A Revision of the Pennsylvanian‐Aged Eremopteris ‐Bearing Seed Plant

Abstract

Historically collected specimens of the Pennsylvanian pteridosperm Eremopteris artemisiaefolia have been reinvestigated to provide detailed information on its morphology and cuticular anatomy and to enable some of the external features of the plant to be reconstructed. The stem bore a distal crown of helically arranged compound leaves that show evidence they were actively abscissed from the plant. The blade consists of a main rachis that is straight or curved and may be undivided or show a single overtopped or occasionally dichotomous division. Secondary foliar segments are once or twice divided, with ultimate segments consisting of an elongate lanceolate blade, sometimes with one or two pairs of suboppositely arranged basal lobes or subsegments. Dense parallel veins run along the ultimate segments and only rarely fork. Stomata occur on both surfaces of the blade but more densely on the abaxial surface; papillae surrounded the abaxial stomata but not the adaxial ones. Remains of platyspermic bicornute ovules with a commissural rib in the minor plane are repeatedly associated with the leaves; these conform to the fossil genus Cornucarpus. Ovule cuticles include an inner integumentary cuticle, a nucellar cuticle that is attached only to the base of the inner integumentary cuticle, and a seed megaspore membrane. The nucellar apex comprises a small pollen chamber and extended nucellar beak consistent with cardiocarpalean ovules. Ovate structures on the main rachis and proximal parts of the secondary rachises are consistent in size with the ovule chalaza and are interpreted as the place of ovule attachment. Inclusion of our reconstruction of E. artemisiaefolia in a cladistic analysis of lignophytes suggests that it represents a distinct clade within a paraphyletic complex of basal pteridosperms and that it diverges after hydrasperman and medullosalean taxa and before the Callistophytales. The family Cornucarpaceae is redefined for this clade.

Full text

A REVISION OF THE PENNSYLVANIAN-AGED EREMOPTERIS-BEARING SEED PLANT
Christopher J. Cleal,1
,
* Cedric H. Shute,yJason Hilton,zand Julian Carter*
*Department of Biodiversity and Systematic Biology, Amgueddfa Cymru–National Museum Wales, Cathays Park, Cardiff CF10 3NP, United
Kingdom; yDepartment of Palaeontology, Natural History Museum, Cromwell Road, London SW7 5BD, United Kingdom; and zSchool
of Geography, Earth and Environmental Sciences, University of Birmingham, Edgbaston, Birmingham B15 2TT, United Kingdom
Historically collected specimens of the Pennsylvanian pteridosperm Eremopteris artemisiaefolia have been
reinvestigated to provide detailed information on its morphology and cuticular anatomy and to enable some of the
external features of the plant to be reconstructed. The stem bore a distal crown of helically arranged compound
leaves that show evidence they were actively abscissed from the plant. The blade consists of a main rachis that is
straight or curved and may be undivided or show a single overtopped or occasionally dichotomous division.
Secondary foliar segments are once or twice divided, with ultimate segments consisting of an elongate lanceolate
blade, sometimes with one or two pairs of suboppositely arranged basal lobes or subsegments. Dense parallel
veins run along the ultimate segments and only rarely fork. Stomata occur on both surfaces of the blade but more
densely on the abaxial surface; papillae surrounded the abaxial stomata but not the adaxial ones. Remains of
platyspermic bicornute ovules with a commissural rib in the minor plane are repeatedly associated with the leaves;
these conform to the fossil genus Cornucarpus. Ovule cuticles include an inner integumentary cuticle, a nucellar
cuticle that is attached only to the base of the inner integumentary cuticle, and a seed megaspore membrane. The
nucellar apex comprises a small pollen chamber and extended nucellar beak consistent with cardiocarpalean
ovules. Ovate structures on the main rachis and proximal parts of the secondary rachises are consistent in size
with the ovule chalaza and are interpreted as the place of ovule attachment. Inclusion of our reconstruction of
E. artemisiaefolia in a cladistic analysis of lignophytes suggests that it represents a distinct clade within a
paraphyletic complex of basal pteridosperms and that it diverges after hydrasperman and medullosalean taxa and
before the Callistophytales. The family Cornucarpaceae is redefined for this clade.
Keywords: Eremopteris, Carboniferous, seed plant, phylogeny, evolution.
Introduction
The plexus of Pennsylvanian (late Carboniferous) paleotropi-
cal gymnosperms with dissected leaves, often informally (and
sometimes formally) referred to as pteridosperms, are among
the most intensively studied Paleozoic plants. Their foliage,
which are the parts of these plants most commonly found fossil-
ized as adpressions, has been the subject of numerous mono-
graphic studies, such as by Buisine (1961), Laveine (1967),
Wagner (1968), Boersma (1972), and van Amerom (1975). The
consensus today is that they belong to one or other of three or-
ders, the Lyginopteridales, Medullosales, and Callistophytales.
There is, however, one Carboniferous pteridosperm-like plant
that remains enigmatic and cannot readily be classified—
Eremopteris Schimper. This distinctive plant has been known since
at least the early nineteenth century (von Sternberg 1825; Brong-
niart 1830; Lindley and Hutton 1832), mainly from British spec-
imens from Northumberland and Durham (fig. 1). Originally,
they were assigned to the fossil genus Sphenopteris Brongniart,
but Go
¨ppert (1836) recognized that the leaves were not exclu-
sively pinnate and so were atypical for that genus. He placed
them in a new fossil genus Gleichenites, which he established
for fossil fronds with a dichotomy (as in extant Gleichenia)but
for which the fructifications are unknown (see also Unger 1845;
Go
¨ppert in Bronn 1848). However, Go
¨ppert unambiguously re-
garded the lyginopteridalean Gleichenites linkii as the most ‘‘typ-
ical’’ species of this genus. Therefore, it is difficult to see how
the genus name could also be used for the much smaller Brit-
ish fronds with only irregularly developed dichotomies.
The generic nomenclature was eventually settled by Schimper
(1869), who assigned the fossils to a new fossil genus, Eremopteris,
but their systematic position remained problematic (Schimper’s
name for them was derived from the Greek for ‘‘fern without
analogy’’). The circumscription of the genus was also not clear,
as the diagnosis offered by Schimper (1869) failed to mention
the most important criterion for distinguishing the leaves from
other Pennsylvanian pteridosperms—the irregular dichotomies
of the main rachis (German and English translations of this di-
agnosis were subsequently published by Schimper [1879] and
Lesquereux [1880]). As a result, over the following years, a va-
riety of different leaves became incorporated within the fossil
genus (table 1).
An improved diagnosis was provided by Zeiller (in Renault
and Zeiller 1888), who mentioned that the leaf is variably di-
chotomous (a feature also discussed by Potonie
´[1897, 1898]
and Gothan in Potonie
´[1921]). Kidston (1924) gave a detailed
discussion on the genus and included the first extensive photo-
graphic documentation of specimens from the British coalfields,
and for many years, this was regarded as the definitive statement
1Author for correspondence; e-mail: chris.cleal@museumwales
.ac.uk.
Manuscript received February 2008; revised manuscript received February 2009.
666
Int. J. Plant Sci. 170(5):666–698. 2009.
Ó2009 by The University of Chicago. All rights reserved.
1058-5893/2009/17005-0009$15.00 DOI: 10.1086/597799

on the taxon. He concluded that it was clearly a seed plant
rather than a fern, as earlier authors had suggested, but was
uncertain about to which group it belonged. Emberger (1968)
proposed that Eremopteris be placed in a new family (Eremop-
teridaceae) but gave no formal diagnosis, and so the proposal
was invalid (International Code of Botanical Nomenclature
[ICBN] articles 34, 41; McNeill et al. 2006). In subsequent
years, the only detailed treatment was by Delevoryas and Taylor
(1969), based mainly on American specimens of Eremopteris
zamioides (Bertrand) Kidston, but here again no opinion was
given regarding their systematic position beyond that they were
pteridosperms. However, Delevoryas and Taylor (1969) were
thefirsttonotethatEremopteris leaves were apparently actively
abscissed from the stem, a feature that is unknown in any other
Middle Pennsylvanian–age tropical pteridosperm. Delevoryas
(1982) later argued that Eremopteris may be related to cycads
on the basis of its leaf morphology, an argument developed by
Crane (1988) following the recognition that platyspermic ovules
were plesiomorphic within cycads; hence, primitive cycads were
more like Eremopteris. Meyen (1987) and Doweld (2001) tenta-
tively suggested callistophytalean affinities on the basis of Dele-
voryas and Taylor’s (1969) supposition that the bicornute ovules
frequently found in association with Eremopteris foliage (also
noticed in Brongniart 1830; Arber 1914; Kidston 1924; Corsin
1928) were borne on leafless pinnate branching structures (‘‘poly-
sperms’’). However, the evidence that we present in this article
shows that the Eremopteris-bearing plant was different from
the typical members of the Callistophytales (Rothwell 1975, 1980;
Galtier and Be
´thoux 2002) and that its inclusion within that or-
der is inappropriate.
In this article, we reexamine what is known about the type
species of Eremopteris—Eremopteris artemisiaefolia (Sternberg)
Schimper. We have looked at the morphology of the leaves and
stem and the ovules that are regularly associated with them. We
report for the first time cuticles of both the foliage and the ovules,
and we show how this provides evidence about where the ovules
were attached on the plant. In view of a number of recent studies
on Carboniferous cuticle biochemistry (Lyons et al. 1994; Zo-
drow and Mastalerz 2002; Zodrow et al. 2003; Ps
ˇeni
cka et al.
2005), we have also investigated it for E. artemisiaefolia by using
infrared spectrometry. Finally, we have examined the geographi-
cal and stratigraphical distribution of this species. We have drawn
together these various strands of evidence to propose a new re-
construction, reassessed its likely phylogenetic position in the
context of recently published cladistic analyses of seed plants plus
their progymnosperm outgroups, and attempted to interpret the
novel features that it shows in terms of the ecology of the plant.
Material and Methods
This study is based largely on specimens in the collections of
the Natural History Museum (London), the Hancock Museum
Fig. 1 Earliest illustrated records of Eremopteris artemisiaefolia leaves. Left, von Sternberg (1825, pl. 54, fig. 1), holotype. Right, Brongniart
(1830, pl. 46).
667
CLEAL ET AL.—THE PENNSYLVANIAN SEED PLANT EREMOPTERIS

(Newcastle-upon-Tyne), and the Sedgwick Museum (Cam-
bridge). The specimens were photographed using cross-polar il-
lumination (Crabb 2001), and the resulting digital images were
traced using CorelDraw 12.
Natural History Museum. Fifteen specimens, mostly la-
beled simply as having originated from the ‘‘Coal Measures,’’ ei-
ther from Newcastle, Northumberland (OR.10404, OR.40597,
OR.52600, V.1714, V.9087, V.9093, V.15866, V.15867); York-
shire (V.65204); Carluke, Lanarkshire (OR.22404, V.65202); or
unlocalized other than ‘‘Britain’’ (V.65205, V.34124, V.34125,
V.52419).
Hancock Museum. Five specimens from the Hutton Collec-
tion: G73.47 and G73.63 (both from shale above Bensham Seam,
Jarrow, County Durham), G76.67 (shale above High Main Seam,
Table 1
Currently Accepted Systematic Position of the Species Historically Included within Eremopteris
Name Currently accepted name Systematic position Authority
E. aldrichii White, 1900 Nomen nudum ... This article
E. artemisiaefolia (Sternberg) Schimper,
1869 E. artemisiaefolia Cornucarpaceae Generic type
E. australis M. White, 1965 ?Genselia sp. Unknown This article
E. bilobata White, 1899 ?Eusphenopteris sp. indet. Lyginopteridales This article
E. cheathamii Lesquereux, 1884a Diplothmema cheathami (Lesquereux)
White
Lyginopteridales White 1943
E. courtinii Zeiller; Renault and Zeiller,
1888 Cf. Rhacopteris sp. ? Doubinger 1956
E. crassinervia (Goeppert) Schimper, 1869 ?Autunia naumannii (Gutbier) Kerp ?Peltaspermales This article
E. crenulata Lesquereux, 1876 ?Palmatopteris sp. Lyginopteridales This article
E. decipiens Lesquereux ex White, 1900 Mariopteris decipiens (Lesquereux)
White
Lyginopteridales White 1893
E. dissecta Lesquereux, 1876 ?Palmatopteris sp. Lyginopteridales This article
E. elegans (Ettingshausen) Lesquereux, 1880 Rhacopteris asplenites (Gutbier)
Schimper
Unknown Ne
ˇmejc 1928
E. erosa (Morris ex Brongniart)
Kidston, 1882 Sphenopteris erosa Morris ex
Brongniart
Unknown Brongniart in Murchison
et al. 1845
E. flexuosa Lesquereux, 1876 ?Palmatopteris sp. Lyginopteridales This article
E. golondrinensis Archangelsky, 1958 ?Fedekurtzia sp. ? Archangelsky 1981
E. gracilis White, 1943 ?Palmatopteris sp. Lyginopteridales This article
E. lincolniana White, 1900 ?Palmatopteris sp. Lyginopteridales This article
E. lucensis de Stefani, 1901 ?Sphenocallipteris sp. ?Peltaspermales This article
E. lyratifolia (Goeppert) Bergeron, 1889 Rhachiphyllum lyratifolia (Goeppert)
Kerp
?Peltaspermales Kerp and Haubolt 1988
E. macconochii Kidston, 1883 Sphenopteridium macconochii
(Kidston) Kidston
Lyginopteridales Kidston 1923
E. marginata Andrews, 1875 ?Megalopteris sp. Unknown This article
E. michiganensis Arnold, 1949 ?Rhacopteris sp. Unknown This article
E. microphylla Lesquereux, 1880 Diplothmema microphylla (Lesquereux)
White
Lyginopteridales White 1943
E. missouriensis Lesquereux, 1880 ?Palmatopteris cf. furcata (Brongniart)
Potonie
´
?Lyginopteridales White 1899
E. moyseyi Arber, 1911 Rhacopteris moyseyi (Arber) Kidston Unknown Kidston 1923
E. neesii (Goeppert) Schimper, 1869 Odontopteris sp. ?Medullosales Kerp and Haubold 1988
E. neffii White, 1943 E. artemisiaefolia Cornucarpaceae This article
E. peruvianus Berry, 1922 Nothorhacopteris argentinica (Geinitz)
Archangelsky
Unknown Jongmans 1954
E. sanjuarina (Kurtz) Frenguelli, 1941 Triphyllopteris collombiana Schimper Unknown Jongmans 1954
E. solida (Lesquereux) White; Adams
et al., 1903 Indet. fragments ? This article
E. stricta (Sternberg) Romanovski, 1890 E. artemisiaefolia Cornucarpaceae Kva
cek and Strakova
´1997
E. strigosa White, 1943 ?Palmatopteris sp. ?Lyginopteridales This article
E. subelegans White, 1943 ?Palmatopteris sp. ?Lyginopteridales This article
E. trappensis White, 1943 Indet. fragments ? This article
E. vasconcellosii de Lima ex Teixeira, 1941 Indet. fragments ? This article
E. warragulensis (McCoy) McCoy;
Stirling, 1892 Sphenopteris warragulensis McCoy Ferns Douglas 1973
E. whitei Berry, 1922 Sphenopteris whitei (Berry) Jongmans Unknown Jongmans 1954
E. zamioides (Bertrand) Kidston, 1924 E. zamioides Cornucarpaceae Kidston 1924
668 INTERNATIONAL JOURNAL OF PLANT SCIENCES

