Two new thalassemyd turtles from the Cretaceous of Arkansas

41

SB

CM

LI B RARY

OF THE

U N I VLRSITY

Of ILLINOIS

550-5

FI

v.7-9

GEOLOGY

UNIVERSITY OF

ILLINOIS LIBRARY

AT URBANA-CH AMPAIGN

GEOLOGY

Totes? oi S . b0 ° k 0n or bef °re the
latest Dote stamped below

GEOLOGY LIBRARY

*-

c

5> l THE LIBRARY OF THE

FEB 22 1944

UNIVERSITY Of ILLINOIS

GEOLOGICAL SERIES

OF
FIELD MUSEUM OF NATURAL HISTORY

Volume 8 Chicago, January 21, 1944 No. 11

TWO NEW THALASSEMYD TURTLES FROM
THE CRETACEOUS OF ARKANSAS

By Karl P. Schmidt

Chief Curator, Department of Zoology

Discoveries of vertebrate fossils from the marine Cretaceous of
the Mississippi Embayment have been relatively few, as the nature
of the deposits and the relative rarity of vertebrate remains in them
has made it impractical for field parties from museums or universi-
ties to undertake organized collecting in the scattered outcrops.
Fortunately for the progress of paleontology in this region, Mr. C.
M. Barber, of Hot Springs, Arkansas, has interested himself in
fossil collecting and has spent week ends and other available time
in a systematic search of the Cretaceous outcrops in his state
for vertebrate remains. His finds include occasional mosasaurs and
plesiosaurs, fragments of crocodilians and fishes, and remains of
turtles. A most notable fossil turtle, Podocnemis barberi, was
described by the writer from his collections in 1940. It is gratifying
to be able to describe two more Cretaceous turtles from Arkansas
collected by Mr. Barber for the Museum of Comparative Zoology,
and acquired by Field Museum in exchange with that institution.
Mr. Barber has supplemented this specimen by gifts of more frag-
mentary turtles to Field Museum, which thus becomes the principal
repository for fossil turtles from the Arkansas segment of the Cre-
taceous of the Mississippi Embayment.

As in the case of the fossil Podocnemis previously described, Mr.
Barber has returned to the sites of his finds in successive seasons,
and has thus found additional pieces, sometimes even five years
after the original discovery, that fit neatly into gaps in the assembled
turtle shells. It is usually impossible to find the specimen in place;
fossil fragments are found washed out into the gullies of the pastured

No. 547 63

U.0FIO.IK*.

64 Field Museum of Natural History — Geology, Vol. 8

hillsides, and in the case of a good specimen it is desirable to return
at frequent intervals, especially after rains, to search for additions
to the fragmentary shell that has meanwhile been pieced together.
In the case of the interesting smaller turtle, described below as a
new genus, the specimen might have been much more complete
had not the owner of the farm on which it was found been seized
with the delusion that Mr. Barber was searching her pastures for
diamonds, and forbidden him to trespass on her property.

The assembling and mounting of the shells of the specimens here
described were done in Field Museum's Laboratory of Vertebrate
Paleontology by Mr. James H. Quinn, Chief Preparator. I am much
indebted to him for aid in interpreting certain elements of the shell
and in tracing the course of the sutures of the horny shields. His
skill in this tedious kind of preparation, in which patience is rewarded
by notably accurate restoration of the specimen, forms the basis for
the relatively simple task of comparison and description. The
geological and paleontological problems involved have been dis-
cussed with Mr. Bryan Patterson, Curator of Paleontology.

