed plants by the Triassic and theyhave maintained this morphological diversitythrough to the present.

Complex vascular characteristics that sug-gest departures from strictly marginal leafgrowth have evolved repeatedly, and these ve-nation patterns are found at low frequency atall times after their first appearance in thePermian gigantopterid seed plants (Fig. 4).Other seed plants to have independentlyevolved these leaf characteristics are the Tri-assic peltasperms, as well as the angiospermsand Gnetales with leaf macrofossil recordsthat begin in the Cretaceous. These leaf traitsoccur in all three extant fern orders: in Ophiog-lossum of the Ophioglossales, in Christensenia

of the Marattiales, and in several filicaleangroups (Fig. 5). Filicales include the Dipteri-daceae (marking the Triassic first appearanceof these characteristics among ferns) and atleast four lineages in the large clade encom-passing polypod and dryopterid ferns: Blech-naceae, Thelypteridaceae, Lomariopsidaceae/Dryopteroideae, and Polypodiaceae/Grami-tidaceae. The last two lineages are large, het-erogeneous groups in which internallydirected veins likely evolved many times.Some of these characteristics also are found infossils, such as the Triassic genus Sanmiguelia(Cornet 1986), of unknown or controversial af-finity (Trivett and Pigg 1996).

The Morphological Divergence of Fernsand Seed Plants

After the Paleozoic, the seed plants pro-gressively lost leaf morphologies that persistamong the ferns. This pattern corresponds, atleast superficially, to extinction patterns with-in the two groups. Among lineages with a fos-sil record, 16 out of 24 filicalean families areextant, with only three lost since the Paleo-zoic, compared with only 4 of 19 non-angio-sperm seed plant orders (compiled respec-tively from Collinson 1996; Taylor and Taylor1993). Morphological patterns therefore mightreflect the persistence of diverse morphologiesamong the ferns as a result of the persistenceof basal lineages, versus progressive loss ofseed plant morphological diversity as a con-sequence of the loss of higher-level groups.However, the evolution of leaf morphology isnot sufficiently conservative for extinction tobe the cause of these morphological patterns.Among extant ferns, the entire range of mor-phologies consistent with strictly marginalgrowth is covered by the genera of a variety ofindividual families (Fig. 2A). Furthermore,linking morphological diversity to extinctionrates would not address the absence of fernsfrom the morphological range to which theseed plants became restricted (Figs. 1A, 2B).

The leaves of several early diversifying seedplant groups, such as the Medullosales andLyginopteridales, are as morphologically di-verse as those of the ferns, but later groups ex-hibit a much more restricted range of mor-phologies. DiMichele and Aronson (1992) em-

120 C. KEVIN BOYCE

FIGURE 1. Principal coordinates analysis of global Paleozoic and Mesozoic fossils and extant fern and gymnospermgenera. Stratigraphic distribution of morphologies for all taxa (A) and phylogenetic distribution of morphologiesfor fossil and extant seed plants segregated based on Paleozoic or Mesozoic time of greatest diversity (B, C). ‘‘Ferns’’includes Filicales, Marattiales, Ophioglossales, and Zygopteridales. ‘‘Paleozoic pteridosperms’’ includes Lyginop-teridales, Medullosales, and Callistophytales. ‘‘Mesozoic pteridosperms’’ includes Peltaspermales, Corystosper-males, and Caytoniales.


FIGURE 2. Phylogenetic distribution of morphologies among extant ferns (A), based on principal coordinates anal-ysis of Figure 1 plotted against a black background of all extant fern genera and among all extant (B) plants plottedagainst a black background of all fossils included in the analysis. (Phylogeny based on Hasebe et al. 1994 and Pryer1995. Taxon names available in Figure 5.)

phasized that drier environments foster theevolution of morphological novelty and doc-umented that many higher-level groups orig-inated in drier extrabasinal settings before lat-er moving into basinal environments as wet-lands contracted late in the Paleozoic. This‘‘paleophytic/mesophytic’’ transition (re-viewed in Knoll 1984) corresponds to the re-placement seen here of seed plant groups withmorphologically diverse fernlike leaves bythose with a more restricted range of leaf ve-

nation patterns and morphologies. Therefore,dry, peripheral environments may well havefostered the evolution of the innovations in re-productive morphology that are given themost weight in establishing ordinal-level seedplant groups, but these environments appearto have fostered convergence in the evolutionof leaf morphologies. This morphological con-vergence may well be adaptive in drier envi-ronments. The theoretically ideal morphologyfor maximizing water transport efficiency

122 C. KEVIN BOYCE

FIGURE 3. A, Pairwise dissimilarity for the entire data set and for individual groups analyzed separately. B, Con-tribution of different groups to the variance of the overall data set. Seed plant subgroups are not expected to bemonophyletic. Jumps in variance values between the Late Cretaceous and the Recent reflects the larger number oftaxa available in the living record and the fact that all living taxa can be assigned unambiguously to ferns or seedplants, whereas a large proportion of taxa are unassignable to either group in all other time intervals.

(Givnish 1979) is approached by these seedplant groups: a leaf with a single order ofstraight, unbranched veins radiating from thepoint of attachment.

The morphologically diverse leptosporan-giate ferns, however, are also among thegroups thought to have originated in drier, pe-ripheral environments (Scott and Galtier 1985;DiMichele and Aronson 1992). The dual rolefern leaves typically play in both photosyn-thesis and reproduction might lead to a largerrange of morphologies than would be expect-ed by selection for photosynthetic function

alone. The need for greater supply of waterand photosynthate to specific areas of sporeproduction rather than to a more homoge-neous photosynthetic surface would likely re-sult in different optimal vein patterns. Sup-porting the mass of reproductive structuresmay in some cases favor alternative vein pat-terns as well.

This conflict of optimizing for both photo-synthetic and reproductive function might ex-plain the overlap between the morphologicalrange of the ferns and the Paleozoic pterido-sperms, known in several cases to have borne


FIGURE 4. Stratigraphic distribution of basic venationcharacteristics in fossil and extant leaves plotted onprincipal coordinates analysis of Figure 1.