Gosforth, County Durham, 32 specimens), and G157.01 (HC
38b/37c, unlocalized).
Sedgwick Museum. Five specimens: M.703 (from Coal Mea-
sures, near Durham), M.704 (from ‘‘Coal Measures’’ but no fur-
ther details given), M.705 (Fawdon Colliery, probably from
shale above High Main Seam), M.1081 (from roof of Main
Coal, Blinkbonny Pit, Canonbie), and M.1121 (unlocalized).
Foliar cuticles were prepared using the standard methods.
The phytoleims were released from the matrix using hydro-
fluoric acid and then macerated in Schulze’s solution (nitric acid
and potassium chlorate). After washing the cuticles in ammo-
nium hydroxide and then distilled water, they were separated
and mounted on glass slides with glycerine jelly containing safra-
nin stain. The ovules were macerated in a similar way but were
mounted in unstained glycerine jelly. The foliar and ovule cuticles
were photographed using both film and digital technology with
a Leitz Ortholux II microscope, mostly with Normarski Phase
Contrast optics. The resulting slides are in the paleontology col-
lections of the Natural History Museum, London.
The biochemistry of a single piece of prepared cuticle was
investigated using Fourier transform infrared spectroscopy (FTIR).
Analysis was performed on a Perkin Elmer Spectrum One spec-
trometer, equipped with a deuterated triglycine sulfate detector,
using the Universal ATR sampling accessory. Spectral data were
captured as an accumulation of 10 scans at a resolution of 4
cm
1
. The infrared signal was recorded in the region of 400–
4000 cm
1
wave numbers. Five spectra were collected from dif-
ferent areas of the cuticle piece examined.
Foliar Morphology
The leaf blade can vary in shape from ovate to obovate and
was borne by an elongate petiole (figs. 2, 3). No complete
leaves have been found, but they seem to have varied somewhat
in size. The largest known specimen, which is incomplete at its
proximal end, has a leaf blade 112 mm wide and at least 240
mm long (fig. 1, right; fig. 2a). The smallest essentially complete
leaf blade (fig. 3f) is 60 mm wide and 105 mm long, but there
are other fragments (e.g., fig. 2e), suggesting a blade as narrow
as 45 mm wide. Complete petioles are rare; the only examples
that we have seen are 52 mm long and 2.5–3 mm wide (fig. 2e),
but this is from a small blade, and we do not know whether it
is typical for the species. A broken fragment of another petiole
is 60 mm long and 4 mm wide (fig. 2g). The petioles are parallel
sided for most of their length but flare at their proximal end. In
a number of cases, the proximal end of the petiole shows a con-
cave surface with a thickened rim, suggestive of an abscission
surface (fig. 4b).
The leaf blade is dissected in an essentially pinnate manner,
but as the branching is not strictly pinnate, we refer to the vari-
ous divisions as (leaf) segments rather than pinnae. The larger
blades have three orders of segments (e.g., fig. 2a,2c;figs.3c,
5b) and the smaller ones essentially two orders (fig. 3e,3f;fig.
5c), although they may still tend to be three times divided in the
proximal part of the blade (fig. 3d). The long axis of the blade
is marked by a main rachis, which forms a continuation of the
petiole. The main rachis can vary from 2 to 4 mm in width. In
some blades, the main rachis continues straight for its entire
length (e.g., fig. 3d,3e). In other cases, the main rachis curves
in the distal part of the blade, with one of the secondary seg-
ments attached to the convex side of the curved main rachis be-
ing enlarged (fig. 2c,2h;figs.4a,5b). This process can go even
further, resulting in a bifurcation of the main rachis (fig. 2a;fig.
3a–3c,3f; fig. 5a,5c). Even in such bifurcated blades, however,
one arm of the bifurcation is often somewhat longer than the
other, and the blade thus continues to have a curved aspect. In
fully bifurcating blades, the angle of bifurcation can vary from
25°to 40°. A straight main rachis is normally found only in smaller
blades, and a curved main rachis occurs only in larger blades, but
bifurcatemainrachisescanoccurinbothsmallandlargeblades.
The secondary segments are obliquely attached to the main
rachis. The most proximal secondary segments are mostly op-
positely attached at 38°–42°, occasionally at up to 50°.Inthe
most distal parts of the blade, the secondary segments are alter-
nately or suboppositely attached at more oblique angles of 25°–
35°. In the proximal and middle parts of the blade, the secondary
segments are rhomboid and up to 60 mm long and 20–25 mm
wide; in more distal positions, they are more oblong to linear.
Tertiary segments, up to 25 mm long and 7–8 mm wide, are
obliquely attached to the secondary rachis at 25°–35°. In dis-
tal positions, the tertiary segments are entire and elongately
lanceolate, with a rounded to obtuse apex. In more proximal
positions, the tertiary segments develop a pair of opposite lat-
eral incisions and then in more proximal positions again these
become a pair and sometimes two pairs of lobes. The basiscopic
tertiary segments are often enlarged, especially in the most
proximal pair of secondary segments, where they become asym-
metrically rhomboid. In such basiscopic tertiary segments, the
lobes may themselves start to lobe, thus representing a further
level of division of the blade. In all cases, the terminal lobe of
the tertiary segments is larger than the lateral lobes.
The ultimate divisions can vary in different blades from grac-
ile (e.g., fig. 2c;fig.4a,4b;fig.5a,5b,5d)torobust(fig.3c),
but the ultimate divisions in most blades are of an intermediate
lanceolate form (e.g., fig. 2a,2d,2f;fig.3a,3d,3e;figs.4d,5c).
The ultimate divisions are widest in their middle part or some-
what distally in the more robust forms and have a rounded to
bluntly acuminate apex.
The rachises have numerous longitudinal striations (;10/mm)
that seem to retain their individuality for some distance and sug-
gest that they may have been discrete vascular bundles. Each ulti-
mate division is provided with one slender vein that detaches
itself from one of the rachial strands at some distance from where
the ultimate division occurs. The vein then forks once or twice in
the proximal part of the ultimate segment, the resulting branches
then running along the length of the segment, sometimes forking
again, especially in the wider segments (fig. 6). There are usually
3.0–3.5 veins/mm across the ultimate segment width.
In a number of specimens (fig. 7), there are small but marked
oval structures along the main rachis and sometimes in the proxi-
mal parts of the secondary rachises. The structures vary consider-
ablyinsizefrom0.3mmlongand0.1mmwidetoupto1.1mm
long and 0.7 mm wide and have a markedly raised rim. They prob-
ably correspond to the oval structures found on the cuticles of the
main rachis (see below).
Comparisons
Most other Paleozoic pteridosperms of the paleotropical wet-
lands had leaves much larger than those of Eremopteris;even
669
CLEAL ET AL.—THE PENNSYLVANIAN SEED PLANT EREMOPTERIS

Fig. 2 Tracings of gracile form of leaves dealt with in this study. Scale bar ¼20 mm. a, Brongniart (1830, pl. 46). b, BMNH V.34124. c,
Hancock Museum Specimen G157.01a. d, BMNH V.9087. e, BMNH V.52610. f, Brongniart (1830, pl. 47, fig. 2). g, BMNH V.15866. h,i,
BMNH V.15867. j, BMNH OR.10404. k, BMNH V.65202. l, BMNH V.65204.

the relatively small-leaved lianescent lyginopteridaleans, such
as Mariopteris and Karinopteris, had fully developed leaves
that were at least twice as large as the largest Eremopteris leaves.
Moreover, lyginopteridalean, callisophytalean, and medullosa-
lean fronds all had a more rigidly programmed architecture, al-
most invariably with a proximal dichotomy of the main rachis
producing two branches with at least a superficially pinnate-
looking construction (Laveine 1967; Boersma 1972; van Amerom
1975; Cleal and Shute 1991; although see also Laveine 1997
for an alternative architectural interpretation of this ‘‘pinnate’’
branching). Eremopteris, in contrast, appears to have had a less
canalized leaf architecture, in which a dichotomy of the main ra-
chis was only sometimes developed, and the ultimate divisions
within the leaf blade tend to be irregularly developed. Finally,
none of these other pteridosperms show evidence of active leaf
abscission as seen in Eremopteris (Thomas and Cleal 1999).
The venation of Eremopteris also differs from that of most of
these other Paleozoic pteridosperms. Each ultimate segment of
Eremopteris has just a single order of veins, which occasionally
dichotomize and lie longitudinally along the segment (fig. 6).
The ultimate segments of the other pteridosperms, in contrast,
have two orders of veins: a midvein that arises from the rachis
and subsidiary, or lateral, veins that arise from the midvein.
There are some species of these pteridosperms that appear to
have no midvein and the subsidiary veins arise directly from the
rachis. However, where these cases have been examined in detail
Fig. 3 Tracings of robust leaves dealt with in this study. Scale bar ¼10 mm. a, BMNH V.34125. b, BMNH OR.40597. c, von Sternberg (1825,
pl. 54, fig. 1), holotype. d, Kidston (1924, pl. 111, fig. 1). e, Brongniart (1830, pl. 47, fig. 1). f, Kidston (1924, pl. 111, fig. 2).
671
CLEAL ET AL.—THE PENNSYLVANIAN SEED PLANT EREMOPTERIS

Fig. 4 Eremopteris artemisiaefolia leaves, all natural size. a, Small leaf with a subdichotomy of the main rachis and numerous associated
ovules. BMNH V.34124, Coal Measures, Britain. b, Proximal part of leaf showing petiole with abscission layer at base. BMNH OR.52600, Coal
Measures, Newcastle, United Kingdom. c, Two leaves, one small and one large, associated with numerous ovules. BMNH V.15867, Coal
Measures, Newcastle, United Kingdom. d, Robust form of leaf. BMNH V.9087, Coal Measures, Newcastle, United Kingdom.
672

Fig. 5 Eremopteris artemisiaefolia leaves, all natural size. a, Gracile form of leaf with dichotomy of main rachis. BMNH OR.40597,
Newcastle, United Kingdom. b, Gracile form of leaf. Hancock Museum G.157.01. c, Leaf showing bifurcation of main rachis. BMNH V.34125,
Coal Measures, United Kingdom. d, Fragment of gracile form leaf. BMNH V.65204, Coal Measures, Yorkshire, United Kingdom.
673

(Bochenski 1960; Saltzwedel 1969; S
ˇimu
˚nek and Cleal 2004), it
has become evident that there is always a midvein, albeit some-
times developed in the very proximal part of the pinnule and
thus difficult to see, and that the subsidiary veins arise only from
that midvein and never directly from the rachis. This is clearly a
fundamental difference between Eremopteris and these other
pteridosperms, although it must be remembered that small frag-
ments of some lyginopteridalean pinnules (e.g., Palmatopteris
Potonie
´) may superficially seem to have only one order of veins
but more complete portions of pinnae are seen to have two or-
ders of veins.
Among other Paleozoic pteridospermous gymnosperms, the
foliage most similar to Eremopteris are the fronds of the Late
Carboniferous to Early Permian putative peltasperm Autunia
Krasser (Kerp and Haubold 1988). Both frond types are essen-
tially two to three times divided, without the main basal dichot-
omy that characterizes most fronds of the other two main Late
Paleozoic pteridospermous gymnosperms, the Lyginopteridales
and Medullosales. Like typical Autunia fronds (Kerp 1988),
many of the Eremopteris leaves are obovate with an obtuse
apex and significantly show some curvature and overtopping of
the primary rachis (e.g., V.34124). Some Eremopteris leaves
have a more fully developed dichotomy of the primary rachis,
in the middle part of the frond (Brongniart 1830, pl. 46;
V.34125; OR.40597), but even here the two arms produced by
this fork tend to be asymmetrical and can clearly be related to
the overtopping seen in Autunia. However, Eremopteris shows
no evidence of the intercalated pinnules on the primary rachis
that characterize Autunia and Rhachiphyllum fronds (Kerp
1988; Kerp and Haubold 1988). Also, the architecture of Ere-
mopteris fronds is far more variable than that of Autunia;some
leaves tend to a more strictly pinnate branching pattern (Brong-
niart 1830, pl. 47; Kidston 1924, pl. 111, fig. 1) more compara-
ble to that seen in Rhachiphyllum (Kerp 1988).
Morphology of ‘‘Anomalous’’ Leaves
A leaf from the collections of the Sedgwick Museum associ-
ated with ‘‘normal’’ examples of Eremopteris artemisiaefolia
shows a number of unusual characteristics (fig. 8f). It has a
main rachis that is very similar to that seen in E. artemisiaefo-
lia, with irregular longitudinal striae and tapering distally in
width from 4 to 3 mm. It is preserved for a length of 37 mm be-
fore dichotomizing at an angle of ;20°. The resulting branches
are 37 mm long and just over 1 mm wide for most of their
length, tapering noticeably only in their most distal parts.
The main rachis and the branches produced by the dichot-
omy bear secondary segments. These are oppositely arranged in
the proximal part of the leaf blade, becoming alternately ar-
ranged in more distal positions. The longest complete lateral seg-
ment (attached just below the dichotomy) is 20 mm long and
has a rachis ;1 mm wide. The angle of attachment varies from
;40°in the most proximal preserved part to 20°in the most
distal part of the leaf blade.
Delicate tertiary segments are attached to the secondary ra-
chises and to the main rachis intercalated between the second-
ary segments. The angle of attachment varies from 40°to 60°.
The smaller tertiary segments, in the distal parts of the second-
ary segments, are 3–4 mm long and 1–2 m wide and consist of
three lobes. The larger tertiary segments, which can be up to 6
mm long and 3–4 mm wide, are themselves short, pinnate struc-
tures bearing short, trilobed ultimatesegments.Theultimatelobes
tend to be linear or slightly barreled, with an acute or blunt apex.
A number of other detached fragments of similar tertiary seg-
ments are preserved on the block.
Smaller fragments of similar, apparently anomalous, leaves
were found in the Hancock Museum collections (e.g., fig. 8g).
They have a main rachis ;4mmwidewithlongitudinalstriae
and oppositely to suboppositely arranged, twice-divided sec-
Fig. 6 Traces of ultimate segments of Eremopteris artemisiaefolia leaves showing venation. a, BMNH V.15867. b, BMNH V.15866. c, BMNH
V.65205.
674 INTERNATIONAL JOURNAL OF PLANT SCIENCES

Fig. 7 Traces of Eremopteris artemisiaefolia leaves showing positions of small oval scars on the rachises (marked by gray circles). Scale bar ¼
20 mm. a, BMNH V.15867. b, BMNH V.65202. c, BMNH V.15866. d, BMNH V.34124. e, BMNH V.9087.

