Download as pdf or txt
Download as pdf or txt
You are on page 1of 15

American Journal of Botany 96(1): 237–251. 2009.

SEED FERNS FROM THE LATE PALEOZOIC AND MESOZOIC:


ANY ANGIOSPERM ANCESTORS LURKING THERE?1
Edith L. Taylor2 and Thomas N. Taylor
Department of Ecology and Evolutionary Biology, and Natural History Museum and Biodiversity Research Center, University of
Kansas, Lawrence, Kansas 66045-7534 USA

Five orders of late Paleozoic–Mesozoic seed ferns have, at one time or another, figured in discussions on the origin of angio-
sperms, even before the application of phylogenetic systematics. These are the Glossopteridales, Peltaspermales, Corystosper-
males, Caytoniales, and Petriellales. Although vegetative features have been used to suggest homologies, most discussion has
focused on ovulate structures, which are generally interpreted as megasporophylls bearing seeds, with the seeds partially to almost
completely enclosed by the megasporophyll (or cupule). Here we discuss current information about the reproductive parts of these
plants. Since most specimens are impression-compression remains, homologizing the ovulate organs, deriving angiospermous
homologues, and defining synapomorphies remain somewhat speculative. Although new specimens have increased the known
diversity in these groups, a reconstruction of an entire plant is available only for the corystosperms, and thus hypotheses about
phylogenetic position are of limited value. We conclude that, in the case of these seed plants, phylogenetic analysis techniques
have surpassed the hard data needed to formulate meaningful phylogenetic hypotheses. Speculation on angiosperm origins and
transitional stages in these fossils provides for interesting discussion, but currently it is still speculation, as the role of these groups
in the origin of angiospermy continues to be cloaked in Darwin’s mystery.

Key words: angiosperm ancestors; caytonialeans; corystosperms; fossil plants; glossopterids; Mesozoic; peltasperms; Perm-
ian; seed plant phylogeny.

The Permian glossopterids, along with the three major groups resolution as the late Paleozoic pteridosperms, many of which
of late Paleozoic–Mesozoic seed ferns (peltasperms, corys- are understood, at least in part, in anatomical detail. Neverthe-
tosperms, caytonialeans), all represent gymnosperms that have less, some of these “younger” seed ferns possess morphological
historically figured prominently in discussions on the origin of features that have been used to suggest affinities with some of
angiosperms (e.g., Krassilov, 1977b). The Glossopteridales are the Paleozoic seed ferns (Petriella, 1981). As research has con-
considered to represent a monophyletic group, and all members tinued, our knowledge of the diversity within some of these
have similar foliage, i.e., Glossopteris Brongniart and Ganga- groups has increased, which often makes it difficult to define
mopteris McCoy, although more recent work suggests some the group according to its original concept. Nevertheless, this
diversity in reproductive organs (e.g., Taylor et al., 2007; additional diversity needs to be accommodated in seed plant
Prevec et al., 2008). They represent the dominant elements in phylogenies.
Permian floras from the supercontinent of Gondwana (includ- Historically, and for obvious reasons, many of these gymno-
ing Antarctica, Australia-New Zealand, parts of Africa and sperms have been the subject of considerable interest as poten-
South America, and peninsular India) and appear to have gone tial angiosperm progenitors. These assumptions have been
extinct around the Permo–Triassic boundary (McManus et al., based on a number of features, most notably the enclosure of
2002). The other, post-Carboniferous seed ferns, however, were the ovules in some type of leaflike structure that has generally
widespread from the Permian to the Cretaceous, and unlike the been termed a cupule. In the sections that follow, we provide a
Carboniferous forms, appear to have few morphological char- discussion of several seed fern groups and offer our perspective
acters in common. Rather, they are included as a collection of on the relationships of these fossil groups and possible homolo-
either orders or families in which the relationships among the gies with the angiosperms.
various taxa are uncertain (Taylor et al., 2006). In most seed The general concept of the seed ferns was proposed by Oli-
plant phylogenies, each of these groups is included as a single ver and Scott (1904) for Paleozoic (Carboniferous) plants that
terminal branch, often a composite terminal. Although a few of possessed fernlike foliage, but with seeds and pollen sacs, rather
these seed plants are known from anatomically preserved speci- than sporangia, attached to the leaves. The initial concept of the
mens, most are represented by impression and compression fos- Pteridospermae was based on structurally preserved organs of
sils and therefore are not understood with the same degree of the Carboniferous lyginopterids, but was subsequently ex-
panded to other Carboniferous–Permian orders, and today also
includes the late Paleozoic Glossopteridales. In Mesozoic rocks,
1 Manuscript received 19 June 2008; revision accepted 27 October 2008. four groups have been delimited. One of these, the Corystosper-
The authors thank the following colleagues for kindly supplying figures: males (Upper Permian–Cretaceous), is known in some detail,
J. M. Anderson, C. P. Daghlian, H. Kerp, S. McLoughlin, H. Nishida, P. based on permineralized and compression fossils from several
Ryberg, and M. Zavada, as well as students and postdoctoral scholars in continents. The Peltaspermales (Pennsylvanian–Triassic) and
our laboratory who contributed to this work. This research was partially
supported by the National Science Foundation (OPP-0229877, ANT-
the Caytoniales (Triassic–Cretaceous) are known only from im-
0635477). pression and compression specimens; both structurally pre-
2 E-mail: etaylor@ku.edu; tntaylor@ku.edu served and compressed fossils have been assigned to the
Petriellales (Triassic). Although foliage and other vegetative parts
doi:10.3732/ajb.0800202 of these seed plants are known, we have limited the discussion
237
15372197, 2009, 1, Downloaded from https://bsapubs.onlinelibrary.wiley.com/doi/10.3732/ajb.0800202 by CAPES, Wiley Online Library on [11/03/2024]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
238 American Journal of Botany [Vol. 96

that follows to the reproductive parts of these plants because found organically attached to axes or leaves, anatomical simi-
they have served as the focal point in discussions linking them larities between of the megasporophylls and vegetative leaves
to the flowering plants. of Glossopteris suggest that they were borne on the same plant
(see Nishida et al., 2007). Attached to one surface of the fleshy
megasporophyll are numerous ovoid seeds with the edges of the
GLOSSOPTERIDALES megasporophyll slightly enrolled and partially covering the
seeds. Now named Homevaleia gouldii Nishida et al. (2007),
There is a considerable body of literature on the Glossop- the megasporophyll lacks the sclerenchymatous hypodermal fi-
teridales, beginning with the description of the vegetative leaf bers found in the vegetative leaf of G. homevalensis. Each or-
Glossopteris by Adolphe Brongniart in 1828. Today, the group thotropous ovule is 1.2–1.3 mm long by 0.8–0.9 mm wide and
is known from stem, root, and leaf genera (Fig. 1), that are pre- attached by a small stalk to the adaxial surface of the megasporo-
served in a variety of modes. The reproductive organs of the phyll, based on the anatomy. The integument is thickened in the
glossopterids, however, have remained a continuing source of micropylar region, and the outer layer (sarcotesta) extends out
controversy since the first report of “linear sori” on some from the ovules to form a meshwork of filaments between the
Glossopteris leaves (Feistmantel, 1886). seeds. Gould and Delevoryas (1977) suggested that this mesh-
work may have been involved in some way in directing pollen
Ovulate organs— Today there are more than 50 different flow, possibly aided by a pollination droplet. Another sugges-
types of ovulate reproductive structures that have been de- tion is that perhaps the spongy meshwork served to insulate the
scribed (McLoughlin, 1990), with additional ones continually ovules in the cool, temperate environment in which this plant
being reported or reexamined (e.g., Adendorff, 2005; Prevec et lived (Nishida et al., 2007). Many of the ovules contain well-
al., 2008). A few reproductive organs have been found attached preserved megagametophytes, each with a single archegonium
to the parent plant in association with vegetative leaves (Fig. 2) (Fig. 6). Bisaccate pollen of the Protohaploxypinus Samoilo-
(e.g., Pant and Singh, 1974). Other reproductive organs are at- vitch type (Fig. 7) has been reported in the pollen chambers of
tributed to the group based on venation patterns of the me- these ovules (Gould and Delevoryas, 1977; Nishida et al., 2004).
gasporophyll that are similar to those of the vegetative leaves Especially significant is the report of pollen tubes in various
and on co-occurrence at the same locality. One of the major stages of releasing flagellated sperm in the region of the arche-
obstacles in the interpretation of glossopterid reproductive or- gonium (Nishida et al., 2003). Sperm are top shaped (Fig. 8),
gans has been the fact that most are preserved as impression- small (~12 µm in diameter), and characterized by spiral bands
compression (e.g., Fig. 3) or mold-cast specimens and thus, of dark dots at one end that are interpreted as corresponding to
interpreting the homology of these organs has been controver- the positions of the basal bodies of numerous flagella aligned
sial. In attempting to document diversity, some authors have along a multilayered structure (MLS). Interestingly, the glossop-
used an artificial system of morphofamilies based on macro- terid sperm are morphologically like those of cycads and
morphological features (e.g., Arberiaceae, Dictyopteridiaceae, Ginkgo, but much smaller (Nishida et al., 2007). In general or-
Lidgettoniaceae, Rigbyaceae) to classify and categorize these ganization, the Homevaleia megasporophyll shows some simi-
structures (see Adendorff, 2005, for a history of glossopterid larities to the impression-compression ovulate structures
classification). Others have used different names for higher- Dictyopteridium Feistmantel ex Zeiller or Plumsteadia Rigby
level taxonomy in the mistaken belief that families of fossil (Fig. 9), in part based on the arrangement of ovules in the per-
plants should be named after an ovulate organ rather than the mineralized and compressed forms (Nishida et al., 2007).
oldest validly published genus in the group. Anatomically preserved glossopterid megasporophylls with
Plumstead (1956) initially described the ovulate fructifica- attached ovules are also known from permineralized peat in the
tions of the glossopterids and thus demonstrated that they were Late Permian Buckley Formation of Antarctica (Taylor and
gymnosperms and not ferns. She erected the genus Hirsutum Taylor, 1992; Taylor et al., 2007). One of these is ~6.0 mm
Plumstead for shield-shaped, “bivalvate” structures that she in- wide and 1.0 mm thick and is thought to have been produced by
terpreted as bisexual (Plumstead, 1956, 1958). Plumstead sug- the same plant that bore Glossopteris schopfii Pigg (1990a)
gested that the fructifications produced hairlike, pollen-producing leaves, based on similar anatomy (Taylor and Taylor, 1992).
organs on one of the two valves. She also interpreted flat, bract- The seeds are attached to the adaxial surface of the megasporo-
like structures in Scutum Plumstead (Fig. 4) as pollen bearing phyll (Fig. 10) (Taylor and Taylor, 1992). This character is
(for details, see Prevec et al., 2008). Thus far, no glossopterid based on the position of the tissues that make up the vascular
reproductive organs have been conclusively demonstrated as bundles and contrasts with the suggested abaxial attachment of
containing both pollen and seeds (Pant, 1987). Despite the nu- ovules described from impression fossils (Adendorff, 2005).
merous ovulate structures attributed to the glossopterids, many Another permineralized specimen from the Late Permian of
represent variations on a basic theme. These consist of a me- Antarctica shows some similarity to several of the so-called cu-
gasporophyll bearing seeds; the megasporophyll has been vari- pulate glossopterid reproductive structures (Surange and Chan-
ously termed a capitulum, cupule, fertiliger, or cladode. In dra, 1975), such as Arberia White or Rigbya Lacey et al. (Fig.
instances where the reproductive structure is attached to the 4). A preliminary analysis suggests that it is a branching struc-
parent plant, it appears in an axillary position (e.g., Holmes, ture bearing at least four uniovulate cupules. Each cupule is
1987) or fused or adpressed to the petiole or lamina of a ~3.0 mm long and contains a sessile ovule with two flattened
Glossopteris leaf. wings extending from the integument. The stalk at the base of
Perhaps the most important discovery contributing to under- the organ contains a C-shaped vascular strand like that pro-
standing the homology of the glossopterid ovulate reproductive duced in some leaves (Taylor et al., 2007).
organs is the report of Gould and Delevoryas (1977) of ana- Dispersed ovules that show details of embryo development
tomically preserved Glossopteris megasporophylls with at- have been described from the Antarctic permineralized peat.
tached ovules (Fig. 5). Although the megasporophylls were not Although they occur in close association with Glossopteris
15372197, 2009, 1, Downloaded from https://bsapubs.onlinelibrary.wiley.com/doi/10.3732/ajb.0800202 by CAPES, Wiley Online Library on [11/03/2024]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
January 2009] Taylor and Taylor—Seed ferns from the late Paleozoic–Mesozoic 239