FIGURE 5. Phylogenetic distribution of venation char-acteristics among extant ferns. Fern families can behighly variable, and labeled taxa often include membersthat do not conform to the indicated morphologicalcharacteristics.

reproductive structures on the lamina of fo-liage leaves (medullosans [Halle 1929]; gigan-topterids [Li and Yao 1983]; glossopterids[Surange and Maheshwari 1970; Surange andChandra 1972]; others [Delevoryas and Taylor1969; Galtier and Bethoux 2002]). The mor-phological ranges of later seed plant groupswould then be free to diverge from that pre-viously shared with the ferns when the re-quirements of reproductive function were seg-regated into separate structures in later seed

plants. This pattern is shown in full withinfossils that have been interpreted as cycads:the earliest have seeds borne directly on leavesof taeniopterid morphologies (Mamay 1976;Gillespie and Pfefferkorn 1986; Axsmith et al.2003), whereas Mesozoic cycads with repro-ductive organs segregated into independentstructures tend towards leaf morphologiesmore typical of extant Cycas and Zamiaceae(e.g., Triassic Leptocycas [Delevoryas and Hope1971]).

Some ferns do have morphologically dis-tinct fertile and nonfertile leaves. This dimor-phism often consists only of a less extensivelamina of the same morphology, but can in-volve complete loss of the lamina in entirefronds or portions thereof. However, the ex-tent of dimorphism can be variable within in-dividual genera and families and, even in themore extreme cases, fertile and vegetativeleaves typically share the same frond architec-ture. These distinctions suggest that much less

124 C. KEVIN BOYCE

developmental and genetic divergence be-tween these forms is present than would beexpected with any of the modifications be-tween fertile and vegetative structures exhib-ited by extant seed plants.

If seed plant lineages with and without leaf-borne reproductive structures are consideredseparately, two distinct dynamics can be rec-ognized in the overall seed plant pattern ofgradual decline following a Paleozoic high inmorphological disparity. First, seed plantswith leaf-borne reproductive structures flour-ish in the Paleozoic before disappearing earlyin the Mesozoic. Second, seed plants with leaf-independent reproductive structures main-tain a smaller, stable morphological rangethroughout their history through to the pre-sent (Fig. 3). The first appearance of this latter,more derived class is with the mid-Carbonif-erous Cordaitales, which had simple, strap-shaped leaves with a single order of parallel,open veins running to the distal margin thatwere remarkably different from the large, pin-nately compound fronds with a diverse arrayof pinnule venation patterns that were borneby Carboniferous pteridosperms.

In living seed plants, peak reproductiveand photosynthetic activity are typically tem-porally segregated, even in the Tropics (re-viewed in Burnham 1993). Staggered sched-ules of vegetative and reproductive organ pro-duction confer potentional advantages forboth biotic and abiotic pollination syndromes(Fenner 1998; Sakai 2001), each of which hasbeen inferred for different Paleozoic seedplants (Taylor and Millay 1979; Taylor 1988;Crepet 2001). Exploitation of such advantageswould be precluded if seed and pollen organsare borne on leaves. Furthermore, direct hy-draulic and energetic competition of photo-synthetic and reproductive activities mayhave made Paleozoic pteridosperms more vul-nerable to fluctuations in water supply. Seg-regation of these processes into independentorgans may have been important for the oc-cupation of drier and more seasonal habitatswhere many later seed plant groups appear tohave originated.

Segregation of photosynthetic and repro-ductive functions would allow the specializa-tion of leaves for hydraulic supply and the

morphologies typical of mesophytic foliage.However, several groups likely to have leaf-in-dependent reproductive structures at leastpartially deviate from this pattern: the peltas-perms, corystosperms, Caytoniales, and Pen-toxylon. Although their architecture often isunknown, these plants, which maintain amore diverse array of leaf morphologies, tendto be smaller, subcanopy plants (Harris 1983;Bose et al. 1985) for which hydraulic speciali-zation of the leaves may have been less im-portant. Where either canopy dominance ordrier, more open environments are likely inthese groups, there is a trend towards moremesophyte-like morphologies, e.g., Xylopterisin the Corystosperms. Such partial approxi-mations are also found among larger trees inlineages with leaf-borne reproductive struc-tures, e.g., the glossopterids, and the largestMedullosan trunks in the earliest Permian (re-viewed in Taylor and Taylor 1993) coincidentwith abundant Odontopteris foliage.

Evolutionary constraints are often evokedwith regard to morphological limitations (re-viewed in Wagner 2001), but the patterns dis-cussed here suggest that such constraints canhave unexpected outcomes. A single optimalform may exist for a particular function, butoptimizing for multiple functions results inmany coequal, suboptimal forms (Niklas1994). The evolution of a narrow range of leafmorphologies may have been permissible inseed plants, owing to the segregation of re-productive and photosynthetic functions. Theferns have always had to balance reproductiveand vegetative leaf functions and have alwayshad a diverse range of leaf morphologies. Inthis case, any evolutionary constraints appearto have been acting upon the morphologicallydiverse group. Presumably, these later seedplants would be more morphologically di-verse, rather than less, if all of morphologyand anatomy were taken into account, insteadof just leaves. However, the possibility of suchcounterintuitive relationships between evolu-tionary constraint and its expression in thefossil record of morphological diversityshould be considered, because paleontologistsmust always work with such partial records,regardless of study organisms.


