BINDING LIST FEB 1 5 T923

7vr

NEW REPTILES AND STEGOCEPHALIANS FROM THE UPPER
TRIASSIC OF WESTERN TEXAS.

BY

E. C. CASE

Professor of Historical Geology and Palaeontology in the University of Michigan

Published bt the Carnegie Institution- of Washington
Washington, October, 1922.

QE

C37

CARNEGIE INSTITUTION OF WASHINGTON
Publication No. 321

I Copies Oil IksImwI
OCT 19 1922

TBCHMICAL PRESS
WASHINGTON, D. C.

CONTENTS.

Page

Introduction 7

The Upper Triassic Beds of the Borders of the Staked Plains 7

The Literature of the Triassic Vertebrates of North America 12

A New Genus of the Stegocephalia, Buettneria perfecta 13

Description of Desmatosuchus spurensis and the New Suborder Desmatosuchia 26

A New Parasuchian, Promyslriosuchus ehlersi 49

New Parasuchians, Leptosuchus crosbiensis and Leptosuchus imperfecta, from Crosby County,

Texas 61

Description of Isolated Bones of Parasuchians 70

Description of the Remains of Dinosaurs ...:.. 78

Coprolites ; .' 83

Incertae sedis 84

3

ERRATA.

Page 21 , Fig. 3, in placo of x' read x, and in place of x read x'.

Page 28, line 24, in place of 7 A read 7B.

Page 44, fifteenth line from bottom, in place of "shown in figure 16A and B",

read "shown in figure 16E and F."
Page 46, line 5, in place of 18C read figure 18A.
Page 61, about middle of page, in place of 25A read 25B.
Plate 7, K and I,, in place of anterior dorsals, read anterior caudals.

NEW REPTILES AND STEGOCEPHALIANS FROM THE UPPER
TRIASSIC OF WESTERN TEXAS

By E. C. Case

NEW REPTILES AM) STEGOCEPHALIANS FROM THE UPPER TRIASSIC

OF WESTERN TEXAS.

By E. C. Case.

INTRODUCTION.

By aid of grants from the Carnegie Institution of Washington, the author
has been for a number of years continuously at work upon the investigation of
the vertebrate fauna of the Permo-Carboniferous beds of North America. The
sudden and complete disappearance of this fauna (followed, after an interval
represented by barren beds, by a highly specialized Upper Triassic fauna) has
always been a tantalizing problem — the more so as the barren beds between the
two widely different faunae are practically the same, in all the implications as to
climate, mode of deposition, and terrestrial conditions, as the fossiliferous beds
above and below.

The interval of time represented by the barren beds is just the interval in
which was developed the wonderful Permian and Triassic reptilian life of South
Africa, Russia, and western Europe. A few very uncertain remains, briefly dis-
cussed in the section of this paper dealing with the Upper Triassic beds of western
Texas, suggest the possibility that something of the Lower Triassic life of the
Eastern Hemisphere reached North America. For these reasons the barren beds
have been followed and searched with extreme care wherever they occur, but so
far no trace of vertebrate life has been found in them. While engaged in such a
search, the author learned of a small area in Crosby County, Texas, which has
yielded a considerable amount of Upper Triassic vertebrate material, most of
which is new. The present paper contains descriptions of this new and interesting
fauna.

The author takes this opportunity to thank the Carnegie Institution of Wash-
ington and its officers for the continued support which has made the work possible.
Also, he wishes to acknowledge his obligation to Mr. Clifford Jones, manager of
the Spur Farm Lands Company, and the other officers of the Swenson properties,
for permission to go upon the lands under their control and for many acts of kind-
ness and courtesy which rendered the task of collection much easier and con-
tributed largely to the success of the work.

THE UPPER TRIASSIC BEDS OF THE BORDERS OF THE

STAKED PLAINS.