leaves, none have been found attached. They are considered to suggests that there is some function for the wing. Protection of
represent glossopterid seeds based on the overwhelming domi- the developing ovules may be one obvious adaptation. Another
nance of this group in the peat and in nearby floras of compa- might involve dispersal, with the entire megasporophyll being
rable age (e.g., Cúneo et al., 1993). Seeds of Plectilospermum shed as a unit (Adendorff et al., 2002), although this is unlikely
Taylor and Taylor are ~5.0 mm long and platyspermic with the because most impression-compression specimens appear to
nucellus and integument attached only at the base (Taylor and have already shed their seeds. The presence of some dispersed
Taylor, 1987). Many seeds have a cellular megagametophyte seeds with integumentary wings is also not consistent with the
with two archegonia, each showing evidence of a multicellular idea of the entire unit being shed.
embryo, confirming the existence of polyembryony (Fig. 11).
Development had progressed in some embryos so that the pres- Pollen organs— Unlike the ovulate organs, there is little
ence of a suspensor could be distinguished (Smoot and Taylor, variability in the pollen organs of the Glossopteridales. The two
1986). most common genera are Glossotheca Surange and Mahesh-
As noted, one of the problems in discussing the ovulate struc- wari (1970) and Eretmonia Du Toit (1932), although it has been
tures in the glossopterids is the use of nonstandard terms to de- suggested that they actually represent a single taxon (Pant,
scribe their morphology. Schopf (1976) suggested the term 1987). In both, clusters of pollen sacs are borne on two stalks,
fertiliger for the megasporophyll, a designation that Pant (1987) which are attached to a subtending leaf. In Eretmonia, which is
also used. Schopf (1976) proposed that the ovulate complex in more completely known, the pollen sacs are elongate (~1.0 mm
the glossopterids represented an axillary shoot and suggested a long), and the subtending leaf is small and rhombohedral. Arbe-
distant relationship with the cordaites, an idea that was termed riella Pant et Nautiyal is used for isolated pollen sacs similar to
the cladode concept by Retallack and Dilcher (1981). Perminer- those of Glossotheca Surange and Maheshwari. Pollen of the
alized reproductive structures that have been described to date, glossopterids is bisaccate and striate (taeniate). It has generally
however, clearly indicate that the ovules are borne on the adax- been compared to the sporae dispersae genus Protohaploxypi-
ial surface of a megasporophyll, although reconstructions based nus, although Lindström et al. (1997) found four different mor-
on impression fossils suggest attachment was abaxial. What re- phogenera within a single pollen sac of Arberiella.
mains to be determined is whether the megasporophyll is at-
tached to a reduced and modified axillary branch system or to
the base of the subtending vegetative or scale leaf (Pigg and CAYTONIALES
Trivett, 1994). It is also possible that the structures that make up
the glossopterid ovulate complex do not involve the same ho- The Caytoniales are a small group of Triassic–Cretaceous
mologies in every case. This hypothesis is further strengthened seed ferns (Krassilov, 1977a) that was erected in 1925 by the
by the variety of morphologic forms, including those that have British paleobotanist H. H. Thomas, based on compression
multiple ovules (Fig. 12) and those with apparently solitary specimens from the Middle Jurassic plant-bearing beds along
ovules. Seed morphology is also variable, with both radially the coast of Cayton Bay in Yorkshire, Great Britain. These
and bilaterally symmetrical forms known. The presence or ab- plants possessed such interesting features that Thomas regarded
sence of wings on some ovules may be related to dispersal syn- them as a new group of angiosperms in his initial description
dromes. Moreover, some glossopterid ovules have only a single (Thomas, 1925). Since that time, the seed-bearing structure
archegonium, while in others, two are present. The situation Caytonia Thomas has captured the imagination of paleobota-
may in part parallel the extraordinary variation demonstrated nists and has been variously used as a link to the angiosperm
among the seeds of Paleozoic seed ferns, which were produced carpel (e.g., Gaussen, 1946; Doyle, 1978, 1996, 2006). As a
on leaves, in cupules, and on naked branching systems. Thus, result, the Caytoniales have figured prominently in a number of
the glossopterids, despite the evident uniformity of their vege- phylogenetic analyses, despite the fact that they continue to be
tative organs, apparently represent a rather diverse group of late one of the most poorly known groups of Mesozoic seed ferns.
Paleozoic seed ferns that may ultimately rival the Carbonifer-
ous taxa in the variability of their ovulate reproductive struc- Ovulate organs— Caytonia consists of an axis ~5.0 cm long
tures. Because the group dominated Gondwana for more than bearing stalked, multiovulate cupules in subopposite pairs (Fig.
30 Myr, perhaps it is not surprising that the glossopterids are 13) (Thomas, 1925; Harris, 1964). This structure, which has
more diverse than their vegetative morphology would suggest. been interpreted as a megasporophyll, bears scars along the sur-
While at least some of the glossopterid ovulate structures are face, suggesting that the cupules were shed at maturity. Each
homologous with a megasporophyll, with ovules attached to the cupule is compressed to be nearly circular in outline (Fig. 14)
adaxial surface, we echo the comments of others who have ex- and recurved with a liplike projection directed toward the point
amined and discussed the glossopterid reproductive structures of attachment (Harris, 1940). A single cupule contains from 8–
in noting that additional permineralized specimens will be criti- 30 seeds, depending on the species, and each seed is attached by
cal in understanding the homologies of these unusual reproduc- a delicate stalk in an orthotropous position (Fig. 15). Seeds are
tive organs. ~2.0 mm long and radially symmetrical with an integument
One interesting feature is the apparent absence of any wing- composed of an outer, uniseriate epidermis that covers a row of
like structure on the megasporophyll in the permineralized radially aligned, thick-walled cells (e.g., Harris, 1958). Harris
specimens, a structure that is commonly found in impression- (1951) suggested that perhaps the outer portion of the integu-
compression forms (Fig. 4). It is doubtful that the enrolled ment was fleshy.
edges of the megasporophyll of Homevaleia gouldii could cre- In the initial description of Caytonia, pollen grains were
ate such a structure if compressed. The absence of the wing found on the liplike portion of the cupule (Fig. 15), a structure
may reflect a developmental stage of the megasporophyll or a that Thomas (1925) termed the stigmatic flap or region. This was
difference in taxa between South Africa and Australia. The no doubt the principal reason that the cupule was described as
presence of this structure on many of the impression specimens an angiospermous fruit containing numerous seeds. According
15372197, 2009, 1, Downloaded from https://bsapubs.onlinelibrary.wiley.com/doi/10.3732/ajb.0800202 by CAPES, Wiley Online Library on [11/03/2024]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
240 American Journal of Botany [Vol. 96