The Repeated Evolution of Alternatives toStrictly Marginal Growth

Leaf venation with many orders, extensivereticulation, and vein endings dispersedthroughout the leaf is often considered a hall-mark of the angiosperms, and the presence ofthis syndrome, or a subset of its characteris-tics, has been used to argue angiosperm affin-ity for a variety of fossil plants (e.g., Surange1966; Melville 1969; Asama 1985; Cornet1986). These leaf characteristics have also beenevoked as a part of several alternative hypoth-eses concerning the ancestral environmentand ecology of the angiosperms. Recognizingthe developmental novelty of angiospermleaves in comparison to the ancestral marginalleaf growth, it has been suggested that theseleaves represent a complete reinvention of alaminate leaf after passage through an evolu-tionary phase in which the lamina was lostthrough adaptation to an arid, or perhapsaquatic, environment (Doyle and Hickey1976). It has also been argued that details ofangiosperm leaves support an herbaceous or-igin for the group along mesic, frequently dis-turbed stream margins (Taylor and Hickey1996). Other suggested advantages of differ-ent aspects of angiosperm leaf venation pat-terns, without necessarily making claims re-garding angiosperm origins, include in-creased structural support of a large thin lam-ina (Givnish 1979), the maximization ofphotosynthetic rate with high light and nutri-ent availability (Bond 1989), a neotenic accel-eration of the life cycle (Takhtajan 1976), andthe ensuring of an adequate water supply un-der arid conditions (reviewed in Roth-Nebel-sick et al. 2001). The obvious complication isthat hypotheses concerning the site of angio-sperm origination and the initial selective ad-vantage of various aspects of angiosperm bi-ology are poorly constrained because angio-sperms now dominate most terrestrial envi-ronments. However, any patterns of ecologicaldistribution among the unrelated groups thathave convergently evolved angiosperm-likeleaves may perhaps indicate the initial ecolog-ical conditions from which the angiospermsradiated.

Though less frequent than other morpho-

logical transitions, such as from open to retic-ulate venation, the evolution of leaves with in-ternally directed veins and nonmarginal veinendings has occurred at least four times in theseed plants and seven times among ferns. Al-though these leaf morphologies are often dis-tinctive and unlikely to be confused withthose of angiosperms, they are all similar insuggesting departures from the strictly mar-ginal laminar growth inferred to be the an-cestral condition in both ferns and seed plants(Doyle and Hickey 1976; Wagner 1979; Boyceand Knoll 2002). The relatively frequent evo-lution of such characteristics suggests thatthey are not an adequate indication of angio-spermous affinity for problematic fossils with-out other lines of evidence. Also, passagethrough a nonlaminate evolutionary stagemay be likely for the Gnetales (Doyle andHickey 1976) if linear-leaved Ephedra is basalwithin the group and if they are indeed close-ly related to or even nested within the conifers(Chaw et al. 1997, 2000; Winter et al. 1999;Bowe et al. 2000; Magallon and Sanderson2002), but the frequent evolution of these leafcharacteristics among the ferns indicates thatsuch a scenario should not be considered nec-essary for the angiosperms.

Ferns represent the majority of independentevolutions of leaf morphologies that suggestdepartures from strictly marginal growth.The center of diversity of these lineages is inpredominantly shaded areas of warm wet en-vironments (compiled from Kramer andGreen 1990), as is the case for ferns in generaland the first appearance of many lineageswithin the Filicales (Skog 2001). The Triassicfirst occurrence of complex reticulation andinternally directed veins within the ferns is inthe dipterids, known from China (Li 1995),which was a series of equatorial islands withhigh rainfall at this time (Shangyou et al.1990), and from northern (Harris 1926) andsouthern (Retallack 1977; Anderson and An-derson 1985) midlatitudes, which were warm,high-rainfall zones in hothouse climate re-gimes (Wing and Sues 1992; Ziegler et al.2003). The pre-Cenozoic record of the clade in-cluding the dryopterid and polypod ferns iscontroversial (Collinson 1996), and neitherOphioglossum nor Christensenia has a fossil re-

126 C. KEVIN BOYCE

cord, but the Cenozoic radiation of these lin-eages appears to be coincident with the rise ofmodern, angiosperm-dominated, tropical rainforests.

Fossil seed plant examples of the evolutionof more angiosperm-like leaves largely corre-spond to the conditions described for ferns;they tend to be smaller plants in warm, at leastseasonally wet environments, although not al-ways likely to have been heavily shaded. Gi-gantopterid leaves are known from the Perm-ian of China and the southwestern UnitedStates, although the two may not be closely re-lated (Asama 1962, 1985). The diverse Chinesematerial allows reconstruction of the plant asa liana (Li and Taylor 1999) in a wet, tropicalforest (Ziegler 1990), although such lycopod-dominated forests are likely to have had anopen canopy (DiMichele and DeMaris 1987).North American gigantopterids are found inriparian environments that would have beenthe wettest areas in a landscape at least sea-sonally dry (DiMichele and Hook 1992). Tri-assic Sanmiguelia is found on levees in season-ally wet environments (Ash 1987). The peltas-perms are thought to have been woody, be-cause of the deciduous nature of their leavesand reproductive organs (Crane 1985). Com-plex vascular characteristics are found in agroup of Triassic foliage taxa that have beenrelated to the peltasperms by cuticular char-acteristics and interpreted as an understorycomponent of conifer-dominated, wet, warmtemperate forests of mid latitude Eurasia(Dobruskina 1975, 1995).

Morphological and ecological divergencewithin the living Gnetales hampers attemptsto infer ancestral states for the group (Crane1996). The living representatives with nonlin-ear leaves are Gnetum, a liana to small tree intropical forests, and the desert plant Welwit-schia. In the fossil record, Drewria from theEarly Cretaceous Potomac Group of Virginia(Crane and Upchurch 1987) had a slenderstem with no evidence of secondary growth orbud dormancy and the plant is reconstructedas an herbaceous to perhaps shrubby memberof streamside, early-successional vegetation ina mesic environment. A diverse gnetalean as-semblage, including broadleaf examples, thathas recently been discovered in the Early Cre-

taceous of Brazil (Rydin et al. 2003) shouldprovide a more robust understanding of theearlier history of the group.

The evidence presented here concerning thepreferential ecology and environment of ori-gin of angiosperm-like leaf traits in non-an-giospermous plants is consistent with severallines of evidence from the early history of theangiosperms and their basal living members.The paucity of Early Cretaceous angiospermwood relative to other angiosperm organs andto conifer wood suggests that the angio-sperms were not initially large trees (Wingand Tiffney 1987). Recent molecular phylog-enies have converged on a phylogeny of an-giosperms (Mathews and Donoghue 1999;Qiu et al. 1999; Soltis et al. 1999; Barkman etal. 2000) with basal branches that consist ofsmaller woody plants of shaded, wet, tropicalenvironments that are frequent sites of distur-bance (Feild et al. 2003a,b). In this setting, theevolution of more angiosperm-like vein pat-terns may be involved in providing physicalsupport (Givnish 1979) and equitable distri-bution of water (Zwieniecki et al. 2002) acrossthe larger lamina common in shaded environ-ments, perhaps also aiding in the exploitationof light flecks. The remarkable aspect of an-giosperm leaves may not be so much theircomplex morphology, as such complexity hasevolved repeatedly in smaller, understoryplants in warm, humid environments, butrather the ability of angiosperms to export thiscomplex morphology to so many other envi-ronments.