Fig. 8 Eremopteris artemisiaefolia.a–e, Stem bearing leaves G73.47 (HC98a, 43 b), Bensham Seam, Jarrow, Durham, United Kingdom. a,
Distal part of stem showing helically arranged petioles, 32. b, General view of stem and attached petioles, 31. c, Close-up of stem surface showing
leaf scars and persistent leaf bases, 36. d, Part of one of the leaves attached to the stem showing characteristic segment morphology for E.
artemisiaefolia,36. e, Close-up of leaf scar on stem surface, 312. f,g, ‘‘Anomalous’’ leaves, 31. f, Leaf showing normal development in proximal
and central part but more delicate pinnules in distal positions. Sedgwick Museum Specimen M 704, ‘‘Coal Measures,’’ no further locality details,
31. g, Near-distal part of leaf showing delicate form of pinnules. Hancock Museum Specimen G76.67a, shale above High Main Seam, Gosforth,
County Durham, 31.

ondary segments. The ultimate segments are a little larger than
those of the Sedgwick Museum specimen but are otherwise sim-
ilar in general form.
Interpretation
The foliar structures that we termed ‘‘anomalous’’ leaves in the
descriptions are clearly the same as what Delevoryas and Taylor
(1969) described as mega- and microsporophylls. However, Dele-
voryas and Taylor found no evidence of ovules or pollen organs
attached to these structures, and neither have we. In fact, we will
present later in this article evidence to suggest that the ovules
were attached directly to the main rachises of ‘‘normal’’ leaves.
The apparently irregular branching in the larger specimen
(fig. 8f) might suggest a teratological explanation. However,
several specimens of this type are known from both Britain and
North America, which makes this explanation unlikely. A re-
sponse to arthropod predation is also not borne out by the ab-
sence of any obvious damage to the leaves.
Delevoryas and Taylor (1969) considered that they might
have been young vegetative leaves that had not fully developed.
They decided against this interpretation because these leaves
look so ‘‘conspicuously different’’ from what were ‘‘obviously
immature fronds’’ found in the same bed. However, they had
available only distal fragments of these ‘‘anomalous’’ leaves,
and the larger specimen that we document in this article (fig.
8f) shows such distal fragments attached to a main rachis indis-
tinguishable from a ‘‘normal’’ eremopterid main rachis. In our
view, the most likely explanation is that these were young
leaves preserved in the early stages of their ontogeny and that it
does not take much of a leap of imagination to relate them to
the ‘‘immature vegetative fronds’’ illustrated by Delevoryas and
Taylor (1969, fig. 10). It is notable that the epidermal structure
shown by the cuticle from one of the ‘‘anomalous’’ leaves (fig.
10d) is similar to that of the costal fields of ‘‘normal’’ leaves.
If these are indeed young leaves, then their ontogeny would
seem to be quite different from that seen in the medullosalean
and lyginopteridalean pteridosperms, where young fronds formed
in a crosier, similar to many modern-day ferns (Huth 1912;
Crookall 1976; Cleal and Laveine 1988).
Cuticles
Leaf Blade Lamina
Although well-preserved cuticles were obtained from the
blade of Eremopteris artemisiaefolia, no complete cuticles from
an ultimate segment or even segment lobe could be obtained.
This was because the phytoleim consistently fragmented during
digestion in hydrofluoric acid. Consequently, interpreting the
distribution of features such as stomata and papillae across the
segments had to be determined by indirect means.
Both adaxial and abaxial surfaces of the lamina are thinly cu-
tinized, and both have an epidermis clearly differentiated into
costal and intercostal fields (fig. 9). On the abaxial epidermis,
the intercostal fields are consistently 225–250 mm wide, but the
costal fields vary from 350 mm in the medial part of the blade
to 55 mm wide near the lobe apices.
Costal epidermal cells are elongate, subrectangular to sub-
rhomboidal, aligned parallel to veins; on the adaxial surface, they
are 30–120 mm long and 5–20 mmwide(typically;60 312 mm),
and on the abaxial surface, they are slightly larger, 40–150 mm
long and 9–30 mm wide (typically ;90 315 mm). Intercostal cells
are irregularly polygonal, more or less isodiametric or just slightly
elongate parallel to the veins, and especially on the adaxial surface
often have sinuous anticlinal walls; on the adaxial surface, they
are 30–100 mm long and 6–80 mmwide(typically;55 330 mm),
and on the abaxial surface, they tend to be a little smaller,
30(48)70 mm long and 20–50 mm wide (typically ;50 330 mm).
Many epidermal cells have a thickly cutinized papilla in the
middle of their periclinal wall (fig. 9b,9d,9e). They are mostly
10–20 mm in diameter, but on the abaxial costal fields, they are
alittlelarger,15–25mm (very rarely up to 30 mm) in diameter.
Papillae are unevenly distributed but tend to be denser in the
abaxial intercostal fields, where they often obscure the anticli-
nal walls of the epidermal cells. The areas of abaxial cuticle
from more distal positions in the lobes of the foliage are nearly
always densely papillate, whereas the areas of large cuticle pre-
sumably from more proximal positions in the lobes often (but
not always) have no or only a few papillae. From this, we surmise
that papillae tend to be more concentrated toward the margins of
the blade. The narrower papillae are bluntly rounded, 7–11 mm
in height. The wider papillae on the abaxial intercostal fields
have more thickly cutinized margins and a flattened apex and are
6–7 mm high. The papillae mostly have a circular cross section,
but the larger ones have an oval cross section with a long axis
parallel to the vein.
Stomata are restricted to the intercostal fields on both sur-
faces. On the abaxial surface, there are usually 90–180 (typically
125–140)/mm
2
(fig. 9e), although some areas seem to have vir-
tually no stomata; on the adaxial surface, they are very sparse
and irregularly distributed, making it difficult to give a meaning-
ful density measurement (fig. 9a,9c). The abaxial stomata appear
to be surrounded by a single ring of six subsidiary cells, between
three and six of which have a prominent papilla that overarches
the stomatal pore (fig. 9f). In contrast, the adaxial stomata
have no subsidiary cells or overarching papillae (fig. 9c). The
stomata are orientated approximately parallel to the veins and
are not noticeably sunken. The guard cells are 20–36 mm long and
;5mm wide.
Rachial Part of Leaf Blade
Cuticles were prepared from the rachis, and these showed
epidermal structures essentially similar to those of the lamina.
The adaxial epidermis is clearly divided into 70–300-mm-wide
costal fields and 70–100-mm-wide intercostal fields (fig. 10a).
The costal epidermal cells are elongate, subrectangular, sub-
rhomboidal, rarely polygonal, 45–90 mm long and 10–15 mm
wide (typically ;70 312 mm), aligned parallel to the line of
the costal fields. The intercostal cells have less thickly cutin-
ized anticlinal walls but in places can be seen to be more isodi-
ametric than the costal cells, polygonal to rounded, 20–35 mmin
size. Stomata occur in single files along the intercostal fields.
The abaxial cuticle shows mostly polygonal, isodiametric,
or only slightly elongate cells, 25–50 mminsize(typically
;40 mm) and with generally weakly cutinized anticlinal
walls (fig. 10b). However, there are also some areas of much
more elongate, subrectangular to subrhomboidal cells, 40–
90 mm long and 8–25 mm wide (typically ;65 315 mm), and
677
CLEAL ET AL.—THE PENNSYLVANIAN SEED PLANT EREMOPTERIS

thickly cutinized anticlinal walls. The relationship between
the fields with these two cell types varies; in some places, the
elongate cells predominate; in others, they are restricted to
narrow bands.
Thickly cutinized papillae are mainly concentrated on the
areas with elongate cells, but some also occur in the areas with
polygonal, weakly cutinized walls (fig. 10b). The papillae are mostly
25–30 mm in diameter, often apparently with an oval trans-
Fig. 9 Cuticles from laminate parts of Eremopteris artemisiaefolia leaves. a, Adaxial cuticle, showing costal and intercostal fields. BMNH
V.15867$1, 3100. b, Adaxial cuticle showing more prominent papillae. BMNH V.15867$2, 3100. c, Close-up of adaxial intercostal field showing
nonpapillate stoma and papillate structure and also adhering Florinites pollen (contaminant). BMNH V.15867$1, 3400. d, Close-up of small
papillae on adaxial surface. BMNH V.15867$2, 3250. e, Abaxial cuticle showing two costal and two intercostal fields. BMNH V.15867$2, 3100.
f, Abaxial intercostal field showing papillate stomata. BMNH V.15867$2, 3250.
678 INTERNATIONAL JOURNAL OF PLANT SCIENCES

verse section, but this may have been due to taphonomic com-
pression. Most, where preserved side-on, are short and blunt
with a rounded apex, ;20–35 mm high. However, there also
appear to be some that were originally more elongate but that
are now broken. Where the papillae occur in areas of polygonal
cells, they occur on a narrow ridge of more thickly cutinized
cells, one cell wide and two to three cells long. There are typi-
cally 35–40 papillae/mm
2
.
Fig. 10 a–c, Cuticles from rachial parts of Eremopteris artemisiaefolia leaves. a, Adaxial cuticle showing nonpapillate stomata. BMNH
V.1714$1, 3100. b, Abaxial cuticle showing papillae. BMNH V.15867$4, 3100. c, Scars on abaxial surface of rachial parts. BMNH V.15867$3,
325. d, Cuticle from ‘‘anomalous’’ leaf showing faint cell structure. BMNH V.15867$2, 325. e, Cuticle from surface of stem shown in fig. 8. Slide
G73.47[HM1(2)], 3100. f, Cuticle from proximal end of petiole stub attached to stem shown in fig. 8. Slide G73.47[HM1(1)], 3100. g, Cuticle
from distal part of petiole stub attached to stem shown in fig. 8. Slide G73.47[HM2], 3100.
679
CLEAL ET AL.—THE PENNSYLVANIAN SEED PLANT EREMOPTERIS

Stomata occur in single or rarely double files in 70–100-mm-
wide discontinuous bands in the areas with mainly isodiametric
epidermal cells. The stomata have no evidence of subsidiary
cells, but there are, in some cases, weakly developed papillae
partially overarching the pore. The guard cells are 25–40 mm
long and 5 mm wide, thickly cutinized with prominent lips. The
distribution of the stomata is very irregular; in some places, the
stomatal density may be up to 125/mm
2
, but in others, quite
substantial areas are totally devoid of stomata.
Along the rachis, cuticles from the abaxial surface (i.e., cuti-
cles with the thickly cutinized papillae) in some specimens show
a number of oval structures, 500 mm long and 220–300 mmwide
(fig. 10c). They appear to be irregularly distributed but with their
long axes always parallel to each other, presumably longitudinally
along the rachis. They consist of a concentric ring of two rows of
thickly cutinized, blocky subrectangular cells, 25–35 mm in size.
Sometimes, there is virtually no cuticle inside this thickly cutin-
ized ring. In others, there is some cuticle inside the ring, but even
in the center of this, there is a hole 315 390 mminsize.Assug-
gested earlier in this article, these probably correspond to the
small oval structures visible on the surface of the main rachis in
certain specimens.
Petiole
Cuticles from both adaxial and abaxial sides of the petiole, be-
low the proximal-most segments, show epidermal structure that
is similar to that of the adaxial rachial surface; nothing equivalent
to the abaxial cuticles was found. Elongate to isodiametric subrect-
angular cells are arranged in longitudinal files along the petiole.
‘‘Anomalous’’ Leaf
One ‘‘anomalous’’ leaf (fig. 8g) yielded a thin, flexible cuticle
(fig. 10d). Faint evidence of cell pattern of the underlying epi-
dermis can be seen in places. The cells are up to 50 mmlong
and 10–20 mm wide, subrectangular to subrhomboidal, and ar-
ranged in rows along the axis. No other structures are visible.
Cuticle Biochemistry
Figure 11ashows the full absorbance spectra typically ob-
tained from the E. artemisiaefolia cuticle. There is a broad hy-
droxyl peak at 3355 cm
1
, with small peaks at 2922 and 2851,
indicating aliphatic carbon content. Small peaks at 869 and
827 also indicate the presence of aromatic carbon in the cuticle.
A number of other broad bands at 1595, 1340, and 1088 cm
1
also dominate the spectrum. The presence of shoulders and the
widths of these bands indicate a merging of bands.
Spectral enhancements were carried out using the deconvolu-
tion (fig. 11b) and second derivative (fig. 11c) functions of the
Spectrum 5.0 software used to operate the spectrometer. These
functions allowed a more detailed evaluation of the spectral
bands present, used to support the main spectral peak assign-
ments listed in table 2.
The FTIR data demonstrated that a range of organic chemi-
cal bonds is present in the sample. This includes aliphatic hy-
drocarbon, alkene groups, hydroxyl groups, carbonyl groups,
esters, and aromatic ring modes. The analysis in this study was
limited to a single cuticle and does not allow detailed com-
parison with previous studies looking at cuticle biochemistry.
However, FTIR analysis has shown that a range of organic mol-
ecules are present in the cuticle, and the FTIR spectral data ob-
tained are quite different from those obtained in other studies
on fossil plant remains (Lyons et al. 1994; Zodrow and Masta-
lerz 2002; Zodrow et al. 2003; Ps
ˇeni
cka et al. 2005).
Stem
There is just one known example of what we interpret as a
stem of Eremopteris (fig. 8a–8e). The axis is preserved for a length
of ;90 mm and is 9 mm wide in its proximal part, tapering
slightly to ;7 mm in its most distal part. It is essentially straight
for most of its length, but in its most distal part, it curves mark-
edly to the left.
Along the length of the axis are attached short, apparently
rigid protuberances at 30°–50°. They have a bulbous basal part,
2–3 mm wide, and a tapered distal part. In the more proximal
part of the axis, the protuberances are straight or slightly curved
basiscopically, but toward the distal end of the axis, they are mark-
edly curved acroscopically. The protuberances are best seen in pro-
file along the visible margins of the axis, but they can also be seen
flattened against the surface of the axis. From this, they can be
seen to have been originally in a helical configuration around
the stem. In some places on the stem surface are elliptical scars,
3 mm wide and 1 mm high. These protuberances are interpreted
as persistent leaf bases.
The remains of six or possibly seven 2-mm-wide petioles are
attached to the distal part of the stem. The more distal petioles
are attached at 20°–35°, the more proximal ones at 40°–60°.
One of the petioles has part of an elongate-ovate, twice-divided
Eremopteris leaf in organic attachment. The leaf (including pet-
iole) is preserved for a length of 90 mm and, from the way it ta-
pers distally, is almost but not quite complete. The secondary
segments are partly obscured by mineral growth but in some
cases can be seen to bear ultimate segments of the gracile type
discussed above (fig. 8d). A second but less complete (35 mm
preserved length) blade is attached to another of the petioles, and
this also has gracile-type ultimate segments. The other petioles/
main rachises attached to the distal end of the stem either are
broken or have no lamina still preserved.
Cuticles prepared from the surface of the stem (fig. 10e)and
the proximal part of the persistent petiole bases (fig. 10f)are
very similar, showing little evidence of cell structure. However,
along the length of the persistent leaf bases, the cuticles start to
show evidence of blocky cells (fig. 10f), and at the distal end of
the petiole bases, this blocky cell structure is markedly devel-
oped. This blocky cell structure is quite different from that seen
on the petioles of the detached leaves (fig. 10a), which consists
of more elongate and delicate longitudinally arranged cells.
Interpretation
The specimen that we have interpreted as part of the Ere-
mopteris stem was originally interpreted by Seward (1917) as a
rhizome. Seward gave no reason for this conclusion, but we be-
lieve that he may have been influenced by the fact that the axis
was shown horizontal in the illustration that he gave (we suspect
that he never actually saw the specimen and was dependent on
the drawing provided by L. D. Sayers). In this orientation, it is
easy to envisage the axis as being a creeping rhizome, with a se-
680 INTERNATIONAL JOURNAL OF PLANT SCIENCES