The Upper Triassic beds exposed on the borders of the Staked Plains in
Texas and New Mexico present the same puzzling complex of terrestrial and
fresh-water deposits that they do wherever they appear in the other Plains and
Rocky Mountain States. No definite determination of continuous horizons is
possible, as the sequence alters rapidly within a short distance and there are many

S NKW KK1TILKS AM) STEGOCEPHALIANS I-'KO.M

lenses and intercalated beds of small extent. Invertebrate fossils are few, limited
to Unios, which are generally so poorly preserved as to be indeterminate specific-
ally ; frequently there is an abundance of fossil wood in the form of larger or smaller
fragments, but none has been determined. An attempt to determine the wood
revealed that it was badly rotted before fossilizatioii and so deeply impregnated
with gypsum as to destroy the cell structure. Knowlton's catalogue of the Meso-
zoic and Cenozoic plants of North America mentions no Triassic plants from Texas,
and those listed from New Mexico are from the central and southern portions.
The age of the beds must be determined by the vertebrate fossils, and these indicate
an I'pper Triassic stage, approximately equivalent to the Keuper of Europe.

On the borders of the Staked Plains, as in the other regions of the United
States where the Triassic is exposed, there seems to be no possibility of determining
the boundary between the Permo-Carboniferous beds and the Triassic. The Red
Beds are apparently continuous across the interval, and the connecting beds are, so
far as known, entirely devoid of any evidence of life. It is in this connecting series
of beds, representing the period of the great development of reptilian life in the
late Permo-Carboniferous and the early Triassic in South Africa, that some repre-
sentation of that life should appear if it were present in North America. Because
of this fact the transition beds have been repeatedly searched with the greatest
care, but as yet nothing has been found. The only remains that suggest the pres-
ence of anything like the South African forms in North America are the prob-
lematical fossil found in West Virginia,1 200 feet below the base of the Pittsburgh
coal bed and the base of the Monongahela Series, in Braxton County, West Vir-
ginia, named Pareiasaurus(f) henningi by I. C. White, and the forms described
by Williston, from a few poorly preserved bones, as Evbraehio%aurv& and Brachy-
brnchium, from the Popo Agie beds near Landor, Wyoming. The first of these
Williston5 regarded as belonging near to Tapinocephalus or Phdcosaurus, and Dr.
Broom, in conversation with the author, stated his opinion that it was Deino-
eephalian in its affinities. Here, also, should be mentioned the humerus, described
by Lucas as Placerias hesternus, from the Triassic near Tanner's Crossing, Little
Colorado River, Arizona.1 These remains are far too little known to permit of
any conclusion being drawn from them, and they do not occur in the transition
beds, but below and above them.

The Triassic beds of the borders of the Staked Plains have been divided by
Drake4 into three parts, and his divisions are recognizable in a broad way, but
rarely can any definite boundaries be assigned to the divisions, nor can the pro-
visional separation made in one place be carried satisfactorily for any distance.
Drake described his three divisions as follows:

fatal nil tin- 1*1. — 1 1 <l< - Evidence of n Pareia*aur-like Reptile in the Conemaugh Series of West
Virginia, West Virginia Geological Burvey, Braxton .'11111 Clay County Report, p. 803, 1917.
■ Williston. s. w., Notice of Boom New Reptiles from the Upper Triassic of Wyoming, Journal of Geology,
vol. xii. p. 690, 1904.
Lucas, l l(.. ProeaedJnai U. 8. National Museum, vol. _'7. p. ml, pi. tv, 1904.
1 Drake, X. !•'., Stratigraphy of the Triassic Formation of Northwest Texas, Third Annual Report Texat
■loifi'nl Burvey, p. 227, 1x92.