The turtles in question appear to be referable to the family Tha-
lassemydidae (as defined by Hay), and one of them to the genus
Catapleura Cope. Catapleura has hitherto been known only from
very fragmentary remains as Catapleura repanda and C. ponderosa
from the uppermost Greensands of New Jersey. While the New
Jersey Greensands are in part of Eocene age, as remarked by the
writer in comparing Podocnemis barberi with the pleurodire turtles
of New Jersey (Schmidt, 1940, p. 10), Mr. Patterson has pointed out
to me that the undoubtedly Eocene Manasquan Marl and Rancocas
group (including the Vincenttown Sand and the Hornerstown Marl)
are underlain by the Monmouth group, the Matawan group and the
Magothy formation, and that these are upper Cretaceous. Without
renewed and critical collecting of the Greensands vertebrates, which
will be extremely difficult unless active exploitation of the marl-pits
should be resumed, it is difficult to allocate many of the forms thus
far known correctly to the Cretaceous or Eocene. Thus, the presence
of Catapleura in the exactly determined Cretaceous horizon in
Arkansas bears on the stratigraphic problem in New Jersey. The
Thalassemydidae, appearing in the Jurassic of Europe, reach a con-
siderable development in the New Jersey Greensands, where no less
than five genera are represented. There is considerable divergence
of opinion as to the systematic position of these Cretaceous and
presumably Eocene forms.

FT

Arkansas Fossil Turtles 65

Only a few hundred yards from the find of the Catapleura, and
in the same year, Mr. Barber made an even more interesting turtle
discovery in a gullied field on the farm of Mrs. Cox, of Arkadelphia,
Arkansas. When the specimen in question finally came into the
possession of Field Museum, Mr. Quinn found the carapace nearly
complete in the elements essential to its restoration, yet exasper-
atingly incomplete in the region of the suprapygals and eighth
costals. Mr. Barber was accordingly urged to make a renewed
search for turtle-shell fragments at the original locality. After the
lapse of five years, these had become further dispersed and lost, but
by means of extensive sieving of the soil along the gully and from the
fan at its mouth, a considerable number of additional fragments were
obtained, including the complete seventh neural, parts of the second
suprapygal, and the base of the left seventh costal, together with
other pieces of costal.

The assembly of the great number of fragments of this smaller

specimen into a convincingly restored turtle shell has been a labor

of love, begun by Mr. Barber in 1938, continued by myself and Mr.

Quinn, with a few days' aid from my son John, on furlough, and with

j a final session during a visit to the Museum by Mr. Barber, in 1943.

j The resulting turtle shell is made up of curiously thin and dense
plates of apparently silicified bone. It has much larger fenestrae,
between the disk of the carapace and the peripherals, than in Cata-
pleura, resembling in this respect Lytoloma. It differs from Cata-
pleura in the suprapygal region, and from Lytoloma in the complete
contact of the first peripheral with the first costal. I have accord-
ingly distinguished it as a new genus.

Order Testudinata

Suborder Cryptodira

Family Thalassemydidae

- Phyllemys gen. nov.

Diagnosis. — A thalassemyd turtle with a very flat shell, with the
bones of the carapace and plastron very thin; carapace as wide as
long; eighth costals narrowed medially to a point; and large fenes-
trae between the peripherals and disk of the carapace.

Allied to Catapleura with which it agrees in the contact of the
first costal with the first and second peripheral; differing from
Catapleura in the broad free edge of the first suprapygal (adjacent

66 Field Museum of Natural History — Geology, Vol. 8

to the pygal); the medial narrowing of the eighth costals; the large,
nearly rectangular first neural; and in having the xiphiplastra
movably articulated to the hypoplastra in a straight joint instead
of by a firm angular suture. Distinguished from Lytoloma by the

Fig. 20. Carapace of Phyllemys barberi gen. et sp. nov.

arrangement of the first and second peripherals and the broad con-
tact of the first suprapygal with the eleventh peripheral on each side.
Type. — Phyllemys barberi sp. nov.

Phyllemys barberi sp. nov.

Holotype. — F.M. No. P27047, a carapace with about half the
elements preserved and a plastron complete except for the epiplastra
and entoplastron. Collected by Charles M. Barber, 1938.

Arkansas Fossil Turtles

67

Horizon and type locality. — Marlbrook Marl, Gulf Series, upper
Cretaceous, on the "Widow Cox farm," about one mile northeast of
the junction of the Hollywood-Okolona Road to Arkadelphia. This
is on the N.W. \i, sec. 28, T. 7 S., R. 20 W.

Fig. 21. Plastron of Phyllemys barberi gen. et sp. nov.

Diagnosis. — The characters of the genus serve to distinguish the
species from all known forms; the anterior border of the carapace is
little emarginate, the pygal not notched or slightly, fenestrae between
costal disk of carapace and peripherals large; shell notably thin and
dense.