Figs. 1–12. Glossopterid vegetative and reproductive organs. 1. Glossopteris leaf, Permian of Antarctica, showing anastomosing venation (courtesy
H. Kerp). Bar = 2.5 cm. 2. Suggested reconstruction of Ottokaria zeilleri Pant and Nautiyal showing relationship between vegetative leaf and megasporo-
phyll (redrawn from Pant and Nautiyal, 1984). 3. Rigbya arberioides Lacey et al. showing lobes and cupules (courtesy P. Ryberg). Bar = 5.0 mm. 4. Scutum
rubidgeum Plumstead (courtesy S. McLoughlin). Bar = 5.0 mm. 5. Megasporophyll of H. gouldii Nishida et al. showing several ovules (courtesy P. Ry-
berg). Bar = 1.0 mm. 6. Seed of Homevaleia gouldii with archegonium (arrow) in megagametophyte tissue (courtesy P. Ryberg). Bar = 0.25 mm. 7. Bisac-
cate pollen grain. Bar = 25 µm. 8. Sperm from H. gouldii ovule (courtesy H. Nishida). Bar = 100 µm. 9. Plumsteadia semnes Rigby showing numerous
15372197, 2009, 1, Downloaded from https://bsapubs.onlinelibrary.wiley.com/doi/10.3732/ajb.0800202 by CAPES, Wiley Online Library on [11/03/2024]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
January 2009] Taylor and Taylor—Seed ferns from the late Paleozoic–Mesozoic 241

to this interpretation, fertilization would have taken place via a cate (Zavada and Crepet, 1986; Osborn, 1994). Another inter-
pollen tube that grew from the pollen grain on the stigmatic pretation is that the grains are protosaccate and the
surface to the pollen chamber of an ovule. Fine strands of cuti- endoreticulations are continuous between the saccus wall and
cle that extended from the cupule to the seeds were interpreted surface of the corpus (Pedersen and Friis, 1986). The ultrastruc-
as the remnants of pollen tubes or as extensions of the seed ture of the exine suggests that the sexine is alveolate, and there
micropyles. Later research, however, revealed that pollen grains is a conspicuous sulcus on the distal surface (Zavada and Crepet,
were actually present inside the cupule, perhaps drawn in by 1986).
pollination droplets that originated at the micropylar end of
each seed. The next step in elucidating the reproductive biology
of Caytonia was provided by Maria Reymanówna (1973). Her CORYSTOSPERMALES
detailed work indicated that each seed was associated with an
elongate canal extending from the seed micropyle to the outer The Corystospermales have a worldwide distribution based
lip of the cupule and that the number of these canals was equal on foliage and, in many cases, reproductive structures. In Gond-
to the number of seeds produced in each cupule. Today, Cayto- wana, foliage of Dicroidium Gothan is often the dominant ele-
nia is interpreted as a multiovulate cupule and not an angiosper- ment in Middle–Late Triassic floras, and the genus is known
mous fruit as originally thought. from localities in Antarctica, South Africa, Australia, Argen-
Caytonia demonstrates at least one method in which seeds tina, Tasmania, and India. In Laurasia, foliage of Pachypteris
became enclosed and in this feature parallels the morphology, Brongniart (Harris, 1964) and Thinnfeldia von Ettingshausen is
although not the function, of an angiosperm carpel (Krassilov, assigned to the corystosperms. Umkomasia Thomas and Pteru-
1977a). Doyle (1978, 2006) has supported the idea, proposed chus Thomas are known from the Jurassic of Europe associated
earlier by Gaussen (1946), that the Caytoniales represent some with Thinnfeldia foliage (Kirchner and Müller, 1992). Recently,
grade of angiosperm ancestors. According to this theory, there three species of Dicroidium leaves have been described from
is a reduction in the number of ovules in Caytonia to one per the Upper Permian of Jordan (Abu Hamad et al., 2008), sug-
cupule. The cupule wall would be homologous with the second gesting that the group may have originated in the late Paleozoic
integument in an angiosperm ovule and the cupule-bearing axis and spread into Gondwana through Africa. The family was es-
would develop to grow over and enclose the cupule and is thus tablished by Thomas in 1933 to include Dicroidium foliage,
homologous with the carpel. This idea requires that the main Umkomasia (and other) ovulate organs, and Pteruchus pollen
Caytonia axis develop into a flattened, dorsiventral structure, organs, with the familial name based on the characteristic hel-
and cuticular differences on the upper and lower sides of the met shape of the uniovulate cupules. Based on the morphologi-
axis have been used to support this scenario. Because all cayto- cal diversity within the corystosperms, Archangelsky (1996)
nialean fossils represent flattened compressions and no internal suggests that they were a diverse and rapidly evolving group
anatomy is known, however, it is difficult to extrapolate the during the Triassic. The plants were probably small to large
three-dimensional morphology of the reproductive structures. woody shrubs and trees and bore pinnate leaves with open di-
In addition, in some specimens of Caytonia, the ovules are chotomous venation. Although the various plant parts have not
borne on long stalks (e.g., Fig. 13), a configuration which would been found attached, their consistent occurrence in the same
presumably make it difficult for overgrowth of the stalk. beds, similarity in epidermal anatomy, and identical pollen
found in both pollen organs and seeds have been used to indi-
Pollen organs— Caytonanthus Harris (originally Antholi- rectly associate the various organs (Petriella, 1983). When per-
thus) is a pollen-bearing structure that was also found in the mineralized, the plants have been reconstructed based on the
Middle Jurassic Yorkshire flora in association with Caytonia occurrence of unique secretory cells in the various organs (Tay-
cupules and foliage of Sagenopteris Presl in Sternberg. Thomas lor, 1996).
included it in the Caytoniales on the basis of the pollen, which
was identical to that found in Caytonia. Caytonanthus consists Ovulate organs— One of the interesting aspects of the corys-
of a slender axis bearing flattened, pinnate lateral branches (Fig. tosperms is that the reproductive organs are morphologically
16); each branch bears from 1–3 elongate synangia (Osborn, consistent within the group. The most common seed-containing
1994). Synangia are about 1.0 cm long, pointed at the distal end structure is Umkomasia (Fig. 18) (Thomas 1933), which is
(Fig. 17), and contain 1–4 pollen sacs (originally termed loc- known widely from Gondwanan deposits (Taylor, 1996), but is
ules) that are arranged around a central zone of tissue (Harris, also known from Europe (Kirchner and Müller, 1992) and
1941). These structures (which have been termed anthers in China (Figs. 19, 20) (Zan et al., 2008). It consists of a fertile
some treatments) are circular in cross section, with dehiscence branch, more than 15 cm long in some taxa, e.g., U. quadripar-
taking place toward the center of each synangium. The epider- tita Anderson and Anderson (2003), which produces laterals.
mis is composed of delicate fusiform cells, perhaps with thick- Each lateral bears one to several pairs of recurved, helmet-like,
er-walled fibrous cells beneath. generally uniovulate cupules. Some specimens have a pair of
Pollen grains of Caytonanthus are small and bisaccate, and if subtending bracts at the base of the branch (Fig. 21); in others,
found dispersed would be included in Vitreisporites Leschik. they are not present and may have been shed (Holmes, 1987).
Grains are 25–40 µm in diameter and contain endoreticulations The branching in Umkomasia was originally interpreted to be
lining the interior of the sacci, thus making these grains eusac- flattened in a single plane (Thomas, 1933), but permineralized


seeds (courtesy P. Ryberg). Bar = 1.7 cm. 10. Suggested reconstruction of a Glossopteris megasporophyll with seeds attached to adaxial surface (from
Taylor and Taylor, 1992). 11. Cross section of Plectilospermum seed showing two archegonial chambers (arrows) containing embryos. Bar = 850 µm. 12.
Diagrammatic reconstruction of Denkania indica (redrawn from Surange and Chandra, 1975).
15372197, 2009, 1, Downloaded from https://bsapubs.onlinelibrary.wiley.com/doi/10.3732/ajb.0800202 by CAPES, Wiley Online Library on [11/03/2024]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
[Vol. 96
American Journal of Botany
242
15372197, 2009, 1, Downloaded from https://bsapubs.onlinelibrary.wiley.com/doi/10.3732/ajb.0800202 by CAPES, Wiley Online Library on [11/03/2024]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
January 2009] Taylor and Taylor—Seed ferns from the late Paleozoic–Mesozoic 243

specimens from the Middle Triassic of Antarctica indicate Some seeds of the corystosperms possessed a slightly curved,
that the cupulate branch has stem-like anatomy and produces bifid micropylar canal, which often extended beyond the cupule
paired traces to the laterals. The cupule-bearing branches of U. margin (Fig. 25) (Thomas, 1933). Macerations suggest that the
resinosa Klavins et al. (2002) are helically arranged and cu- integument contained no fibrous cells and that the seed epider-
pules contain either one or two ovules attached to the abaxial mis was covered with a thick cuticle. Embedded in the pollen
surface (Fig. 22) of the cupule wall. Ovules of this species are chambers are numerous saccate pollen grains similar to those
small, possess a single integument, and are characterized by a found in the pollen-producing organ Pteruchus (Townrow,
bifid micropylar extension at the distal end. 1962).
Cupules preserved as compressions and impressions often
appear urn-like in morphology with two prominent lobes (Fig. Pollen organs— The most common pollen organ attributed
19). The surface of the cupule is wrinkled, suggesting that it to the Corystospermales is Pteruchus (Thomas, 1933) (Fig. 26).
may have been fleshy (Thomas, 1933). On well-preserved spec- The genus is known throughout Gondwana (Townrow, 1962;
imens, stomata are present on both surfaces of the cupule. The Taylor, 1996), but there are also scattered reports from Laur-
seeds of U. macleanii are small (5.0 mm long) and borne either asia, e.g., P. septentrionalis from the Rhaeto–Liassic (Late
singly or in pairs in each cupule (Thomas, 1933). Compressed Triassic–Early Jurassic) of Germany (Kirchner and Müller,
cupules have also been found attached to short shoots, which in 1992). Most specimens have alternately arranged microsporo-
turn arise from longer branches bearing D. odontopteroides phylls attached to an axis about 4.0 cm long (Thomas, 1933;
leaves (Fig. 23) (Axsmith et al., 2000). Cupules of U. uniramia Townrow, 1962), although in P. indicus the microsporophylls
Axsmith et al. (2000) are borne in whorls of five to eight. The appear to be helically arranged (Pant and Basu, 1973). Each
cupules attached to short shoots (Fig. 24) and leaves attached to microsporophyll terminates in a flattened head (Fig. 27) that
long shoots from the Upper Triassic of Antarctica has perplexed bears numerous, elongate pollen sacs on the abaxial surface;
some who challenge whether long shoots would retain leaves these are partially protected by the tissue of the head. The num-
(e.g., Artabe and Brea, 2003; Holmes and Anderson, 2005). In ber of sacs per head is variable (20–200), and dehiscence is
modern Ginkgo L., however, leaves are borne on relatively longitudinal. The cuticle on all parts shows epidermal cells with
large stems as well as on short shoots (Axsmith et al., 2007). It slightly undulate margins and few stomata (Townrow, 1962).
is also important to note that this Dicroidium plant grew at high Permineralized specimens of Pteruchus are known from peat
polar latitudes, and thus the parameters affecting growth were deposits from the lower Middle Triassic of Antarctica (Yao et
no doubt much different than they are in temperate regions to- al., 1995). Pteruchus fremouwensis Yao et al. (1995) consists of
day (Axsmith et al., 2007). The report of Umkomasia cupules a branch bearing helically arranged microsporophylls (Fig. 28).
(Figs. 19, 20) in association with Thinnfeldia-type foliage from Each microsporophyll has a flattened head bearing ~20 elon-
the Upper Triassic of China (Zan et al., 2008) further expands gate, sessile, unilocular pollen sacs. Pollen sacs are ≤2.0 mm
the diversity within the corystosperms. Umkomasia has also re- long and possess walls that contain secretory cells (Fig. 29) like
cently been described from the Upper Permian of India (Chan- those in the leaves of D. fremouwensis (Pigg, 1990b) from the
dra et al., 2008), providing some support for the idea that the same peat deposit (Pigg and Taylor, 1990). If found dispersed,
corystosperms arose in the late Paleozoic (Kerp et al., 2006; the pollen of Pteruchus would be assigned to Alisporites Daugh-
Abu Hamad et al., 2008). erty, Pteruchipollenites Couper, or Falcisporites Leschik. Pol-
Several additional genera of cupulate structures are often in- len is bisaccate, with the sacci slightly inclined and partially
cluded in the corystosperms. These include Spermatocodon covering the distal sulcus. Pollen of P. dubius ranges from 80–
Thomas (1933), Pilophorosperma Thomas (1933) (Fig. 25), 115 µm in the primary plane (Taylor et al., 1984). The sacci
and Karibacarpon Lacey (1976). Judging from the number of have endoreticulations on their inner surfaces, but the outer sur-
species described, Pilophorosperma is perhaps the most com- faces are smooth (Zavada and Crepet, 1985).
mon seed-producing member of this group (Thomas, 1933). Another pollen organ believed to belong to the corystosperms
Some cupules are less hemispherical in outline than are those in is Pteroma Harris. Pteroma is known to date only from the
Umkomasia. Specimens of P. granulatum have three cupules at Middle Jurassic of Yorkshire, where it is associated with Pac-
the end of each branch, while two are common in P. gracile. On hypteris foliage (Harris, 1964). In Pteroma thomasii, the mi-
the outer surface of the cupule are short papillae, and the inner crosporophyll is flattened and bears two rows of pollen sacs
surface is covered with long, pointed hairs. In P. crassum the from what is interpreted as the lower surface. Each head is
cupules are borne in an overlapping fashion along the axis so shield-shaped, with the pollen sacs partially embedded in the
that the fertile branch appears compact. Holmes (1987) places tissue. Pollen is bisaccate and up to 107 µm long. It is not known
Pilophorosperma and Karibacarpon in the genus Umkomasia. whether these grains are protosaccate or eusaccate like those of
On the basis of consistent co-occurrence in the Middle Triassic Pteruchus (Osborn and Taylor, 1993). Other putative corys-
of eastern Australia, he suggests that U. feistmantelii was the tosperm pollen organs include Nidiostrobus from India (Bose
cupulate organ of the plant that produced Dicroidium zuberi and Srivastava, 1973) and Kachchhia from India and Antarctica
foliage. (Bose and Banerji, 1984; Gee, 1989).