ries of leaves growing out of the distal end and curving upward.
However, an examination of the specimen shows this to be un-
likely, as there are clear, helically arranged leaf bases around the
axis. Furthermore, the distal part of the axis (fig. 8a) shows peti-
oles being emitted in a helical arrangement, and the apparent
presence of leaves on just one side of the axis is just due to poor
preservation. The arrangement of the leaves is far more compati-
ble with the axis being a short, upright stem than a horizontal
creeping rhizome.
Delevoryas and Taylor (1969) recognized that the leaves of
Eremopteris zamioides ‘‘dropped in their entirety,’’ were a little
distorted, and had an abrupt truncation at the base of the peti-
ole. Leaves of Eremopteris artemisiaefolia show the same char-
acteristics, in addition to an expanded base to the petiole. The
leaves of both species show evidence of having been abscissed.
The cauline specimen described in this article shows that the
leaves of E. artemisiaefolia were abscissed twice. The first ab-
scission occurred a short distance distal from the stem, leaving
a persistent stubby leaf base. These persistent leaf bases are
strongly reminiscent of those seen in many living cycads, where
the leaves are abscissed some distance away from the stem,
leaving a layer (‘‘armor’’) of protective tissue around the young
part of the stem. The leaf bases are subsequently abscissed in
cycads, and we have evidence of a second leaf base abscission
in Eremopteris leaving a rhomboidal scar on the stem (fig. 8e).
Cuticle preparation from the distal end of the persistent leaf
base (fig. 10g) shows thick-walled, blocky cells very different in
appearance to the cell structure near the base of the abscissed
petiole (fig. 10a,10b). The blocky cellular structure of the per-
sistent leaf base could be interpreted either as an indication of
an abrupt structural change, leading to the mechanical abscis-
sion of the leaf, or as a protective zone of cells sealing the apex
after abscission. There is no anatomical evidence of a dedicated
abscission layer, such as a band of transversely flattened cells to
facilitate the fracturing as seen in many extant and fossil plants
(Thomas and Cleal 1999). The high frequency of seeds and de-
tached leaves occurring in close proximity seen in collections of
both species of Eremopteris might indicate that leaf abscission
occurred shortly after seed dispersal had taken place.
The leaves attached to this Eremopteris stem are small. This
could be either because they were young leaves or because they
were leaves of a young plant. However, they are rather different
from the other leaves that we have interpreted as juvenile (fig.
8f,8g). Moreover, they are not attached to the distalmost part
of the stem, as would be expected if they were juvenile. Instead,
we believe that this was part of a relatively young, presumably
immature plant. However, it was clearly not a seedling, as there
are so many leaf bases preserved along the length of the seed. If
there is an approximate correlation between leaf size and the
size of the whole plant, then the fully grown plant may have
had a stem 20–25 mm wide and 0.2 m high.
Ovules
Brongniart (1830) reported several ovoid bicornate ‘‘fruits’’
together with his specimens of Sphenopteris artemisiaefolia but
regarded the association as accidental. Duns (1872) is often
credited as the first to suggest that the ovules and leaves were
originally parts of the same plant (Seward 1917; Crookall 1970),
although he was discussing fossils from the Lower Carboniferous
Burdiehouse Limestone that he had misidentified as S. artemi-
siaefolia. Howse (1890) appears to have been the first person
to make the connection between the Late Carboniferous Brit-
ish leaves and the ovoid bodies (although he interpreted them
as spore capsules). The correlation has also been noted in North
American specimens of Eremopteris zamioides (Delevoryas and
Taylor 1969), and the similarity between these two sets of ma-
terials with eremopterid foliage and horned ovules is striking.
In the material that we have studied, the association was found
in at least two localities, and at those localities, no other leaves
were present: those labeled as ‘‘Newcastle’’ and ‘‘Carluke.’’ Al-
though organic connection between the leaves and the ovules is
still to be found, in our view, the consistent evidence of associa-
tion is compelling that they were parts of the same parent plants.
Morphology
The ovule morphology has been determined from numerous
specimens preserved in close association with leaves of Eremop-
teris. The ovules, which are consistent in their morphology and
clearly belong to a single fossil species, have a 180°rotational
symmetry and are cordate with a small pedicel and two in-
wardly recurved apical integumentary extensions (figs. 12,
13a). Individual ovules vary from 7.2 to 7.7 mm long, are 5.3–
6.1 mm wide, and attain their maximum width approximately
halfway along the ovule length. The apical extensions vary in
length from 0.85 to 1.0 mm, are prominently tapered, and are
typically 0.8–1.1 mm apart. The area between the apical exten-
sions forms a U-shaped recess in which a narrow micropylar
opening can be seen in some specimens (fig. 12b,Mi).
Ovules comprise two distinct parts when seen as adpressions:
(1) a thin outer zone that includes the pedicel and apical exten-
sions and that is typically lighter colored and (2) an inner ovate
part that is darker in color and has a prominent centrally posi-
tioned longitudinal rib (fig. 13a,R). In some cases, the inner
part has become partially (fig. 12a, left) or completely (fig. 12a,
right) detached from the outer, and the central part of complete
specimens, in many cases, exhibits a darkened zone of tissue to-
ward the chalaza (fig. 13a,NC).Whereboththeinnerand
outer zones are present, the outer zone is contiguous with the
margin of the inner, is of more or less even width around the
ovule, and typically ranges from 0.3 to 0.5 mm wide. This is al-
ways less well preserved than the inner zone (e.g., fig. 13a)and
typically has an irregular surface that follows the surface of the
underlying sediment. We interpret this outer part of the ovule
as being a thin integumentary wing positioned in the major
plane of the ovule. The apical extensions are therefore part of
the integumentary wing.
A centrally positioned longitudinal rib extends over the entire
length of the inner zone but does not appear to extend into the
pedicel. This rib is always higher in sediment than the sur-
rounding ovule tissues, although in several instances some of
the tissues of the rib have become detached (e.g., ovule to right
in fig. 12a).
In some of the better-preserved specimens, variation can be
seen in the composition of the integumentary wing in which nar-
row (;0.03–0.04 mm wide) darker-colored lines are visible, ly-
ing parallel to the outside of the ovule (fig. 13a,?VB). In the
ovule shown in figure 13a, these faint lines run from the pedicel
681
CLEAL ET AL.—THE PENNSYLVANIAN SEED PLANT EREMOPTERIS

toward the base of the nucellus, diverge at the chalaza in the
major plane, and then pass into the integumentary wing. In
the integumentary wing, they then pass through the center of
the wing’s width, where they pass distally through the wing and
extend for at least two-thirds of the ovule length. These are present
on both sides of the ovule (fig. 13a) and appear to be positioned
more or less symmetrically to either side of the major plane.
Ovule Cuticles
Features of the inner part of the ovule are best characterized
from isolated cuticle preparations that preserve cellular details
of the ovular cuticles. In each case, the macerated remains are
smaller than the premacerated specimen; superimposing macer-
ations onto the premacerated specimens (fig. 13a–13c)reveals
that the outer parts of the ovule, including the pedicel, integu-
mentary wing, and apical integumentary extensions, do not
have cuticular preservation. The remaining cuticle consists of
three distinct layers: (1) an outermost thin integumentary cuti-
cle; (2) a middle layer that corresponds to the nucellar cuticle,
with a nucellar apex comprising a pollen chamber and nucellar
beak; and (3) the innermost layer, the seed megaspore. The size
and shape of the outermost of these layers (the integumentary
cuticle) correspond to the margin of the inner dark-colored unit
seen in the unmacerated specimens (figs. 12a,13a); this is
shown for a single ovule, pre- and postmaceration, in figure
13a–13c. As this integumentary cuticle occurs inside the posi-
tion of the integumentary wing (see above), it represents an in-
ner integumentary cuticle. It is uncertain whether this species
possessed an outer integumentary cuticle. Of the eight ovule cu-
ticles prepared, six are complete, while two have lost the mega-
spore and retain only the integumentary cuticle and vestiges of
the nucellus at the chalaza (e.g., fig. 13g).
Integumentary cuticle. The inner integumentary cuticle is
thin and typically orange to yellow orange. When complete
(fig. 13c–13f) it is cordate, 5.3–6.2 mm long, and with a maxi-
mum width of 3.8–4.7 mm. It has an irregular distal apex (fig.
13b-13f) that forms the inner part of the micropyle and lines
up with the micropylar opening observed between the apical in-
tegumentary extensions. The distal extension of the integumen-
tary cuticle is 0.2–0.35 mm wide and typically 0.5–0.9 mm
long.
In the cuticular preparations, a prominent ring-shaped struc-
ture occurs at the chalaza. This corresponds to a darkened area
of tissue observed in the premacerated specimens (fig. 13a,13c,
13d,13f,13g,NC). In most of the preparations, the ring-
shaped structure is preserved oriented perpendicular to the
plane of compression (e.g., fig. 14a–14c), from which it is un-
certain whether this structure is part of the inner integumentary
cuticle, the underlying nucellus, or both. Fortunately, a single
specimen has its chalazal region more obliquely preserved and
shows an oval-circular gap at the base of the integumentary cu-
ticle through which the base of the nucellus is visible (fig. 14d).
In this specimen, the inner integumentary cuticle ends as it comes
into connection with the ring-shaped structure, resulting in a
circular chalazal gap in the inner integumentary cuticle. Vari-
ation in this structure comes through the relative position of the
nucellus and megaspore (see below) with respect to the integu-
mentary opening; in some cases, the megaspore protrudes into
the chalazal region (e.g., fig. 14a,14b), whereas in other speci-
mens, the base of the nucellus and megaspore coincides with
the ring-shaped structure (fig. 14c).
Distal from this ring-shaped structure, a darkened zone arises
within the cuticle that corresponds with the longitudinal rib
noted on the compressions (fig. 13a). As this feature passes
beyond the apex of the nucellus, it is clearly part of the inner
integumentary cuticle and not a feature of the nucellus seen
through the semitransparent integumentary cuticle.
Nucellar cuticle. The nucellar cuticle is free from the integ-
umentary cuticle, except where it is attached to the ring-shaped
structure, which we interpret as a vascular nucellar pad (fig. 14;
see below). This is best seen in an incomplete specimen that
lacks the megaspore but that shows the nucellus as a folded
mass free from the inner integumentary cuticle (fig. 14e). The
nucellar cuticle is generally cordate in shape and has an outline
that follows the outer margin of the inner integumentary cuticle
(fig. 13c–13e) but in one case is considerably narrower (fig. 13f)
and more elongate. Typically, the nucellus is adnate to the mega-
spore, with the megaspore filling the nucellar cavity (fig. 13c–
13e), although the specimen with a narrower nucellus (fig. 13f)
also has a smaller megaspore and shows the nucellar cells to be
subrectangular, ;54 mm long and 22 mmwide(figs.13f,14a).
The apex of the nucellar cuticle comprises a proximal pollen
chamber and distally from it a short tubular nucellar beak (fig.
15). The pollen chamber varies considerably from compara-
tively narrow, tall, and domed, ;1.2 mm wide at the base and
0.65 mm in maximum height (e.g., figs. 13f,15a), to compara-
tively short and wide, typically having a rounded apex, ;1.6–
2.0 mm wide at the base and 0.4–0.7 mm high (e.g., fig. 13e;
fig. 15b,15d), to one that is intermediate between these two ex-
tremes (fig. 13c). The nucellar beak is located centrally on the
Table 2
Summary of the Infrared Band Assignments for the
Eremopteris Cuticle Examined
Chemical bond Wave number cm
1
assignments
OH 3350, 1415, 1346, 656
NH 3300, 1622
CH 1393
CH
3
2960, 2870, 1358
CH
2
2921, 2852, 718
C555O 1774, 1751, 1741, 1717, 1659
C555C 1735, 1642, 1614, 641
CO 1122, 1046
CCO 1227, 1122, 1089, 1023, 822
CN 1250, 742
Aromatic ring modes 1603, 1582, 1574, 695, 685
Fig. 11 a, Typical attenuated total reflectance–derived spectrum for Eremopteris cuticle (BMNH V.15867). b, Improved spectral resolution
using deconvolution to reduce the width of individual components. c, Further enhancement of the band resolution using the derivative function
showing a section of the second derivative spectrum. Band position is now marked by the downward-pointing peaks.
683
CLEAL ET AL.—THE PENNSYLVANIAN SEED PLANT EREMOPTERIS

top of the pollen chamber and consistently has a cuticle that is
darker colored than the rest of the nucellus. The nucellar beak
ranges from 300 to 370 mm long, gently tapers distally, and has
a distal opening that varies from 63 to 73 mm in diameter (fig.
15a–15d). The apex of the nucellar beak is, in most cases,
aligned with the micropylar tube (fig. 15).
Seed megaspore. The megaspore is dark brown to orangey
brown and is, in most cases, adnate to the nucellus such that
the two distinct layers are not easily distinguished (e.g., fig.
13c–13e). However, in one case, the megaspore is considerably
smaller than the nucellus (figs. 13f,14a) while in other speci-
mens the megaspore can be seen in narrow zones at the margins
of the nucellus (fig. 13d,13e). The megaspore is generally ovate
to obovate, varies considerably in size from 2.3 mm long and
1.6 mm wide to 4.1 mm long and 3.3 mm wide (cf. fig. 13c–
13f), and has a rounded base and a flattened apex beneath the
pollen chamber. In two specimens, the apex of the megaspore is
more irregular and appears to be centrally domed (fig. 15a,ar-
rows), but, as it is covered by the nucellus and integumentary
cuticle, this is difficult to characterize accurately.
Interpretation
Arber (1914) established the generic name Cornucarpus for
these ovules, noting that, although they are regularly associated
with Eremopteris,nonehavebeenfoundattached.Tidwell
(1967) subsequently emended the diagnosis so that the fossil ge-
nus could encompass all ovules with prominent horn- or whip-
like apical projections. However, the relationship between the
type species of Cornucarpus (Cornucarpus acutus [Lindley and
Hutton] Arber) and Eremopteris is now so firmly established
that this widening of the circumscription of the genus seems un-
warranted.
Cornucarpus is platyspermic, with 180°rotational symmetry
(Rothwell 1986). We interpret the longitudinal rib seen in the com-
pressions as representing the position of a commissural rib in the
minor plane of the ovule, as seen in some permineralized cardio-
carpalean taxa such as Mitrospermum Arber (Taylor and Stewart
1964; Hilton et al. 2003). The coincidental darkened zone seen in
the inner integumentary cuticle may also mark the position of a
commissural rib or may represent the position of a vascular bundle
in the minor plane of the ovule, which arises from the nucellar vas-
cular pad and passes distally to the micropyle.
We are uncertain of the origin of the darker-colored lines some-
times visible on the integumentary wings. They may merely rep-
resent folds in the soft tissue of the wing. In our view, however,
they are more likely to represent the position of vascular bundles
in the integument. Similarly positioned vascular bundles are known
in the integumentary wing in several permineralized Paleozoic
cardiocarpalean taxa, including species of Mitrospermum Arber
and Cardiocarpus Brongniart (Taylor and Stewart 1964; Hilton
et al. 2003).
As the ring-shaped gap in the inner integument is distal from
the pedicel and integumentary wing in the unmacerated speci-
mens, it cannot be an attachment/abscission scar as it does not
represent the margin of the dispersed ovule. We instead interpret
it to be the base of the integumentary cuticle that marks the
point below which the nucellus and integument were congenitally
fused. As this feature is ring shaped, we interpret it to mark the
position of a vascular nucellar pad at the base of the nucellus.
This interpretation is consistent with the size and shape of vas-
cular nucellar pads in a number of permineralized cardiocarpa-
lean ovules (see Hilton et al. 2003 for summary), and, as such,
this interpretation is consistent with known features from com-
parable taxa.
The macerated specimens show no evidence of a central column
within the pollen chamber. However, in other Paleozoic pterido-
sperm species that have a central column, it consists of parenchyma-
tous tissue, which would not survive maceration for cuticles.
The megaspore appears to be unexpanded, and there is no
evidence of pollen grains within the pollen chamber. From this
we conclude that the ovules studied had not been pollinated be-
fore they were shed.
Comparisons
These ovules share features with a number of cardiocarpalean
ovule taxa, many of which are of unknown affinity. They are
closest in morphology to the Cornucarpus ovules reported by
Delevoryas and Taylor (1969) in close association with Eremop-
teris zamioides. Ovules of both species have a prominent pedicel
Fig. 12 Ovules of Eremopteris artemisiaefolia before maceration. a,
Two ovules showing their cordate shape and paired apical integumen-
tary extensions. The ovule to the left has a lighter-colored outer area
that represents an integumentary wing (W) and darker central area that
represents the nucellus (N). This ovule is shown macerated in fig. 13d.The
specimen to the right displays a prominent centrally located longitudinal
rib (R), but the nucellar cuticle has been lost, leaving only tissues of the
integumentary wing (W) preserved. Scale bar in 1-mm divisions. b,En-
largement of micropylar region (Mi), with micropylar canal and apical
integumentary extensions shown in a. Scale bar ¼0.5 mm.
684 INTERNATIONAL JOURNAL OF PLANT SCIENCES

