Content uploaded by R. A Stockey
All content in this area was uploaded by R. A Stockey on Jan 10, 2020
Content may be subject to copyright.
Content uploaded by Michael G Riley
All content in this area was uploaded by Michael G Riley on Nov 15, 2016
Content may be subject to copyright.
See discussions, stats, and author profiles for this publication at: https://www.researchgate.net/publication/249158097
Cardstonia tolmanii gen. et sp. nov. (Limnocharitaceae) from the Upper
Cretaceous of Alberta, Canada
ArticleinInternational Journal of Plant Sciences · September 2004
Some of the authors of this publication are also working on these related projects:
Taxodiaceous seed cones from Japan, Cornales morphological matrix, Pinaceous leaves from the Cretaceous of Hokkaido, Araucarian cones from Hokkaido and Argentina,
etc. View project
Conifer Evolution View project
Michael G Riley
Aeon Paleontological Consulting Ltd., Sherwood Park, Alberta, Canada
1 PUBLICATION21 CITATIONS
R. A Stockey
Oregon State University
250 PUBLICATIONS4,822 CITATIONS
All content following this page was uploaded by Michael G Riley on 15 November 2016.
The user has requested enhancement of the downloaded file.
CARDSTONIA TOLMANII GEN. ET SP. NOV. (LIMNOCHARITACEAE) FROM THE
UPPER CRETACEOUS OF ALBERTA, CANADA
Michael G. Riley and Ruth A. Stockey1
Department of Biological Sciences, University of Alberta, Edmonton, Alberta T6G 2E9, Canada
Several new broad-leaved monocots were identiﬁed in gray siltstones and ﬁne-grained sandstones from the
Upper Cretaceous (Campanian-Maastrichtian) St. Mary River Formation near Cardston, Alberta, Canada.
Specimens are compression/impressions of long-petiolate aquatic plants that were probably buried in situ. Leaf
blades are entire, ovate to elliptic, with deeply cordate bases; leaf blades are 5–12 cm long and 3.5–8.5 cm
wide. Petioles are at least 4 mm wide with ﬁve to seven primary veins that enter the leaf blade. Venation is
campylodromous, resulting in 23–27 primary veins with three medial veins that remain unbranched to the
apex. Major and minor secondary veins (ABAB pattern) diverge at angles of 45°–65°near the midveins and 90°
near the leaf margin. Tertiary veins are usually unbranched, but occasional dichotomies and anastomoses
occur. No freely ending veinlets are visible. The Cardston specimens are compared with extant leaves of
Alismatales and show closest similarities to those of Limnocharis L., Hydrocleys Rich., and Butomopsis
Kunth. These leaves are also similar to fossil leaves of Haemanthophyllum Budantsev, in particular H.
cordatum Golovneva from the Maastrichtian-Danian deposits of the Koryak Highlands, Russia. A
reexamination of the genus Haemanthophyllum, the generitype (H. kamtschaticum Budantsev), the holotype
for H. cordatum, and the Cardston specimens results in the description of a new genus, Cardstonia tolmanii
gen. et sp. nov. (Limnocharitaceae). This study points to the need for reexamination of the remaining species in
the genus Haemanthophyllum, which appear to represent a diverse assemblage of leaves of varying
morphology that are probably not a natural group.
Keywords: Alismatales, Aponogeton, aquatic, Butomopsis, Cretaceous, Echinodorus,Haemanthophyllum,
The well-preserved fossil plants from the Upper Cretaceous
Cardston Flora of southern Alberta have yielded 32 taxa of
predominantly aquatic plants (Riley and Stockey 1999).
These plants are compression/impressions that in some cases
are nearly complete because of rapid burial in ﬁne-grained
sediments. Geologic data indicate that they were buried in
situ in shallow oxbow lakes or ponds during a period of
rapid sedimentation (Riley and Stockey 2000). Only two
plants have been described in detail and reconstructed as
whole plants: Hydropteris pinnata Rothwell and Stockey
(1994), a heterosporous fern, and Quereuxia angulata (New-
berry) Krysht. ex Baikovskaja (Stockey and Rothwell 1997),
a ﬂoating, rosette-forming dicot.
The Cardston site is remarkable in its preservation of
large-leaved monocots. Broad leaves of these herbaceous
plants are usually rare in the fossil record (Herendeen and
Crane 1995). At least six types of monocot leaves (in addi-
tion to those of a sabaloid palm) are present in sediments of
the St. Mary River Formation at Cardston. One of these
monocot leaf types has been compared with leaves described
as Haemanthophyllum Budantsev (1983) from the Late
Paleocene–Lower Eocene of western Kamchatka (Riley and
Stockey 1998). This leaf type was originally assigned to the
Amaryllidaceae, but further studies of leaf venation and
comparisons to extant and fossil monocots have led several
workers to suggest relationships to Alismataceae, Potamoge-
tonaceae, and Aponogetonaceae (Heer 1868; Golovneva
1987, 1997; Boulter and Kva
cek 1989). A number of leaves
of differing morphology have been assigned to this genus
(table 1), but their afﬁnities remain in doubt (Golovneva
Several broad-leaved monocots generally conforming to
Haemanthophyllum have been identiﬁed in the Cardston
Flora (Riley and Stockey 1998). In this study we compare
leaf morphology and detailed venation patterns of these Cre-
taceous leaves with fossils in the genus Haemanthophyllum
and those of extant Aponogetonaceae, Alismataceae, Hydro-
charitaceae, Stemonaceae, Potamogetonaceae, and Limno-
charitaceae. We describe these remains as Cardstonia
tolmanii Riley et Stockey gen. et sp. nov. and suggest that
they have afﬁnities to Limnocharitaceae.
Material and Methods
Fifty compression/impression specimens of these leaves
have been collected from three sites on the banks of the St.
Mary River below the reservoir and spillway (ﬁg. 1).
1Author for correspondence; e-mail email@example.com.
Manuscript received March 2003; revised manuscript received March 2004.
Int. J. Plant Sci. 165(5):897–916. 2004.
Ó2004 by The University of Chicago. All rights reserved.
Comparative Morphology of Haemanthophyllum Budantsev Species
Species Locality Age
(cm) Blade Base Apex NLV NCV
Haemanthophyllum zhilinii (Pneva)
Golovneva Kazakhstan Late Oligocene–Early
20–40 ;11 Elliptical Cuneate Acuminate 11–13 10–20
Budantsev (Generitype) Russia, Kamchatka
Late Paleocene/Eocene ? 7.5
? Cordate ? ;21 (24–35)
Budantsev (original description) Russia, Kamchatka
Late Paleocene/Eocene 12–35 7–17 Broadly
Cordate Attenuate to
Haemanthophyllum sp. 1 Russia, Amur
Paleocene 5 2 Elliptical Cuneate ? 15 14–17 (25)
Haemanthophyllum sp. 2 Northern Ireland,
? 2.2–2.8 Oblong ? ? 13 (17)6–8
(Heer) Boulter et Kva
cek Norway, Spitsbergen Early Paleocene/ Eocene 7–20 4–10 Elliptical
15–21 7–9 (14)
Haemanthophyllum sp. 3 United States, North
Dakota, South Dakota
Paleocene ;30 ;15 Broadly
? Acuminate 25 9–12
Haemanthophyllum sp. 4 United States, Alaska Paleocene ;11 ;5 Ovate Rounded to
Rounded 17–25 ?
Haemanthophyllum sp. 5 Canada, Saskatchewan Early Paleocene ;12 ;7 ? ? Acuminate 27 15–17
Golovneva Russia, Koryak
4–18 3–14 Ovate Cordate Rounded (13)17–25 13–18 (7–25)
Cardston fossil Canada, Alberta Late Cretaceous 5–12 3.5–8.5 Ovate Deeply cordate Convex to
Note. NLV ¼number of longitudinal veins (1°veins; does not include the ﬁmbrial [marginal] vein); NCV ¼number of cross veins (2°veins) per centimeter; boldface ¼difference in observed number of veins; ta-
ble modiﬁed from Golovneva (1997).
aEstimate of leaf width.
Outcrops are part of the St. Mary River Formation, Upper
Cretaceous, and are Late Campanian–Early Maastrichtian in
age (Nadon 1988; Hamblin 1998; A. R. Sweet, personal com-
Fossils come from alternating layers of siltstones and ﬁne-
grained sandstones that were deposited on the ﬂoodplains or
the abandoned channels of an anastomosing river system
(Nadon 1988, 1994). Most plants were preserved by rapid
sedimentation in oxbow lakes or ponds. The best-preserved
leaves are found in the siltstones. Specimens were prepared
by degagement, photographed using a MicroLumina digital
scanning camera (Leaf Systems), and processed using Adobe
All specimens are housed in the University of Alberta Pa-
leobotanical Collections (UAPC-ALTA) and carry specimen
numbers S50,939; S50,940; S50,947; S50,955–S50,958;
S50,988; S50,989; S52,263–S52,283; S52,292–S52,295;
S55,129–S55,135; S55,137–S55,150; S55,154.
Leaves of extant monocots from a large number of families
(many previously suggested to have afﬁnities to Haemantho-
phyllum Golovneva 1997) were examined in the herbarium of
the Missouri Botanical Garden; the Royal Botanic Gardens,
Kew; the Munich Botanical Gardens; the University of Alberta
Herbarium; and the published literature (Tomlinson 1982;
van Bruggen 1990; Haynes and Holm-Nielsen 1992; Kassel-
mann 1995, 2003; Cook 1996a, 1996b; Mayo et al. 1997;
Kubitzki 1998) for similarities in morphology and growth
habit to the fossils described here. Those of Alismataceae,
Amaryllidaceae, Aponogetonaceae, Hydrocharitaceae, Lim-
nocharitaceae, Potamogetonaceae, and Stemonaceae that
showed the closest similarities were examined during this
study from fresh and herbarium specimens from the Munich
Botanical Gardens, Herbarium of the Northern Territory, Dar-
win, Australia (DNA), and University of Alberta Vascular
Plant Herbarium (ALTA) (table 2). Illustrations that appear in
this article are only those leaves that showed the closest simi-
larities to the fossils described here; however, additional taxa
appear in table 2. Photographs were taken using both trans-
mitted and reﬂected light (ﬁgs. 5a–5h,6a–6h).
Genus—Cardstonia Riley et Stockey gen. nov.
Generic diagnosis. Leaves simple, ovate to elliptic, mar-
gin entire; apex convex to rounded; base cordate. Venation
campylodromous, primary veins originating at or near base,
running in strongly recurved arches that converge apically,
merging at or near apex; apical pore present. Prominent ﬁm-
brial vein present. Secondary veins diverging at low angles
near midrib and higher angles near margin; alternating major
and minor secondary veins, with ABAB pattern, occasionally
anastomose and dichotomize. Tertiary (transverse) veins few,
with straight to slightly curved courses, forming moderately
developed (irregular shape, variable-sized) areolae.