Description of type. — General form nearly round, with little
anterior emargination ; shell very low, its depth little more than a
sixth of its length; plastron essentially flat. Nuchal wide. First
neural subquadrangular, meeting the second in a straight suture;

68 Field Museum of Natural History — Geology, Vol. 8

succeeding neurals 3-6 with concave anterior and convex posterior
border; seventh and eighth neurals reduced, as wide as long. Costals
strongly emarginate terminally, with a projecting rib-end to the
peripherals, which have deep corresponding pits.

Peripherals, except the first and second, deeply grooved on their
inner faces; rib-ends entering pits in peripherals 3-9, the pit for the
eighth rib being between peripherals ten and eleven.

The only trace of marginal grooves for horny shields is a trans-
verse groove on the third neural and adjacent portions of the third
costals.

Fig. 22. Shell of Phyllemys barberi, from left side.

In the absence of the eighth costals, eighth neural, and portions
of the second suprapygal, the restoration of this region of the cara-
pace is governed (1) by the adjustment of the peripherals to the
only arc in which they can fall; (2) by the arch of the costals and
neurals, which indicates a very flat shell; and (3) by the base of the
seventh left costal, which outlines part of the eighth neural, and con-
vinces us that the eighth costal is proximally much reduced. The
direction of its rib-end is shown accurately by the pit between the
tenth and eleventh costals.

First suprapygal (complete), with a broad thin free edge at each
side, broadly in contact with the eleventh peripheral on each side,
and meeting the second suprapygal in a broad convex arc. Second
suprapygal (partially restored) with thin latero-posterior border,
broadly notched in front for the eighth neural.

Plastron extremely like that of Porthochelys laticeps (Hay, 1908,
pi. 31, fig. 3) in general form, and like that of Catapleura arkansaw
(fig. 24), but apparently peculiar in having the xiphiplastra much less
firmly and perhaps movably united to the hypoplastra. No trace
of epiplastra or entoplastron was found.

Remarks. — The possible relation of the present turtle to Desmato-
chelys, distinguished as a distinct family by Williston, has been con-
sidered. Only fragments of the shell of Desmatochelys are known
(associated with an excellent skull). The pygal of Desmatochelys

Arkansas Fossil Turtles 69

lowi is very much thinner than in Phyllemys. Williston regards the
humerus as indicating a paddle-like forelimb, while the shell of
Phyllemys does not exhibit any antero-lateral excavation for a
paddle. The shell characters of Phyllemys, in general, point plainly
to relations with Catapleura. The typical material of the shell of
Desmatochelys lowi was made available to me for comparison with
the new form through the courtesy of Dr. C. H. Hibbard, of the
University of Kansas.

MEASUREMENTS

All measurements are made by caliper in millimeters, estimated
dimensions ending in zero, i.e., to the nearest centimeter

Length of carapace on mid-line 460

Depth of anterior emargination 5

Greatest width of carapace (at fourth costal) 460

Greatest width of carapacial disk (without rib-ends) 310

Greatest depth of shell 80

Length of nuchal 53

Anterior width of nuchal 90

Greatest width of nuchal 108

Length and width of neurals

First 46, 35

Second 44, 30

Third 45,31

Fourth 45, 28

Fifth 41,29

Sixth 40, 30

Seventh 23, 27

Eighth 21, 22

Length and width of second suprapygal 35, —

Length and width of first suprapygal 47, 77

Length and width of pygal 33, 54

Antero-mesial point of hypoplastron to tip of xiphiplastron 280

Width of plastron 370

Length of xiphiplastron on outer border 110

Antero-posterior width at hyoplastral-hypoplastral constriction,

left and right 71, 73

Length of peripherals on outer border, left and right

First —,51

Second — , 52

Third —,46

Fourth —,49

Fifth —,60

Sixth 59, 60

Seventh 66, 64

Eighth 65, 68

Ninth *. . .67, 67

Tenth 62, 61

Eleventh 55, 55

70 Field Museum of Natural History— Geology, Vol. 8

Catapleura arkansaw sp. nov.