and possess a centrally positioned longitudinal rib, which Dele-
voryas and Taylor interpreted as representing the position of a
vascular bundle or a commissure dividing the integument into
two valves. However, the ovules associated with E. zamioides
are consistently larger, varying from 10 to 15 mm long and
from 4 to 6 mm wide, and have more prominent apical integu-
mentary extensions. Although the absence of cuticular evidence
from the North American specimens hinders further compari-
sons, there can be little doubt that the two plants are taxonomi-
cally distinct and that we are dealing with the remains of two
quite separate plant species, albeit probably of the same genus.
As noted previously, there is also some comparison with the
anatomically preserved ovules known as Mitrospermum.How-
ever, all species of Mitrospermum have multiple integumentary
bundles in the integument, lack apical integumentary projec-
tions, and form a micropylar canal that extends to the apex of
the integument (the Mississippian ovule Mitrospermum bulbo-
sum Long lacks these features and was thus excluded from the
genus by Hilton et al. [2003]). Furthermore, true Mitrospermum
is repeatedly associated with the cordaitean coniferophyte cones
Gothania and stems Mesoxylon (Trivett and Rothwell 1985,
1988, 1991; Hilton et al. 2003). The large strap-shaped leaves
organized in a helical phylotaxy normally associated with cordai-
teans are very different from the compound Eremopteris leaves.
Cornucarpus compares with a number of other adpression
ovule species with bicornute integumentary apices, including
Cardiocarpus bicornutus Lesquereux (Lesquereux 1880), Cor-
nucarpus minutus Halle, Cornucarpus incurvus Halle, Cornu-
carpus apertus Halle, and Samaropsis sinensis Halle (Halle
1927). However, in each case, these species lack sufficient de-
tails to enable accurate comparisons, and their cuticle structure
is unknown. None of these species are of known affinity, and
none are associated with eremopterid foliage.
In terms of its external morphology, Cornucarpus resembles
the platyspermic Mississippian ovule Lyrasperma scotica Long
(Long 1960). While the shape of the integument is similar, Ly-
rasperma has a weakly developed commisure that separates the
integument into two valves rather than a commissural rib as
noted in Cornucarpus.Furthermore,Lyrasperma is a hydra-
sperman pteridosperm that has an exposed nucellar apex with
large salpinx that distally protrudes through the integument,
very different from the small nucellar apex positioned within
the integumentary cavity in Cornucarpus.
Ovule Attachment
Despite the frequency with which the detached Cornucarpus
ovules are associated with Eremopteris leaves, no example of
attached ovules has been reported. Delevoryas and Taylor
(1969) suggested that they were attached to pinnate structures
(‘‘megasporophylls’’), sometimes found associated with eremop-
terid leaves. However, no ovules were found attached to these
structures, and there is no evidence of attachment scars; as noted
in this article, we believe that these pinnate structures are parts
of young and underdeveloped leaves.
We instead suggest that the ovules were probably attached to
the small oval structures observed both on the compression sur-
face and in the cuticles of the main rachis and proximal parts of
the secondary rachises (fig. 10c). This is based on two observa-
tions. First, the sizes of the attachment area at the chalazal end of
the detached ovules and of the oval structures on the leaf rachises
are very similar. Second, there is a good correlation between a
leaf having ovules associated with it on the same block and these
oval structures on the leaf. No example was found in which the
oval structures were present, but no ovules were preserved. Just
two specimens (NHM V.9093 and V.34125) have associated
ovules but no evidence of the oval structures; however, in both
cases, the surface of the fossil is poorly preserved, and the oval
structures could well have been obliterated during fossilization.
We can find no convincing alternative interpretation for these
oval structures. There is a superficial resemblance to fungal wound
reaction, reported in certain medullosalean fronds (Zodrow
and Cleal 1998). However, such ‘‘cork warts’’ usually have a
more circular outline and have much thicker cuticle in the cen-
ter of the structure. Insect damage is also, in our view, unlikely,
partly for the same reasons and partly because of the relatively
consistent distribution of the scars along the main rachis. It is
perhaps noteworthy that very little evidence of surface damage
attributable to insects has been reported in other Carboniferous
leaves. Glandular structures, such as the resin bodies reported
in callistophyte foliage (Rothwell 1975), also tend to have a more
circular basal outline, and it seems unlikely that they would all
have been lost before fossilization.
If these oval structures are ovule attachment scars, then the
ovules were directly attached to the abaxial side of the leaf blade,
mainly along the main rachis or the proximal parts of the sec-
ondary rachises. As far as we can make out, the arrangement
is variable; in some cases, the oval structures are widely but reg-
ularly spaced (e.g., fig. 7d), but in others, they are more tightly
but irregularly clustered in one part of the leaf blade (e.g., fig.
7b). However, these scars can be difficult to recognize on the
compression surface, which may be hampering us in determining
the full details of their arrangement pattern.
Evidence for how other Pennsylvanian pteridosperms bore their
ovules is generally poor. The only group known to have had
ovules attached abaxially to the leaf blade lamina is the Callis-
tophytales (Rothwell 1980). At least one fossil genus of lyginop-
teridalean fronds (Palmatopteris) also bore ovules directly attached
to the pinnules but in a more marginal position (Corsin 1931).
Other members of the Lyginopteridales had the ovules attached
to trusses of terete axes (van Amerom 1975), but where these
trusses were borne on the plant is not known. The Medullosales
are widely thought to have also borne their ovules attached to
the fronds, either replacing terminal or lateral pinnules. How-
ever, in those medullosalean specimens that have been pre-
sented as supporting this interpretation (Buisine 1961, fig. 4;
Arnold 1935, 1937), organic attachment cannot be confirmed
(L. Seyfullah, personal communication, 2008). Where medul-
losalean ovules have been found in organic attachment, it is to
terete axes and arranged in trusses (Drinnan et al. 1990; Doubinger
et al. 1995, figs. 247–248; see also Cleal et al., forthcoming).
Pollen and Pollen-Bearing Organs
No pollen organs have been reported attached to Eremopte-
ris artemisiaefolia. However, associated with the leaves are a
few small (5–7 mm long and 2 mm wide) capsulate bodies with
an oblong to lanceolate outline and a short stalk or pedicel (fig.
16), which may represent pollen organs. Most occur as isolated
685
CLEAL ET AL.—THE PENNSYLVANIAN SEED PLANT EREMOPTERIS

Fig. 13 Ovules of Eremopteris artemisiaefolia before and after maceration. a–c, Different images of specimen V.34124. a, Ovule before
maceration with cordate-ovate shape, small pedicel (P), diminutive integumentary wing (W), apical integumentary extensions (E), prominent

bodies, but one (fig. 16c) shows what may be a cluster in lateral
view. Corsin (1928) interpreted stellate clusters of smaller lan-
ceolate objects associated with Eremopteris zamioides as possible
pollen organs. Delevoryas and Taylor (1969) reported similar
stellate structures associated with the Pennsylvania E. zamioides
leaves but interpreted them as probably the result of capsulate
pollen bodies having split open to release the pollen (they were
unable to extract any pollen grains from them). It is possible,
therefore, that the British capsulate bodies were preserved be-
fore dehiscence; it is noteworthy that the lanceolate structures
(e.g., fig. 16c) do not look as though they were ‘‘valves’’ produced
by the dehiscence of a capsulate body. However, the British speci-
mens also yielded no pollen grains, and so this interpretation
must, for the time being, remain speculative. It is clearly nec-
essary to find better-preserved examples before we can finally
establish the form of the pollen organs of the Eremopteris-
bearing plant.
We similarly found no pollen in the pollen chambers of the
ovules associated with the Eremopteris leaves. All that we can
say about the pollen, therefore, is that the grains were presum-
ably no more than ;60 mm in diameter for them to pass through
the distal opening of the nucellar beak of the ovule.
Distribution
The best-documented occurrences of this species are from the
Northumberland and Durham Coalfield in northeastern En-
gland. The following records are in ascending stratigraphical se-
quence based on stratigraphical data given by Land (1974):
Northumberland Low Main Seam (lower Duckmantian, up-
per Modiolaris Chronozone), Cramlington, 12 km north of
Newcastle-upon-Tyne (Kidston 1922).
Bensham Seam (also known as Maudlin or Stone Seam, lower
Duckmantian Substage, base of Lower Similis-Pulchra Chrono-
zone), Jarrow Colliery, Jarrow, 8 km east of Gateshead (Kidston
1922); Springwell Colliery, ;5 km southeast of Gateshead (Kid-
ston 1922); Chopwell Colliery, 4 km northeast of Ebchester
(Kidston 1922).
High Main Seam (middle Duckmantian Substage, middle Lower
Similis-Pulchra Chronozone), Fawdon, Newcastle-upon-Tyne
(von Sternberg 1825), and Gosforth, 3 km north of Newcastle-
upon-Tyne (Kidston 1922).
There are also records from Scotland, again in ascending strati-
graphical order:
Kiltongue Seam (uppermost Langsettian Substage, middle Modi-
olaris Chronozone), Souterton, near Coatbridge, Lanarkshire
(Kidston 1901).
Stranger Seam (lower Duckmantian, upper Modiolaris Chro-
nozone), Grange Colliery, Kilmarnock (Kidston 1893).
Main Seam (lower Duckmantian, lower Lower Similis–Pulchra
Chronozone), Blinkbonny Pit, Rowanburn, Canonbie Coalfield
(Kidston 1903).
Chemiss Seam (middle Duckmantian Substage, middle Lower
Similis–Pulchra Chronozone), Leven and Durie Collieries, Leven,
Fife Coalfield (Kidston 1924).
Main Seam (upper Duckmantian Substage, upper Lower
Similis–Pulchra Chronozone), pit near Carmyle, ;1.5 km north
of Cambuslang, Lanarkshire (Kidston 1924).
Ell Seam (upper Duckmantian Substage, upper Lower Similis–
Pulchra Chronozone), pit ;1 km east of Baillieston, Lanarkshire
(Kidston 1924).
Although Kidston’s records quoted above were often unil-
lustrated, they were based mainly on specimens that can be ex-
amined in the collections of the British Geological Survey,
Keyworth, Nottingham, and a number were illustrated by
Kidston (1924).
There are also unillustrated records from southern Britain,
such as from the South Wales (Kidston 1894), North Stafford-
shire (Kidston 1891), and South Staffordshire (Kidston 1914)
coalfields, but these are based on tiny fragments and must be
treated with considerable doubt. On the basis of observations
on specimens in the Sedgwick Museum, the records from Kent
(Arber 1909; Kidston 1919) appear to refer to specimens of
Eremopteris zamioides (it is notable that Kidston [1924] did
not include these Kent records in his inventory of British Ere-
mopteris artemisiaefolia specimens, although neither did he in-
clude them under E. zamioides).
To date, there have been no reliable European records of this
species from outside of northern Britain. Sauveur (1848, pl. 20,
figs. 1, 2) figured specimens from the Westphalian Stage of Bel-
gium, together with a copy of part of Brongniart’s (1830, pl.
46) British specimen. Kidston (1886, p. 65) and Gothan (1913,
p. 245) both argue that these Belgian specimens in fact belong
to the fern Urnatopteris tenella (Brongniart) Kidston. Gothan
also dismissed the specimens that he had seen in German mu-
seums identified as this species from the Waldenburg and Wet-
tin floras; in his view, this was almost certainly the result of
British specimens having been mislabeled as coming from here.
Boulay (1876, pl. 1, fig. 6) figured photographs of two speci-
mens from the Westphalian Stage of northern France as E. arte-
misiaefolia. However, they are small fragments that show none
of the diagnostic features of Eremopteris and may easily belong
to Eusphenopteris or even Karinopteris; one of the Boulay spec-
imens, for instance, shows what might be drip tips.
longitudinal central rib (R), micropylar area (Mi), and nucellar cup (NC). BMNH V.34124. b, Macerated cuticle superimposed on unmacerated
specimen showing loss of integumentary wing and outer integumentary cuticle through maceration. BMNH V.34124$1 (cuticles) superimposed
on BMNH V.34124 (compression). c, Macerated cuticle showing chalazal nucellar pad, position of central longitudinal rib, inner integumentary
cuticle (IC), megaspore (M), pollen chamber of the nucellar apex (PC), and a small tabular nucellar beak (Nb) extending from the end of the
pollen chamber toward the micropylar opening. BMNH V.34124$1. d, Macerated ovule cuticle showing features similar to those in cbut with
nucellus and pollen chamber floor (F) visible. BMNH V.15866$4. e, Small ovule with short, wide pollen chamber, prominent nucellus, and oval
gap in the inner integumentary cuticle at the chalaza (IO). BMNH V.52419$2. f, Macerated cuticle from cordate-shaped ovule displaying small
megaspore within a larger nucellus, tall and thin pollen chamber, prominent pollen chamber floor, and nucellar beak leading to the micropyle.
BMNH V.15866$1 þBMNH V.15866$2. g, Macerated ovule cuticle from fig. 12ain which only the chalazal region of the nucellus and nucellar
cup remain but with the integumentary cuticle and position of the longitudinal central rib observed. BMNH V.15867$1.
687
CLEAL ET AL.—THE PENNSYLVANIAN SEED PLANT EREMOPTERIS

Fig. 14 Macerated ovule cuticles from Eremopteris artemisiaefolia showing chalaza features; all scale bars ¼200 mm. a, Enlargement from fig.
13fshowing thin inner integumentary cuticle (IC), thick nucellus (N), and megaspore (M) with prominent chalazal bulge. BMNH V.15866$2. b,
Ovule with rounded megaspore protruding into area of the nucellar cup (NC) of the nucellus. BMNH V.15866$1. c, Enlargement of chalazal part
of ovule from fig. 13band 13cshowing ring-shaped nucellar cup with irregular cellular features at the junction of nucellus and megaspore. BMNH
V.34124$1. d, Chalaza enlarged from fig. 13fshowing obliquely flattened nucellar cup with prominent integumentary opening (IO). In this
specimen, the nucellus is fused basally to the megaspore and surrounded by thin inner integumentary cuticle. BMNH V.52419$1. e, Enlargement
from fig. 13dshowing cellular details of the nucellar cup surrounded by the inner integumentary cuticle, with vestigial remains of the nucellus after
the megaspore and the other parts of the nucellus have been lost. BMNH V.52419$2.