Species.Cardstonia tolmanii Riley et Stockey sp. nov.
Holotype. UAPC-ALTA S55138 (ﬁg. 2a).
Paratypes. UAPC-ALTA S50947, S52263, S52279,
S52268, S52266, S52272, S50989.
Speciﬁc diagnosis. Leaves simple, ovate to elliptic, 5–12
cm long, 3.5–8.5 cm wide; l : w ratio 1.4 : 1; margin entire;
apex convex to rounded; base cordate. Petiole 3–4 mm wide,
at least 2 cm long, petiolar attachment marginal. Venation
campylodromous, primary veins 23–27, with 3–5 medial
veins originating at or near base; outer primaries running in
strongly recurved arches that converge apically, merging at
or near apex; apical pore present. Prominent ﬁmbrial vein
present. Secondary veins diverging at angles 45°–60°near
midvein and 90°near margin; alternating major and minor
secondary veins, with variable ABAB pattern, occasionally
anastomose and dichotomize. Tertiary (transverse) veins few,
Fig. 1 Aerial photo of collecting locality in southern Alberta, Canada. Photo shows St. Mary Reservoir (R), associated dam (D), and both
spillways. Locations of two fossil sites along banks of St. Mary River (arrows). Scale bar ¼300 m.
RILEY & STOCKEY—CARDSTONIA TOLMANII GEN. ET SP. NOV.
Comparison of Morphological Characters of Cardstonia tolmanii with Selected Extant Taxa
of 2°veins Major
presentTaxon Family Blade Base Apex 1°veins Center Margin
Riley et Stockey Ovate
23–27 N Y 10–40 45–60 90 Y Y Y Y
Alisma subcordatum Raf. Alismataceae
Decurrent Convex 9 Y N 5–12 35–45 70–90 N N Y Y
Caldesia Parl. sp. Alismataceae
Ovate Cordate Acuminate 13–15 Y N 30 60 90 Y Y Y N
11–12 Y N 6–10 70–80 90 Y Y Y Y
(Chamisso et Schlechtendal)
Ovate Cordate Convex 15 Y N 8–11 60 90 Y Y Y Y
(Mart.) Griseb. Alismataceae
Ovate Convex Straight
9 Y N 8–9 50–55 50–55 Y Y Y Y
cariensis L. f. Aponogetonaceae
Oblong Decurrent Emarginate 13 N Y 5–6 80–90 80–90 N N N N
(Planch.) Walp. Hydrocharitaceae
Oblong Cuneate Straight 7–19 Y N 3–4 60–90 60–90 Y Y
(D. Don) Kunth Limnocharitaceae
Elliptical Cuneate Straight
7–9 N Y 10–12 40–50 80–85 Y Y Y Y
Hydrocleys martii Seubert Limnocharitaceae
Elliptical Cordate Rounded 11 N Y
6–20 70 90 Y Y Y Y
Limnocharis ﬂava (L.)
Elliptical Cordate Rounded 11–13 N Y 11–13 60–80 90 Y Y Y Y
Oblong Cuneate Acuminate 11–13 N Y 8–11 65–75 75–90 Y Y Y Y
Potamogeton lucens L. Potamogetonaceae
Elliptic Decurrent Acuminate
9 N Y 6–10 40–60 60–80 N Y Y Y
Oblong Cuneate Acuminate 21 N Y 11–23 60 80 Y
Stemona tuberosa Lour. Stemonaceae
Ovate Cordate Acuminate 11 N
N 28–30 90 90 Y Y Y N
Note. 1°veins ¼number of primary veins in total (not including ﬁmbrial); 2°veins per cm ¼number of secondary veins (both major and minor) per centimeter.
aOrder Alismatales (APG 1998).
dOrder Asparagales (APG 1998).
eOrder Pandanales (APG 1998).
fNo ﬁmbrial vein present.
with straight to slightly curved courses, forming moderately
developed (irregular shape, variable-sized) areolae.
Etymology. The genus is named after the nearby town of
Cardston, Alberta. The speciﬁc epithet is proposed in honor
of Shayne Tolman of Cardston, Alberta, who brought the lo-
cality to our attention and spent many hours working with
us in the ﬁeld.
Locality. The locality is 26 km NE of the town of Card-
ston, Alberta (T5 R254 S12 SE 1/4 W4M; 49°229100N,
113°069100W; UTM Grid 12: 5470600N, 347300E).
Stratigraphic occurrence. St. Mary River Formation.
Age. Late Campanian–Early Maastrichtian, Upper Creta-
Of the 50 specimens of this leaf type available, ﬁve are
complete or nearly complete, and Morphotype Quality
Index ¼2 (Leaf Architecture Working Group 1999). Leaves
are simple and ovate to elliptic with an entire margin (ﬁgs.
2a–2c). Several leaves have attached petioles that are up to
2 cm long (dictated by the size of the collected block). Petioles
show ﬁve to seven longitudinal vascular bundles that lack
transverse septa (ﬁg. 2d). Petiolar attachment is basal, i.e.,
‘‘marginal’’ (Leaf Architecture Working Group 1999) and the
petioles are always bent down sharply into the siltstone ma-
trix (ﬁgs. 2a–2d,3e). The leaves were probably emergent and
likely had long petioles with blades borne at right angles, or
nearly so, to the petiole. Leaf blades are 5–12 cm long and
3.5–8.5 cm wide with a length/width ratio of 1.4:1 (table 2).
Leaf apices are convex to rounded and the bases are deeply
cordate (ﬁg. 2a–2d; ﬁg. 3a,3b).
Venation is campylodromous, with 23–27 primary veins (ﬁg.
2a,2b). Five to seven primary veins enter the base of the leaf,
the outermost primaries branching to give rise to the most basal
primary veins (ﬁgs. 2d,3e). The three medial primary veins are
more or less parallel to one another for the entire length of the
leaf (ﬁg. 2a,2b;ﬁg.3c). The outermost primary vein becomes
the ﬁmbrial (marginal) vein that dichotomizes several times in
the cordate base and gives rise to additional primaries (ﬁg. 2d;
ﬁg. 3d,3e). The additional veins anastomose just beneath the
apex of the leaf (ﬁgs. 2a,3b). All of the remaining primary
veins converge around an apical pore (ﬁg. 3b).
Secondary veins (‘‘cross’’ veins of Golovneva 1997) diverge
from the closely spaced medial veins at angles of 40°–45°in
most of the leaf blade to 90°near the leaf base (ﬁg. 3c,3e). To-
ward the leaf margin, secondary veins typically arise at angles
of 90°or nearly so. There are from 10 to 40 of these secondary
veins per cm, but 30–35 are common in most of the lamina.
Their density is greatest near the leaf margin (ﬁg. 2d; ﬁg. 3c,
3d); there are fewer secondary veins near the midrib. In some
of the best-preserved leaves there appear to be alternating ma-
jor and minor (thick and thin) secondary veins (ﬁg. 3f–3h)in
an ABAB pattern (Hickey and Peterson 1978). This pattern is
not always regular, however, and occasionally two thin veins
occur between two thick veins (ﬁg. 3f), ABABBAB. Or thin
veins may be absent, ABAAAB (ﬁg. 3d). Secondary veins occa-
sionally dichotomize (ﬁg. 3d,3f) or anastomose (ﬁg. 3f–3h).
Very few of what might be termed tertiary veins (using the
terminology of the Leaf Architecture Working Group 1999
or transverse veins of Hickey and Peterson 1978) are seen in
the fossil leaves. These are distinctly thinner veins that occur
at right angles or nearly so to the secondary veins (ﬁg. 3f,
3g). These veins typically run a straight or slightly curved
course and terminate in the adjacent secondary vein.
In some of the best-preserved leaves there are polygonal
patterns preserved on the surface of the fossils that have split
between the two leaf surfaces (ﬁg. 3i). While these polygonal
patterns may represent tertiary venation, when examined
closely, these hexagonal patterns overlap with the secondary
veins and appear to be superimposed on them (ﬁg. 3i). In
some parts of the leaf they are preserved in sediment between
the two leaf surfaces. It is probable that these patterns are
the outlines of underlying aerenchyma due to their shape and
position and not veins. The thickness of the dark material
that outlines the polygonal pattern is about 50 mm in diame-
ter, about the size of parenchyma cells.
Russian Fossil Haemanthophyllum Leaves
Haemanthophyllum kamtschaticum Budantsev
Budantsev ﬁrst described the genus Haemanthophyllum in
1983 from the Paleocene–Eocene of western Kamchatka. The
holotype of Haemanthophyllum kamtschaticum Budantsev
(generitype) was reexamined here (ﬁg. 4a,4b; table 1). In ad-
dition to the type specimen, two other specimens were exam-
ined from the same locality (ﬁg. 4c–4e). The specimen in
ﬁgure 4cis illustrated by Budantsev (1983) in the original
article. The other specimen (ﬁg. 4d,4e) is from the same locality
but is not illustrated by Budantsev (1983). In addition to the
specimens illustrated here, two other specimens (Budantsev
1983, pl. 63, ﬁgs. 1, 4) were not available for examination.
In table 1 we show two rows of data for H. kamtschaticum;
the ﬁrst is from the original holotype, the second is Budant-
sev’s (1983) concept of the species as shown in Golovneva’s
table (1997). This is necessary because there is only one holo-
type, but workers have often used a species concept derived
from several fragments that, in our opinion, probably do not
represent the same taxon.
The holotype of H. kamtschaticum is a leaf fragment with
entire margin, cordate base, ﬁmbrial vein, up to nine re-
curved primary veins, and 24–35 secondary veins per centi-
meter (table 1; ﬁg. 4a). Note that the number of secondary
veins per centimeter counted by us is markedly higher than
the nine to 13 originally reported by Budantsev (1983) in the
generic description. Secondary veins remain unbranched or
dichotomize and/or anastomose between adjacent primary
veins (ﬁg. 4b). There are certain areas on the holotype where
secondary veins show several closely spaced anastomosing di-
chotomies between the adjacent primaries (ﬁg. 4b). Three
closely spaced medial veins are not visible (ﬁg. 4a). The sec-
ond specimen Budantsev illustrated has a long, thickened pet-
iole (>7 cm in length) with multiple veins that continue as
a thickened costa into the blade (ﬁg. 4c). The primary veins
branch into the lamina from the midvein, with the most basal
veins recurving into a cordate base that is slightly concave-
convex (ﬁg. 4c). Three closely spaced medial veins are not
visible (ﬁg. 4c). The third leaf fragment studied here (not il-
lustrated by Budantsev) shows a ﬁmbrial vein, dense
RILEY & STOCKEY—CARDSTONIA TOLMANII GEN. ET SP. NOV.
secondary veins (ﬁg. 4d), and possible polygonal aerenchyma
between secondary veins (ﬁg. 4e).