Figs. 13–22. Reproductive organs of Caytoniales and Corystospermales. 13. Caytonia cupules attached to axis. Bar = 5.0 mm. 14. Cupule of Caytonia
sewardii Thomas emend. Harris. Bar = 2.5 mm. 15. Suggested reconstruction of Caytonia cupule showing attachment of seeds and “stigmatic lip” (arrow).
16. Reconstruction of Caytonanthus arberi (Thomas) Harris showing synangia attached to pinnate branches (from Crane, 1985). 17. Branch bearing several
Caytonanthus synangia. Bar = 3.0 mm. 18. Several Umkomasia cupules (C) attached to branch. Bar = 1.0 mm. 19. Suggested reconstruction of U. asiatica
Zan et al. (courtesy B. Axsmith). 20. Axis of U. asiatica bearing cupules at multiple nodes. Bar 2.0 mm. (courtesy B. Axsmith) 21. Diagrammatic recon-
struction of U. uniramia (from Axsmith et al., 2000). 22. Cross section of U. resinosa cupule (C) showing ovule (O) and cupule stalk (B). Note distinctive
secretory cavities, e.g., below “C” (from Klavins et al., 2002). Bar = 1.0 mm.
15372197, 2009, 1, Downloaded from https://bsapubs.onlinelibrary.wiley.com/doi/10.3732/ajb.0800202 by CAPES, Wiley Online Library on [11/03/2024]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
244 American Journal of Botany [Vol. 96

Figs. 23–33. Reproductive organs of corystosperms and Petriellales. 23. Short shoot (SS) of a corystosperm bearing several pedicels terminating in
cupules (arrows) of Umkomasia uniramia. Bar = 1.0 cm. 24. Dicroidium axis showing leaf attached to stem and short shoot (SS), which bears cupules of
U. uniramia. Bar = 2.0 mm. 25. Suggested reconstruction of Pilophorosperma geminatum Thomas (redrawn from Thomas, 1933). 26. Branch bearing
several Pteruchus sp. microsporophylls (courtesy C. P. Daghlian). Bar = 5.0 mm. 27. Pteruchus microsporophyll showing attached pollen sacs (arrows)
15372197, 2009, 1, Downloaded from https://bsapubs.onlinelibrary.wiley.com/doi/10.3732/ajb.0800202 by CAPES, Wiley Online Library on [11/03/2024]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
January 2009] Taylor and Taylor—Seed ferns from the late Paleozoic–Mesozoic 245

PETRIELLALES 1933; Harris, 1937). As defined by Townrow (1960), the group


included bipinnate foliage of Lepidopteris, peltate, ovulate re-
This order of seed ferns was defined based on a single genus productive structures of Peltaspermum, and pinnately branched
of permineralized cupule containing seeds from the lower Mid- pollen organs of Antevsia, which bore small, oval pollen grains.
dle Triassic of Antarctica (Taylor et al., 1994). A new family, Townrow (1960) was able to put the three parts together be-
Kannaskoppiaceae, based on impression-compression speci- cause L. ottonis Schimper, A. zeilleri (Nathorst) Harris, and P.
mens, was later included in the order (Anderson and Anderson, rotula Harris all bore characteristic blisters on the epidermis
2003). and had the same stomatal structure. A unique feature of the
foliage of the peltasperms is the presence of intercalary pin-
Ovulate organs— Cupules of Petriellaea Taylor et al. (1994) nules on the rachis, termed Zwischerfiedern. Because of the
are bilateral, elongate, and borne in clusters on a dichotomizing peltate megasporophylls, the group was at one time thought to
axis (Fig. 30). A vascular bundle extends along the midrib of be morphologically isolated from other seed ferns. Today, how-
the cupule with a trace given off to each ovule. The cupule wall ever, the order encompasses a much wider range of morpholo-
is thin, but there is no evidence of a preformed suture. Each gies and reproductive biologies than in the original concept.
cupule contains from 2–6 small (1.0 mm in diameter), triangu- The geographic and geologic range of the group has also ex-
lar ovules (Fig. 31), each attached to the adaxial surface of the panded, and peltasperms are now known from North America,
cupule surface at differing levels (Fig. 32). Ovules are orthotro- Europe (Kerp, 1988; Kerp et al., 2001), and the Russian plat-
pous with the integument thickened only in the corners; extend- form (Gomankov and Meyen, 1986; Naugolnykh and Kerp,
ing from the distal end of the seed is a short micropylar tube. 1996), and from the Pennsylvanian (Kerp et al., 2001) into the
Morphologically, there is some superficial similarity between Triassic. As currently circumscribed, the group was probably
the cupules of Petriellaea and those of Caytonia. While the more widespread during the Permian than the Triassic.
structure of the Caytonia cupules remains unknown, in Petriel-
laea the organization suggests that the cupule has evolved by a Ovulate organs—The ovulate organs of the peltasperms dem-
proximal-distal folding of the megasporophyll (Fig. 32) (Taylor onstrate a wide range of morphological variability. Those from
et al., 1994), rather than by enrolling or folding laterally along the Rotliegend (Upper Pennsylvanian–Lower Permian) of Eu-
the midrib, the hypothesized ancestral conduplicate carpel or- rope are unlike the structures for which the group was erected.
ganization sometimes attributed to the earliest angiosperms. Based on repeated co-occurrences of material from the Rotlieg-
Nothing is known about pollen organs that were associated with end of central Europe, the well-known foliage type Callipteris
the Petriellaea cupules. conferta (Sternberg) Brongniart and the ovule-bearing reproduc-
The Kannaskoppiaceae includes impression fossils of pollen tive organ Autunia Krasser (Fig. 34) were combined into Autunia
organs (Kannaskoppianthus), ovulate structures (Kannaskop- conferta (Sternberg) Kerp (Kerp, 1982, 1988). Autunia was orig-
pia) (Fig. 33), and foliage (Kannaskoppifolia) from the Upper inally introduced for strobili with helically arranged, fan-shaped
Triassic Molteno Formation of South Africa (Anderson and megasporophylls, each bearing one or two ovules on the lower
Anderson, 2003). Leaves are cuneate to flabellate and lack a surface (Figs. 35, 36). Pollen-producing organs of Arnhardtia
petiole; the lamina is entire to deeply divided into segments (Zeiller) Haubold et Kerp are believed to belong to A. conferta
with anastomosing venation. Similar leaves from the Upper and consist of 5–9 pollen sacs borne on a peltate microsporophyll
Triassic of Argentina and Chile have been described as Rochip- (Barthel, 2006). Unlike previously described peltasperms, how-
teris by Herbst et al. (2001). The seed-bearing organs Kan- ever, these pollen sacs contain monosaccate grains assignable to
naskoppia are termed strobili with each ~2.0 cm long and are Vesicaspora Schemel; grains are up to 50 µm in diameter.
attached in groups of two or three in the axils of the leaves. The genus Peltaspermum, for which the group was estab-
Each consists of a short, proximal axis segment, which then lished, is known from both Gondwana and Laurasia, and in-
forks to form two secondary axes. Each of these bears two rows cludes ovulate organs (Harris, 1937) that consist of elliptical to
of megasporophylls, with each sporophyll containing a single radial, peltate, umbrellalike discs (Figs. 37, 38), which are
cupulate ovule (Fig. 33); at maturity, the cupule splits into three sometimes termed peltoids (e.g., Meyen 1984; Naugolnykh,
lobes. The pollen organ Kannaskoppianthus is up to 45 mm 2001). The discs (megasporophylls) are up to 1.0 cm in diame-
long and constructed of two rows of microsporophylls arising ter and helically arranged into simple linear-cylindrical strobili
from a forked axis, each with five recurved, longitudinal pollen (Fig. 39); ovules are borne in a ring on the lower surface of the
sacs in a distal concavity protected by an operculum. megasporophylls (Townrow, 1960). The number of ovules per
disc is variable.
Based on association, a concept which was expanded by
PELTASPERMALES Meyen (1987) and termed assemblage genera, Poort and Kerp
(1990) emended the genus Peltaspermum, so that foliage and
The peltasperms were initially described from the Upper Tri- ovulate structures are both circumscribed by the “natural” ge-
assic of Greenland and South Africa based on associated foli- nus Peltaspermum. “Natural” taxon concepts similar to the one
age (Lepidopteris Schimper), pollen organs (Antevsia Harris), suggested for A. conferta have been proposed for several other
and seed-bearing organs (Peltaspermum Harris) (Thomas, Early Permian taxa, including Autunia naumannii (Gutbier)