Holotype. — F.M. No. P27045, a nearly complete carapace and
plastron. Collected by Charles M. Barber, August 25, 1934.

Horizon and type locality. — Marlbrook Marl, Gulf Series, upper
Cretaceous. Gather Brothers farm, one mile northeast of Okolona,
Clark County, Arkansas; in the N.E. H, sec. 29, T. 73 S., R. 20 W.

Fig. 23. Carapace of Catapleura arkansaw sp. nov.

Diagnosis. — A species of Catapleura with the generic characters
evident, i.e. dorsal disk entirely free from the peripherals, except
for contact of the latter with the first costal; first peripheral nearly
triangular. The fontanelles between the disk and peripherals narrow
and the projecting rib-ends correspondingly short; carapace broadly
notched anteriorly; nuchal broad; first peripheral resembling that

Arkansas Fossil Turtles

71

of Catapleura repanda, with a broad external and narrow internal
border; nine neurals, the last three much shortened; first supra-
pygal without free edge; second with broad free edges at the sides.
Plastron with a rounded central fontanelle; hyoplastra, hypoplastra,
and xiphiplastra not in contact on the mid-line; entoplastron free at
the sides; xiphiplastra suturally connected with the hypoplastra;

I ~ \ *'

Fig. 24. Plastron of Catapleura arkansaw sp. nov.

hyoplastron and hypoplastron on each side meeting in a long suture.
Horny scutes essentially in agreement with those of Osteopygis.

Differing from the species of the New Jersey Greensands in the
strongly concave anterior margin of the carapace.

Description of type. — Carapace cordate, the anterior notch a
smooth concave curve, slightly arched, the highest part at the third
to fifth neurals; plastron nearly flat.

72 Field Museum of Natural History — Geology, Vol. 8

Nuchal extremely wide, with a concave anterior border; twenty-
two peripherals, grooved on their inner faces, numbers one and four
to eleven represented in the present specimen on each side; first
peripheral with an extremely short inner border, making it nearly
triangular, suturally in contact with the first costal; sutural contact
of first costal extending about halfway along the second peripheral
(as shown by the costal). Fourth (and presumably the third) to
ninth peripherals each with a deep pit for the rib-end and an eighth
pit between the tenth and eleventh peripherals. Pygal with a slight
median external tubercle; a similar tubercle on the eleventh periph-
eral, with more weakly developed ones on the sixth to tenth. Nine
neurals, the first with a convex anterior and concave posterior
border, the second convex at both ends, the third to sixth roughly
six-sided and notched or concave anteriorly, the seventh to ninth
much shortened and with nearly straight anterior and posterior
borders. Costals eight, only the first in sutural contact with the
peripherals (i.e. with the nuchal and first and second peripherals).
The first, third, and seventh costals much widened at the outer end;
the second distinctly narrowed. Second costal, meeting the reduced
second neural, with an angular process both anteriorly and posteri-
orly, meeting both first and third neurals; remaining costals each in
contact with only two neurals. Third suprapygal small, transverse;
second with a convex anterior border, its outer edges free, and
strongly concave posterior border. First suprapygal 1 without free
border, entering the fenestrum between the eleventh peripheral and
the eighth costal at a point; its posterior border angulate, in broad
contact with the pygal.

Plastron with nine elements. The epiplastra similar to those
elements in Eretmochelys. Entoplastron with grooved lateral edges,
free posteriorly. Hyoplastron and hypoplastron of each side with
a long, nearly straight sutural contact; xiphiplastra suturally con-
nected with the hypoplastra. The several paired elements except
the entoplastron separated medially and digitate toward the median
line. Outer borders of hyoplastra and hypoplastra strongly digitate;
these two elements much as in Porthochelys.

The horny scutes agree in general disposition and number with
those of Osteopygis gibbi, i.e. with a nuchal and five vertebral scutes,
four costal scutes, and twelve marginals on each side. The course
of the sutures of the plastral scutes is shown in figure 24.

1 1 have numbered the suprapygals forward from the pygal.