THE UPPER TRIASSIC OF WESTERN TEXAS. 9

"Stratigraphy. — The following classification or grouping is not intended as a cor-
relation with any other Triassic beds, but only to apply to the Dockum beds over the
area examined. The Dockum may be divided into three beds, though some localities
show more, that are more or less well marked. * * * These three main beds are as
follows: A lower bed of sandy clay, which is from 0 to 150 feet thick; a central bed or
beds of sandstone, conglomerate, and some sandy clay, which is from 0 to 235 feet thick;
an upper bed of sandy clay and some sandstone, which is from 0 to 300 feet thick. While
these groups represent the different geological horizons over most of the Triassic area,
there is nevertheless at some places a thinning out of the one, and a thickening of another,
which shows that at the same time the conditions of deposition were somewhat different
at different localities. The same geological horizon is, therefore, more or less represented
in other beds than that which generally represents it. Then, while these beds do not
absolutely represent geological horizons, they do so approximately and are so well
marked as to be of stratigraphical value."

On the east side of the Staked Plains the lower beds are red sandy shales which
may be traced from the Canadian River southward to where they disappear
beneath the Cretaceous and younger deposits south of Big Springs, in Howard
County. These beds shade downward indefinitely into the Double Mountain beds
of the Permo-Carboniferous. The author has repeatedly crossed the line, in
every county of Texas where it lies, from the Canadian River on the north as far
south as the Triassic appears, and has been unable to draw any line that can be
used to separate the two formations. In passing from the towns Seymour, Vernon,
Haskell, Anson, and Abilene westward to the Triassic at the base of the Plains,
the same indefinite boundary is crossed. The red beds grow more shaly, and in
general there is an increase in the amount of gypsum, which may occur in beds
from a few inches to as much as 3 feet thick, or may be in fine seams running in
all directions through the red sandy clay. The same condition is found in the
breaks of the Canadian River and in the great canyons which penetrate into the
Plains in Randall. Armstrong, Swisher, and Briscoe Counties. Similar conditions
prevail wherever these two formations come together in New Mexico, Arizona, or
Wyoming and the adjacent States.

In the vicinity of Dickens, in Dickens County, where the transition beds are
beautifully exposed in the Croton Breaks, 2 miles directly east of the town, the
intermediate series is terminated above by a few feet of yellowish and bluish clay
which is overlain by a considerable thickness of grit and conglomerate, such as is
described by Drake in his middle division of the Dockum beds. The conglomerate
and the resulting gravel, which in many places covers the surface, are generally very
easily distinguished from the grits and gravel derived from the overlying Tertiary
beds of the Plains by the large amount of browaiish, semiangular, indurated clay
as opposed to the dominant whitish, well-rounded quartzite of the upper beds.
The Triassic conglomerate is typically shown at Dickens, but can be traced from
that point as far as the Canadian River on the north and beyond Big Springs on
the south. At Dickens in Dickens County, Roaring Springs and Matador in
Motley County, and on the section through Quitaque to Tulia in Hall, Briscoe,
and Swisher Counties, this grit and conglomerate lies directly beneath the Tertiary

10 NEW REPTILES AND STEGOCEPHALIANS FROM

cap-rock. At Spur, a few miles south of Dickens, it has the same position, forming
the hill, in the north part of the town, upon which the water-tower is located, and
is easily traceable in the small elevations which can be seen in all directions for
niilcs from that point. West of Spur it is the surface rock, appearing at intervals
through the surficial clays and sands to within a few miles of the Blanco or Catfish
River, where the lower formations are again visible in the breaks. On the east
side of the river the Triassic is much obscured by the accumulation of wind-blown
sand, but on the western side the bluffs are prominent and stand out as conspicu-
ous features similar to the bluffs at Dickens. Such a prominence is Cedar Moun-
tain, between the forks of Sand Creek. From the summit of these bluffs the hand-
level shows that the tops of the hills are approximately on the same level. The
beds below the conglomerate cap in this region are decidedly different from those
shown in the C rot on Breaks east of Dickens; there is much less of the deep-red
clay and relatively little gypsum. The beds are composed of light-red or yellowish
clay in most of the exposures. Though there is considerable continuity in the
beds, there is a decided irregularity of deposition exposed in the breaks of Holmes,
Sand, and Davidson Creeks near Cedar Mountain; here there are frequently
lenses and intercalated beds of light cream-colored clay, light-bluish clay, and
light-red clay, with abundant irregular concretions and nodules. By far the
greater number of the beds, and uniformly those which are at all regular in their
deposition, are totally barren of fossils. It is only in the irregular beds which
were evidently deposited in stream-channels and local pools that any remains are
found. It is evident that this area, which is so different from that of the Croton
Breaks, is the site of some great stream-channel and flood-plain.