(courtesy C. P. Daghlian). Bar = 2.0 mm. 28. Suggested reconstruction of Pteruchus fremouwensis fertile axis (from Yao et al., 1995). 29. Section of per-
mineralized Pteruchus fremouwensis pollen sac containing bisaccate grains. Bar = 250 µm. 30. Suggested reconstruction of Petriellaea triangulata T.
Taylor et al. showing position of cupules on fertile axis. 31. Cross section of P. triangulata cupule containing three triangular seeds. Bar = 1.0 mm. 32. Dia-
grammatic cutaway of P. triangulata cupule showing seed attachment and proximal-distal folding of megasporophyll to form cupule (from Taylor et al.,
1994). 33. Branches bearing numerous Kannaskoppia vincularis Anderson and Anderson cupules (courtesy J. M. Anderson). Bar = 2.0 cm.
15372197, 2009, 1, Downloaded from https://bsapubs.onlinelibrary.wiley.com/doi/10.3732/ajb.0800202 by CAPES, Wiley Online Library on [11/03/2024]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
246 American Journal of Botany [Vol. 96

Figs. 34–42. Reproductive organs of peltasperms. 34. Suggested reconstruction of Autunia conferta ovuliferous organ (from Kerp, 1988). 35. Autunia
conferta ovulate organ showing megasporophylls (courtesy H. Kerp). Bar = 1.8 cm. 36. Suggested reconstruction of two A. conferta megasporophylls (from
Kerp, 1988). 37. Suggested reconstruction of Peltaspermum rotula megasporophyll showing several ovules (from Crane, 1985). 38. Peltaspermum martin-
sii (Kurtze) Poort et Kerp peltate megasporophyll showing scalloped margin (courtesy H. Kerp). Bar = 4.5 mm. 39. Suggested reconstruction of Peltasper-
mum thomasii axis bearing numerous megasporophylls with attached ovules (redrawn from Dilcher, 1979). 40. Suggested reconstruction of Peltaspermopsis
polyspermis (from Naugolnykh, 2001). 41. Suggested reconstruction of Lepidopteris frond with pollen organs of the Antevsia-type at the tip (courtesy M.
Zavada). 42. Suggested reconstruction of Antevsia zeilleri pollen organ showing pinnate axis bearing clusters of pollen sacs (from Crane, 1985).

Kerp (ovulate organs), Rhachiphyllum schenkii (Heyer) Kerp Peltaspermopsis polyspermis Gomankov is a natural genus
(foliage), and Arnhardtia scheibei (Gothan) Haubold et Kerp concept used for Late Permian peltasperms from Russia. The
(pollen organs) (e.g., Kerp, 1988; Kerp and Haubold, 1988; plant includes vegetative stems, some with conspicuous
Barthel, 2001, 2006). Peltaspermum reproductive organs, de- nodes, which are interpreted as evidence of seasonal growth
scribed from a site in Morocco, have certain features, i.e., resin- interruptions. Lanceolate leaves assigned to Pursongia
ous bodies, that also occur on Rhachiphyllum Kerp foliage from Zalessky are also considered to be part of the same plant,
the same site (Kerp et al., 2001). along with reproductive organs in the form of seed-bearing
15372197, 2009, 1, Downloaded from https://bsapubs.onlinelibrary.wiley.com/doi/10.3732/ajb.0800202 by CAPES, Wiley Online Library on [11/03/2024]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
January 2009] Taylor and Taylor—Seed ferns from the late Paleozoic–Mesozoic 247

discs and their “racemose” aggregations (Fig. 40) (Naugol- Nevertheless, at one time or another, taxa in all of these groups
nykh, 2001). have been suggested as demonstrating morphological stages of
The genus Meyenopteris Poort et Kerp has also been sug- angiospermy based on the implied homologies of the reproduc-
gested to be a “natural” genus of peltasperms (but see Karasev tive organs or as showing some intermediate stage in the trans-
and Krassilov, 2007), using the concept of Poort and Kerp formation of an angiosperm organ, usually a carpel. Space does
(1990), and includes the form taxa Lepidopteris natalensis and not permit a discussion of what does or does not constitute a
Peltaspermum thomasii. Lepidopteris natalensis foliage is typi- unique angiosperm feature, and all are familiar with the limita-
cally bipinnate, with small blister-like swellings on the axes tions imposed on defining such a feature in the fossil record.
and intercalary pinnules (Zwischerfiedern), characters that are The problem of real or implied homology using fossils is exac-
diagnostic for the genus. The blister-like swellings also occur erbated by the absence of structurally preserved fossils in some
on the ovule-bearing organs of P. thomasii, which made it pos- of the groups and by the use of composite terminals in some
sible to connect the two plant parts. In this species, two ovules phylogenetic analyses, rather than whole plants reconstructed
(3.0 mm long) occur on the lower surface of each disc-shaped on the basis of attachment, shared anatomical characters, or
megasporophyll. Lopadangium Zhao et al. emend. Gomankov consistent and exclusive co-occurrence.
et Meyen is used for peltaspermaceous ovulate organs that can- Like many other Paleozoic and Mesozoic pteridosperms, the
not be correlated with foliage or other organs. glossopterids have been suggested as possible ancestors to the
angiosperms. Perhaps the earliest advocate of this relationship
Pollen organs— In the Mesozoic forms, such as Antevsia, was Plumstead (1956), who interpreted the glossopterid ovulate
pollen sacs are borne on branches. Antevsia includes branched organs as bisexual. The glossopterid-angiosperm connection
axes that bear lateral groups of 4–12 elongate pollen sacs at has periodically gained support as additional information about
their distal tips (Fig. 41). Townrow (1960) described Antevsia the ovulate organs has been published and hypotheses advanced
as pinnate, with the main axis bearing alternate laterals in one that attempt to homologize these organs with those of other
plane and secondary branches produced irregularly. Although seed plants (e.g., Melville, 1983). For example, Retallack and
these branches are described as microsporophylls, the branches Dilcher (1981) championed the idea initially suggested by Steb-
are not expanded at their tips (Fig. 42). Branches show the char- bins (1974) of the glossopterids as flowering plant progenitors.
acteristic blister-like swellings that also occur on Lepidopteris They proposed a transformational series in which the number of
fronds. Pollen sacs are up to 5 mm long and dehisce longitudi- ovules on the megasporophyll is reduced to one. The me-
nally. Like the majority of the Mesozoic Gondwanan pel- gasporophyll would then enclose the ovule to form a second
tasperms, pollen grains are small (23–40 µm) and oval, with a integument, and the subtending leaf would be transformed into
distal sulcus; if dispersed, these grains would be assigned to the carpel. Doyle (2006) also discussed this idea as a way to
Cycadopites Wodehouse ex Wilson and Webster. evolve the angiosperm carpel, with perhaps Caytonia as an in-
In the Late Pennsylvanian–Early Permian peltasperms, the termediate. Geologic time remains an obstacle to the acceptance
pollen organs are not known in great detail. Perhaps the best- of glossopterids as angiosperm ancestors, however, because the
known form is Callipterianthus arnhardtii Roselt (1962) from glossopterids disappeared around the Permian–Triassic bound-
the Rotliegend (Pennsylvanian–Early Permian) of Germany. ary (~251 million years ago [Ma]), while the first recognizable
This fossil has a frond-like (pinnate) organization with pollen angiosperm megafossils do not occur until the Lower Creta-
sacs attached to reduced pinnules in the distal part of the frond. ceous (Barremian–Aptian, ~145 Ma; Friis et al., 2006).
Sterile pinnae occur in the proximal portion of the frond, and Phylogenetic analyses have treated the glossopterids as a
intercalary pinnules are present on the rachis. Although the single, composite terminal, reconstructed of various organs, of-
specimen is partially sterile, a correlation with sterile foliage is ten from different sites, thus providing little resolution of the
difficult. Callipterianthus arnhardtii has been suggested to rep- group’s diversity in the analysis (e.g., Nixon et al., 1994; Doyle,
resent the pollen-producing organ of either Autunia naumannii 2006; Hilton and Bateman, 2006). There has been some attempt
or Arnhardtia scheibei (see Kerp, 1996). Barthel (2006), how- to relate isolated organs based on associations and anatomy,
ever, assigned pollen organs of the Pterispermostrobus type, and in some cases, the vegetative parts of plants have been re-
which is composed of elongate pollen sacs attached to a peltate constructed, e.g., Glossopteris schopfii Pigg and G. skaarensis
microsporophyll, to Arnhardtia scheibei. Pigg from Antarctica (Pigg and Taylor, 1993) and G. homeva-
lensis Pigg et McLoughlin from Australia (Pigg and Nishida,
2006). A permineralized ovulate structure from the Late Perm-
DISCUSSION ian of Antarctica could be attributed to G. schopfii leaves based
on similar anatomy (Taylor and Taylor, 1992), and Nishida et
We have provided information about the reproductive parts al. (2007) suggested that Homevaleia gouldii ovulate organs
of the glossopterids and the major groups of late Paleozoic– belong to leaves of G. homevalensis. These studies represent a
Mesozoic seed ferns so as to provide base-level data that can be first step in taking what may be a heterogeneous group and de-
used in discussing either real or inferred relationships with the veloping a framework that can be used to test subsequent hy-
angiosperms. Although there are, of course, still gaps in our potheses of relationships, not only within the glossopterids, but
knowledge of these groups, a large amount of data has been with other seed plants. Until homologies can be established,
gathered in the last ten years, especially for the peltasperms and there is currently insufficient evidence to link the glossopterids
the corystosperms. Except for the Petriellales, these groups with any other major group of plants, either as progenitors or
have been considered as possible angiosperm ancestors because descendants.
all have ovules attached to a more or less leaflike megasporo- The late Paleozoic–Mesozoic seed ferns have never repre-
phyll. Some groups have cupules that partially enclose their sented a natural group, and whether each of the four orders dis-
ovules, i.e., the corystosperms, Petriellales, and caytonialeans. cussed represents a monophyletic group is also far from settled,
Others, such as the peltasperms, do not have enclosed ovules. because no complete phylogeny has been presented for any of
15372197, 2009, 1, Downloaded from https://bsapubs.onlinelibrary.wiley.com/doi/10.3732/ajb.0800202 by CAPES, Wiley Online Library on [11/03/2024]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
248 American Journal of Botany [Vol. 96