Haemanthophyllum cordatum Golovneva
The holotype of Haemanthophyllum cordatum Golovneva
(1987) from the late Maastrichtian-Danian sediments (Raryt-
kin Series) of the Koryak Highlands in Russia was reex-
amined but was difﬁcult to study because the specimen is
preserved as a black impression on a black matrix (ﬁg. 4f–
4h). The leaf has an entire margin, convex to rounded apex,
cordate base, and a ‘‘triple midvein’’ (three closely spaced
medial veins). The number of primary veins appears to be
25, with a ﬁmbrial vein visible in some marginal areas. The
margin was not preserved or was slightly degaged away in
parts of the specimen, making it difﬁcult to trace the path of
the outermost veins. Golovneva (1987, pl. I, 1) originally il-
lustrated the outermost primary veins merging with the ﬁm-
brial vein; however, we cannot observe the actual leaf margin
with certainty in many areas of the specimen. It appears that
several of the primary veins do merge together at the apex of
the leaf. Golovneva reports 18 secondary veins per centime-
ter in this species; we count seven to 25 secondaries.
Extant Monocot Leaves
Leaves of Alisma subcordata Raf. (Alismataceae) are ob-
ovate to elliptical with decurrent bases and convex to acumi-
nate apices (table 2). Some leaves show an emarginate apex
with an acuminate leaf tip. There are nine primary veins, the
outer two of which merge with the ﬁmbrial vein midway up
the leaf. The primary veins do not merge beneath the leaf
apex but extend into the tip itself. There are ﬁve to 12 sec-
ondary veins per centimeter that do not anastomose but
sometimes dichotomize (table 2). Secondary veins, all of
more or less equal thickness, diverge at angles of 35°–45°
near the midvein and 70°–90°near the leaf margin (table 2).
Tertiary veins are present and occasionally dichotomize, and
they have a wavy or sinuous course in the lamina, often aris-
ing at right angles. Golovneva (1994) illustrates similar sec-
ondary venation in Alisma plantago-aquatica L. These veins
form the irregular, rectangular areolae.
Caldesia Parl. (Alismataceae) leaves (ALTA 49675) were
examined from a plant growing at the Technische Hoch-
¨rich (ﬁg. 5c,5d). They are ovate with cordate bases
and acuminate apices, with 13–15 primary veins (ﬁg. 5c; ta-
ble 2). (Note: Caldesia parnassifolia [Bassi ex L.] Parl. [MO
2326418] examined at the Missouri Botanical Garden Her-
barium had 17 primary veins.) All of the primary veins merge
with the ﬁmbrial vein (ﬁg. 5c, arrows) except the medial
vein, which enters the leaf tip. There are 30 secondary veins
per centimeter that arise at angles of 60°–70°near the mid-
vein and 90°near the leaf margin. There are major and mi-
nor secondary veins that anastomose and dichotomize (ﬁg.
5d). Tertiary veins appear to be absent (ﬁg. 5d).
We examined eight different species of Echinodorus Rich.
ex Engelm. (Alismataceae). Of these, Echinodorus inpai Ra-
taj, Echinodorus osiris Rataj, Echinodorus longiscapus
Arech., Echinodorus decumbens Kasselmann, and Echinodo-
rus uruguayensis Arech. were not put in table 2 as they
greatly differ from the fossil leaves in being narrow, oblong
leaves with very few primary veins. Echinodorus glaucus Ra-
taj (Alismataceae) leaves are ovate to oblong with cordate ba-
ses and convex to rounded apices (ﬁg. 5a; table 2). There are
11–13 primary veins, the outermost of which merge with the
ﬁmbrial vein (ﬁg. 5a, arrows), while the inner three veins en-
ter the leaf tip. There are six to 10 secondary veins per centi-
meter diverging at 70°–80°near the center of the leaf and
80°–90°near the margin (table 2). Secondary veins have
a curving course, and some appear straight or slightly sinusoi-
dal. Major and minor (thick and thin) secondary veins occur,
show a very irregular course, and can anastomose and dichot-
omize (table 2). Tertiary veins are present, and tertiary and
possible quaternary veins form irregular, polygonal areolae.
Leaves of Echinodorus grandiﬂorus (Chamisso et Schlech-
tendal) Micheli (Alismataceae) are ovate with cordate bases
and convex apices (table 2). They have 15–17 primary veins,
the outermost of which merge with the ﬁmbrial vein, whereas
the inner three extend into the leaf apex (table 2). The most
basal primary veins are very weak in this taxon but run the
same course as the stronger primary veins. There are eight to
11 secondary veins per centimeter diverging at angles of 60°
near the midvein and 80°–90°near the leaf margin (table 2).
Secondary veins have a slightly curving course, and some are
straight or sinusoidal. Major (ﬁg. 5b,atx) and minor (ﬁg.
5b,ato) secondary veins are present. The minor secondary
veins often show a sinuous course, while the stronger sec-
ondary veins have a straighter course (ﬁg. 5b). Tertiary veins
usually arise perpendicular to secondary veins, often di-
chotomize, and produce irregular areoles (ﬁg. 5b). Quaternary
veins do occur but are rare (table 2).
Echinodorus subalatus (Mart.) Griseb. (Alismataceae)
leaves are ovate with convex bases and straight to slightly con-
vex apices (table 2). There are nine primary veins, the outer
three on each side merge with the margin. There are eight or
nine secondary veins per centimeter, which diverge at 50°–55°
Fig. 2 Cardstonia tolmanii Riley et Stockey leaves. Scale bar ¼0:5 cm. a, Holotype showing entire margin, rounded apex, cordate base, and
three closely spaced medial veins. All primary veins appear to diminish in thickness toward apex. Note primary veins recurved in base and
continuing to apex without merging with marginal vein. Overlapping lobes of the lamina (arrow) are probably an artifact of deposition. Holotype
S55138A. b, Large lamina showing entire margin, cordate base with single lobe, and three closely spaced medial veins. Note steep angle of
divergence of secondary veins near midvein and marginal vein dichotomizing in base (arrow) giving rise to additional primary veins; possible
herbivory (*). S52263A. c, Large lamina showing petiole/lamina interface descending into matrix between cordate lobes. Note basally thickened
primary veins on each side of three closely spaced medial veins; arrow indicates dichotomy of primary medial vein. S50947A. d, Cordate base
showing petiole descending into matrix at 30°, recurved primary veins radiating near petiole/lamina interface, and large number of secondary
veins per centimeter. Note marginal vein dichotomizing and giving rise to primary veins (arrows). S52279A.
RILEY & STOCKEY—CARDSTONIA TOLMANII GEN. ET SP. NOV.
near the center and margins of the leaf. These secondaries
are often not straight but run a sinusoidal course in the leaf
lamina. Major and minor secondary veins are present, di-
chotomize frequently, and anastomose occasionally. Minor
secondary veins occur irregularly and frequently curve api-
cally, terminating in the above major secondary before reach-
ing the adjacent primary. Tertiary veins are perpendicular, or
nearly so, to the secondaries. They have a straight to sinuous
course and dichotomize toward the midvein. Fourth-order
veins are present (table 2) and form irregular, rectangular to
rarely triangular areolae.
Leaves of Aponogeton madagascariensis L. f. (Aponogeto-
naceae) are oblong to slightly obovate with asymmetrical de-
current bases and emarginate apices (table 2). The leaves
examined here were submerged and have a fenestrate lamina.
There are 13 primary veins, all of which merge with other pri-
mary veins from the outside in, but not with the marginal
vein. This anastomosing sequence starts with the outermost
primaries merging with adjacent primaries just below the apex
continuing admedially. There are three closely spaced medial
veins that run the entire length of the lamina. The number of
secondary veins per centimeter is ﬁve or six (table 2). Second-
ary veins are of uniform thickness, do not anastomose or di-
chotomize, and diverge at 80°–90°near the midvein and the
margins. There are no tertiary veins in this species.
Ottelia ulvifolia (Planch.) Walp. (Hydrocharitaceae) have
leaves that are oblong, somewhat linear, with cuneate bases
and straight apices (table 2). All the leaves of this species that
we examined were growing submerged. There are seven to
19 primary veins, the outermost of which merge with the
ﬁmbrial vein. The inner primary veins continue to the apex
and join in the leaf tip. There are only three or four second-
ary veins per centimeter (table 2). The angle of divergence of
secondary veins is 60°–90°in the center of the leaf and 60°–
90°near leaf margins. There are major and minor secondary
veins; however, secondary veins rarely anastomose or dichot-
omize. Third- and possible fourth-order veins make up rect-
angular to elongate polygonal areoles, especially near the
midvein. Areoles are larger near the midvein and decrease in
size toward the leaf margin.
Leaves of Butomopsis latifolia (D. Don) Kunth (Limno-
charitaceae) are elliptical with cuneate bases and straight to
acuminate apices (table 2; ﬁg. 6e,6g). There are seven to
nine primary veins, the outermost of which merge with the
adjacent primary vein near the leaf apex (ﬁg. 6g). There are
10–12 secondary veins per centimeter that diverge at angles
of 40°–50°near the midvein and 80°–85°near the margin
(table 2; ﬁg. 6f–6h). Major and minor secondary veins are
present (ﬁg. 6g). Major secondary veins run a fairly straight
course, dichotomize occasionally, and anastomose rarely. Mi-
nor secondary veins appear irregularly between major sec-
ondary veins and run a weak sinusoidal course (ﬁg. 6h);
these commonly dichotomize and anastomose, with branches
occasionally curving toward the leaf apex and terminating in
the adjacent secondary. Tertiary veins form well-developed
four to many-sided polygonal areoles (ﬁg. 6f,6h). Areoles
are vertically elongate near the center of the leaf (ﬁg. 6f) and
horizontally elongate near the leaf margins (ﬁg. 6h). Buto-
mopsis leaves, like those of Limnocharis (below) are ﬂeshy,
especially near the center.
Hydrocleys martii Seubert (Limnocharitaceae) leaves are el-
liptic with cordate bases and rounded apices (table 2; ﬁg. 5e).
There are 11 primary veins that continue to the apex but do
not merge with adjacent primaries or the ﬁmbrial vein (ﬁg.
5e; table 2). Leaves show three closely spaced medial veins
(ﬁg. 5f). The two adjacent primaries thin toward the leaf apex
and do not reach the apex. One leaf showed two primaries
anastomosing beneath the apex (ﬁg. 5e, arrow). The primary
veins merge at the apex, forming an apical ring (ﬁg. 5e).
There are six to 20 secondary veins per centimeter (far fewer
secondary veins near the midrib) that arise at angles of 70°
near the midvein and 90°at the margin (table 2). Major and
minor secondary veins are present (table 2), and both are
more or less straight. Tertiary veins are mostly perpendicular
to the secondary veins and form regular, polygonal areoles
that are elongated vertically near the midrib (ﬁg. 5f) and hori-
zontally near the leaf margins. There are no quaternary veins.