Unfortunately, only a small area is exposed in the breaks, and a careful search
of the whole eastern side of the Plains has revealed no similar locality. Water-
worn fragments of bone frequently appear in the conglomerate; some were collected
as far south as the vicinity of Slaughter's Ranch, 18 miles southwest of Post City
in Garza County, and traces of bone were found east of Big Springs, but it is evi-
dent that the areas where vertebrate fossils may be collected in any quantity and
in a usable state of preservation are very limited on the eastern side of the Plains.

The exposures of the Triassic in the breaks of the Canadian River, in the
northern part of the Plains, show a more or less uniform series of red beds, clays,
and sandy shales, with occasional bands and veins of gypsum beneath the conglom-
erate. These beds can not be exactly located in Drake's series, but appear to be
the ones located in his upper division, though Permo-Carboniferous vertebrates
have been reported from the vicinity of Plenums in Hutchins County, and the
Texas Geological Survey has mapped Permo-Carboniferous in the western part of
Oldham ( ounty.

On the western side of the Staked Plains a splendid view of the exposures of
the Triassic can be had from the edge of the cap just west of Adrian in Oldham
County. From this point it is possible to see a large area of the breaks in the
upper part of the Canadian Valley and much of the area to the south, almost to
Glen Rio on the Texas-New Mexico line. The land is roughly rolling and much

THE UPPER TRIASSIC OF WESTERN TEXAS. 11

of it is grassed, so the exposures are limited in extent; but they are uniformly of
red clay and red sandy shale, with occasional layers of more or less heavy brown
sandstone and thin beds of grit or conglomerate. Throughout this area fragmen-
tary remains of Phytosaurs and Stegocephalians occur in small quantity. The
rolling country continues westward to and beyond Tucumcari in Quay County.
About 5 miles west of San Jon, in Quay County, there is an area of river deposits
with beds of conglomerate, heavy sandstone, loose gravel, and clay; the change in
the character of the beds is reflected in the nature of the surface, for the region
is locally known as the Bad Lands. In this limited area remains of Phytosaurs
and Stegocephalians occur, similar to those found in Crosby County, Texas, but
in far smaller number and in a poorly preserved condition.1 Beyond San Jon the
surface of the country again becomes more regular, reflecting the more uniform
character of the beds beneath.

At Mount Tucumcari and along the west front of the Plains there is, in gen-
eral, a more uniform sequence of the beds than on the east side. A heavy mass of
red clay and red sandy shales is overlain by a second heavy bed of sandstone which
may be white, as directly west of Montoya, or brownish farther to the west and
south, typically exposed at Cuervo, in Guadalupe County. These dominant heavy
beds of clay and sandstone are frequently interrupted by locally developed beds
of lighter-colored clays and sands, white, yellow, or blue, in which poorly preserved
fragments of bones occur in limited amount. The same character of the beds is
found as far west as Santa Rosa, in Guadalupe County, and as far south as beyond
Roswell. East of Roswell the red beds contain much more gypsum than those
farther north, and no remains of vertebrates have been found in them.