these plants. It is clear from our current state of knowledge of cept of Thomas when he described the Caytoniales, subsequent
these groups that the discovery of additional specimens, espe- theories have used the cupule to form a second integument of
cially those in different modes of preservation, and research the seeds, and other tissue to derive the carpel. All these sce-
that focuses directly on these groups is what is needed in the narios suggest that megasporophyll closure is longitudinal, a
years ahead if more robust relationships are to be fully realized. theoretical basis that is supported by conduplicate folding in
In almost all phylogenetic analyses of seed plants published to certain extant flowering plants and even in some Early Creta-
date, each of these seed fern orders has been treated as a clade ceous angiosperms (e.g., Archaeanthus Dilcher et Crane,
and is often represented as a single, composite terminal, even Lesqueria Crane et Dilcher, Protomonimia Nishida et Nishida).
when reconstructions have been based only on association. What is interesting about Petriellaea is that megasporophyll
Some reconstructions of “whole” plants have used organs from closure may have proceeded in a different pattern in which the
different continents! In the case of the corystosperms, which tip of the megasporophyll is bent back toward the base or peti-
currently represent the best-known group overall, the remark- ole (Taylor et al., 1994). This may also represent the pattern in
able uniformity of both pollen and ovulate organs throughout Caytonia.
Gondwana suggests that at least the Gondwanan representa- Beginning in the late Paleozoic and continuing into the Me-
tives may constitute a single clade. Artabe and Brea (2003) con- sozoic, almost all seed-plant groups demonstrate increasing
clude that the group is not a natural one, but their analysis was levels of cupule and megasporophyll closure. Historically, this
based primarily on stem anatomy, a character that is variable has been interpreted as a response to predation or other selec-
and may simply reflect habitat. It is interesting to note that the tive advantages associated with ovule protection. While this
Dicroidium plants reconstructed from South America and Ant- structure–function relationship may have some validity, another
arctica suggest the presence of two quite different growth archi- hypothesis suggests that intercalating sterile tissue between
tectures (based on stem anatomy), although their reproductive ovules and pollen was a key physiological innovation in the
organs and foliage are very similar (Petriella, 1978; Taylor, transition from gymnospermy to angiospermy (Zavada and
1996). Taylor, 1986a). In this context, closure of the megasporophyll
The Caytoniales are perhaps the most poorly known of the was directly linked to the evolution of self incompatibility and
Mesozoic seed ferns, yet it is the cupule-bearing megasporo- a change from postzygotic to prezygotic selection. There would
phyll of Caytonia that has received the greatest amount of at- have been strong selective pressure on pollen to develop alter-
tention in scenarios leading to the origin of the angiosperm native mechanisms to insure compatibility with the stigma in
carpel and seeds with two integuments (e.g., Doyle, 2006). angiosperms. More efficient pollen transfer, the evolution of
While important details about the reproductive biology of the accessory floral parts to influence potential pollinator behavior,
Caytoniales are known, there is nothing to suggest that they and the loss of sacci may have been some of the evolutionary
possessed anything other than a distinctly gymnospermous stimuli in the transition to angiospermy. Interestingly, to the
seed-plant reproductive system. As is the case with some of the extent where pollen for a group is known, all the late Paleozo-
peltasperms, the Caytoniales produced saccate pollen, a mor- ic–Mesozoic seed ferns, including the glossopterids, possess
phological character state that both groups share with the Paleo- saccate pollen, with the exception of Cycadopites grains recov-
zoic callistophytalean seed ferns. It is important to point out ered from the peltasperm Antevsia (Townrow, 1960). This is in
that prior to the recognition of the Pennsylvanian Callisto- contrast to the situation in angiosperms where sacci are found
phytales, pollen of this morphotype would have implied affini- only in the extant genus Lactoris Philippi (Zavada and Taylor,
ties with some conifer or cordaite seed-plant group. If in fact any 1986b)
of these groups are distant progenitors of flowering plants, then While the morphological features showing a transition to the
the fact that many possess saccate pollen may simply reflect that angiosperm carpel may be difficult to resolve based on fossils,
the pollination syndrome did not keep pace with developmental the fossil record can sometimes offer another dimension—the
changes in the function of the mature microgametophyte. The gametophyte phase—in deciphering the transition from gymno-
primitive nature of the male gametophyte is also underscored spermy to angiospermy. The discovery of male gametes similar
by the discovery of flagellated sperm in the glossopterid Hom- to those in cycads and Ginkgo in the glossopterid Homevaleia
evaleia that, aside from their reduced size, morphologically re- (Nishida et al., 2003) contributes important details about one
semble the male gametes of cycads and Ginkgo. group of plants thought to represent potential flowering plant
At the present time, the Peltaspermales are perhaps the most progenitors. The inclusion of this character state in a data ma-
perplexing of all of these late Paleozoic–Mesozoic seed plants. trix may suggest glossopterid affinities with the cycads and
The morphology of the Autunia ovulate organ is distinct from Ginkgo—a conclusion that is predictable, but may not be cor-
that of the Peltaspermum known from numerous other sites, rect if the glossopterids are more heterogeneous than previously
including Siberia and Gondwana; however, the absence of any thought.
anatomical features significantly impedes progress in resolving We believe that our understanding of all these fossil groups
this seed-plant group. This is also the case for the Petriellales, will significantly advance when anatomically preserved speci-
although structurally preserved cupules and seeds are known, mens are found and when there are better reconstructions that
no other plant parts have been found to date. These isolated either combine organs unique to a site or that link organs based
organs are included within the assemblage known as seed ferns on distinctive synapomorphies (Taylor et al., 2006). With the
both for convenience and necessity, because all produce ovules exception of the Petriellales and Corystospermales, which are
on a modified, leaflike structure. known from permineralized and compressed fossils, we know a
We have noted earlier that many of the ideas regarding the great deal about the morphology and even the epidermal anat-
origin of the flowering plants have been based on scenarios in omy of some groups like the Caytoniales and Peltaspermales,
which megasporophylls become transformed so that the lamina but nothing about the internal anatomy. These data will be criti-
covers the ovule(s), and this stage has been cited as a precursor cal in characterizing homologies with other seed plant groups.
to the evolution of a carpel. Although this was the original con- In addition, details about where the reproductive organs are
15372197, 2009, 1, Downloaded from https://bsapubs.onlinelibrary.wiley.com/doi/10.3732/ajb.0800202 by CAPES, Wiley Online Library on [11/03/2024]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
January 2009] Taylor and Taylor—Seed ferns from the late Paleozoic–Mesozoic 249