Leaves of Limnocharis ﬂava (L.) Buchenau (Limnocharita-
ceae) are elliptical with cordate bases and rounded apices
(ﬁg. 5g,5h). There are 11–13 primary veins, 10–12 of which
merge with adjacent primaries near the apex (normally
within 1 mm of the apex). The midvein is the only primary
vein that does not merge with other primaries. The primary
veins do not merge with the ﬁmbrial vein. There are 11–13
secondary veins per centimeter that diverge at angles of 60°–
80°near the midline and 90°near the margin. Major and mi-
nor secondary veins are present and consistently alternate
with each other in the ABAB pattern (Hickey and Peterson
1978). The major secondary veins run a straight course and
occasionally dichotomize but do not appear to anastomose
Fig. 3 Cardstonia tolmanii Riley et Stockey leaves. a, Lamina showing convex to rounded leaf apex. Scale bar ¼5 mm. S52268A. b, Apex
showing primary veins converging at what appears to be dark apical gland. Note primary veins occasionally anastomosing with adjacent primary
near apex (arrows), but absence of primary veins merging with marginal vein. Scale bar ¼2 mm. S52266A. c, Middle laminar area showing three
closely spaced medial veins (bracket), angle of divergence of secondary veins, and ﬁnely preserved internal tissue (arrow), possibly aerenchyma.
Note increase in number of secondary veins per centimeter toward margin. Scale bar ¼1 mm. S52263A. d, Leaf margin showing secondary veins
that anastomose and dichotomize (arrows), that appear to alternate irregularly between major and minor veins, and that run perpendicular to the
primary veins. Note marginal vein. Scale bar ¼1 mm. S52272. e, Base of leaf (ﬁg. 2c) showing three closely spaced medial veins and apparent
palmate radiation of primary veins (arrow). Scale bar ¼1 mm. S50947A. f, Venation from cshowing alternation between major and minor
secondary veins. Major veins can be either opposite or alternate between primaries. Note irregular branching pattern of tertiary veins (arrows).
Scale bar ¼1 mm. S52263A. g, Venation showing anastomosing and dichotomizing of major and minor veins. Note rare square-shaped areole
(arrow). Scale bar ¼1 mm. S52263A. h, Venation showing extensive anastomosing and dichotomizing of primarily minor veins. Note tertiary
veins crossing between adjacent secondary veins (arrows). Scale bar ¼1 mm. S50989. i, Possible aerenchyma. Note several lacunae appear to
overlap secondary veins (arrows). Scale bar ¼0:5 mm. S52263A.
RILEY & STOCKEY—CARDSTONIA TOLMANII GEN. ET SP. NOV.
(ﬁg. 5h). The minor secondary veins regularly dichotomize
and anastomose, running a weak sinusoidal course (ﬁg. 5h).
Tertiary veins are present and frequently dichotomize and
anastomose (ﬁg. 5h), forming a network of polygonal areoles
that appear to be concentrated on the adaxial side of the leaf
lamina. Near the center of the leaf on the abaxial surface is
another series of tertiary veins forming regular, polygonal, re-
ticulate areolae. In herbarium specimens it is necessary to use
both transmitted and reﬂected light of various intensities to
see these vein patterns clearly. Leaves of L. ﬂava are ﬂeshy
and full of aerenchyma, especially near the center.
Limnocharis laforestii Duchessaing (Limnocharitaceae)
leaves are oblong with cuneate bases and acuminate apices
(ﬁg. 6a,6c; table 2). There are 11–13 primary veins, with the
three adjacent primary veins on each side of the three medial
veins merging near the apex (ﬁg. 6a–6c). The primary veins
do not merge with the margin. The two medial veins that
run adjacent to the midvein are noticeably thinner (ﬁg. 6b)
and do not reach the apex (ﬁg. 6c). There are eight to 11 sec-
ondary veins per centimeter that diverge at angles of 65°–75°
near the midvein (ﬁg. 6c) and 75°–90°near the margin (ﬁg.
6d; table 2). Major and minor secondary veins regularly al-
ternate and run straight to slightly sinusoidal courses while
occasionally dichotomizing but rarely anastomosing (ﬁg. 6c).
Tertiary veins appear random reticulate; i.e., they anastomose
with other tertiary or secondary veins at random angles (Leaf
Architecture Working Group 1999), forming well-developed
polygonal areolae (ﬁg. 6b,6d). As in L. ﬂava, there is a sec-
ond set of overlapping regular polygonal reticulate veins in
the center of the leaf near the abaxial leaf surface (ﬁg. 6d).
These veins arise from the primary and secondary veins.
Quaternary veins are present, forming freely ending veinlets
that appear unbranched and one-branched (ﬁg. 6d).
Leaves of Potamogeton lucens L. (Potamogetonaceae) are
elliptical with decurrent bases and acuminate to convex api-
ces (table 1). There are nine primary veins that merge with
adjacent primaries near the leaf apex. There is a single mid-
vein, and the adjacent primaries are widely spaced from this
medial vein. This species shows minute serrations at the leaf
margin and a distinct ﬁmbrial vein. Other species have been
described with clearly serrate margins (e.g., Potamogeton
crispus L.) or what appear to be entire margins with serra-
tions that are not visible to the naked eye (e.g., Potamogeton
perfoliatus L.) (Cook 1996a). Primary veins do not merge
with the ﬁmbrial vein (table 1). The number of secondary
veins per centimeter is only six to 10, and the angles of diver-
gence of the secondaries are 40°–60°near the midvein and
60°–80°near the leaf margin (table 1). There are no major
and minor secondary veins. Secondary veins often anasto-
mose and dichotomize, and distinct tertiary veins occur at
right angles to the secondary veins. They often show a sinu-
ous course in between secondary veins but always arise per-
pendicular to the adjacent secondary vein. Tertiary veins also
Haemanthus katherinae Baker (Amaryllidaceae) leaves are
generally oblong with cuneate bases and acuminate tips (ta-
ble 1). There are 21 primary veins that do not merge with
the ﬁmbrial vein, and many of the primary veins merge near
the leaf apex. There are 11–23 secondary veins per centi-
meter with an angle of divergence of 60°near the center in-
creasing to 80°near the margin. Secondary veins often
dichotomize and sometimes anastomose. There is a common
occurrence of what could be termed ‘‘intersecondary veins’’
(Leaf Architecture Working Group 1999) that do not reach
the adjacent primary or the margin. These are slightly thinner
than the normal secondary veins, and they usually end
blindly. Haemanthus leaves also do not show a regular alter-
nation of thin and thick secondary veins (an ABAB pattern).
Occasional slightly thinner veins continue between the adja-
cent primary veins, but these are rare. Regular tertiary veins
occur rarely (table 1).
Stemona tuberosa Lour. (Stemonaceae) have ovate leaves
with cordate bases and acuminate tips (table 1). There are 11
primary veins; the outer three on each side merge with the leaf
margin. These primary veins continue along the margin and
then end in the margin. The outer primaries do not merge
with each other at the margin but disappear prior to reaching
the leaf apex. There is a single medial primary vein, and the
adjacent primaries are widely spaced from this medial vein.
These four adjacent primaries continue to the apex of the leaf
and enter the acuminate tip without merging. The number of
secondary veins per centimeter is 28–30, and the angles of di-
vergence of secondary veins are 90°both in the center of the
leaf and near the leaf margin (table 1). There appear to be ma-
jor and minor secondary veins in regular alternation between
adjacent primaries. These secondary veins can dichotomize
and anastomose. No tertiary veins were observed (table 1).
Comparison with Fossil Leaves
An entire margin, parallel petiolar veins, and campylodro-
mous primary venation (Leaf Architecture Working Group
Fig. 4 Haemanthophyllum kamtschaticum Budantsev (a–e) and Haemanthophyllum cordatum Golovneva (f–h) leaves. a,H. kamtschaticum,
generitype, leaf fragment showing cordate base, recurved primary veins, dense secondary veins, and marginal vein (arrow). Scale bar ¼0:5 cm.
960-1/2091 BIN RAS. b,Haemanthophyllum kamtschaticum, generitype, enlarged lamina from ashowing secondary veins dichotomizing and
anastomosing. Scale bar ¼1 mm. 960-1/2091 BIN RAS. c,‘‘Haemanthophyllum kamtschaticum’’ leaf fragment showing long thickened petiole,
cordate base, and large laminar size. Scale bar ¼3 cm. 960-1/2092 BIN RAS. d,Haemanthophyllum kamtschaticum, generitype, lamina from
aenlarged to show secondary vein density and marginal vein. Scale bar ¼2 mm. 960-2/2091 BIN RAS. e,Haemanthophyllum sp. (from type
locality) lamina showing areolae. Scale bar ¼1 mm. 960-1/3017 BIN RAS. f,Haemanthophyllum cordatum, holotype, lamina showing rounded
apex and single lobe of cordate base. Scale bar ¼2 cm. 967A/101 BIN RAS. g,Haemanthophyllum cordatum leaf apex from f. Scale bar ¼0:5
cm. 967A/101 BIN RAS. h,Haemanthophyllum cordatum leaf base from fshowing three closely spaced medial veins (bracket), recurved primary
veins, and angle of divergence of secondary veins. Note radiating primary veins near petiole attachment (arrow). Scale bar ¼0:5 cm. 967A/101 BIN
RILEY & STOCKEY—CARDSTONIA TOLMANII GEN. ET SP. NOV.
1999) of the fossil leaves from Cardston indicate that this
plant is a monocotyledon. The ovate blade with a large num-
ber of primary veins originating at or near the leaf base and
converging at or near the apex, dense secondary veins, and
thin, higher-order veins forming polygonal areolae (Golov-
neva 1997) are similar to those in leaves described as Hae-
manthophyllum Budantsev (1983). Budantsev (1983) erected
Haemanthophyllum based on 20 fossil leaves collected from
several localities in Kamchatka, Russia. These localities range in
age from Late Paleocene/Early Eocene (Anadyrka Formation)
to Middle to Late Eocene (Tkaprorayam Formation). The speci-
mens Budantsev illustrated, and the one that we studied that he
did not illustrate, were incomplete fragments. From these frag-
ments of apices, bases, and margins of different ages, Budantsev
(1983) constructed the concept of the genus Haemanthophyl-
lum. Subsequent authors placed other species of varying geo-
logic ages and localities in this genus due to morphological
similarities (table 1). This resultedinadiversearrayoftaxabe-
ing assigned to Haemanthophyllum based on this questionable
association of leaf parts.