Discussion

For the terrestrial biota, the Paleozoic hasbeen described as a time of ecosystem assem-bly fundamentally different from later time(DiMichele and Hook 1992), an assessmentbroadly mirrored by the patterns describedhere. After the initial evolution of develop-mental mechanisms in Paleozoic plants, eco-logical and architectural specializations werepartitioned in the post-Paleozoic, includingthe phylogenetic segregation of morphologiesderived from marginal growth and the re-peated evolution of departures from strictlymarginal growth.

A great deal of developmental diversity is


found among plants with nonmarginal leafgrowth (e.g., comparison of Liriodendron andQuiina in Foster 1952). For example, some fos-sil morphologies that are here lumped withdiffuse growth, such as the gigantopterids Ev-olsonia and Gigantopteridium, suggest a com-bination of marginal and internal growth indiscrete intercalary zones. Because of this di-versity, the repeated evolutions of alternativesto marginal growth are unlikely simply toreplicate a fixed evolutionary sequence ofmorphologies, as was seen with the evolutionof marginal growth during the Paleozoic(Boyce and Knoll 2002); however, a much moredetailed analysis would be needed to deter-mine whether regularities exist. The angio-sperm leaves in the Cretaceous PotomacGroup show a progression toward more high-ly organized venation and more-discrete dif-ferences between successive vein orders(Doyle and Hickey 1976), a pattern with whichother groups can be compared. Despite anydevelopmental diversity subsumed into theclassification of alternatives to marginalgrowth, these leaves may be physiologicallyconvergent. Vessels have been shown to playa key role in dicot leaf hydraulic function andin the equitable distribution of water acrossthe broad laminar surface (Zwieniecki et al.2002); this role is consistent with the two otherknown evolutions of vessels among the seedplants being in the only examples of alterna-tives to marginal leaf growth for which anat-omy is available, the Gnetales and gigantop-terids (Boyce in press).

In the same way that the enormous radia-tion of angiosperms has obscured the environ-ment and ecology of their origin, the angio-sperm radiation has likely obscured a varietyof patterns concerning the former ecologicaldistributions of other plant lineages (Wingand Sues 1992) and of leaf morphologies de-rived from strictly marginal growth. Futurelocality-based investigations may reveal moredetailed patterns of morphological occupationin pre-angiosperm floras, including latitudi-nal and ecotype gradients as well as furthersegregation than has been shown here be-tween plants of different overall habit. Fur-thermore, the nearly complete overlap of liv-ing fern and gymnosperm leaf morphologies

with those of fossils (Fig. 2B) suggests thatmany hypotheses concerning developmentaland physiological implications of differentleaf morphologies seen in the fossil record canbe tested with study of living analogues.

Acknowledgments

Software for creation of the dissimilaritymatrix and analysis of variance was written byR. Lupia (available at http://geosci.uchicago.edu/paleo/csource). B. Craft provided pro-gramming assistance. A. Knoll, N. M. Hol-brook, D. Pfister, C. Marshall, and D. Sunder-lin provided helpful discussion. This workwas supported by a National Science Foun-dation grant (ERA 0106816) and a NationalResearch Council/NASA Astrobiology Insti-tute associateship.

Literature CitedAnderson, J. M., and H. M. Anderson. 1985. Palaeoflora of

southern Africa. Balkema, Rotterdam.Asama, K. 1962. Evolution of the Shansi flora and origin of the

simple leaf. Science Reports of the Research Institute of To-hoku University, series 2, special volume 5:247–273.

———. 1985. Permian to Triassic floral change and some prob-lems of the paleobiogeography, parallelism, mixed floras, andthe origin of the angiosperms. Pp. 199–218 in K. Nakazawaand J. M. Dickens, eds. The Tethys. Tokyo University Press,Tokyo.

Ash, S. R. 1987. The Upper Triassic red bed flora of the ColoradoPlateau, Western United States. Journal of the Arizona-Neva-da Academy of Science 22:95–105.

Axsmith, B. J., R. Serbet, M. Krings, T. N. Taylor, E. L. Taylor,and S. H. Mamay. 2003. The enigmatic Paleozoic plants Sper-mopteris and Phasmatocycas reconsidered. American Journal ofBotany 90:1585–1595.

Barkman, T. J., G. Chenery, J. R. McNeal, J. Lyons-Weiler, W. J.Ellisens, G. Moore, A. D. Wolfe, and C. W. dePamphilis. 2000.Independent and combined analyses of sequences from allthree genomic compartments converge on the root of flow-ering plant phylogeny. Proceedings of the National Academyof Sciences USA 97:13166–13171.

Beck, C. B. 1976. Origin and early evolution of angiosperms. Co-lumbia University Press, New York.

Berleth, T., J. Mattson, and C. S. Hardtke. 2000. Vascular conti-nuity and auxin signals. Trends in Plant Science 5:387–393.

Bond, W. J. 1989. The tortoise and the hare: ecology of angio-sperm dominance and gymnosperm persistence. BiologicalJournal of the Linnean Society 36:227–249.

Bose, M. N., P. K. Pal, and T. M. Harris. 1985. The Pentoxylonplant. Philosophical Transactions of the Royal Society of Lon-don B 310:77–108.

Bowe, L. M., G. Coat, and C. W. dePamphilis. 2000. Phylogenyof seed plants based on all three genomic compartments: ex-tant gymnosperms are monophyletic and Gnetales’ closestrelatives are conifers. Proceedings of the National Academyof Sciences USA 97:4092–4097.