produced, how they are attached, and how much diversity they Archangelsky [ed.], Compte Rendu, Congrès International de Géologie
demonstrate within each group are important areas that require du Carbonifère-Permian, Buenos Aires, Argentina, vol. 2, 13–40.
additional attention. This is especially true of the Glossopterid- Dilcher, D. L. 1979. Early angiosperm reproduction: An introductory re-
ales. As to the question posed in the title of this paper—the port. Review of Palaeobotany and Palynology 27: 291–328.
Doyle, J. A. 1978. Origin of angiosperms. Annual Review of Ecology and
same question that has been addressed countless times since Systematics 9: 365–392.
Darwin’s declaration regarding the “abominably perplexing” Doyle, J. A. 1996. Seed plant phylogeny and the relationships of
problem of angiosperm origins (Crepet, 2000)—from our per- Gnetales. International Journal of Plant Sciences 157 (6 Supplement):
spective, the mystery remains intact. We would offer one ca- S3–S39.
veat, however. In our opinion, it will be more productive to Doyle, J. A. 2006. Seed ferns and the origin of angiosperms. Journal of
attempt to solve Darwin’s mystery if there were greater atten- the Torrey Botanical Society 133: 169–209.
tion directed at mining the rock record in the hope of discover- Du Toit, A. L. 1932. Some fossil plants from the Karroo System of South
ing more informative and new specimens, than to continue to Africa. Annals of the South African Museum 28: 369–393.
construct new phylogenies using the same, often ambiguous Feistmantel, O. 1886. The fossil flora of the Gondwana System. Part 2.
characters. The fossil flora of some of the coal fields in western Bengal. Memoirs
of the Geological Survey of India: Palaeontologia Indica, 12th series,
4: 1–71.
LITERATURE CITED Friis, E. M., K. Raunsgaard Pedersen, and P. R. Crane. 2006. Cretaceous
Abu Hamad, A., H. Kerp, B. Vörding, and K. Bandel. 2008. A Late angiosperm flowers: Innovation and evolution in plant reproduction.
Permian flora with Dicroidium from the Dead Sea region, Jordan. Palaeogeography, Palaeoclimatology, Palaeoecology 232: 251–293.
Review of Palaeobotany and Palynology 149: 85–130. Gaussen, H. 1946. V: Les autres cycadophytes. L’origine cycado-
Adendorff, R. 2005. A revision of the ovuliferous fructifications of phytique des angiospermes. Travaux du Laboratoire Forestier de
glossopterids from the Permian of South Africa. Ph.D. dissertation, Toulouse, Tome II, Etudes dendrologiques, Première section, Les
University of the Witwatersrand, Johannesburg, South Africa. gymnospermes genéralités, vol. I, Les gymnospermes actuelles et fos-
Adendorff, R., S. McLoughlin, and M. K. Bamford. 2002. A new siles, partie 2, fascicule 3, chapitre 5: 1–26.
genus of ovuliferous glossopterid fructifications from South Africa. Gee, C. T. 1989. Revision of the Late Jurassic/Early Cretaceous flora
Palaeontologia Africana 38: 1–17. from Hope Bay, Antarctica. Palaeontographica 213B: 149–214.
Anderson, J. M., and H. M. Anderson. 2003. Heyday of the Gomankov, A. V., and S. V. Meyen. 1986. Tatarina-flora (composi-
Gymnosperms: Systematics and Biodiversity of the Late Triassic tion and distribution in the Late Permian of Eurasia). Akademii Nauk
Molteno fructifications (Strelitzia 15). South African National SSSR. Ordena Trudovogo Krasnogo Znameni Geologicheskii Institut
Botanical Institute, Pretoria, South Africa. 401: 1–174.
Archangelsky, S. 1996. Aspects of Gondwana paleobotany: Gould, R. E., and T. Delevoryas. 1977. The biology of Glossopteris:
Gymnosperms of the Paleozoic–Mesozoic transition. Review of Evidence from petrified seed-bearing and pollen-bearing organs.
Palaeobotany and Palynology 90: 287–302. Alcheringa 1: 387–399.
Artabe, A., and M. Brea. 2003. A new approach to Corystospermales Harris, T.M. 1937. The fossil flora of Scoresby Sound East Greenland.
based on Triassic permineralized stems from Argentina. Alcheringa Part 5: Stratigraphic relations of the plant beds. Meddelelser om
27: 209–229. Grønland 112(2): 1–114.
Axsmith, B. J., E. L. Taylor, and T. N. Taylor. 2007. The “new ap- Harris, T. M. 1940. Caytonia. Annals of Botany 4: 713–734.
proach to Corystospermales” and the Antarctic fossil record: A cri- Harris, T. M. 1941. Caytonanthus, the microsporophyll of Caytonia.
tique. Ameghiniana 44: 223–230. Annals of Botany 5: 47–58.
Axsmith, B. J., E. L. Taylor, T. N. Taylor, and N. R. Cúneo. 2000. New Harris, T. M. 1951. The relationships of the Caytoniales. Phytomorphology
perspectives on the Mesozoic seed fern order Corystospermales based 1: 29–39.
on attached organs from the Triassic of Antarctica. American Journal Harris, T. M. 1958. The seed of Caytonia. Palaeobotanist 7: 93–106.
of Botany 87: 757–768. Harris, T. M. 1964. The Yorkshire Jurassic flora. II. Caytoniales,
Barthel, M. 2001. Arnhardtia scheibei (Gothan) Haubold & Kerp Cycadales and pteridosperms. British Museum (Natural History),
und andere Rotliegend-Peltaspermaceen. Wege und Irrwege ihrer London, UK.
Erforschung. Hallesches Jahrbuch für Geowissenschaften 23: 1–8. Herbst, R., A. Troncoso, and S. Gnaedinger. 2001. Rochipteris nov.
Barthel, M. 2006. Makroflora und zugehörige “in situ” Sporen. In W. gen., hojas incertae sedis (= Chiropteris pro parte) del Triásico supe-
Reichel and M. Schauer [eds.], Das Döhlener Becken bei Dresden— rior de Argentina y Chile. Ameghiniana 38: 257–269.
Geologie und Bergbau, (Bergbau in Sachsen 12), 28–40. Sächsisches Hilton, J., and R. M. Bateman. 2006. Pteridosperms are the backbone
Landesamt für Umwelt und Geologie (LfUG), Dresden, Germany. of seed-plant phylogeny. Journal of the Torrey Botanical Society 133:
Bose, M. N., and J. Banerji. 1984. The fossil floras of Kachchh. I. 119–168.
Mesozoic megafossils. Palaeobotanist 33: 1–189. Holmes, W. B. K. 1987. New corystosperm ovulate fructifications from
Bose, M. N., and S. C. Srivastava. 1973. Some micro- and megas- the Middle Triassic of eastern Australia. Alcheringa 11: 165–173.
trobili from the Lower Triassic of Gopad River Valley, Nidpur. Holmes, W. B. K., and H. M. Anderson. 2005. The Middle Triassic
Geophytology 3: 69–80. megafossil flora of the Basin Creek Formation, Nymboida Coal
Brongniart, A. T. 1828. Prodrome d’une histoire des végétaux fossiles. Measures, New South Wales, Australia, part 4, Umkomasiaceae.
Dictionnaire des Sciences Naturelles 57: 16–212. Dicroidium and affiliated fructifications. Proceedings of the Linnean
Chandra, S., K. J. Singh, and N. Jha. 2008. First report of the fertile Society of New South Wales 126: 1–37.
plant genus Umkomasia from Late Permian Beds in India and its bio- Karasev, E. V., and V. A. Krassilov. 2007. Late Permian phyllo-
stratigraphic significance. Palaeontology 51: 817–826. clades of the new genus Permophyllocladus and problems of the
Crane, P. R. 1985. Phylogenetic analysis of seed plants and the origin of evolutionary morphology of peltasperms. Paleontological Journal
angiosperms. Annals of the Missouri Botanical Garden 72: 716–793. 41: 198–206.
Crepet, W. L. 2000. Progress in understanding angiosperm history, suc- Kerp, H. 1996. Post-Variscan late Palaeozoic Northern Hemisphere gym-
cess, and relationships: Darwin’s “abominably perplexing phenom- nosperms: The onset to the Mesozoic. Review of Palaeobotany and
enon.” Proceedings of the National Academy of Sciences, USA 97: Palynology 90: 263–285.
12939–12941. Kerp, H., A. Abu Hamad, B. Vörding, and K. Bandel. 2006. Typical
Cúneo, N. R., J. Isbell, E. L. Taylor, and T. N. Taylor. 1993. The Triassic Gondwanan floral elements in the Upper Permian of the pa-
Glossopteris flora from Antarctica: Taphonomy and paleoecology. In S. leotropics. Geology 34: 265–268.
15372197, 2009, 1, Downloaded from https://bsapubs.onlinelibrary.wiley.com/doi/10.3732/ajb.0800202 by CAPES, Wiley Online Library on [11/03/2024]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
250 American Journal of Botany [Vol. 96