Five species of Haemanthophyllum are currently recognized
by Golovneva (1997): Haemanthophyllum kamtschaticum
Budantsev, Haemanthophyllum zhilinii (Pneva) Golovneva,
Haemanthophyllum nordenskioldii (Heer) Boulter et Kva
Haemanthophyllum cordatum Golovneva, and Haemantho-
phyllum sp. from Ireland (table 1). The remaining species as-
signed to Haemanthophyllum resemble H. kamtschaticum
(table 1, Haemanthophyllum sp. 3, 5) or H. nordenskioldii
(table 1, Haemanthophyllum sp. 1, 4) (Golovneva 2000).
The holotype for H. kamtschaticum (the generitype) is
a fragment of one lobe of a cordate base from the Anadyrka
Formation of the Late Paleocene/Early Eocene (ﬁg. 4a,4b,
4d). The thickness, branching pattern, and density of the pri-
mary and secondary veins are similar to our leaves; however,
without an attached apex, midrib, or petiole, it is difﬁcult to
compare H. kamtschaticum with our leaves morphologically.
And as our study shows, widely disparate monocot families
may show similar venation patterns if only portions of one
leaf are examined. The second fragmentary base with a petiole
from the same locality, assigned to H. kamtschaticum (Bu-
dantsev 1983), is also without an apex and most of the lam-
ina (ﬁg. 4c). The primary veins are barely visible, but they
appear to be fewer in number, and the angle at which they re-
curve is shallower than in the holotype specimen. Although
a drawing of the secondary veins of this specimen appears in
Golovneva (1997, ﬁg. 2a), we were unable to observe the sec-
ondary veins in the actual specimen. Because of the state of
preservation, it is not possible to determine whether the holo-
type and this second specimen belong to the same taxon.
The remaining leaf fragments illustrated by Budantsev
(1983, pl. 63, ﬁgs. 1, 4) and Golovneva (1997, ﬁg. 2b–2d)
are from localities within the Tkaprovayam Formation that
are younger (Middle to Late Eocene) than the previously dis-
cussed specimens from the Anadyrka Formation. These speci-
mens are also unconnected leaf fragments of varying shape
with a ﬁmbrial vein and numerous primary and secondary
veins (Budantsev 1983; Golovneva 1997). The fragmentary
nature of the specimens, and the locality and age differences
preclude the assignment of these specimens to H. kamtschati-
cum. It appears to us that there may be several types of
monocot leaves present in these localities and better speci-
mens are needed to assess their afﬁnities.
Leaf fragments from three different localities (Ashutas
Mountain, Zhayremsk Quarry, and Kinyak) of comparable
ages (Late Oligocene to Early Miocene) from Kazakhstan
have been suggested to represent a single species, H. zhilinii
(Golovneva 1997). Previous workers have assigned fragmen-
tary specimens from Ashutas Mountain to Alisma macrophyl-
la Heer (Kryshtofovich et al. 1956) and Aponogeton zhilinii
Pneva (1988). Zhilin (1974) described a small laminar frag-
ment from Kinyak as Aponogeton, which Pneva sp. (1988)
later included in A. zhilinii. All of the illustrated fossils show
some similarities in primary and secondary venation and den-
sity to Haemanthophyllum, and certain specimens appear sim-
ilar to H. kamtschaticum (Golovneva 1997, ﬁg. 5a,5c).
However, the fragmentary nature of the material and lack of
clear morphological characters (i.e., laminar characters, base
and apex shape, and overall leaf size) leave their taxonomic
afﬁnities in doubt.
Golovneva, after reexamining Heer’s specimens in Stock-
holm, suggests that leaf specimens from the early Paleocene
Barentsburg Formation and the Eocene Skilvika Formation
of Spitsbergen be assigned to H. nordenskioldii (Heer) Boul-
ter et Kva
cek (Golovneva 1997; our table 1). Heer (1868)
originally assigned fragmentary specimens from the Bare-
ntsberg Formation (Heer 1868, pl. 30, ﬁgs. 1b,5b,6a, 7; Go-
lovneva 1997, ﬁg. 1b) and several fragmentary specimens
from the Skilvika Formation (Heer 1876, pl. 27, ﬁgs. 1, 2,
3a)toPotamogeton nordenskioldii based on small laminae
(relative to Alisma macrophylla) and convex (blunt) apices.
Fig. 5 Selected extant taxa of Alismatales. a,Echinodorus glaucus Rataj leaf showing round apex, cordate base, recurved primary veins in
base, and secondary veins nearly perpendicular at center. Note outer primary veins merge with margin before reaching apex (arrows). Scale
bar ¼3 cm. b,Echinodorus grandiﬂorus, abaxial surface, lamina showing many linear secondary veins (x’s) and fewer irregular sinusoidal
secondary veins (o’s). Note numerous linear and branched tertiary veins running perpendicular to and terminating in adjacent secondary veins.
Scale bar ¼5 mm. c,Caldesia sp., leaf showing convex apex, deeply cordate base, recurved primary veins. Note outer primary veins merging with
marginal vein subapically (arrows). Scale bar ¼1 cm. d,Caldesia sp., leaf showing major (x) and minor (0) secondary veins. Note dichotomizing
of major secondary veins (arrows). Scale bar ¼1 mm. e,Hydrocleys martii Seubert, leaf showing rounded apex, cordate base, and recurved
primary veins terminating in an apical gland. Note adjacent primary veins anastomosing (arrow); primary veins do not merge with marginal vein.
Scale bar ¼5 mm. f,Hydrocleys martii, abaxial surface of leaf base showing three closely spaced medial veins and angle of divergence of
secondary veins. Note numerous tertiary veins forming distinct areoles. Scale bar ¼2 mm. g,Limnocharis ﬂava (L.) Buchenau, abaxial surface of
leaf showing rounded apex, cordate base, and recurved primary veins terminating in an apical gland. Note three closely spaced medial veins in
base of leaf. Scale bar ¼4 cm. h,Limnocharis ﬂava, abaxial surface, lamina showing major (x) and minor (o) secondary veins branching and
irregular branching of tertiary veins. Scale bar ¼3 mm.
RILEY & STOCKEY—CARDSTONIA TOLMANII GEN. ET SP. NOV.
He assigned the remaining specimens from the Skilvika For-
mation to A. macrophylla based on large laminae and acumi-
nate leaf tips (Heer 1876, pl. 26, ﬁgs. 1–6; pl. 27, ﬁgs. 3b,
3c, 4–7; Golovneva 1997, ﬁg. 1a,1c). Golovneva (1997) sug-
gested that both leaf types belong to the same species of Hae-
manthophyllum [now called H. nordenskioldii (Heer)
Boulter et Kva
cek]. However, the assignment of these frag-
mentary plant remains of varying morphology and age to
Haemanthophyllum is tenuous. In addition, the leaves from
Spitsbergen, unlike our leaves, have larger laminae, cuneate
bases, attenuate to convex apices, lack three closely spaced
medial veins in the leaf base, and have only seven to nine
cross veins per centimeter (table 1).
Boulter and Kva
cek (1989, ﬁg. 22D, 22E; table 1, Haeman-
thophyllum sp. 2) originally assigned fragmentary specimens
with narrow oblong leaves and parallelodromous primary ve-
nation from the Late Paleocene/Early Eocene of Ireland to H.
nordenskioldii. Golovneva (1997) concluded that the leaf
characters were sufﬁciently different from the Spitsbergen
material and recommend that they be removed from H. nor-
denskioldii but recognized that they may represent a new
species of Haemanthophyllum (Golovneva 1997; L. B.
Golovneva, personal communication, 1999). Again, without
an apex or base it is difﬁcult to classify the Irish material.
The Irish leaves differ in morphology from the Cardston
leaves but resemble fragmentary remains described by Zhilin
(1974) as Aponogeton tertiarius Zhil. These leaves may show
afﬁnities to extant Aponogetonaceae or Potamogetonaceae.
Other leaf fragments sometimes assigned to Haemantho-
phyllum include a specimen with a strong ‘‘triple midvein,’’
representing the middle to basal part of a leaf from the Early
Paleocene of the Tsagajan Formation in the Amur Region of
Russia. This leaf appears similar to specimens described as
H. nordenskioldii (Krassilov 1976, pl. 11, ﬁgs. 3, 4; table 1,
Haemanthophyllum sp. 1). The specimen is fragmentary,
lacking an apex and base (Golovneva 1997 suggests the base
is cuneate). The secondary veins appear to be as numerous as
25 per centimeter, slightly greater than the 14–17 reported by
Golovneva (1997), and diverge at nearly 90°near the mid-
vein. The difference in the angle of divergence of secondary
veins and the lack of a cordate base with recurved primary
veins distinguish these specimens from the Cardston leaves.
There are several known occurrences of Haemanthophyl-
lum-like leaves from North America (table 1). Brown (1962,
pl. 15, ﬁgs. 1, 4, 6) described fragmentary specimens of
monocot leaves with large blades from the Paleocene Fort
Union Formation in North and South Dakota as Alismaphyl-
lites grandifolius (Penhallow) Brown. These leaves from two
different localities were combined by Golovneva (1997, ta-
ble) and described as Haemanthophyllum sp. 3. An acumi-
nate leaf apex, described as an unidentiﬁed monocot by
McIver and Basinger (1993) from the Paleocene Ravenscrag
Formation of Saskatchewan, Canada, has been referred to as
Haemanthophyllum sp. 5 by Golovneva (1997; our table 1).
Two leaf fragments also similar to Haemanthophyllum have
been described from the Paleocene Sagavanirktok Formation
of northern Alaska (Spicer et al. 1994). One is a fragmentary
cuneate leaf base (Spicer et al. 1994, ﬁg. 3.2); the other is
a fragment of a rounded apex (Spicer et al. 1994, ﬁg. 3.3).
These specimens are referred to as Haemanthophyllum sp. 4
(Golovneva 1997; our table 1). It is quite possible that these
are fragments of two different taxa. All of the specimens are
too poorly known to assign them with certainty to Haeman-
thophyllum or the fossil leaves described in this article, and
are in need of reinvestigation once more specimens are found.
The Cardston fossil leaves are most similar to H. cordatum
(Golovneva 1987). The overall shape, size, number of pri-
mary veins, merging of several primary veins below the apex,
and angles of divergence of secondary veins are similar; how-
ever, there are several notable differences. Golovneva (1987,
ﬁg. 1) illustrates the outermost primary veins merging with
the margin. In the Cardston leaves the primary veins con-
tinue to the apex without merging with the margin. This is
a key character that distinguishes the Russian specimens
from our fossils. An apical pore is present in the Cardston
leaves, but because of poor preservation, its presence or ab-
sence is unknown in the Russian material. A larger number
of secondary veins per centimeter is seen in the Cardston
leaves (up to 40) than in H. cordatum (13–18). Even consid-
ering variability in preservation and our counts of seven to
25 veins per centimeter on the holotype specimen of H. cor-
datum, it appears that the secondary veins have a greater
density in the Cardston leaves (table 1). In addition to these
morphological differences, the Cardston leaves are older than
those of H. cordatum by at least 5 m.yr., based on Russian
stratigraphy (table 1).