A reconnaissance from Montoya westward through Isidor, Buxton, and Cabre
Springs and up the Conchas Canyon to Las Vegas showed that the beds retained
a similar character until they disappeared beneath the Cretaceous. It is evident
that the great bulk of the Triassic beds revealed on the borders of the Staked Plains
and in eastern and central New Mexico were deposited under conditions unfavorable
to the preservation of vertebrate fossils. The more uniformly deposited beds of
clay and shale were apparently laid down in deep water or in water far from the
shores; it is only in the disturbed beds, which bear evidence of having been de-
posited by great flood washes, that the remains of animals and plants are found.
Such remains are usually badly broken and water- worn ; occasionally good specimens
will turn up, as evidenced by the presence of a skull of a Phytosaur preserved in
the University of Chicago, which was obtained from the School of Mines at Socorro,
New Mexico, and said to come from near Santa Rosa,2 and from some remains
secured by the author near Carthage, Socorro County. It is only in such rare
occurrences as the section of some large river flood-plain, such as occurs in eastern
Crosby County, Texas, that good material will be found. The rarity of such occur-
rences is shown by the experience of the author, who, in four trips in the regions
mentioned, has found only the single exposure.

1 For a more detailed description of the Bad Lands west of San Jon, see an article by the author in the Journal

of Geology, vol. xn, No. 3, 1914.

2 Mehl, M. G., A New Phytosaur from the Trias of Arizona, Journal of Geology, vol. xxx, p. 144, 1922.

12 m;\v ki:imii.i> \\d 8TBGOCBPHALIAN8 PROM

THE LITERATURE OF THE TRIASSIC VERTEBRATES OF

NORTH AMERICA.

Our knowledge of the Triassic vertebrate life of North America is so limited
that it is not yet time to attempt a summary statement of the material at hand,
but it will not be amiss, as a step in progress, to list the main articles which have
from time to time partially reviewed the work done up to the time of their appear*
ance. The bibliographies in these articles will lead to the original publications.

1902. Eat, O. 1*. Bibliography and catalogue of the fossil vertebrates of North America, Bulletin No. 179, United

States Geo! Survey.

1905. Bhansox. E. B. Structure and relationships of American Labyrinthodontida;. Journal of Geology, vol.

xm, Xo. 7.

Contains critical review of known Triassic Stegocephalians from North America and description of a
new genus, Anaachisma.
190(3. McObboob, J. H. The Pkytoaavria, with especial reference to Mystriowchua and Rkyttdodon, Memoirs
American Mus, Nat. Hist,, vol. i.\. pt. 11.

Contains a list and critical review of genera and species of North American l'hytosauria, and a bib-
liography.

1906. lit knk. F. v. Uebei die Dmosaurier ties Auseereuropaeisohen Trias. Geologisclu- und paleontologische

Abhandiungen, N. P., Bd vm, lift. 2.

Contains a critical review of the genera and species of the Triassic Dinosaurs of North America, with
a liililiography.
1911. Bums, F. v. Beitrage aur Kenntnis und Beurtheilung der Parasuchier. Geologische und paleontologiscb (

Abhandiungen, X. !'.. Bd x, lift. 1.
1911. BaSTKAN, C. II. Triassic fishes of Connecticut, Bulletin Is, Geological and Natural History Survey,
State of ( 'onnecticut.

Contains a critical revision of the Triassic fishes of the northeastern United States.
191.5. Mf.hi., M ('■. The l'hytosauria of the Trias. Journal of Geology, vol. xxm, Feb. M.n .

Contains a description of new forms, a review of known genera and species, and a bibliography.
1915. LULL, K. 8. Triassic life of the Connecticut Valley. Bulletin 24, Geological and Natural History Survey,
State of Connecticut.

Contains a critical discussion of the vertebrate fossils of the Triassic of the northeastern United
States, and a bibliography.
1920. Cask, K. C. Preliminary description of a new suborder of phytosaurian reptiles, with a description of a new
species o! Phytosaur ux. Journal of Geology, vol. xxviu, Sept. Oct.

Contains a description of the new suborder DtsnuUosuchia, the genus and species Desmatosuchus
tpurensit and Phytosmmu doughlyi.