Boyce, C. K. In press. The evolutionary history of roots andleaves. In M. A. Zwieniecki and N. M. Holbrook, eds. Longdistance transport processes in plants. Elsevier, Amsterdam

128 C. KEVIN BOYCE

Boyce, C. K., and A. H. Knoll. 2002. Evolution of developmentalpotential and the multiple independent origins of leaves inPaleozoic vascular plants. Paleobiology 28:70–100.

Burnham, R. J. 1993. Time resolution in terrestrial macrofloras:guidelines from modern accumulations. In S. M. Kidwell andA. K. Behrensmeyer, eds. Taphonomic approaches to time res-olution in the fossil record. Short Courses in Paleontology 6:57–78. Palentological Scoeity, Knoxville, Tenn.

Chaw, S.-M., A. Zharkikh, H. M. Sung, T. C. Lau, and W. H. Li.1997. Molecular phylogeny of extant gymnosperms and seedplant evolution: analysis of nuclear 18s rRNA sequences. Mo-lecular Biology and Evolution 14:56–68.

Chaw, S.-M., C. L. Parkinson, Y. Cheng, T. M. Vincent, and J. D.Palmer. 2000. Seed plant phylogeny inferred from all threeplant genomes: monophyly of extant gymnosperms and ori-gin of Gnetales from conifers. Proceedings of the NationalAcademy of Sciences USA 97:4086–4091.

Collinson, M. E. 1996. ‘‘What use are fossil ferns?’’ 20 years on:with a review of the fossil history of extant pteridophyte fam-ilies and genera. Pp. 349–394 in M. J. Camus, M. Gibby, andR. J. Johns, eds. Pteridology in perspective. Royal Botanic Gar-dens, Kew, England.

Cornet, B. 1986. The leaf venation and reproductive structuresof a Late Triassic angiosperm, Sanmiguelia lewisii. Evolution-ary Theory 7:231–309.

Crane, P. R. 1985. Phylogenetic analysis of seed plants and theorigin of the angiosperms. Annals of the Missouri BotanicalGardens 72:716–793.

———. 1996. The fossil history of Gnetales. International Journalof Plant Science 157(Suppl. to No. 6):S50–S57.

Crane, P. R., and G. R. J. Upchurch. 1987. Drewria potomacensisgen. et. sp. nov. an Early Cretaceous member of the Gnetalesfrom the Potomac Group of Virginia. American Journal of Bot-any 74:1722–1736.

Crepet, W. L. 2001. Plant-animal interactions: insect pollination.Pp. 426–429 in D. E. G. Briggs and P. R. Crowther, eds. Pa-laeobiology II. Blackwell Science, London.

Delevoryas, T., and R. C. Hope. 1971. A new Triassic cycad andits phyletic implications. Postilla 150:1–21.

Delevoryas, T., and T. N. Taylor. 1969. A probable pteridospermwith eremopterid foliage from the Allegheny Group of north-ern Pennsylvania. Postilla 133:1–14.

DiMichele, W. A., and R. B. Aronson. 1992. The Pennsylvanian-Permian vegetational transition: a terrestrial analogue to theonshore-offshore hypothesis. Evolution 46:807–824.

DiMichele, W. A., and P. J. DeMaris. 1987. Structure and dynam-ics of a Pennsylvanian-age Lepidodendron forest: colonizers ofa disturbed swamp habitat in the Herrin (No. 6) coal of Illi-nois. Palaios 2:146–157.

DiMichele, W. A., and R. W. Hook. 1992. Paleozoic terrestrialecosystems. Pp. 205–325 in A. K. Behrensmeyer, J. D. Damuth,W. A. DiMichele, R. Potts, H.-D. Sues, and S. L. Wing, eds.Terrestrial ecosystems through time. University of ChicagoPress, Chicago.

Dobruskina, I. A. 1975. The role of peltaspermacean pterido-sperms in Late Permian and Triassic floras. Paleontology Jour-nal 9:536–548.

———. 1995. Keuper (Triassic) Flora from Middle Asia (Mady-gen, Southern Fergana). New Mexico Museum of Natural His-tory and Science Bulletin 5:1–49.

Dolan, L., and R. S. Poethig. 1998. Clonal analysis of leaf de-velopment in cotton. American Journal of Botany 85:315–321.

Doyle, J. A., and L. J. Hickey. 1976. Pollen and leaves from themid-Cretaceous Potomac Group and their bearing on earlyangiosperm evolution. Pp. 139–206 in Beck 1976.

Feild, T. S., N. C. Arens, and T. E. Dawson. 2003a. The ancestralecology of angiosperms: emerging perspectives from extant

basal lineages. International Journal of Plant Science 164:S129–S142.

Feild, T. S., P. J. Franks, and T. L. Sage. 2003b. Ecophysiologicalshade adaptation in the basal angiosperm, Austrobaileya scan-dens (Austrobaileyaceae). International Journal of Plant Sci-ence 164:313–324.

Fenner, M. 1998. The phenology of growth and reproduction inplants. Perspectives in Plant Ecology, Evolution and System-atics 1:78–91.

Foote, M. 1993. Contribution of individual taxa to overall mor-phological disparity. Paleobiology 19:403–419.

———. 1995. Morphological diversification of Paleozoic cri-noids. Paleobiology 21:273–299.

Foster, A. S. 1952. Foliar venation in angiosperms from an on-togenetic standpoint. American Journal of Botany 39:752–766.

Galtier, J., and O. Bethoux. 2002. Morphology and growth habitof Dicksonites plunkenetii from the Upper Carboniferous ofGraissessac (France). Geobios 35:525–535.

Gillespie, W. M., and H. W. Pfefferkorn. 1986. Taeniopterid lam-ina on Phasmatocycas megasporophylls (Cycadales) from theLower Permian of Kansas, U.S.A. Review of Palaeobotany andPalynology 49:99–116.

Givnish, T. 1979. On the adaptive significance of leaf form. Pp.375–407 in O. T. Solbrig, S. Jain, G. B. Johnson, and P. H. Raven,eds. Topics in plant population biology. Columbia UniversityPress, New York.

Gower, J. C. 1966. Some distance properties of latent root andvector methods used in multivariate analysis. Biometrika 53:325–338.

Hagemann, W., and S. Gleissberg. 1996. Organogenetic capacityof leaves: the significance of marginal blastozones in angio-sperms. Plant Systematics and Evolution 199:121–152.