Comparison with Extant Monocots
Budantsev (1983) based the name Haemanthophyllum on
similarities of the fossil leaves to leaves of Haemanthus L.
(Amaryllidaceae). Several workers concluded that these
leaves were more closely related to Aponogeton (Zhilin
1974; Golovneva 1987, 1997; Pneva 1988; Boulter and Kva
1989) or Potamogeton (Heer 1868, 1876; Krassilov 1976;
Boulter and Kva
cek 1989). Heer (1876), Krystofovich et al.
Fig. 6 Leaves of Alismatales. a,Limnocharis laforestii Duchessaing, whole leaf. Scale bar ¼2 cm. b,Limnocharis laforestii, adaxial surface of
leaf base showing three closely spaced medial veins and random reticulate tertiary veins (arrow). Scale bar ¼5 mm. c,Limnocharis laforestii, leaf
apex showing primary veins merging near apex (arrow) and major (x) and minor (o) secondary veins. Scale bar ¼1 cm. d,Limnocharis laforestii,
abaxial leaf surface showing abaxial tertiary veins (arrowheads) that are slightly darker than abaxial tertiary veins. Note freely ending veinlets (4°
veins) on adaxial side (arrow) and marginal vein. Scale bar ¼2 mm. e,Butomopsis latifolia (D. Don) Kunth, whole leaf. Scale bar ¼3 cm. DNA
1759. f,Butomopsis latifolia, adaxial leaf surface showing secondary veins (arrow) and tertiary veins with vertical elongate areoles over midvein
(arrowhead). Scale bar ¼3 mm. DNA 23463. g,Butomopsis latifolia, adaxial leaf apex showing primary and secondary vein patterns. Scale
bar ¼5 mm. DNA 23463. h,Butomopsis latifolia, leaf margin showing major and minor secondary veins and tertiary veins with horizontally
elongate areoles. Note marginal vein. Scale bar ¼2 mm. DNA 23463.
RILEY & STOCKEY—CARDSTONIA TOLMANII GEN. ET SP. NOV.
(1956), Brown (1962), and Golovneva (1997) also compared
fossil leaf fragments with those of Alismataceae. Our com-
parisons further include Hydrocharitaceae, Limnocharita-
ceae, and Stemonaceae (table 2).
Fossil leaves from Cardston differ from those of Haeman-
thus in general leaf shape. Haemanthus leaves are oblong
with cuneate bases and acuminate tips, whereas the fossils
are ovate with cordate bases and convex to rounded apices.
A number of characters are, however, shared by these two
taxa, including the number of primary veins and primary
veins that merge near the leaf apex and not with the margin.
The number of secondary veins per centimeter overlap be-
tween these two taxa, but the fossil leaves have greater num-
bers of secondary veins per centimeter, especially near the
margin. The angles of divergence of the secondary veins are
similar; however, angles near the center in the fossil leaves
can be more acute (table 2). The fossil leaves show regular
major and minor secondary veins. Minor secondary veins in
Haemanthus are rare. The midvein is undulating in Haeman-
thus, unlike that seen in the fossil leaves. Secondary veins in
Haemanthus may dichotomize and anastomose as in the fos-
sil leaves, but common intersecondary veins occur in this ge-
nus. These do not occur in the fossil leaves. Tertiary veins are
very rare in Haemanthus leaves but are relatively more com-
mon in the fossil leaves from Cardston.
Leaves of Stemona are ovate with cordate bases like those
of the fossils, but with acuminate apices rather than convex to
rounded, as in the fossils (table 2). The number of primary
veins is small (11) compared to the 23–27 veins seen in the
fossil leaves. The primary veins in Stemona reach the margin
and extend partway up the leaf in the margin (becoming what
might be considered a ﬁmbrial vein). However, in Stemona the
primary veins never merge in the margin. The lowermost pri-
mary disappears just as the inner adjacent primary vein ap-
proaches the margin and continues toward the apex as
a ﬁmbrial vein. The number of secondary veins per centimeter
is comparable between Stemona and Cardstonia, and their an-
gles of divergence near the margin are both 90°. However,
leaves of Stemona show angles of 90°throughout the leaf,
even near the midvein, while those of the fossil leaves are 45°–
60°near the center of the leaf. The secondary veins are similar
to those of Cardstonia with regular alternation of major and
minor secondaries that dichotomize and anastomose. Tertiary
veins, however, were not observed in leaves of S. tuberosa.
Leaves of Potamogeton lucens are elliptic with decurrent
bases and acuminate to convex tips. Leaf shape in the family
Potamogetonaceae is highly variable and often heteromor-
phic, with ﬂoating leaves that are broad and petiolate and
submerged leaves that are very thin to lanceolate or elliptic
(Cook 1996b). There are about 80–90 species of Potamoge-
ton, some of which show minute teeth and serrated margins
(Cook 1996a). The number of primary veins, however, is
much smaller than in the fossil leaves described here (table 2).
The primaries, like those in Cardstonia leaves, do not merge
with the ﬁmbrial vein but continue to the leaf apex. The num-
ber of secondary veins per centimeter in P. lucens is very
small, and there are no minor secondary veins. In addition,
the tertiary veins in P. lucens diverge at right angles and show
a sinuous course between secondary veins, and they rarely di-
Van Bruggen (1990) studied Aponogeton leaves in some
detail. We closely examined Aponogeton madagascariensis
(table 1) and Aponogeton ulvaceous in this study. Aponoge-
ton leaves are more oblong (linear) than those of the Card-
ston fossil taxon. While A. madagascariensis has a decurrent
base, the shape can vary in Aponogeton with some species
(e.g., Aponogeton cordatus Jumelle) developing a shallow
cordate base (van Bruggen 1990). The fossil leaves have
deeply cordate bases with lobes that extend for some distance
below the point of petiolar attachment, unlike any extant
species of Aponogeton. The leaf apex shape in Aponogeton
can be straight, convex, rounded, or emarginate (van Brug-
gen 1990). The fossil leaves show convex to rounded apices
that are similar to some Aponogeton species, e.g., Aponoge-
ton jacobsenii van Bruggen (van Bruggen 1990).
The 13 primary veins in A. madagascariensis are signiﬁ-
cantly fewer than the 23–27 observed in the fossil leaves.
This is also consistent with the typical 11–15 primary veins
reported by Tomlinson (1982) for the genus Aponogeton.
The primary veins do not merge with the ﬁmbrial vein but
do merge with adjacent primaries in Aponogeton and Card-
stonia.InAponogeton, the ﬁrst veins to merge near the apex
are the outermost pair of primary veins, each of which
merges with an adjacent primary. This pattern continues until
most of the primaries join at the leaf apex. Although adjacent
primaries do occasionally merge in the fossil leaves, there
does not appear to be a successive admedial pattern of
merger as seen in Aponogeton. The number of secondary
veins per centimeter is ﬁve to six in A. madagascariensis,
while the Cardston fossil leaves show a much greater density
(10–40). Secondary veins diverge at angles of 80°–90°in A.
madagascariensis at the center and margins of the leaf, while
those in Cardstonia leaves diverge at angles of 45°–60°in the
center and 90°at the margins. Aponogeton leaves lack major
and minor secondary veins, whereas there is a consistent al-
ternation of major and minor veins in the fossil leaves. Al-
though secondary veins are not seen dichotomizing or
anastomosing in A. madagascariensis var. madagascariensis,
they do dichotomize or anastomose occasionally in A. mada-
gascariensis var. henkelianus (van Bruggen 1990). Third-
order veins are normally absent in submerged leaves of
Aponogeton (Tomlinson 1982) but can form well-developed
regular areolae in emergent leaves, e.g., Aponogeton junceus
Lehm. (¼Aponogeton spathaceum E. Meyer). Third-order
veins in the Cardston fossil leaves are few in number and
form irregular areolae, unlike those of Aponogeton. An api-
cal pore, as in the Cardston fossil leaves, has been reported
in old leaves of Aponogeton (Tomlinson 1982).
Leaves of Ottelia ulvifolia (Hydrocharitaceae) differ in
shape from the ovate leaves from Cardston (with cordate
bases and convex to rounded apices) in being oblong to lin-
ear with cuneate bases and straight apices. The genus Ottelia
Pers. contains about 21 species (Cook 1996b), however, and
cordate bases have been reported in some taxa (Tomlinson
1982). There are fewer primary veins (seven to 19) than in
Cardstonia. The outer veins in O. ulvifolia merge with the
margin, unlike those in Cardstonia, which continue to the
apex. There are fewer secondary veins per centimeter (three
or four) than the 10–40 seen in the Cardston leaves. Angles
of divergence of secondary veins in Ottelia are 60°–90°and
912 INTERNATIONAL JOURNAL OF PLANT SCIENCES
45°–60°in Cardstonia in the center of the leaf and 60°–90°
and 90°near the margins, respectively. Major and minor sec-
ondary veins occur in both (table 2); however, secondary
veins rarely anastomose or dichotomize as they do in the
Cardston fossil leaves. The rectangular to elongate polygonal
areoles that are formed by third- and fourth-order veins in
O. ulvifolia are absent in the Cardston leaves.
The Cardston fossil leaves are more similar to those of
Alismataceae and Limnocharitaceae than to leaves of any of
the other families examined here. Ovate leaf shapes with cor-
date bases occur in Alismataceae, as in Cardstonia (table 2;
Meyer 1935b). Leaves of Alisma differ in leaf base shape,
lack of major and minor secondary veins, and the secondary
veins do not anastomose (table 2; Golovneva 1994). Those
of Caldesia have more acuminate tips and lack tertiary veins
(table 2; Mayr 1943). Similar to the fossil leaves, apices can
be convex to rounded in Alisma and some Echinodorus spe-
cies. There are fewer primary veins in Alismataceae (nine to
15) than in the fossils that show (23–27). In all the Alismata-
ceae examined, the outer primary veins merge with the ﬁm-
brial vein. In the fossil leaves, all primary veins continue to
the leaf apex before fusing. The number of secondary veins
per centimeter is far fewer in most taxa of Alismataceae, but
Echinodorus species and Alisma can reach the lower limits
seen in the fossils. Secondary venation in Caldesia is very
similar to that in the fossil leaves. Secondary veins in Echino-
dorus have a more curved course in the leaf than those seen
in the fossils. The species of Echinodorus that we examined
have fourth-order veins, and the areoles (formed by tertiary
and sometimes quaternary veins) are vertically elongate, un-
like the horizontally elongate areoles in the fossil leaves.