1920. ('ask, E. C. On ■ very perfect thoracic shield of a large labyrinthodont in the geological collections of the

Fniversity of Michigan. Occasional Papers of the .Museum of Zoology, University of Michigan, Ann
Arlx.r, Mich. Xo. B2.

Contains a description of the clavicles and interclavicle of MetopOSaWU* jontti.

1921. Cask. Iv ( '. A new species of Ceratodvs from the upper Triassic of western Texas. Occasional I'apers of the

Museum of Zoology. Fniversity of Michigan. Aim Arbor, Mich. Xo. 101.

Contains a description of a tooth of a new species ,,f Certttodut C. dorotheCB.

1922. Mmi., M. (;. A new I'hvtosaur from the Trias of Arizona. Journal of Geology, vol. XXX, p. III. 1922.

Contains a description of a new form Phytoaaurtu, Macharoprosopus andersom.

The few papers published in 1920, 1921, and 1922 are later than any summary
papers and are inserted to bring the literature up to date.

CASE

PLATE 1

!>'

o

CO

d
X

5.

=q

J4

»\

* ^

THE UPPER TRIASSIC OF WESTERN TEXAS. 13

A NEW GENUS OF THE STEGOCEPHALIA, BUETTNERIA PERFECTA.

The specimen here described, No. 7475, University of Michigan, was found in the
breaks of Sand Creek just south of Cedar Mountain, in Crosby County. It lay in a
dark-red, mud-lump conglomerate with some finer material, the deposit of an old river-
wash. The undistorted skull is unique in the perfect preservation of the bones and the
minutiae with which the osteological details may be traced. The matrix was readily
removed from the bones of the lower surface, leaving them clean and white, except
where stained red or brown by iron. The rugose upper surface was less readily freed,
but the matrix came away very clean, revealing the pits and ridges and all the details
of the sutures and the slime-canals. There is no distortion of the bones of the upper
and lower surfaces, but the edges of some of the slender bones which form the walls of
the brain-case are slightly crumpled and injured by decay. No parts are missing
from the upper surface except the major part of the left squamosal, the extremities of
the tabulare, and the extreme posterior tip of the right maxillary. On the lower surface
a part of the distal end of the left pterygoid is missing,- there is no trace of the stapes,
and the otic opening is extremely large; for reasons given in the body of the paper it
is believed that this region was largely cartilaginous.

The upper surface of the skull. — The form and arrangement of the various elements,
the position of the slime-canals, and the character of the sculpture are shown in figure
1 a, and plate 1, fig. a, and do not need extended discussion. The general resemblance
to the skull of Anaschisma from the Popo Agie beds of Wyoming is apparent, but the
arrangement of the teeth and the bones of the lower surface show that the two forms
can not be placed in the same genus, and render it doubtful whether they should be
placed in the same family. A comparison with Branson's figures1 shows that the skull
was a little broader, proportionately, than in Anaschis?na and that the orbits were a
little farther forward. Branson was unable to trace all the sutures on the upper surface
of his specimen, but, so far as he was able to determine them, the position and relations
of the bones correspond as well as could be expected in animals in which there was so
much of individual variation. A comparison of his figures with figure 1, a and b, of
this paper will show the position of the sutures he was unable to follow, outlining in whole
or in part the lachrymals, the quadratojugals, the jugals, and the maxillaries. In
Buettneria the lachrymals are small elements not reaching to the nares; the maxillaries
extend inward anterior to the orbits and form a part of the posterior and lateral bound-
aries of the nares; posteriorly they lie upon the sides of the skull, and the postorbital
portions are only visible from above as narrow bands. The premaxillaries, nasals,
frontals, and parietals exhibit a decided asymmetry. In both the figures and the plates
it will be seen that the extremities of the tabulare have been restored; it may easily
be that these have been made too large, as the projections are very slight in Anaschisma.