Arkansas Fossil Turtles 73

measurements

All measurements are made by caliper in millimeters, estimated
dimensions ending in zero, i.e. to the nearest centimeter

Length of carapace on mid-line 750

Greatest length of carapace 780

Greatest width of carapace 730

Length of plastron on mid-line 534

Greatest width of plastron 550

Length of nuchal on mid-line 71

Anterior width of nuchal 122

Greatest width of nuchal 231

Length of neurals on mid-line

First 64

Second 61

Third 70

Fourth 62

Fifth 64

Sixth 60

Seventh 39

Eighth 35

Ninth 35

Length of third suprapygal 26

Length of second suprapygal 54

Length of first suprapygal 59

Suture between epiplastra 55

Suture between hyoplastron and hypoplastron 175

Length of epiplastron 175

Length of xiphiplastron on inner border 143

Remarks. — With the nearly complete specimen here described,
the genus Catapleura takes its place with Osteopygis and Lytoloma
among the more adequately known genera of this group of extinct
turtles. Associated skull and limb bones remain to be discovered.
Some fragmentary elements of the pelvis are represented in the pres-
ent specimen ; they resemble the corresponding bones in the modern
sea turtles, the Chelonidae.

Discussion. — Definitive classification of turtles requires knowl-
edge of the carapace and plastron, the skull, the neck vertebrae, and
the limbs and limb girdles. Thus we are still far from the possibility
of an adequate revision of the fossil cryptodire marine turtles, and
equally far from a clarification of the phylogeny of the existing
genera of Chelonidae. Linkage between Lytoloma and Catapleura
on one hand and the Dermochelidae (the leather-back turtles) on
the other, postulated by Case (1897), is not evident.

Examination of the forms here described serves to define a prob-
lem in the phylogeny of the turtle group. The disk of the carapace is

74 Field Museum of Natural History — Geology, Vol. 8

connected with the peripherals composing the rim of the shell by
the projecting rib ends. Among modern turtles such a condition
may be found in the young of various forms in which the periph-
erals and costals are suturally joined in the adults. On the other
hand, development of fenestrae, and thus of increased flexibility of
the shell, appears to be clearly correlated with aquatic habits — as
shown by the loss of the peripherals and development of fontanelles
in the Trionychidae, by the persistence of the fontanelles between
peripherals and carapacial disk in Caretta, and by the large vacuities
in the shells of such evidently marine forms as Toxochelys. In the
two turtles at hand, the proportionate extent of these fontanelles is
widely different. It is evident that in the absence of series of speci-
mens of different ages, of the extinct forms, the direction of evolution
in the shell in marine turtles can not readily be established. Thus,
the genera and species of the Thalassemydidae may be presumed to
be in need of further critical revision. The remarkable Testudo
tornieri, a land turtle adapted to life in rock crevices by reduction
and fenestration of the shell, proves that both ontogenetic and phy-
logenetic increase of such fenestration (and thus of flexibility of shell)
can take place, and the plasticity of the turtle shell is still further
evidenced by the remarkable thinning of the carapace in the Gala-
pagos species of Testudo. Thus it appears that the extent of fenes-
tration is a decidedly unstable character, subject to evolutionary
change in contrary directions. The interpretation of this condition
in isolated fossil specimens is accordingly extremely difficult, and
series of specimens of assorted age are not yet available. Within the
Thalassemydidae (as defined by Hay) turtles like Catapleura and
Phyllemys represent a degree of fenestration essentially equivalent
to that of Caretta, while the much deeper fenestration of Lytoloma is
found in a shell equivalent in length to the unfenestrated Osteopygis.
The architecture of the turtle carapace is a topic of great interest,
much in need of renewed comprehensive study.

REFERENCES
Case, E. C.

1897. Osteology and Relationships of Protostega. Journ. Morph., 14, pp. 21-
60, pis. 3-6.

Hay, O. P.

1908. The Fossil Turtles of North America. Carnegie Inst. Wash. Publ.,
No. 75, iv + 568 pp., 704 figs., 113 pis.

Schmidt, K. P.

1940. A New Turtle of the Genus Podocnemis from the Cretaceous of Arkansas.
Field Mus. Nat. Hist., Geol. Ser., 8, pp. 1-12, figs. 1-5.