‘‘Weakly developed’’ apical pores have been reported in older
leaves of Echinodorus (Tomlinson 1982) but were not ob-
served by us in any of the leaves that we examined. Internal
aerenchyma tissues have been described in this family by
Meyer (1934, 1935a) and Stant (1964), and large lacunae are
present in the leaf midrib with much smaller air spaces in the
The fossil leaves from Cardston show the greatest similari-
ties to those of Limnocharitaceae. Hydrocleys has leaves that
are nearly ovate to suborbicular (Stant 1967; Cook 1996b).
In ‘‘dicot’’ leaf terminology (Leaf Architecture Working
Group 1999), the leaf illustrated in ﬁgure 5eis classed as el-
liptic. Bases are cordate as in the fossil leaves (Stant 1967).
The number of primary veins is only 11 in Hydrocleys com-
pared to 23–27 in Cardstonia. The primary veins do not
merge with the ﬁmbrial vein but do merge with other pri-
mary veins near the leaf apex as in Cardstonia. These leaves
show three closely spaced medial veins as in Cardstonia, but
the two veins on either side of the midvein do not reach the
leaf apex. The number of secondary veins per centimeter
overlaps that seen in the Cardston leaves but never reaches
the 40 per centimeter maximum seen in the fossils. Angles of
divergence of secondary veins are similar to those in the
Cardston leaves but generally greater near the center of the
leaf. Major and minor secondary veins are present that anas-
tomose and dichotomize as in the fossils, and the secondary
veins also have a more or less straight course. The tertiary
veins in Hydrocleys arise at nearly right angles from the sec-
ondary veins, and there is much more regular areolation than
in the fossils. There is an apical pore in Hydrocleys,asin
Cardstonia (Stant 1967).
Leaves of Butomopsis are elliptical with cuneate bases,
straight to acuminate apices, and are markedly different in
shape than leaves from Cardston (table 2; ﬁg. 6e). The num-
ber of primary veins in Butomopsis latifolia is only seven to
nine, the lowest count for the family and more similar to
some Alismataceae than the fossil leaves. The primary veins,
like those in the fossils and other Limnocharitaceae, do not
merge with the margin but merge at or near the apex. Angles
of divergence of secondary veins are very similar to those of
the fossil leaves. Major and minor secondary veins with
a more or less straight course that dichotomize and anasto-
mose (although rarely) are similar to the Cardston leaves.
Tertiary veins form somewhat irregular but well-developed
polygonal areoles. As in Hydrocleys, the areoles are more
elongate toward the center of the leaf than the margins.
Limnocharis laforestii leaves are oblong with cuneate bases
and acuminate apices and therefore differ in general shape
from the Cardston leaves (table 2). As in Hydrocleys, there
are 11–13 primary veins that do not merge with the margin,
and three closely spaced medial veins are seen near the leaf
base, but the two veins on either side do not reach the leaf
apex. The number of secondary veins per centimeter (eight to
11) overlaps the minimal number seen in the fossil leaves (ta-
ble 2). Angles of divergence of secondary veins are similar to
those seen in the fossil leaves but generally greater near the
center of the leaf. Major and minor secondary veins that run
nearly straight are similar to those seen in the Cardston
leaves. These leaves, however, have randomly reticulate ter-
tiary veins with well-developed polygonal areoles, unlike the
irregular areolation seen in Cardstonia. Quaternary veins
also occur in this species. These form freely ending veinlets,
like those reported in ‘‘dicot’’ leaves (Leaf Architecture Work-
ing Group 1999) and unlike those seen in any of the other
monocot taxa observed by us.
Leaves of Limnocharis ﬂava show the closest similarities to
the Cardston fossil leaves in many characters. The general
leaf shape with a cordate base and rounded apex are similar
between the two. While the number of primary veins is only
11–13 in L. ﬂava, the primary veins do not merge with the
ﬁmbrial vein but do merge with other primary veins beneath
the apex, and an apical pore is present (Stant 1967). The an-
gles of divergence of the secondary veins are very similar,
while slightly higher in the center of the leaf than in the
Cardston fossils (table 2). The number of secondary veins per
centimeter in L. ﬂava is at the low end for Cardstonia. Ter-
tiary veins are very similar to those in the Cardston leaves,
and they frequently dichotomize and anastomose. Irregular
polygonal areolae are present in L. ﬂava on the adaxial sur-
face, similar to those seen in Cardstonia.
We observed internal patterns of aerenchyma in all leaves
of Limnocharitaceae that we examined. These leaves are espe-
cially ﬂeshy near the leaf center, whereas the fossil leaves seem
to be ﬂeshy throughout. In all extant Limnocharitaceae, there
appears to be a set of tertiary veins that form larger areoles
on the abaxial surface of the leaf. These veins are not seen in
reﬂected light when adaxial leaf surfaces are examined. They
are best seen with a combination of transmitted and reﬂected
light on a dissecting microscope. Illustrations of venation of
RILEY & STOCKEY—CARDSTONIA TOLMANII GEN. ET SP. NOV.
Butomopsis (ﬁg. 6h), L. laforestii (ﬁg. 6d), Hydrocleys (ﬁg.
5h) are photographed in this type of light. It is difﬁcult to see
through the paper on herbarium sheets, and this type of ex-
amination of leaf venation is best done on pressed leaves that
are not mounted on sheets. The photograph of L. ﬂava (ﬁg.
5h) was taken with a predominance of reﬂected light. In all of
these leaves there is an abaxial set of tertiary veins that can be
seen to connect to the secondary veins in several places. If
these leaves were found as fossils, the predominant pattern
seen would depend on how the specimen was cracked open
and how close the fracture is to each surface. We illustrate
a similar tertiary vein pattern (ﬁg. 4e) in the Haemanthophyl-
lum sp. fragment from the type locality (one of the specimens
collected by Budantsev).
It should be noted that some authors might be observing
internal aerenchyma patterns as well as a second set of ter-
tiary veins in fossil leaves. Observations of the leaves of
‘‘Haemanthophyllum’’ in the fossil record have had various
interpretations of tertiary venation. Golovneva (1997, pl. 9,
ﬁg. 8) illustrates small polygonal areolae in H. kamtschati-
cum, and we show similar features in the Cardston leaves
(ﬁg 3i). We are interpreting these polygons that can overlap
veins in compressions (ﬁg 3i, arrows) as underlying aeren-
chyma. In some fossil specimens it may be extremely difﬁcult
to tell whether the patterns observed are underlying aeren-
chyma or tertiary venation patterns or both. The regularity
of the nearly hexagonal aerenchyma, if preserved, and its
overlap with overlying veins are seen when there is good fossil
preservation. Furthermore, the size ranges of the cells sur-
rounding the lacunae are about 50 mm, a reasonable size for
While the fossil leaves from Cardston described here have
similar numbers of primary veins to those of the genus Hae-
manthus, their venation details and presence of an apical
pore where the veins converge are most like leaves in Alis-
mataceae and Limnocharitaceae. The merging of primary
veins with the ﬁmbrial vein in most Alismataceae is a major
difference in leaf architecture. Thus, the fossil leaves are
most similar to those of Limnocharitaceae in overall vein
patterns and the presence of an apical pore where the pri-
mary veins converge and unite (Sauvageau 1891, 1893;
Stant 1967). Within Limnocharitaceae, L. ﬂava shows the
closest venation pattern, even to the order of tertiary veins,
to the leaves from Cardston.
The Cardston leaves have a unique combination of charac-
ters, including the large number of primary veins, the three
to ﬁve closely spaced medial veins that are continuous to the
leaf apex, and the irregular areolae formed by tertiary veins.
There are few secondary veins in the midrib region, while the
number dramatically increases toward the leaf margin. Thus,
the Cardston leaves are described as Cardstonia tolmanii sp.
nov., Limnocharitaceae, as they show many of the characters
of leaves in this family. They are most similar to those of L.
ﬂava differing only in the amount of aerenchyma present in
the leaf and the number of primary veins.
Limnocharitaceae have often been included in Butomaceae,
but a number of characters have suggested a relationship to
Alismataceae (Cronquist 1981). In cladistic analyses, Limno-
charitaceae and Alismataceae form a clade, with Butomaceae
either as the sister group or as the basal member of the sister
group (Haynes and Holm-Nielsen 1992; Les and Haynes
1995; Les et al. 1997; Soros and Les 2001, 2002). Morpho-
logical cladistic analyses (Haynes and Holm-Nielsen 1992)
indicate that Limnocharitaceae and Alismataceae may be sis-
ter groups, or that Alismataceae are nested within Limno-
charitaceae (Haynes and Holm-Nielsen 1992). In molecular
studies using rbcL, Limnocharitaceae is embedded within
Alismataceae, indicating that Alismataceae may be paraphy-
letic (Les et al. 1997). Data from the internal transcribed
spacers (ITS-1; ITS-2) of the nuclear ribosomal region, rbcL,
the ﬂanking introns and coding region of chloroplast matK,
and morphology all indicate a sister relationship between
Alismataceae and Limnocharitaceae (Soros and Les 2001,
Our reinvestigation of the genus Haemanthophyllum sug-
gests that this genus should probably be restricted to leaf
fragments that show cordate bases with similar venation pat-
terns to the generitype. Budantsev’s (1983) holotype speci-
men is incomplete. We show in this study that if only leaf
base fragments are known (i.e., the cordate lobes), leaves of
this type might be mistaken for those of Stemona or Calde-
sia, which show similar secondary vein characteristics. How-
ever, when whole leaves are found, these may prove to have
other afﬁnities. The suggestion that some leaves that have
been described as Haemanthophyllum are more similar to
Aponogeton is a valid one (Golovneva 1997), but the Card-
ston material is clearly different from Aponogetonaceae in
several characters. It becomes evident that most of the previ-
ously described species of Haemanthophyllum should be re-
examined, and larger numbers of more complete specimens
are needed from each particular locality before descriptions
are made and family assignments suggested. At the Cardston
locality we have a second monocot with broad leaves that
are elliptic with rounded bases and a distinctly different ter-
tiary venation pattern from that seen in Cardstonia that will
be the subject of another study.
This study also points out the need for a comprehensive
monocot leaf venation terminology. The Manual of Leaf Ar-
chitecture (Leaf Architecture Working Group 1999) covers
‘‘dicotyledonous’’ leaves and net-veined monocotyledonous
angiosperms. However, we ran into several problems in using
these terms for the monocot leaves described in this article.
While primary vein terminology will work with these leaves,
terminology for secondary and tertiary vein patterns is more
problematical. We have tried to describe these patterns
in several ways and to use the dicot terminology when possi-
ble. However, the distinctions between different leaf vena-
tion patterns are not always seen in table format using these
‘‘dicot’’ terms, even though we can distinguish the leaves
The preservation of fossil monocot leaves (other than those
of palms) is rare (Herendeen and Crane 1995). Hickey and
Peterson (1978) discussed the problems in the identiﬁcation
of fossil monocot leaf remains and proposed a terminology
that they used for zingiberalean monocots. Boyd (1992) sug-
gested additional vein patterns in several families of Zingiber-
ales. However, in some species of Alismataceae, venation is
much more complex. We found differing abaxial and adaxial
vein patterns; and we were only able to use this ‘‘bar code’’
terminology to a limited degree in our study for the veins
914 INTERNATIONAL JOURNAL OF PLANT SCIENCES
near the adaxial surface. The abaxial patterns are less like
those of zingiberalean monocots with anastomosing veins
that lack the ABAB pattern.