In North America, there are records mainly from the Appala-
chian Basin by Lesquereux (1880) and White (1943). The most
likely to be correct are by White (1943), as they are stratigraph-
ically compatible with the European records and there are accom-
panying drawings of the specimens: from the Mercer Group,
Blossburg, Pennsylvania, and (?) Lower Kanawha Formation,
roadside at mouth of Rock Fork of Bell Creek, Nicholas, West
Virginia.
Fig. 15 Macerated ovule cuticles from Eremopteris artemisiaefolia showing the micropylar region and nucellar apex; all scale bars ¼200 mm.
a, Enlargement from fig. 13fshowing cellular patterns on the inner integumentary cuticle (IC) and on the pollen chamber (PC). The position of the
micropyle is well defined and follows from the nucellar beak (Nb), and the apex of the megaspore appears to be domed. BMNH V.15866$1. b,
Enlargement from fig. 13cshowing wide and short pollen chamber with prominent nucellar beak, with the position of the micropyle following the
line of the commissural rib (R). In this specimen, the distal apex of the megaspore (M) is abruptly truncated and flat, corresponding with the
position of the pollen chamber floor. BMNH V.34124$1. c, Enlargement from fig. 13dshowing wide, domed pollen chamber (PC) with prominent
nucellar beak and with the distal apex of the megaspore being irregular and centrally domed (arrow). BMNH V.15866$4. d, Ovule with
megaspore with irregular distal extent and with longitudinally elongated cells on the pollen chamber. BMNH V.15866$1. e, Cellular features of
the distal part of the inner integumentary cuticle enlarged from fig. 13fand at focal plane different from that in fig. 14aand with dark central line
of cells marking position of the commissural rib. BMNH V.15866$1.
689
CLEAL ET AL.—THE PENNSYLVANIAN SEED PLANT EREMOPTERIS

Both Lesquereux (1880; Cannelton, OH) and White
(1943; Canfield Cannel Coal, Mahoning, OH) record the
species from the Kittaning Coals in the upper Allegheny For-
mation. None are accompanied by illustrations, and they
are stratigraphically rather younger than the specimens
found in Europe. These records must therefore be treated with
some reservation. The records by Lesquereux (1880) from the
Hollenback Mines, Wilkes-Barre, Pennsylvania; the Hazlegreen
Coal, Morgan County, Kentucky; and the Helena Mines,
Alabama, also cannot be verified by illustration, and their strati-
graphical level is unclear. There are also records from the east-
ern Interior Coal Basin, from Indiana by Lesquereux (1884b,
pl. 15, fig. 5) and the Morrison Coal, Illinois, by White (1943,
unillustrated).
Fig. 16 Eremopteris artemisiaefolia with associated possible pollen-bearing organs. BMNH V. 15866, Coal Measures, Newcastle, United
Kingdom. a, Whole specimen. Natural size. b, Close-up of lanceolate body with pedicel, 36. c, Close-up of oblong body with pedicel and a possible
cluster of such bodies, 36.
690 INTERNATIONAL JOURNAL OF PLANT SCIENCES

Discussion
Reconstruction
Figure 17 shows our interpretation of what an Eremopteris-
bearing stem may have looked like. It is based largely on the
specimen shown in figure 8a–8e. We have assumed that this
specimen represented most of a complete plant rather than a
branch of a plant with a more complex architecture; there is no
evidence of axillary branching, which might have produced a
more complex architecture. We have argued that this specimen
may represent a relatively young plant and that a fully grown
individual may have been larger. However, in view of the rela-
tively small size of even the largest leaves, a plant of such archi-
tecture is unlikely to have been large. If there is an approximate
correlation between leaf size and the overall size of the plant,
we estimate that a fully grown plant was probably somewhere
between 0.2 and 0.3 m high. The only alternative interpretation
would be that this is part of a much larger plant with a branch-
ing stem, but there is no evidence to support such a view.
Phylogenetic Context
Considering the combination of morphological and anatomi-
cal features from the cuticle present in the Eremopteris-bearing
plant, its phylogenetic position within seed plants is difficult to
envisage (in the remainder of this subsection, we shall refer to
this plant using the shorthand term ‘‘Eremopteris’’ ). Eremopte-
ris has not been included in previous morphological cladistic
analyses of either lignophytes (Nixon et al. 1994; Rothwell and
Serbet 1994; Hilton and Bateman 2006) or seed plants (Doyle
1996, 2006), and as such its affinities remain enigmatic. To try
to resolve its phylogentic position, we have added our recon-
struction of Eremopteris to the lignophyte cladistic analysis of
Hilton and Bateman (2006), working on the assumption that
the bicornute ovules were produced by the same plant. We have
chosen this matrix as it is one of the more recent analyses avail-
able and, in comparison with other available matrices, includes
the most comprehensive selection of pteridosperms. It thus pre-
sents the best of the currently available analyses for determining
relationships among basal members of the seed plant lineage.
As characterized here, the Eremopteris-bearing plant is known
from vegetative and ovulate morphology and cuticle, but its cau-
line anatomy is unknown. As discussed earlier, there is some evi-
dence regarding the pollen-bearing organs, but they are not well
characterized and so have been excluded from the analysis. In
this matrix, Eremopteris scores for 30 out of a possible 102 char-
acters (30%), although of the total number of characters, 12 are
not applicable to fossils, and a further 12 relate only to angio-
sperms; thus, Eremopteris scores for 30 out of 78 applicable
characters (38%). This level of known data is comparable to the
level in taxa that Hilton and Bateman (2006) termed ‘‘wildcards’’
and added sequentially to the core taxa in their analysis. Hilton
and Bateman (2006) considered this approach likely to infer the
position and relationships of the wildcard taxa and in no in-
stance did this approach destabilize the previously determined
topology based on the core taxa alone. We use the same protocol
here and consider Eremopteris tobeawildcardinthisanalysis.
However, we do not advocate including it in subsequent analyses
until details of stem and reproductive anatomy are known
for the genus. Individual scores for Eremopteris are shown in
table 3.
In this account, we have used the same search strategy as em-
ployed by Hilton and Bateman and have restricted this analysis
to the 48 core taxa of their analysis (i.e., excluding their wild-
card taxa). The only difference is that this analysis was spawned
from Winclada (Nixon 2002) into NONA, version 2 (Goloboff
1999), using the heuristics search option with maximum trees
to keep (hold) set to 10,000 and number of replications (mult*N)
of 100, using the multiple TBR þTBR (mult*max) option.
The revised matrix generated nine most parsimonious 314-
step trees from a single island of trees with a consistency index
of 0.45 and a retention index of 0.80 (fig. 18), a single step lon-
ger than the core analysis of Hilton and Bateman (2006). Un-
ambiguous character changes on most parsimonious trees for
the branches immediately above and below Eremopteris and
also those above Callistophyton are the same and are here mapped
onto the strict consensus tree (fig. 18).
The resulting topology is largely consistent with that gener-
ated by Hilton and Bateman (2006) but differs in the position of
the medullosaleans Medullosa and Quaestora, which in Hilton
and Bateman’s analysis form a clade between the hydrasperman
pteridosperms and the Callistophytales, whereas in this analysis,
this relationship is reduced to a hydrasperman þmedullosalean
polytomy (fig. 18). Eremopteris diverges from the stem at the
node above the hydrasperman and medullosalean polytomy and
immediately below the callistophytalean Callistophyton.Inthis
arrangement (fig. 18), Eremopteris falls within a paraphyletic
grade of basalmost pteridosperms that also includes hydrasper-
man, medullosalean, and callistophytalean taxa.
However, Eremopteris shows a number of key differences
from most other members of this group of basal pteridosperms.
In Eremopteris, there is a clear organographic differentiation
between the reproductive organs with (1) ovules having been
borne directly on the rachises of fertile leaves, which are mor-
phologically indistinguishable from the vegetative leaves, and
(2) the pollen organs apparently having been borne on a dichot-
omous and terete branching system. In pteridosperms, this kind
of organographic separation is most similar to that of hydras-
perman taxa, in which male and female organs are also borne
on different kinds of branching structures (Hilton and Li 2003).
Medullosalean pteridosperms, in contrast, have radially sym-
metrical ovules borne terminally on dichotomous branching
systems and have pollen organs borne directly on the rachis
(Halle 1927, 1929). The Callistophytales have both ovules and
pollen organs embedded into the leaf lamina (Rothwell 1980).
Furthermore, if correctly interpreted, the young ‘‘abnormal’’ fronds
of Eremopteris display an ontogenetic pattern that is also distinct
from that seen in hydrasperman, medullosalean, and callistophyta-
lean pteridosperms, where young fronds form in a crosier, similar
to many modern-day ferns (Huth 1912; Crookall 1976; Rothwell
1980; Cleal and Laveine 1988).
Eremopteris is evidently distinct from the other previously
recognized pteridosperm groups and thus corroborates Ember-
ger’s (1968) conclusion that it belongs to a separate pterido-
sperm family. In the resultant topology (fig. 18), Eremopteris is
sister to both Callistophyton and all more derived seed plants,
from which we infer that it may be significant to the origin of
Callistophytales, but this placement does not support Eremop-
teris’s membership in that order as advocated by Meyen (1987)
691
CLEAL ET AL.—THE PENNSYLVANIAN SEED PLANT EREMOPTERIS

and Doweld (2001). In this position, Eremopteris also remains
significant to the origin of cycads, albeit perhaps more indi-
rectly than postulated by Delevoryas (1982) and Crane (1988).
The resulting topology places two platyspermic taxa immedi-
ately below cycads with ovule symmetry (character 61) chang-
ing from radial to bilateral below Eremopteris (fig. 18), thus
agreeing with Crane’s (1988) inferences on the plesiomorphic
ovule condition within cycads.
The addition of Eremopteris in this analysis has had the effect
of reducing the long branch leading to Callistophyton noted by
Hilton and Bateman. In their analysis, nine unequivocal charac-
ter changes occurred at the node from which Callistophyton
arose plus an additional character change on the branch leading
to Callistophyton (Hilton and Bateman 2006, fig. 12). In this
analysis, this changes to two character changes at the node
from which Callistophyton arises plus a further two changes on
Fig. 17 Proposed reconstruction of the external appearance of an Eremopteris-bearing stem represented by the specimen in fig. 8a–8e. This
may have been a young plant. We have no evidence of the internal anatomy apart from the epidermal anatomy obtained from the leaf and stem.
Drawing by Annette Townsend (Amgueddfa Cymru–National Museum Wales, Cardiff, UK).
692 INTERNATIONAL JOURNAL OF PLANT SCIENCES

the branch leading to Callistophyton.Eremopteris is therefore
important as it spans the transition between hydrasperman þ
medullosalean pteridosperms and Callistophyton. In the basal
part of the topology that we have generated, long branches are
evident in four parts of the tree: the first between progymno-
sperms and seed plants, the second between basal pteridosperms
(hydrasperman þmedullosalean þEremopteris þCallistophyton)
and cycads, the third immediately above cycads, and a fourth
that leads to Ginkgo. Our conclusion is that each of these rep-
resents areas in which there may be missing taxa; additional
whole plants in any of these parts of the tree could have pro-
found effects on the resultant topologies and will further pro-
gress our understanding of seed plant phylogeny.
Systematic Paleontology
Family—Cornucarpaceae Doweld, 2001, nov. emend.
Emended diagnosis. Monoaxial plants with a distal crown
of helically arranged compound leaves. Platyspermic ovules
were attached to the abaxial surface of the leaf rachises.
Typ e .Cornucarpus Arber, 1914.
Remarks. The family encompasses the fossil genera Ere-
mopteris (foliage) and Cornucarpus (ovules). We have not, at
this time, incorporated evidence of the putative pollen organs
described by Corsin (1928) and Delevoryas and Taylor (1969),
as their configuration and position of attachment to the plant
still have to be verified. As we have discussed earlier in the arti-
cle, there are now two well-documented cases in which Ere-
mopteris foliage and Cornucarpus ovules were borne by the
same plant. Moreover, we know of no other plant that had Ere-
mopteris foliage but bore ovules other than Cornucarpus (as de-
fined here) or vice versa. On the face of it, it might seem sensible
to unite these fossil genera, but we are reluctant to make such a
proposal until the organic connection between the ovules and
leaves has been verified.
The name Cornucarpaceae was predated by Eremopterida-
ceae Emberger (1968), but the latter lacked a formal diagnosis
and so was invalidly published. Doweld (2001) typified the
family by Cornucarpus Arber (1914), which had been estab-
lished specifically to accommodate the ovules found associated
with Eremopteris, and so coincides with the family as we envis-
age it. However, we disagree with one major point in the diag-
nosis given by Doweld, that the ovules are aggregated into
clusters; the evidence that we present suggests that they were
singly attached to the leaves. Nevertheless, the name Cornucar-
paceae was validly published according to ICBN articles 32 and
41 and, as the earliest available name for the family, must be al-
lowed to stand (ICBN article 11), albeit with an emended diag-
nosis. We also disagree with Doweld’s placing of the family
within the Callistophytales. The phylogenetic analysis discussed
earlier shows that the Cornucarpaceae (marked by the position
of Eremopteris) diverge from other seed plants after the Medul-
losales and before the Callistophytales, and in this context, the
family is distinct and does not belong to either the Medullosales
or the Callistophytales. In terms of traditional systematics, we
therefore regard the Cornucarpaceae as a satellite family (sensu
Thomas and Brack-Hanes 1984) within the spermatophytes
and a member of the paraphyletic plexus of seed plants infor-
mally circumscribed as pteridosperms.
Fossil Genus—Eremopteris Schimper, 1869, nov. emend.
Type.Eremopteris artemisiaefolia (Sternberg) Schimper,
designated by Gothan (in Potonie
´1921, p. 87).
Original diagnosis. Frons superne dichotoma pinnata, pin-
nis erecto-patentibus, irregulariter pinnatifidis, laciniis, obovato-
elongatis vel elongate-cuneatis, erecto-patentibus, inferioribus
lacinatis, superioribus subintegris. Nervatio Sphenopteridis vel
Neuropteridis (Schimper 1869, p. 416).
Emended diagnosis. Relatively small leaves, usually no more
than0.3mlong,withanelongate petiole and an irregularly di-
vided blade. The proximal-most part of the petiole is flared
where it was attached to a persistent petiole base. The blade con-
sists of a main rachis, which is straight or curved, and may be
undivided or show a single overtopped or occasionally dichoto-
mous division. Secondary foliar segments are more or less oppo-
sitely attached in the proximal part of the leaf blade, becoming
alternately arranged toward the blade apex; in smaller leaves,
the secondary segments are once divided, in larger leaves becom-
ing twice divided. Tertiary segments consist of an elongate lanceo-
late lamina, sometimes with one or two pairs of suboppositely
arranged basal lobes or subsegments. Dense veins run approx-
imately parallel along the ultimate segments and only rarely
fork. Both abaxial and adaxial surfaces of leaf blade show a
clear differentiation into costal and intercostal fields. Stomata
occur on both surfaces of the ultimate segments of the blade,
are oriented approximately parallel to the veins, and are not no-
ticeably sunken; they are densely distributed on the abaxial sur-
face and are surrounded by a ring of papillate subsidiary cells.
The adaxial stomata are much sparser and lack subsidiary cells
or surrounding papillae. Epidermal hairs are absent.
Remarks. Schimper (1969) included two species in the fossil
genus: E. artemisiaefolia (Sternberg) Schimper and Eremopteris
neesii (Go
¨ppert) Schimper (basionym Gleichenites neesii Go
¨p-
pert 1836, p. 183). Gothan (in Potonie
´1921, p. 87) unequivo-
Table 3
Individual Scoring for Our Hypothesized Eremopteris Plant Based on the Character Codes of Hilton and Bateman (2006)
1 2345678 91011121314151617181920212223242526
1 0010000/10000 ????????? ? ? ???
27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52
???????0 01000?0?????? ? ? ?0?
53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78
?0????21 1?3??0???110? ? ? ?0?
79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102
??????? ?????????????? ? ? ?
693
CLEAL ET AL.—THE PENNSYLVANIAN SEED PLANT EREMOPTERIS

cally designated Sternberg’s species as the type of Eremopteris.
The second species was transferred to Callipteris by Zeiller
(1898) but as noted by Kerp and Haubold (1988), the type (Go
¨p-
pert 1836, pl. 2, fig. 1) is a relatively small, bifurcate frond with
sphenopteroid pinnules. Kerp and Haubold (1988) tentatively
suggested a comparison with Odontopteris (Brongniart) Stern-
berg, but no species normally included today within that genus
would have such consistently lobed pinnules. A closer compari-
son is with the pinnule morphology of Arnhardtia mouretii (Zeil-
ler) Haubold and Kerp in Kerp and Haubold (1988; e.g., as
figured in Doubinger 1956, pl. 11, figs. 3, 4) although the frond
architecture of that species is unknown.
To date, 37 species have been assigned to Eremopteris (details
as summarized in table 1). However, most of these species were
included in Eremopteris in error because of the confusion over
the diagnoses given by Schimper (1869, 1879) and Lesquereux
(1880). White (1943) partly disentangled the problem by trans-
ferring a number of the species in which the leaf architecture
Fig. 18 Strict consensus tree in which the Eremopteris-bearing plant is positioned on the stem among a paraphyletic grade of Paleozoic
pteridosperms above a polytomy with hydrasperman þmedullosalean pteridosperms and below the callistophytalean pteridosperm Callistophyton.
Unambiguous character changes present in all nine most parsimonious trees on the branches leading to Eremopteris and Callistophyton and on the
branch above Callistophyton are here mapped onto the consensus topology that is well resolved in this part of the tree. Numbers relate to character
codes from Hilton and Bateman (2006).
694 INTERNATIONAL JOURNAL OF PLANT SCIENCES