Not only the sutures but the slime-canals are very perfectly shown in the specimen.

fig. 1 a. The course of the sutures and the canals shown in the figures was traced
with a camera lucida and the lines are almost exactly as they appear in the specimen;
the figures have not been diagrammatized to any extent. The edges of the slime-canals
have been traced as straight lines to distinguish them from the sutures.) The course
of the slime-canals is similar to that in Anaschisma, but differs in one or two important
particulars. Adopting the nomenclature proposed by Moodie,2 the anterior com-
missure and the occipital cross-commissure are absent; the course of the supraorbital
canals between the orbits and nares is quite different (compare fig. 14 of Moodie's paper).

' Branson, E. 15., Journal of Geology, vol. xrn, figs. 1 and 7, 1905.
2 Moodie. H. I,., Journal of Morphology, vol. xi\, No. 2, p. 513, 1908.

14

MEW REPT1I.KS AM) STKtiOCKIMIALIANS Klfd.M

The infraorbital canal turns sharply inward across the short lachrymal bone and joins
the supraorbital: the single canal t bus formed runs forward for a short distance and then
bifurcates on the anterior pari of the maxillary just posterior to the narial opening.
The outer branch runs forward for a short distance and terminates in an irregular expan-
sion. The inner branch follows the normal course of the supraorbital canal: there is no
antorbital commissure. Posterior to the orbit the supraorbital canal is complete and
runs backward to join the temporal canal. It will be observed that the supraorbital
canal on the left side lies for a short distance on the frontal: on the righl side it docs not

tab dso

Fro. 1. — BueUneria perfecta, skull of. No. 7I7.">, University of Michigan.

A. Upper surface. B. Lower surface.

('. Posterior surface. D. Lateral surface.

pmx., premaxillary ; n,j\, maxillary; 11.. nasal; ///., prefrontal;/., frontal; /.. lachrymal; pin., posl-
orliital; plf., postfrontal; j., jugal; si., suprateinporal; p., parietal; »}., squamosal; qj.,
quadratojugal; tab., tabulare; </«».. dermsupraoccipital; eo., exoccipital; rn.. vomer;
ps., parasphenoid; trt, transverse; /<"/.. palatine; </.. quadrate; xl, outlet for eleventh

nerve; n.pl.. anterior rising process of pterygoid; />./■'., posterior rising process of ptery-
goid; qj.. quadrate foramen; n.. orliit.

touch that element. The supraorbital and the temporal canals meet at a sharp angle,
and there is a canal connecting them wit 1 1 the infraorbital. Continuing backward, the
temporal canalfl lie upon the suprateinporal bones and turn sharply outward at their

rior ends; on the left side the canal touches the anterior outer corner of the tabulare;
on the right ride it clear- that bone. The jugal canal continues backward to the ex-
tremity of the quadratojugal and, with a slight interruption, turns inward for a short
distance near the posterior edge of the squamosal.

The edges of the canals are irregular and the sculpture is continued on the bottom
of the grooves. It is evident that in this specimen there is a departure from the usual
and ("according to Moodie) morphologically important position of the canals; that i-,

THE UPPER TRIASSIC OF WESTERN TEXAS. 15

they do not lie upon definite bones so certainly as to be entirely dependable criteria for
the determination of the homology of the bones with those of the fish skull. This is
particularly noticeable in the posterior part of the course of the temporal canals and their
relation to the tabulare. In comparison with the course of the canals in Metoposaurus,
as given by Moodie, the differences are quite similar to those noted in comparison with
Anaschisma.