The St. Mary River Reservoir locality near Cardston, Al-
berta, has yielded a large number of extremely well-preserved
broad-leaved monocots in ﬁne-grained sediments. This re-
markable preservation has provided detailed data on the leaf
architecture of many taxa, including aerenchyma patterns
and internal anatomy in some rhizomes. Many plants are
also buried in situ, providing insights into the ancient habi-
tats and environments of the fossil taxa at the time of deposi-
tion. Further work at this site will provide a better
understanding of the diversity of these ancient aquatic com-
munities and their paleoenvironment.
We thank Georgia Hoffman, Trevor Lantz, Stefan Little,
John Priegert, Gar Rothwell, Jesse Rothwell, Rudolph Ser-
bet, Selena Smith, and Shayne Tolman for assistance in the
ﬁeld; Joseph Bogner, Munich Botanical Garden, for provid-
ing specimens and helpful discussion; Sean W. Graham, Uni-
versity of Alberta (ALTA), and Pat Herendeen, George
Washington University, for helpful discussion; Georgia Hoff-
man for sedimentary and stratigraphic interpretation; and
the Northern Territory Herbarium, Darwin, Australia, for
loan of herbarium specimens. This study was supported in
part by a grant from the Natural Sciences and Engineering
Research Council of Canada (A-6908 to R. A. Stockey).
APG (Angiosperm Phylogeny Group) 1998 An ordinal classiﬁcation
for the families of ﬂowering plants. Ann Mo Bot Gard 85:531–553
Boulter MC, Z Kva
cek 1989 The Paleocene ﬂora of the Isle of Mull.
Spec Pap Palaeontol 42:1–149.
Boyd A 1992 Musopsis n. gen.: a banana-like leaf genus from the
early Tertiary of eastern North Greenland. Am J Bot 79:1359–1367.
Brown RW 1962 Paleocene ﬂora of the Rocky Mountains and Great
Plains. Government Printing Ofﬁce, Washington, D.C. 119 pp.
Budantsev LU 1983 History of Arctic ﬂora of the early Cenophytic
epoch. Nauka, Leningrad. 155 pp. (In Russian.)
Cook CDK 1996aAquatic and wetland plants of India. Oxford
University Press, New York. 385 pp.
——— 1996bAquatic plant book.SPB Academic, Amsterdam.228 pp.
Cronquist A 1981 An integrated system of classiﬁcation of ﬂowering
plants. Columbia University Press, New York. 1262 pp.
Golovneva LB 1987 A new species of the genus Haemanthophyllum
from the Rarytkin series of Koryak Upland. Bot Zh 72:1127–1131.
——— 1994 Maastrichtian-Danian ﬂoras of Koryak Upland. Trudy
Botanicheskogo instituta im. VL Komarova, vol. 13. Rossiiskaia
Akademiia Nauk, St. Petersburg. 148 pp. (In Russian.)
——— 1997 Morphology, systematics and distribution of the genus
Haemanthophyllum in the Paleogene ﬂoras of the Northern
Hemisphere. Paleontol Zh 31:197–207.
——— 2000 Aquatic plant communities at the Cretaceous-Palaeogene
boundary in north-eastern Russia. Acta Palaeobot 40:139–151.
Hamblin AP 1998 Edmonton Group/St. Mary River Formation: sum-
mary of literature and concepts. Geol Surv Can Open File 3578:1–29.
Haynes RR, LB Holm-Nielsen 1992 Flora neotropica: monograph
56: the Limnocharitaceae. New York Botanical Garden, New York.
Heer O 1868 Flora fossilis arctica. 1. Die fossile Flora der Polar-
¨nder enthaltend die in Nordgro
¨nland, auf der Melville-Insel, im
Banksland, am Mackenzie, in Island und in Spitzbergen entdeckten
fossilen Pﬂanzen. Schulthess, Zurich. 192 pp.
——— 1876 Flora fossilis arctica. 4.1. Beitra
¨ge zur fossilen Flora
Spitzbergens. K Svenska Vetenskaps-Akademiens Handlingar 14:
Herendeen PS, PR Crane 1995 The fossil history of the monocoty-
ledons. Pages 1–21 in PJ Ruddall, PJ Cribb, DF Cutler, CJ
Humphries, eds. Monocotyledons: systematics and evolution. Vol
1. Royal Botanic Gardens, Kew.
Hickey LJ, RK Peterson 1978 Zingiberopsis, a fossil genus of the
ginger family from Late Cretaceous to Early Eocene sediments of
Western Interior North America. Can J Bot 56:1136–1152.
Kasselmann C 1995 Aquarienpﬂanzen. Ulmer, Stuttgart. 472 pp.
——— 2003 Aquarium plants. Translated by U Kotlenga, ASJ van
Proosdig. Krieger, Malabar, Fla. 518 pp.
Krassilov VA 1976 Tsagajan ﬂora of Amur Province. Nauka,
Moscow. 92 pp.
Kryshtofovich AN, IV Palibin, KK Shaparenko, AV Yarmolenko, TN
Baikovskaya, VI Grybov, IA Ilinskaya 1956 Oligocene ﬂora of the
Ashutas Mountain, Kazakhstan. Tr Bot Inst Akad Nauk SSSR.
Series 8, vol. 1. (In Russian.)
Kubitzki K 1998 The families and genera of vascular plants. IV.
Flowering plants: Monocotyledons: Alismatanae and Commelina-
nae (except Gramineae). Springer, Berlin. 511 pp.
Leaf Architecture Working Group 1999 Manual of leaf architecture:
morphological description and categorization of dicotyledonous
and net-veined monocotyledonous angiosperms. Smithsonian In-
stitution, Washington, D.C. 65 pp.
Les DH, MA Cleland, M Waycott 1997 Phylogenetic studies in
Alismatidae. II. Evolution of marine angiosperms (seagrasses) and
hydrophily. Am Soc Plant Taxon 22:443–463.
Les DH, RR Haynes 1995 Systematics of subclass Alismatidae:
a synthesis of approaches. Pages 353–377 in PJ Rudall, PJ Cribb,
DF Cutler, CJ Humphries, eds. Monocotyledons: systematics and
evolution. Royal Botanic Garden, Kew.
Mayo SJ, J Bogner, PC Boyce 1997 The genera of Araceae. Royal
Botanic Gardens, Kew. 370 pp.
Mayr F 1943 Beitra
¨ge zur Anatomie der Alismataceen: die Blattana-
tomie von Caldesia parnassifolia (Bassi) Parl. Beih Bot Zentbl 62A:
McIver EE, JF Basinger 1993 Flora of the Ravenscrag Formation
(Paleocene), southwestern Saskatchewan, Canada. Palaeontogr Can
Meyer FJ 1934 Beitra
¨ge zur Anatomie der Alismataceen. III, IV. Die
Blattanatomie von Lophotocarpus und Limnophyton. Beih Bot Zbl
¨ge zur Anatomie der Alismataceen. V. Die
Gattungen Damasonium und Alisma im Lichte der Anatomie. Beih
Bot Zbl 54A:156–169.
¨tter: als Beitrag zur Kenntnis der Bedingtheit
und der Leistungen der Leitbu
¨ndelverbindungen. Planta 23:557–592.
Nadon G 1988 Tectonic controls on sedimentation within a foreland
basin: the Bearpaw, Blood Reserve and St. Mary River Formations,
southwestern Alberta. Canadian Society of Petroleum Geologists,
Calgary. 84 pp.
Nadon GC 1994 The genesis and recognition of anastomosed ﬂuvial
deposits: data from the St. Mary River Formation, southwestern
Alberta, Canada. J Sediment Res 64:451–463.
RILEY & STOCKEY—CARDSTONIA TOLMANII GEN. ET SP. NOV.
Pneva GP 1988 A new tertiary species of Aponogeton (Aponogeto-
naceae) from Kazakhstan and Karakalpakia. Bot Zh 73:1597–1599.
Riley MG, RA Stockey 1998 An aquatic broad-leaved monocot from
the Upper Cretaceous (Maastrichtian) St Mary River Formation of
southern Alberta. Am J Bot 85(suppl):79–80.
——— 1999 Upper Cretaceous ﬂora of Cardston, Alberta. XVI Int
Bot Congr Abstr 965:453.
——— 2000 A new aquatic angiosperm with a ﬂoating rosette of
leaves from the St. Mary River Formation of southern Alberta. Am J
Rothwell GW, RA Stockey 1994 The role of Hydropteris pinnata
gen. et sp. nov. in reconstructing the phylogeny of heterosporous
ferns. Am J Bot 81:479–492.
Sauvageau MC 1891 Sur les feuilles de quelques monocotyle
aquatiques. Ann Sci Nat Bot 13:103–296.
——— 1893 Sur la feuille des Butome
´es. Ann Sci Nat Bot 17:295–326.
Soros CL, DH Les 2001 Phylogenetic relationships in the Alismata-
ceae. Botanical Society of America Abstracts in Botany, 2001
Conference, Albuquerque, N.Mex. p. 143.
——— 2002 Phylogenetic relationships in the Alismataceae. Botan-
ical Society of America Abstracts in Botany, 2002 Conference,
Madison, Wis. p. 152.
Spicer RA, KS Davies, AB Herman 1994 Circum-Arctic plant fossils
and the Cretaceous-Tertiary transition. Pages 161–174 in MC
Boulter, HC Fisher, eds. Cenozoic plants and climates of the Arctic.
NATO ASI series I, vol. 27. Springer, Berlin.
Stant MY 1964 Anatomy of Alismataceae. J Linn Soc Lond Bot 59:
——— 1967 Anatomy of Butomaceae. J Linn Soc Lond Bot 60:
Stockey RA, GW Rothwell 1997 The aquatic angiosperm Trapago
angulata from the Upper Cretaceous (Maastrichtian) St. Mary River
Formation of southern Alberta. Int J Plant Sci 158:83–94.
Tomlinson PB 1982 Anatomy of the monocotyledons. VII. Helobiae
(Alismatidae). Clarendon, Oxford. 559 pp.
van Bruggen HWE 1990 Die Gattung Aponogeton (Aponogetona-
ceae). VDA-Arbeitskreis Wasserpﬂanzen, Berlin. 84 pp.
Zhilin SG 1974 The Tertiary ﬂoras of Ust-Urt. Nauka, Leningrad.
121 pp. (In Russian.)
916 INTERNATIONAL JOURNAL OF PLANT SCIENCES