Halle, T. G. 1929. Some seed-bearing pteridosperms from thePermian of China. Kungliga Svenska VetenskapsakademiensHandlingar 6:3–24.

Harris, T. M. 1926. The Rhaetic flora of Scoresby Sound. Med-delelser om Grønland 68:46–147.

———. 1983. The stem of Pachypteris papillosa (Thomas &Bose)Harris. Botanical Journal of the Linnean Society 86:149–159.

Hasebe, M., et al. 1994. rbcL gene sequences provide evidencefor the evolutionary lineages of leptosporangiate ferns. Pro-ceedings of the National Academy of Sciences USA 91:5730–5734.

Hickey, L. J. 1974. A revised classification of the architecture ofdicotyledonous leaves. Pp. 25–39 in C. R. Metcalfe and L.Chalk, eds. Anatomy of the Dicotyledons, Vol. I, 2d ed. Clar-endon, Oxford.

Knoll, A. H. 1984. Patterns of extinction in the fossil record ofvascular plants. Pp. 21–68 in M. Nitecki, ed. Extinctions. Uni-versity of Chicago Press, Chicago.

Kramer, K. U., and P. S. Green. 1990. I. Pteridophytes and gym-nosperms. Springer, Berlin.

Leaf Architecture Working Group. 1999. Manual of leaf archi-tecture: morphological description and categorization of di-cotyledonous and net-veined monocotyledonous angio-sperms. Smithsonian Institution, Washington, D.C.

Li, H., and D. W. Taylor. 1999. Vessel-bearing stems of Vasovineatianii gen. et sp. nov. (Gigantopteridales) from the UpperPermian of Guizhou Province, China. American Journal ofBotany 86:1563–1575.

Li, X., ed. 1995. Fossil floras of China through the geologic ages.Guangdong Science and Technology Press, Guangzhou.

Li, X., and Z. Yao. 1983. Fructifications of gigantopterids fromSouth China. Palaeontographica 185B:11–26.

Lupia, R. 1999. Discordant morphological disparity and taxo-nomic diversity during the Cretaceous angiosperm radiation:North American pollen record. Paleobiology 25:1–28.

Magallon, S., and M. J. Sanderson. 2002. Relationships among


seed plants inferred from highly conserved genes: sortingconflicting phylogenetic signals among ancient lineages.American Journal of Botany 89:1991–2006.

Mamay, S. H. 1976. Paleozoic origin of cycads. United StatesGeological Survey Professional Paper 934:1–48.

Mathews, S., and M. J. Donoghue. 1999. The root of angiospermphylogeny inferred from duplicate phytochrome genes. Sci-ence 286:947–950.

McGhee, G. R., Jr. 1999. Theoretical morphology: the conceptand its applications. Columbia University Press, New York.

Melville, R. 1969. Leaf venation patterns and the origin of theangiosperms. Nature 224:121–125.

Niklas, K. J. 1994. Morphological evolution through complex do-mains of fitness. Proceedings of the National Academy of Sci-ences USA 91:6772–6779.

Poethig, R. S., and I. M. Sussex. 1985a. The developmental mor-phology and growth dynamics of the tobacco leaf. Planta 165:158–169.

———. 1985b. The cellular parameters of leaf development intobacco: a clonal analysis. Planta 165:170–184.

Pray, T. R. 1955. Foliar venation of angiosperms. II. Histogenesisof the venation of Liriodendron. American Journal of Botany 42:18–27.

———. 1960. Ontogeny of the open dichotomous venation in thepinna of the fern Nephrolepis. American Journal of Botany 47:319–328.

———. 1962. Ontogeny of the closed dichotomous venation ofRegnellidium. American Journal of Botany 49:464–472.

Pryer, K. M., A. R. Smith, and J. E. Skog. 1995. Phylogenetic re-lationships of extant ferns based on evidence from morphol-ogy and rbcl sequences. American Fern Journal 85:205–282.

Qiu, Y.-L., J. Lee, F. Bernasconi-Quadroni, D. E. Soltis, P. S. Soltis,M. Zanis, E. A. Zimmer, Z. Chen, V. Savolainen, and M. W.Chase. 1999. The earliest angiosperms: evidence from mito-chondrial, plastid, and nuclear genomes. Nature 402:404–407.

Retallack, G. J. 1977. Reconstructing Triassic vegetation of east-ern Australasia: a new approach for the biostratigraphy ofGondwanaland. Alcheringa 1:247–277.

Roth-Nebelsick, A., D. Uhl, V. Mosbrugger, and H. Kerp. 2001.Evolution and function of leaf venation architecture: a review.Annals of Botany 87:553–566.

Rydin, C., B. Mohr, and E. M. Friis. 2003. Cratonia cotyledon gen.et sp. nov.: a unique Cretaceous seedling related to Welwit-schia. Proceedings of the Royal Society of London B270(Suppl.):S29–S32.

Sachs, T. 1991. Pattern formation in plant tissues. CambridgeUniversity Press, Cambridge.

Sakai, S. 2001. Phenological diversity in tropical forests. Popu-lation Ecology 43:77–86.

Scott, A. C., and J. Galtier. 1985. Distribution and ecology of ear-ly ferns. Proceedings of the Royal Society of Edinburgh B 86:141–149.

Shangyou, N., D. B. Rowley, and A. M. Ziegler. 1990. Constraintson the locations of Asian microcontinents in Palaeo-Tethysduring the Late Paleozoic. Geological Society Memoir 12:397–409.

Skog, J. E. 2001. Biogeography of Mesozoic leptosporangiateferns related to extant ferns. Brittonia 53:236–269.

Soltis, P. S., D. E. Soltis, and M. W. Chase. 1999. Angiospermphylogeny inferred from multiple genes as a tool for compar-ative biology. Nature 402:402–404.

Surange, K. R. 1966. Pre-Cretaceous angiosperms. Proceedingsof the Autumn School in Botany, Pp. 328–331.

Surange, K. R., and S. Chandra. 1972. Fructifications of Glos-sopteridae from India. Palaeobotanist 21:1–17.