Kerp, H., J. Broutin, S. Lausberg, and H. Aassoumi. 2001. Discovery Oliver, F. W., and D. H. Scott. 1904. On the structure of the Palaeozoic
of latest Carboniferous–Early Permian radially symmetrical pel- seed Lagenostoma lomaxi, with a statement of the evidence upon
taspermaceous megasporophylls from Europe and North Africa. which it is referred to Lyginodendron. Philosophical Transactions of
Comptes Rendus de l’Académie des Sciences. Sciences de la Terre the Royal Society of London 197B: 193–247.
332: 513–519. Osborn, J. M. 1994. The morphology and ultrastructure of Caytonanthus.
Kerp, J. H. F. 1982. Aspects of Permian palaeobotany and palynol- Canadian Journal of Botany 72: 1519–1527.
ogy. II. On the presence of the ovuliferous organ Autunia millery- Osborn, J. M., and T. N. Taylor. 1993. Pollen morphology and ultra-
ensis (Renault) Krasser (Peltaspermaceae) in the Lower Permian of structure of the Corystospermales: Permineralized in situ grains from
the Nahe area (F.G.R.) and its relationship to Callipteris conferta the Triassic of Antarctica. Review of Palaeobotany and Palynology
(Sternberg) Brongniart. Acta Botanica Neerlandica 31: 417–427. 79: 205–219.
Kerp, J. H. F. 1988. Aspects of Permian palaeobotany and palynology. X. Pant, D. D. 1987. Reproductive biology of the glossopterids and their
The west- and central European species of the genus Autunia Krasser affinities. Bulletin de la Société Botanique de France. Actualités
emend. Kerp (Peltaspermaceae) and the form-genus Rhachiphyllum Botaniques 134: 77–93.
Kerp (callipterid foliage). Review of Palaeobotany and Palynology Pant, D. D., and N. Basu. 1973. Pteruchus indicus sp. nov. from the
54: 249–360. Triassic of Nidpur, India. Palaeontographica 144B: 11–24.
Kerp, J. H. F., and H. Haubold. 1988. Aspects of Permian palaeobotany Pant, D. D., and D. D. Nautiyal. 1984. On the morphology and
and palynology. VIII. On the reclassification of the west- and cen- structure of Ottokaria zeilleri sp. nov.—A female fructification of
tral European species of the form-genus Callipteris Brongniart 1849. Glossopteris. Palaeontographica 193B: 127–152.
Review of Palaeobotany and Palynology 54: 135–150. Pant, D. D., and R. S. Singh. 1974. On the stem and attachment of
Kirchner, M., and A. Müller. 1992. Umkomasia franconica n. sp. und Glossopteris and Gangamopteris leaves. Part II. Structural features.
Pteruchus septentrionalis n. sp., Fruktifikationen von Thinnfeldia Palaeontographica 147B: 42–73.
Ettingshausen. Palaeontographica 224B: 63–73. Pedersen, K. R., and E. M. Friis. 1986. Caytonanthus pollen from the
Klavins, S. D., T. N. Taylor, and E. L. Taylor. 2002. Anatomy of Lower and Middle Jurassic. In J. T. Møller [ed.], Twenty-five years
Umkomasia (Corystospermales) from the Triassic of Antarctica. of geology in Aarhus, Department of Geology, Aarhus University,
American Journal of Botany 89: 664–676. Aarhus, Denmark. GeoSkrifter 24: 225–267.
Krassilov, V. A. 1977a. Contributions to the knowledge of the Petriella, B. 1978. La reconstrucción de Dicroidium (Pteridospermop-
Caytoniales. Review of Palaeobotany and Palynology 24: 155–178. sida, Corystospermaceae). Obra del Centenario del Museo de la Plata
Krassilov, V. A. 1977b. The origin of angiosperms. Botanical Review 5: 107–110.
43: 143–176. Petriella, B. 1981. Sistemática y vinculaciones de las Corystospermaceae
Lacey, W. S. 1976. Further observations on the Molteno flora of Rhodesia. H. Thomas. Ameghiniana 18: 221–234.
Arnoldia Rhodesia 7: 1–14. Petriella, B. 1983. Sinopsis de las Corystospermaceae (Corystospermales,
Lindström, S., S. McLoughlin, and A. N. Drinnan. 1997. Intraspecific Pteridospermophyta) de la Argentina. III. Troncos y chrono-strati-
variation of taeniate bisaccate pollen within Permian glossopterid spo- grafía. Ameghiniana 20: 41–46.
rangia, from the Prince Charles Mountains, Antarctica. International Pigg, K. B. 1990a. Anatomically preserved Glossopteris foliage from
Journal of Plant Sciences 158: 673–684. the central Transantarctic Mountains. Review of Palaeobotany and
McLoughlin, S. 1990. Some Permian glossopterid fructifications and Palynology 66: 105–127.
leaves from the Bowen Basin, Queensland, Australia. Review of Pigg, K. B. 1990b. Anatomically preserved Dicroidium foliage from
Palaeobotany and Palynology 62: 11–40. the central Transantarctic Mountains. Review of Palaeobotany and
McManus, H. A., E. L. Taylor, T. N. Taylor, and J. W. Collinson. Palynology 66: 129–145.
2002. A petrified Glossopteris flora from Collinson Ridge, Central Pigg, K. B., and H. Nishida. 2006. The significance of silicified plant
Transantarctic Mountains: Late Permian or Early Triassic? Review of remains to the understanding of Glossopteris-bearing plants: An
Palaeobotany and Palynology 120: 233–246. historical review. Journal of the Torrey Botanical Society 133:
Melville, R. 1983. Glossopteridae, Angiospermidae and the evidence 46–61.
for angiosperm origin. Botanical Journal of the Linnean Society 86: Pigg, K. B., and T. N. Taylor. 1990. Permineralized Glossopteris and
279–323. Dicroidium from Antarctica. In T. N. Taylor and E. L. Taylor [eds.],
Meyen, S. V. 1984. Basic features of gymnosperm systematics and Antarctic paleobiology—Its role in the reconstruction of Gondwana,
phylogeny as evidenced by the fossil record. Botanical Review 50: 164–172. Springer-Verlag, New York, New York, USA.
1–111. Pigg, K. B., and T. N. Taylor. 1993. Anatomically preserved Glossopteris
Meyen, S. V. 1987. Fundamentals of palaeobotany. Chapman & Hall, stems with attached leaves from the central Transantarctic Mountains,
New York, New York, USA. Antarctica. American Journal of Botany 80: 500–516.
Naugolnykh, S. V. 2001. A new peltaspermaceous pteridosperm from the Pigg, K. B., and M. L. Trivett. 1994. Evolution of the glossopterid
Upper Permian of the Russian Platform. Palaeobotanist 50: 189–205. gymnosperms from Permian Gondwana. Journal of Plant Research
Naugolnykh, S. V., and H. Kerp. 1996. Aspects of Permian palaeobot- 107: 461–477.
any and palynology. XV. On the oldest known peltasperms with radi- Plumstead, E. P. 1956. Bisexual fructifications borne on Glossopteris
ally symmetrical ovuliferous discs from the Kungurian (uppermost leaves from South America. Palaeontographica 100B: 1–25.
Lower Permian) of the fore-Urals (Russia). Review of Palaeobotany Plumstead, E. P. 1958. Further fructifications of the Glossopteridae
and Palynology 91: 35–62. and a provisional classification based on them. Transactions of the
Nishida, H., K. B. Pigg, K. Kudo, and J. F. Rigby. 2004. Zooidogamy Geological Society of South Africa 61: 52–79.
in the Late Permian genus Glossopteris. Journal of Plant Research Poort, R. J., and J. H. F. Kerp. 1990. Aspects of Permian palaeobot-
117: 323–328. any and palynology. XI. On the recognition of true peltasperms in
Nishida, H., K. B. Pigg, K. Kudo, and J. F. Rigby. 2007. New evidence the Upper Permian of Western and Central Europe and a reclassifica-
of reproductive organs of Glossopteris based on permineralized fos- tion of species formerly included in Peltaspermum Harris. Review of
sils from Queensland, Australia. I. Ovulate organ Homevaleia gen. Palaeobotany and Palynology 63: 197–225.
nov. Journal of Plant Research 120: 539–549. Prevec, R., S. McLoughlin, and M. K. Bamford. 2008. Novel double
Nishida, H., K. B. Pigg, and J. F. Rigby. 2003. Swimming sperm in an wing morphology in a South African ovuliferous glossopterid fructifi-
extinct Gondwanan plant. Nature 422: 396–397. cation. Bifariala intermittens (Plumstead 1958) comb. nov. Review of
Nixon, K. C., W. L. Crepet, D. Stevenson, and E. M. Friis. 1994. A re- Palaeobotany and Palynology 150: 22–36.
evaluation of seed plant phylogeny. Annals of the Missouri Botanical Retallack, G. J., and D. L. Dilcher. 1981. Arguments for a glossop-
Garden 81: 484–533. terid ancestry of angiosperms. Paleobiology 7: 54–67.
15372197, 2009, 1, Downloaded from https://bsapubs.onlinelibrary.wiley.com/doi/10.3732/ajb.0800202 by CAPES, Wiley Online Library on [11/03/2024]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
January 2009] Taylor and Taylor—Seed ferns from the late Paleozoic–Mesozoic 251

Reymanówna, M. 1973. The Jurassic flora from Grojec near Kraków Taylor, T. N., M. A. Cichan, and A. M. Baldoni. 1984. The ultra-
in Poland. Part II. Caytoniales and anatomy of Caytonia. Acta structure of Mesozoic pollen: Pteruchus dubius (Thomas) Townrow.
Palaeobotanica 14: 46–87. Review of Palaeobotany and Palynology 41: 319–327.
Roselt, G. 1962. Untersuchungen zur Gattung Callipteris. I. Callipteris Taylor, T. N., G. M. del Fueyo, and E. L. Taylor. 1994. Permineralized
scheibei Gothan, Aufbau und Habitus des ganzen Gewächses sowie seed fern cupules from the Triassic of Antarctica: Implications
stratigraphisches und geographisches Vorkommen. II. Callipterianthus for cupule and carpel evolution. American Journal of Botany 81:
arnhardtii n. g., n. sp., die erste durch Zusammenhang erwiesene 666–677.
Callipteris-Fruktifikation. Freiberger Forschungshefte, serie C 131: Taylor, T. N., and E. L. Taylor. 1987. Structurally preserved fos-
1–81. sil plants from Antarctica. III. Permian seeds. American Journal of
Schopf, J. M. 1976. Morphologic interpretation of fertile structures in Botany 74: 904–913.
glossopterid gymnosperms. Review of Palaeobotany and Palynology Thomas, H. H. 1925. The Caytoniales, a new group of angiosper-
21: 25–64. mous plants from the Jurassic rocks of Yorkshire. Philosophical
Smoot, E. L., and T. N. Taylor. 1986. Evidence of simple polyembry- Transactions of the Royal Society of London, B 213: 299–363.
ony in Permian seeds from Antarctica. American Journal of Botany Thomas, H. H. 1933. On some pteridospermous plants from the Mesozoic
73: 1079–1081. rocks of South Africa. Philosophical Transactions of the Royal Society
Stebbins, G. L. 1974. Flowering plants: Evolution above the spe- of London, B 222: 193–265.
cies level. Belknap Press of Harvard University Press, Cambridge, Townrow, J. A. 1960. The Peltaspermaceae, a pteridosperm family of
Massachusetts, USA. Permian and Triassic age. Palaeontology 3: 333–361.
Surange, K. R., and S. Chandra. 1975. Morphology of the gymnosper- Townrow, J. A. 1962. On Pteruchus, a microsporophyll of the
mous fructifications of the Glossopteris flora and their relationships. Corystospermaceae. Bulletin of the British Museum (Natural History),
Palaeontographica 149B: 153–180. Geology 6: 289–320.
Surange, K. R., and H. K. Maheshwari. 1970. Some male and female Yao, X., T. N. Taylor, and E. L. Taylor. 1995. The corystosperm
fructifications of Glossopteridales from India. Palaeontographica pollen organ Pteruchus from the Triassic of Antarctica. American
129B: 178–192. Journal of Botany 82: 535–546.
Taylor, E. L. 1996. Enigmatic gymnosperms? Structurally pre- Zan, S.-Q., B. J. Axsmith, N. C. Fraser, F.-X. Liu, and D. Xing.
served Permian and Triassic seed ferns from Antarctica. Review of 2008. New evidence for Laurasian corystosperms: Umkomasia from
Palaeobotany and Palynology 90: 303–318. the Upper Triassic of Northern China. Review of Palaeobotany and
Taylor, E. L., and T. N. Taylor. 1992. Reproductive biology of the Palynology 149: 202–207.
Permian Glossopteridales and their suggested relationship to flower- Zavada, M. S., and W. L. Crepet. 1985. Pollen wall ultrastructure of the
ing plants. Proceedings of the National Academy of Sciences, USA type material of Pteruchus africanus, P. dubius and P. papillatus.
89: 11495–11497. Pollen et Spores 27: 271–276.
Taylor, E. L., T. N. Taylor, H. Kerp, and E. J. Hermsen. 2006. Mesozoic Zavada, M. S., and W. L. Crepet. 1986. Pollen wall structure of
seed ferns: Old paradigms, new discoveries. Journal of the Torrey Caytonanthus. Plant Systematics and Evolution 153: 259–264.
Botanical Society 133: 62–82. Zavada, M. S., and T. N. Taylor. 1986a. The role of self-incompatibility
Taylor, E. L., T. N. Taylor, and P. E. Ryberg. 2007. Ovule-bearing and sexual selection in the gymnosperm–angiosperm transition: A hy-
reproductive organs of the glossopterid seed ferns from the Late pothesis. American Naturalist 128: 538–550.
Permian of the Beardmore Glacier region, Antarctica. USGS Open- Zavada, M. S., and T. N. Taylor. 1986b. Pollen morphology of
File Reports 2007-1047(082): 1–4. Lactoridaceae. Plant Systematics and Evolution 154: 31–39.

You might also like