had been determined to the lyginopteridalean fossil genera Mar-
iopteris Zeiller (1879) or Diplothmema Stur 1877 (some of the
latter species are now included either in Eusphenopteris [Novik
1947] or Palmatopteris [Potonie
´1893]). However, White
(1943) also complicated the issue by including a number of spe-
cies based on small fragments of leaf/frond for which the archi-
tecture is unknown; these species appear to be indeterminate or,
in some cases, fragments of Palmatopteris. A few Permian spe-
cies have been included by authors, but these appear to belong
either to peltasperms or, in one case, to a medullosalean. The
species from Gondwanan Carboniferous floras either belong to
Nothorhacopteris Archangelsky (1983) or ?Genselia Knaus
(1994) or do not show enough of the frond architecture to de-
termine their taxonomic position. When all of these species
have been reassessed, only four can be retained within Eremop-
teris as now defined: E. artemisiaefolia,Eremopteris zamioides,
Eremopteris stricta,andEremopteris neffii, with the latter two
being taxonomic synonyms of E. artemisiaefolia.
Eremopteris artemisiaefolia (Sternberg) Schimper 1869,
p. 416, nov. emend.
Basionym.Sphenopteris artemisiaefolia von Sternberg, 1825,
p. 15, pl. 54, fig. 1.
Taxonomic synonyms.Sphenopteris stricta von Sternberg,
1825,p.15,pl.56,fig.2;E. neffii White, 1943, p. 90, pl. 23,
figs. 1–5.
Original diagnosis. S. fronds pinnata, pinnis pinnatifidis,
laciniis cuneatis lobatis, lobis ovatis seu lanceolatis integris trifi-
disque (von Sternberg 1825, p. 15).
Emended diagnosis.Eremopteris leaves mostly up to 0.3 m
long and 0.15 m wide, with a blade ;0.25 m long. The ulti-
mate segments are lanceolate, up to 25 mm long and 7–8 mm
wide, obliquely attached to the secondary rachis at 25°–35°.
Vein density is the equivalent of 3.0–3.5/mm across the segment
width. Costal epidermal cells are elongate, subrectangular to
subrhomboidal, aligned parallel to veins; on the adaxial sur-
face, they are 30(57)120 mm long and 5(12)20 mm wide, and
on the abaxial surface, they are 40(88)150 mmlongand
9(15)30 mm wide. Intercostal cells are irregularly polygonal,
more or less isodiametric, or just slightly elongate parallel to
the veins and often have sinuous anticlinal walls; on the adaxial
surface, they are 30(55)100 mm long and 6(30)80 mm wide,
and on the abaxial surface, they tend to be a little smaller,
30(48)70 mm long and 20(30)50 mm wide. Epidermal cells of-
ten have prominent papilla. Abaxial stomatal density is usually
90(125–140)180/mm
2
; on the adaxial surface, the stomata are
so sparse that a meaningful density measurement is difficult to
give. The abaxial stomata are surrounded by six subsidiary
cells. The guard cells are 20(25)36 mmlongand;5mm wide.
Typ e . According to Kva
cek and Strakova
´(1997, p. 38), the
holotype of E. artemisiaefolia is lost. The illustration given by
von Sternberg (1825, pl. 54, fig. 1) must therefore be taken as
the effective type (refigured here in fig. 1).
Remarks. The type is a generally more ‘‘lush’’ leaf with
rather broader pinnule lobes than most leaves normally assigned
to E. artemisiaefolia. However, Sternberg figured another speci-
men from the same locality (under the name Sphenopteris stricta),
which is far more typical for this species, and there is little doubt
that they are conspecific.
Distribution. Duckmantian and topmost Langsettian sub-
stages of northern Britain (Northumberland-Durham and Scottish
coalfields) and the United States (West Virginia, Pennsylvania, Illi-
nois, Ohio, Kentucky, and Indiana).
Conclusions
1. Our investigation leads us to characterize the Eremopteris
artemisiaefolia–bearing plant as a pteridosperm with an upright
stem that bears a distal crown of helically arranged compound
leaves. Although not known in organic attachment, we inter-
pret isolated ovules of the Cornucarpus type closely associated
with the foliage to belong to the same plant and interpret ovate
scars on the abaxial surface of the leaf rachis to represent the
position of attachment. Anomalous leaves previously consid-
ered to represent mega- and microsporophylls are here reinter-
preted as young leaves.
2. As currently defined, the genus Eremopteris includes two
bona fide species: E. artemisiaefolia (Sternberg) Schimper and
Eremopteris zamioides (Bertrand) Kidston. Both species are
characterized by their vegetative foliage and, in each case, are
repeatedly associated with bicornute platyspermic ovules. Col-
lectively, the species of Eremopteris are placed in the family
Cornucarpaceae, which is here redefined to include features of
the parent plant rather than only the ovular characters as in its
original protologue.
3. The Eremopteris-bearing plant is only rarely represented
in the fossil record, and its habitat preferences are unknown.
Eremopteris artemisiaefolia has been found only in Duck-
mantian and topmost Langsettian (lower Moscovian) strata in
northern Britain and in North America in the Appalachians
and the Eastern Interior Coal Basins. Eremopteris zamioides is
known only from Duckmantian strata from central and south-
ern Britain, the Nord-Pas-de-Calais Basin in northern France,
and the Appalachian Basin.
4. Eremopteris artemisiaefolia as reconstructed has been in-
cluded in a cladistic analysis of lignophytes (progymnosperms
plus seed plants) as a wildcard species. It forms a distinct clade
within a paraphyletic complex of basal pteridosperms, thus sup-
porting its family-level systematic distinction. In this analysis,
E. artemisiaefolia diverges from the stem after the hydrasper-
man and medullosalean taxa and before the Callistophytales.
The organographically distinct nature of the male and female
branching systems is similar to that known in hydrasperman
pteridosperms and contrasts with the Callistophytales, in which
ovules and pollen organs are both embedded into the leaf lam-
ina. If correctly interpreted as immature fronds, the morphology
of the anomalous leaves presents an ontogenetic model, quite
different from that in hydrasperman and medullosalean pterido-
sperms, in which young fronds form a crosier, generally similar
to many modern-day ferns.
5. Addition of Eremopteris as a wildcard into a cladistic anal-
ysis of lignophytes reduces the previously noted long branch
leading to Callistophyton, thus spanning the phylogenetic space
between hydrasperman þmedullosalean pteridosperms and
Callistophyton. However, long branches remain evident in four
parts of the topology generated, namely: (1) between progym-
nosperms and seed plants, (2) between basal pteridosperms (hy-
drasperman þmedullosalean þEremopteris þCallistophyton)
695
CLEAL ET AL.—THE PENNSYLVANIAN SEED PLANT EREMOPTERIS

and cycads, (3) the branch immediately above cycads, and (4)
the branch leading to Ginkgo. We conclude that each long branch
represents an area in which missing taxa are evident in the topol-
ogy generated. This conclusion agrees with the obvious disparity
between the large numbers of pteridosperm taxa recognized in the
fossil record against the low numbers of available reconstructed
whole plants, which can be included in phylogenetic analyses.
Acknowledgments
We thank Steve McLean and Dan Pemberton for making
available the specimens of Eremopteris from the Hancock Mu-
seum (Newcastle-upon-Tyne) and the Sedgwick Museum (Cam-
bridge), respectively. We thank Bill DiMichele and James Doyle
for reviewing the manuscript.
Literature Cited
Adams GI, GH Girty, D White 1903 Stratigraphy and paleontology
of the Upper Carboniferous rocks of the Kansas section. Bull US
Geol Surv Ser Syst Geol Paleontol 211:1–123.
Andrews EB 1875 Descriptions of fossil plants from the Coal Mea-
sures of Ohio. Pages 415–426 in Report of the geological survey
of Ohio. Vol 2. Geology and palaeontology. Pt 2. Palaeontology.
Geological Survey of Ohio, Columbus.
Arber EAN 1909 On the fossil pants of the Waldershare and
Fredville Series of the Kent Coalfield. Q J Geol Soc Lond 65:21–40.
——— 1911 On the fossil flora of the southern portion of the
Yorkshire Coalfield in north Derbyshire and Nottinghamshire. Proc
Yorkshire Geol Soc 17:132–155.
——— 1914 A revision of the seed impressions of the British Coal
Measures. Ann Bot, NS, 109:81–108.
Archangelsky S 1958 Eremopteris golondrinensis nueva especie de la
Serie La Golondrina, Bajo de La Leona, Santa Cruz. Acta Geol
Lilloana 2:285–289.
——— 1981 Fedekurtzia, a new Carboniferous frond from Gond-
wanaland and its fructification. Am J Bot 68:1130–1138.
——— 1983 Nothorhacopteris, a new generic name for some Car-
boniferous monopinnate fronds of Gondwanaland (¼Rhacopteris
ovata auct. and Pseudorhacopteris Rigby 1973). Rev Palaeobot Paly-
nol 38:157–172.
Arnold CA 1935 Observations on Alethopteris grandifolia Newberry
and its seeds. Contrib Mus Paleontol Univ Mich 15:279–282.
——— 1937 The seeds of Alethopteris, and other pteridosperms from
North America. C R 2e Cong Intl Stratigr Geol Carbon (Heerlen,
1935) 1:41–45.
——— 1949 Fossil flora of the Michigan Coal Basin. Contrib Mus
Paleontol Univ Mich 7:131–269.
Bergeron J 1889 E
´tude ge
´ologique du massif ancien situe
´au sud du
Plateau Central. Ann Sci Geol 22:1–361.
Berry EW 1922 Carboniferous plants from Peru. John Hopkins Univ
Stud Geol 4:9–41.
Bochenski TA 1960 Rozwo
´j unerwienia listeczko
´w u Karbonskich
paproci nasiennych alethopterides (Alethopteris iLonchopteris)oraz
znaczenie analizy ich unerwienia dla diagnozy gatunku. Inst Geol Prac
20:4–42.
Boersma M 1972 The heterogeneity of the form genus Mariopteris
Zeiller: a comparative morphological study with reference to the
frond composition of west-European species. Laboratory of Palaeo-
botany and Palynology, Utrecht University, Utrecht. 172 pp.
Boulay N 1876 Le terrain houiller du Nord de la France et ses
ve
´ge
´taux fossiles. Lefebvre-Ducrocq, Lille. 74 pp.
Brongniart A 1830 Histoire des ve
´ge
´taux fossiles, vol 1, no 4. G
Dufour & E d’Ocagne, Paris, pp 169–208.
Bronn HG 1848 Index palaeontologicus oder u
¨bersicht der bis jetzt
bekannten fossilen Organismen. Erste Abtheilung. A. Nomenclator
palaeontologicus, in alphabetischer Ordnung. 2 vols. E Schwei-
zerbart, Stuttgart. 1381 pp.
Buisine M 1961 Contribution a l’e
´tude de la flore du terrain houiller:
les Ale
´thopte
´ride
´es du Nord de la France. Etud Geol Atlas Topogr
Souterr 1:1–317.
Cleal CJ, JP Laveine 1988 The juvenile frond of the Middle Carbon-
iferous pteridosperm Paripteris Gothan. Geobios 21:245–250.
Cleal CJ, CH Shute 1991 The Carboniferous pteridosperm frond
Neuropteris heterophylla (Brongniart) Sternberg. Bull Br Mus (Nat
Hist) Geol Ser 46:153–174.
Cleal CJ, EL Zodrow, M Mastalerz Forthcoming An association of
Alethopteris foliage, Trigonocarpus ovules and Bernaultia-like
pollen organs from the Middle Pennsylvanian of Nova Scotia,
Canada. Palaeontogr Abt B.
Corsin P 1928 Sur les fructifications et la position syste
´matique de
Sphenopteris zamioides P. Bertrand. Ann Soc Geol Nord 53:222–
230.
——— 1931 Fructifications de Sphenopteris (Diplotmema)alata
(Brgt) Kidston. Ann Soc Geol Nord 56:25–33.
Crabb P 2001 The use of polarised light in photography of macro-
fossils. Palaeontology 44:659–664.
Crane PR 1988 Major clades and relationships in the ‘‘higher’’
gymnosperms. Pages 218–272 in CB Beck, ed. Origin and evolution
of gymnosperms. Columbia University Press, New York.
Crookall R 1970 Fossil plants of the Carboniferous rocks of Great
Britain. 2nd sec. Pt 6. Mem Geol Surv G B Palaeontol 4:793–840.
——— 1976 Fossil plants of the Carboniferous rocks of Great
Britain. 2nd sec. Pt 7. Mem Geol Surv G B Palaeontol 4:841–1004.
Delevoryas TD 1982 Perspectives on the origin of cycads and cycad-
eoids. Rev Palaeobot Palynol 37:115–132.
Delevoryas TD, TN Taylor 1969 A probable pteridosperm with
eremopterid foliage from the Allegheny Group of northern Pen-
nsylvania. Postillia 133:1–14.
De Stefani C 1901 Flore Carbonifere e Permiane della Toscana. Pub-
blicazioni del R Inst Stud Super Prat Perf Firenze, Sez Sci Fisiche
Nat 41:1–212.
Doubinger J 1956 Contribution a
`l’e
´tude des flores Autuno-Ste
´phaniennes.
Mem Soc Geol Fr, NS, 75:1–180.
Doubinger J, P Vetter, J Langiaux, J Galtier, J Broutin 1995 La flore
fossile du bassin houiller de Saint-E
´tienne. Mem Mus Natl Hist Nat
164:1–357.
Douglas JG 1973 The Mesozoic floras of Victoria. Pt 3. Mem Geol
Surv Vic 29:1–185.
Doweld A 2001 Prosyllabus Tracheophytorum. Geos, Moscow. 110
pp.
Doyle JA 1996 Seed plant phylogeny and the relationships of Gneta-
les. Int J Plant Sci 157(suppl):S3–S39.
——— 2006 Seed ferns and the origin of angiosperms. J Torrey Bot
Soc 133:169–209.
Drinnan AN, JM Schramke, P Crane 1990 Stephanospermum kono-
peonus (Langford) comb. nov.: a medullosan ovule from the Middle
Pennsylvanian Mazon Creek flora of northeastern Illinois, USA. Bot
Gaz 151:385–401.
Duns J 1872 On Cardiocarpon. Proc R Soc Edinb 7:692–694.
Emberger L 1968 Les plantes fossiles dans leurs rapports avec les
ve
´ge
´taux vivants. 2nd ed. Masson, Paris. 758 pp.
Frenguelli J 1941 Sobre una flo
´rula Carbonı
´fera. Notas Mus La Plata,
NS, 6:459–478.
Galtier J, O Be
´thoux 2002 Morphology and growth habit of Dick-
696 INTERNATIONAL JOURNAL OF PLANT SCIENCES