Surange, K. R., and H. K. Maheshwari. 1970. Some male and fe-male fructifications of Glossopteridales from India. Palaeon-tographica 129:178–192.

Takhtajan, A. 1976. Neoteny and the origin of flowering plants.Pp. 207–219 in Beck 1976.

Taylor, D. W., and L. J. Hickey. 1996. Evidence for and implica-tions of an herbaceous origin for angiosperms. Pp. 232–266 inD. W. Taylor and L. J. Hickey, eds. Flowering plant origin, evo-lution, and phylogeny. Chapman and Hall, New York.

Taylor, T. N. 1988. Pollen and pollen organs of fossil gymno-sperms. Pp. 177–217 in Beck 1976.

Taylor, T. N., and M. A. Millay. 1979. Pollination biology andreproduction in early seed plants. Review of Palaeobotanyand Palynology 27:329–355.

Taylor, T. N., and E. L. Taylor. 1993. The biology and evolutionof fossil plants. Prentice Hall, Englewood Cliffs, NJ.

Trivett, M. L., and K. B. Pigg. 1996. A survey of reticulate ve-nation among fossil and living plants. Pp. 8–31 in D. W. Taylorand L. J. Hickey, eds. Flowering plant origin, evolution andphylogeny. Chapman and Hall, New York.

Wagner, P. J. 2001. Constraints on the evolution of form. Pp. 147–152 in D. E. G. Briggs and P. R. Crowther, eds. PalaeobiologyII. Blackwell Science, London.

Wagner, W. H. 1979. Reticulate veins in the systematics of mod-ern ferns. Taxon 28:87–95.

Wing, S. L., and H.-D. Sues. 1992. Mesozoic and Early Cenozoicterrestrial ecosystems. Pp. 327–418 in A. K. Behrensmeyer, J.D. Damuth, W. A. DiMichele, R. Potts, H.-D. Sues, and S. L.Wing, eds. Terrestrial ecosystems through time. University ofChicago Press, Chicago.

Wing, S. L., and B. H. Tiffney. 1987. The reciprocal interactionof angiosperm evolution and tetrapod herbivory. Review ofPalaeobotany and Palynology 50:179–210.

Winter, K.-U., A. Becker, T. Muenster, J. T. Kim, H. Saedler, andG. Theissen. 1999. MADS-box genes reveal that gnetophytesare more closely related to conifers than to flowering plants.Proceedings of the National Academy of Sciences USA 96:7342–7347.

Ziegler, A. M. 1990. Phytogeographic patterns and continentalconfigurations during the Permian Period. In W. S. McKerrowand C. R. Scotese, eds. Palaeozoic Palaeogeography and Bio-geography. Geological Society of London Memoir 12:363–379.

Ziegler, A. M., G. Eshel, P. M. Rees, T. A. Rothfus, D. B. Rowley,and D. Sunderlin. 2003. Tracing the tropics across land andsea: Permian to present. Lethaia 36:227–254.

Zurakowski, K. A. and E. M. Gifford. 1988. Quantitative studiesof pinnule development in the ferns Adiantum raddianum andCheilanthes viridis. American Journal of Botany 75:1559–1570.

Zwieniecki, M. A., P. J. Melcher, C. K. Boyce, L. Sack, and N. M.Holbrook. 2002. Hydraulic architecture of leaf venation Lau-rus nobilis L. Plant, Cell and Environment 25:1445–1450.

Appendix 1

Characters used for description of leaf morphologies. Inap-plicable or missing characters were coded as ‘‘?’’.

1. Lamina lobed: 0 yes; 1 no; 2 variable.2. Veins per laminar segment: 0 one; 1 more than one; 2 vari-

able.3. Vein paths: 0 divergent or parallel; 1 convergent; 2 variable.4. Venation: 0 open; 1 always includes reticulation; 2 variable.5. Marginal vein closure: 0 no; 1 yes; 2 variable.6. Endings: 0 all equally marginal; 1 also internally directed;

2 variable.7. Marginal endings: 0 along some margins; 1 along all mar-

gins; 2 variable.8. Internal endings: 0 all marginally directed, but unevenly

truncated; 1 internally directed.9. Midvein: 0 no; 1 yes; 2 variable.

10. Vein orders (except any midvein): maximum number.

130 C. KEVIN BOYCE

11. Laminar attachment: 0 narrow (can include both sessile andpetiolate); 1 broad; 2 variable.

12. Number of entering veins: 0 one; 1 multiple; 2 variable.13. Multiple entering veins: 0 multiple equivalent; 1 multiple

not equivalent; 2 both.14. Vein paths: 0 regular; 1 irregular; 2 both.15. Vein paths: 0 not parallel; 1 extensively parallel.16. Vein paths: 0 no arching; 1 arching concave up; 2 arching

concave down; 3 both.17. Lamina shape: 0 regular; 1 irregular; 2 variable.18. Vein branching: 0 no; 1 yes.19. Vein branching: 0 evenly distributed; 1 uneven arm to arm;

2 origin to edge; 3 both.

Appendix 2

Sources used to compile leaf morphological characters as in-dicated in Appendix 3 and to determine stratigraphic rangesand systematic affinities.

1. Anderson, J. M., and H. M. Anderson. 1985. Palaeoflora ofsouthern Africa. Balkema, Rotterdam.

2. Andrews, H. N., C. A. Arnold, E. Boureau, J. Doubinger, andS. Leclercq. 1970. Traite de paleobotanique. Tome IV, fasc.1. Filicophyta. Masson, Paris.

3. Asama, K. 1959. Systematic study of the so-called Gigantop-teris. Science Reports of the Research Institute of TohokuUniversity, series 2, 31:1–72.

4. ———. 1962. Evolution of the Shansi flora and origin of thesimple leaf. Science Reports of the Research Institute of To-hoku University, series 2, special volume 5:247–273.

5. Ash, S. R. 1970. Dinophyton, a problematical new plant genusfrom the Upper Triassic of the south-western United States.Palaeontology 13:646–663.

6. ———. 1972. Marcouia, gen. nov., a problematical plant fromthe Late Triassic of the southwestern U.S.A. Palaeontology15:424–429.