sonites pluckeneti from the Upper Carboniferous of Graissessac
(France). Geobios 35:525–535.
Goloboff P 1999 NONA (NO NAME). Version 2. Published by the
author, Tucuma
´n, Argentina.
Go
¨ppert H 1836 Systema filicum fossilium. E Weber, Breslau.
486 pp.
Gothan W 1913 Einige bemerkenswerte neuere Funde von Steinkoh-
lenpflanzen in der Dortmunder Gegend. Verh Nat Ver Preuss Rheinl
Westfal 69:239–253.
Halle TG 1927 Palaeozoic plants from central Shanxi. Palaeontol Sin
Ser A 2:1–316.
——— 1929 Some seed bearing pteridosperms from the Permian of
China. K Sven Vetenskapsakad Handl Tredj Ser 6:3–24.
Hilton J, RM Bateman 2006 Pteridosperms are the backbone of
seed-plant phylogeny. J Torrey Bot Soc 133:119–168.
Hilton J, CS Li 2003 Reinvestigation of Nystroemia pectinformis
Halle, an enigmatic seed plant from the Late Permian of China.
Palaeontology 46:29–51.
Hilton J, SJ Wang, BL Tian 2003 Reinvestigation of Cardiocarpus minor
(Wang) Li from the Early Permian Taiyuan Formation of northern
China and an evaluation of cardiocarpalean ovule taxonomy, system-
atics and phylogeny. Bot J Linn Soc 141:151–175.
Howse R 1890 Contributions towards a catalogue of the flora of the
Carboniferous System of Northumberland and Durham. Pt 1. Fossil
plants from the Hutton Collection. Nat Hist Trans Northumberland
Durham Newcastle-on-Tyne 10:19–151.
Huth W 1912 Die fossile Gattung Mariopteris in geologischer und
botanischer Beziehung. Lieferung VIII. Pages 1–13 in H Potonie
´, ed.
Abbildungen und Beschreibungen fossiler Pflanzenreste. Ko
¨niglich
Preußischen Geologischen Landesanstalt, Berlin.
Jongmans WJ 1954 The Carboniferous flora of Peru. Bull Br Mus
(Nat Hist) Geol Ser 2:189–224.
Kerp JHF 1988 Aspects of Permian palaeobotany and palynology. X.
The west- and central-European species of the genus Autunia Krasser
emend. Kerp (Peltaspermaceae) and the form-genus Rhachiphyllum
Kerp (callipterid foliage). Rev Palaeobot Palynol 54:249–360.
Kerp JHF, H Haubold 1988 Aspects of Permian palaeobotany and
palynology. VIII. On the reclassification of the western and central
European species of the form-genus Callipteris Brongniart 1849 non
Bory 1804. Rev Palaeobot Palynol 54:135–150.
Kidston R 1882 Report on fossil plants collected by the Geological
Survey of Scotland in Eskdale and Liddesdale. Trans R Soc Edinb
30:531–550.
——— 1886 Catalogue of the Palæozoic plants in the Department of
Geology and Palæontology, British Museum (Natural History).
British Museum (Natural History), London. 288 pp.
——— 1891 On the fossil flora of the Staffordshire Coal Fields.
Trans R Soc Edinb 36:63–98.
——— 1893 On the fossil plants of the Kilmarnock, Galston, and
Kilwinning Coal Fields, Ayrshire. Trans R Soc Edinb 37:307–359.
——— 1894 On the fossil flora of the South Wales Coal Field, and
the relationship of its strata to the Somerset and Bristol Coal Field.
Trans R Soc Edinb 37:565–614.
——— 1901 The Carboniferous fossil plants of the Clyde Basin.
Pages 468–476 in GFS Elliott, M Laurie, JB Murdoch, eds. Fauna,
flora, and geology of the Clyde area. Local Committee for the
Meeting of the British Association, Glasgow.
——— 1903 The fossil plants of the Carboniferous rocks of Canon-
bie, Dumfriesshire, and parts of Cumberland and Northumberland.
Trans R Soc Edinb 40:741–833.
——— 1914 On the fossil flora of the Staffordshire Coal Fields. Pt
III. The fossil flora of the Westphalian Series of the South Staf-
fordshire Coal Field. Trans R Soc Edinb 50:73–190.
——— 1919 List of the plants from the Coal Measures of the borings
at Bere Farm, Elham, Folkestone, and Lydden Valley, Kent. Summ
Prog Geol Surv G B Mus Pract Geol 1918:46–49.
——— 1922 List of fossil plants from the Upper Carboniferous rocks
of the Northumberland and Durham Coalfield and their bearing on
the age of the coalfield. Summ Prog Geol Surv G B Mus Pract Geol
1921:129–145.
——— 1923 Fossil plants of the Carboniferous rocks of Great Brit-
ain. Pts 1–4. Mem Geol Surv G B Palaeontol 2:1–376.
——— 1924 Fossil plants of the Carboniferous rocks of Great
Britain. Pt 5. Mem Geol Surv G B Palaeontol 2:377–522.
Knaus MJ 1994 Triphyllopteris collombiana: a clarification of the
generic concept based on rediscovered specimens from Kossberg bei
Plauen, Germany, and a reassignment of the North American spe-
cies of Triphyllopteris to Genselia gen. nov. Int J Plant Sci 155:97–
116.
Kva
cek J, M Strakova
´1997 Catalogue of fossil plants described in
works of Kaspar M. Sternberg. National Museum, Prague. 201 pp.
Land DH 1974 Geology of the Tynemouth district. Mem Geol Surv
G B Sheet 15. 176 pp.
Laveine JP 1967 Contribution a
`l’e
´tude de la flore du terrain houiller:
les Neuropte
´ride
´es du Nord de la France. Etud Geol Atlas Topogr
Souterr 1:1–344.
——— 1997 Synthetic analysis of the neuropterids: their interest for
the decipherment of Carboniferous palaeogeography. Rev Palaeo-
bot Palynol 95:155–189.
Lesquereux L 1876 Partial list of coal plants from the Alabama
Fields, and discussion of the geological positions of several coal
seams. Pages 75–82 in EA Smith, ed. Report of Progress for 1875.
Geological Survey of Alabama, Montgomery.
——— 1880 Description of the Coal Flora of the Carboniferous in
Pennsylvania and throughout the United States. Vols 1–2. Second
Geological Survey of Pennsylvania, Report of Progress, Harrisburg.
694 pp.
——— 1884aDescription of the Coal Flora of the Carboniferous in
Pennsylvania and throughout the United States. Vol 3. Second
Geological Survey of Pennsylvania, Report of Progress, Harrisburg,
pp 695–977.
——— 1884bPrinciples of Paleozoic botany. Pages 7–106 in J Col-
lett, ed. Indiana Department of Geology and Natural History. Thirteenth
Annual Report. Pt II, Paleontology. WB Burford, Indianapolis.
Lindley J, W Hutton 1832 The fossil flora of Great Britain. Vol 1, Pt
2. WJ Ridgeway, London, pp 49–166.
Long AG 1960 On the structure of Samaropsis scotica Calder
(emended) and Eurystoma angulare gen. et sp. nov., petrified seeds
from the Calciferous Sandstone Series of Berwickshire. Trans R Soc
Edinb 64:261–280.
Lyons PC, WH Orem, M Mastalerz, EL Zodrow, A Vieth-Redemann,
MB Bustin 1995
13
C NMR, micro-FTIR and fluorescence spectra,
and pyrolysis-gas chromatograms of coalified foliage of late Car-
boniferous medullosan seed ferns, Nova Scotia, Canada: implica-
tions for coalification and chemotaxonomy. Int J Coal Geol 27:
227–248.
McNeill J, FR Barrie, HM Burdet, V Demoulin, DL Hawksworth, J
Marhold, DH Nicolson, et al 2006 International Code of Botan-
ical Nomenclature (Vienna Code). ARG Gantner, Ruggell, Liech-
tenstein. 568 pp.
Meyen SV 1987 Fundamentals of palaeobotany. Chapman & Hall,
London. 432 pp.
Murchison RI, E de Verneuil, A de Keyserling 1845 Ge
´ologie de la
Russie d’Europe et de Montagnes de l’Oural. Vol II. 3 e Partie
Pale
´ontologie. J Murray, London, xxxii þ512 pp, 50 pls.
Ne
ˇmejc F 1928 Revise Karbonske
´a Permske
´kveteny stredocesky
´ch
pavnı
´uhelny
´ch. Palaeontogr Bohem 12:1–82.
Nixon KC 2002 WinClada. Version 1.00.08. Published by the
author, Ithaca, NY.
697
CLEAL ET AL.—THE PENNSYLVANIAN SEED PLANT EREMOPTERIS

Nixon KC, WL Crepet, DW Stevenson, EM Friis 1994 A reevalua-
tion of seed plant phylogeny. Ann Mo Bot Gard 81:484–533.
Novik EO 1947 Klassifikatsiya kamennougol’nikh pteridospermov.
Dok Akad Nauk SSSR 58:277–279.
Potonie
´H 1893 Ueber einige Carbonfarne. III. Theil. Jahrb Ko
¨ni-
glich Preuss Geol Landesanst Berg Berl 12:1–36.
——— 1897 Lehrbuch der Pflanzenpalaeontologie mit besonderer
Ru
¨cksicht auf die Bedu
¨rfnisse des Geologen (1–2). F Du
¨mmler,
Berlin. 208 pp.
——— 1898 Die Metamorphose der Pflanzen im Lichte palaeonto-
logischer Thatsachen. F Du
¨mmler, Berlin. 29 pp.
——— 1921 Lehrbuch der Pala
¨obotanik. Zweite Auflage umgear-
beitet von Prof. Dr W Gothan. Borntraeger, Berlin. 537 pp.
Ps
ˇeni
cka J, EL Zodrow, M Mastalerz, J Bek 2005 Functional groups
of fossil marattialeans: chemotaxonomic implications for Pennsyl-
vanian tree ferns and pteridophylls. Int J Coal Geol 61:259–280.
Renault B, R Zeiller 1888 E
´tudes sur le terrain houiller de Com-
mentry. Livre deuxie
´me. Flore fossile. Socie
´te
´de l’Industrie Mine
´rale,
St. E
´tienne. 746 pp.
Romanowski GD 1890 Materialy dlya geologii Turkestanskogo kraya.
Vypusk trety. Istoricheskaya geologiya: paleontologichesky character
osadochnych obrazovanij zapadnogo Tyan-Shanya I Turanskoy
nizmennosti. Imperatorskoy Akademii Nauk, St. Petersburg. 165 pp.
Rothwell GW 1975 The Callistophytaceae (Pteridospermopsida). Pt
I. Vegetative structures. Palaeontogr Abt B 151:171–196.
——— 1980 The Callistophytales (Pteridospermopsida): reproductively
sophisticated pteridosperms. Rev Palaeobot Palynol 32:103–121.
——— 1986 Classifying the earliest gymnosperms. Syst Assoc Spec
Vol 31:137–162.
Rothwell GW, R Serbet 1994 Lignophyte phylogeny and the evolution
of spermatophytes: a numerical cladistic analysis. Syst Bot 19:443–482.
Saltzwedel K 1969 Revision der Imparipteris ovata (Hoffmann) Gothan.
1. Teil. Typus- und Typoid-Material vom locus typicus. Arg Palaeobot
3:131–162.
Sauveur JJD 1848 Ve
´ge
´taux fossiles des terrains houillers de la
Belgique. Acade
´mie Royale des Sciences, des Lettres et des Beaux-
Arts de Belgique, Brussels. 2 pp.
Schimper WP 1869 Traite
´de pale
´ontologie ve
´ge
´tale. Vol 1. JB
Baillie
`re, Paris. 740 pp.
——— 1879 Handbuch der Palaeontologie. II. Band I. Lieferung. R
Oldenbourg, Munich. 232 pp.
Seward AC 1917 Fossil plants for students of botany and geology.
Vol 3. Pteridospermae, Cycadofilices, Cordaitales, Cycadophyta.
Cambridge University Press, Cambridge. 656 pp.
S
ˇimu
˚nek Z, CJ Cleal 2004 Small-pinnuled odontopterid medullosa-
leans from the middle and upper Stephanian of Central Bohemia
and Saar-Lorraine. Rev Palaeobot Palynol 129:21–38.
Stirling J 1892 Reports on Victorian coalfields. Victoria Geological
Survey, Melbourne.
Stur D 1877 Die Culm-Flora der Ostrauer und Waldenburger
Schichten. Abh Kaiser-Konig Geol Reich Wien 8:107–472.
Taylor TN, WN Stewart 1964 The Paleozoic seed Mitrospermum in
American coal balls. Palaeontogr Abt B 115:51–58.
Teixeira C 1941 W. de Lima e a Flora fo
´ssil do Permo-Carbo
´nico
portugue
ˆs. Comun Serv Geol Port 22:3–17.
Thomas BA, SD Brack-Hanes 1984 A new approach to family group-
ings in the lycophytes. Taxon 33:247–255.
Thomas BA, CJ Cleal 1999 Abscission in the fossil record. Pages
183–203 in MH Kurman, AR Hemsley, eds. The evolution of plant
architecture. Royal Botanic Gardens, Kew.
Tidwell WD 1967 Flora of the Manning Canyon Shale. Pt I. A lower-
most Pennsylvanian flora from the Manning Canyon Shale, Utah, and
its stratigraphic significance. Brigham Young Univ Geol Stud 14:
3–66.
Trivett ML, GW Rothwell 1985 Morphology, systematics and pa-
leoecology of Paleozoic fossil plants: Mesoxylon priapi, sp. nov.
(Cordaitales). Syst Bot 10:205–223.
——— 1988 Diversity among Paleozoic Cordaitales: the vascular
architecture of Mesoxylon birame Baxter. Bot Gaz 149:116–125.
——— 1991 Diversity among Paleozoic Cordaitales. Neues Jahrb
Geol Palaeontol Abh 183:289–305.
Unger F 1845 Synopsis plantarum fossilium. L Voss, Leipzig. 320 pp.
van Amerom HWJ 1975 Die eusphenopteridischen Pteridophyllen
aus der Sammlung des Geologischen Bureaus in Heerlen, unter
besonderer Beru
¨cksichtigung ihrer Stratigraphie bezu
¨glich des
su
¨dlimburger Kohlenreviers. Meded Rijks Geol Dienst Ser C III-1
7:1–202.
von Sternberg KM 1825 Versuch einer geognostisch-botanischen
Darstellung der Flora der Vorwelt. Vol 1, pt 4, Tentamen. E Brenck’s
Wittwe, Regensburg. 48 pp.
Wagner RH 1968 Upper Westphalian and Stephanian species of
Alethopteris from Europe, Asia Minor and North America. Meded
Rijks Geol Dienst Ser C III-1 6:1–188.
White D 1893 Flora of the outlying Carboniferous basins of south-
western Missouri. Bull US Geol Surv 98:11–139.
——— 1899 Fossil flora of the Lower Coal Measures of Missouri.
US Geological Survey, Washington, DC. 467 pp.
——— 1900 The stratigraphic succession of the fossil floras of the
Pottsville Formation in the Southern Anthracite Coal Field,
Pennsylvania. Annu Rep US Geol Surv 20:755–893.
——— 1943 Lower Pennsylvanian species of Mariopteris,Eremop-
teris,Diplothmema and Aneimites from the Appalachian Region.
Prof Pap US Geol Surv 197C:1–103.
——— 1965 Plant fossil collection from the Georgetown/Clarke
River area. Bull Aust Bur Miner Resour Geol Geophys 71:152–165.
Zeiller R 1879 Pre
´sentation de l’Atlas du tome IV de l’Explication de
la Carte ge
´ologique de la France, et note sur le genre Mariopteris.
Bull Soc Geol Fr Ser 3 7:92–98.
——— 1898 Contribution a
`l’e
´tude de la flore pte
´ridologique des
schistes permiens de Lode
`ve. Bull Mus Hist Nat Mars 1:6–69.
Zodrow EL, CJ Cleal 1998 Revision of the pteridosperm foliage
Alethopteris and Lonchopteridium (Upper Carboniferous), Sydney
Coalfield, Nova Scotia, Canada. Palaeontogr Abt B 247:65–122.
Zodrow EL, M Mastalerz 2002 FTIR and py-GC-MS spectra of
true-fern and seed-fern sphenopterids (Sydney Coalfield, Nova
Scotia, Canada, Pennsylvanian). Int J Coal Geol 51:111–127.
Zodrow EL, M Mastalerz, Z S
ˇimu
˚nek 2003 FTIR-derived character-
istics of fossil-gymnosperm leaf remains of Cordaites principalis and
Cordaites borassifolius (Pennsylvanian, Maritimes Canada and
Czech Republic). Int J Coal Geol 55:95–102.
698 INTERNATIONAL JOURNAL OF PLANT SCIENCES