7. ———. 1980. Upper Triassic floral zones of North America.Pp. 153–170 in D. L. Dilcher and T. N. Taylor, eds. Biostra-tigraphy of fossil plants. Dowden, Hutchinson and Ross,Stroudsburg, Pa.

8. ———. 1987. Growth habit and systematics of the UpperTriassic plant Pelourdea poleoensis, southwestern U.S.A. Re-view of Palaeobotany and Palynology 51:37–49.

9. Bose, M. N., and J. Banerji. 1984. The fossil floras ofKachchh. 1. Mesozoic megafossils. Palaeobotanist 33:1–189.

10. Bose, M. N., P. K. Pal, and T. M. Harris. 1985. The Pentoxylonplant. Philosophical Transactions of the Royal Society ofLondon B 310:77–108.

11. Boureau, E., and J. Doubinger. 1975. Traite de paleobota-nique. Tome IV, Fasc. 2. Pteridophylla (Premiere Partie).Masson, Paris.

12. Boyce, C. K., and A. H. Knoll. 2002. Evolution of develop-mental potential and the multiple independent origins ofleaves in Paleozoic vascular plants. Paleobiology 28:70–100.

13. Cornet, B. 1986. The leaf venation and reproductive struc-tures of a Late Triassic angiosperm, Sanmiguelia lewisii. Evo-lutionary Theory 7:231–309.

14. Crane, P. R., and G. R. J. Upchurch. 1987. Drewria potomacen-sis gen. et. sp. nov., an Early Cretaceous member of the Gne-tales from the Potomac Group of Virginia. American Journalof Botany 74:1722–1736.

15. Dobruskina, I. A. 1975. The role of peltaspermacean pteri-

dosperms in Late Permian and Triassic floras. PaleontologyJournal 9:536–548.

16. ———. 1995. Keuper (Triassic) Flora from Middle Asia(Madygen, Southern Fergana). New Mexico Museum of Nat-ural History and Science Bulletin 5:1–49.

17. Harris, T. M. 1926. The Rhaetic flora of Scoresby Sound.Meddelelser om Grønland 68:46–147.

18. ———. 1961. The Yorkshire Jurassic flora, Vol. 1 Thalloph-yta–Pteridophyta. British Museum of Natural History, Lon-don.

19. ———. 1964. The Yorkshire Jurassic flora Vol. 2. Caytoni-ales, Cycadales, and Pteridosperms. British Museum of Nat-ural History, London.

20. Jones, D. L. 1993. Cycads of the world. Smithsonian Insti-tution Press, Washington, D.C.

21. Kramer, K. U., and P. S. Green. 1990. The Families and Gen-era of Vascular Plants. I. Pteridophytes and gymnosperms.Springer, Berlin.

22. Li, X. 1995. Fossil floras of China through the geologic ages.Guangdong Science and Technology Press, Guangzhou.

23. Pal, P. K. 1984. Triassic plant megafossils from the Tiki For-mation, South Rewa Gondwana Basin, India. Palaeobotanist32:253–309.

24. Pant, D. D. 1977. The plant of Glossopteris. Journal of the In-dian Botanical Society 56:1–23.

25. Pant, D. D., and B. K. Verma. 1963. On the structure of leavesof Rhabdotaenia Pant from the Raniganj Coalfield, India. Pa-laeontology 6:301–314.

26. Pant, D. D., and B. K. Verma. 1964. On the structure of Pa-laeovittaria raniganjensis n. sp. from the Raniganj Coalfield,India. Palaeontographica, Abteilung B 115:45–50.

27. Person, C. P., and T. Delevoryas. 1982. The Middle Jurassicflora of Oaxaca, Mexico. Palaeontographica 180B:82–119.

28. Retallack, G. J. 1977. Reconstructing Triassic vegetation ofeastern Australasia: a new approach for the biostratigraphyof Gondwanaland. Alcheringa 1:247–277.

29. ———. 1980. Late Carboniferous to Middle Triassic mega-fossil floras from the Sydney Basin. Geological Survey ofNew South Wales Bulletin 26:384–430.

30. Schopf, J. M., and R. A. Askin. 1980. Permian and Triassicfloral biostratigraphic zones of southern land masses. Pp.119–152 in D. L. Dilcher and T. N. Taylor, eds. Biostratigra-phy of fossil plants. Dowden, Hutchinson and Ross,Stroudsburg, Pa.

31. Surange, K. R., and S. Chandra. 1972. Fructifications ofGlossopteridae from India. Palaeobotanist 21:1–17.

32. Surange, K. R., and H. K. Maheshwari. 1970. Some male andfemale fructifications of Glossopteridales from India. Pa-laeontographica 129:178–192.

33. Taylor, T. N., and E. L. Taylor. 1993. The biology and evo-lution of fossil plants. Prentice Hall, Englewood Cliffs, NJ.

34. Tidwell, W. D. 1998. Common fossil plants of western NorthAmerica. Smithsonian Institution Press, Washington, D.C.

35. Tryon, R. M., and A. F. Tryon. 1982. Ferns and allied plantswith special reference to tropical America. Springer, NewYork.

36. Vakhrameev, V. A. 1991. Jurassic and Cretaceous floras andclimates of the earth. Cambridge University Press, Cam-bridge.

37. van Konijnenburg-van Cittert, J. H. A., and H. S. Morgans.1999. The Jurassic flora of Yorkshire. Palaeontological As-sociation, London.


Appendix 3Phylogenetic affinity, literature sources, and stratigraphic range for all taxa included in the analysis. Affinity of fossil ferns,

seed plants, progymnosperms, and sphenopsids identified to the ordinal level; living ferns assigned to families. An affinity of‘‘Other’’ indicates unknown affinity, known affinity to a group not discussed here, or a form-genus of multiple affinities.Abbreviations: C, Carboniferous; D, Devonian; E, Early; J, Jurassic; K, Cretaceous; L, Late; M, Middle; P, Permian; Tr,Triassic.

132 C. KEVIN BOYCE

Appendix 3. Continued.

134 C. KEVIN BOYCE


136 C. KEVIN BOYCE


138 C. KEVIN BOYCE


140 C. KEVIN BOYCE


patterns of segregation and convergence in the evolution of fern and

Documents