The lower surface of the skull. — In this region the skull shows very decided differences
from Anaschisma. Branson states that in Anaschisma each palatine bone bears only a
single tooth, and this has been confirmed by a reexamination of the skull by Mr. Paul
Miller. It is also stated by Branson that there are four teeth on the anterior edge of
each prevomer (vomer), but no mention is made of a row of teeth on the inner edges of
the narial openings. It may be that these rows are obscured by the condition of the
specimen in the University of Chicago. The vomers are shown in Anaschisma as ter-
minating behind in long points which extend down the side of the parasphenoid. The
transverse bones are shown as large elements. The character of the articulation of the
parasphenoids with the exoccipitals is very different in the two forms. Mr. Miller has
reexamined this region and assures the author that there is a median suture as shown by
Branson and interpreted by him as the meeting of the exoccipitals and that there is a
large union of the parasphenoid with the pterygoids. The parasphenoid-exoccipital
suture is reported as uncertain, but it could not be far from the position given it by Bran-
son. The long median suture shown by Branson is one of the most interesting points
about the skull of Anaschisma; if it is formed by the meeting of the exoccipitals for any
considerable distance (the posterior portion of the skull is restored in plaster), it is
unique among the Stegocephalia, for in no other form is there a meeting of the posterior
ends of the exoccipitals for any considerable distance, if at all.

In Buettneria the premaxillaries are short, with large openings through which a
small part of the external nares can be seen from below. At the point of the union of
the premaxillaries with the vomers there is a small irregular depression in the median
line which may be significant or may be due only to imperfect ossification at the meeting-
point of the four bones. There are 13-13 teeth on the premaxillaries.

The vomers are large flat plates ; the anterior outer part forms the floor of the external
narial opening when seen directly from above. The anterior outer corners support large
tusks on each side; these were placed in pairs in which the tusks were alternately func-
tional; on the left side the anterior one was in use and on the right the posterior one
at the time of the death of the animal. On the anterior edges of the bones between
the tusks there is a nearly straight row of fairly large teeth. There are 5 teeth on each
side, the outer one smaller than the others. On the inner edge of the internal narial
openings there is a row of small teeth, 18 to 20 in number; the rows terminate posteriorly
at the palatine-vomer suture; anteriorly they are separated from the tusks by a short
vacant space. Within the inner edge of the internal nares there is a small cluster of
very small teeth; these point toward the center of the opening in the specimen.

The palatines are separated from the maxillaries by a long, straight suture, and from
the premaxillaries and the vomers by clearly defined sutures, as shown in figure 1 b,
and plate 1, fig. b. External to the posterior half of the internal nares there are large
paired tusks; as in the vomerine tusks, the functional teeth of the pairs are alternate in
position. Just posterior to the internal nares is a small cluster of irregular teeth, 5 to 6
in number. The long row of teeth on the outer edge begins just posterior to the tusks and
is continuous to the extreme posterior end. The teeth are very closely set and diminish
to exceedingly small size towards the rear; many of the teeth are represented by empty
spaces and were apparently not functional at the time of death. There are 36 to 38
teeth on each side. The palatine joins the pterygoid by an oblique suture.

16 NEW REPTILES AM) BTEGOCEPHALIANS KKOM

The transverse bone, if present, i> a small element. On each side there is an uncer-
tain line, indicated in figure 1 b, which seems to mark the separation of a distinct element
from the pterygoid. On the right side the small element is slightly displaced, indicating
its distinct character.

The maxillaries appear on the lower surface as a tooth-bearing edge only. There
are no tusks. The bone of the right side, which is complete, carries 97 teeth and spaces;
allowing for the small pari missing on the left side there would be 98 on that side. These
numbers were certainly not a fixed quantity in the individual or species.

The parasphenoid has a broad, flat processus cultriformis, which terminates at
the anterior end in a sharp point extending far forward between the vomers. In the
specimen the anterior end is raised, so that there is a decided elongate pit on the roof of
the mouth. This seems to be entirely natural, but may be due, in part, to slight post-
mortem changes. The posterior end is moderately expanded and meets the pterygoids
in long, clearly marked sutures. The central part of this expanded portion is marked
by low but distinct rugosities. Near the posterior end there is a low elevation on each
side, which runs from the pterygoids out upon the parasphenoid and then turns sharply
to the rear near the median line. The sutures with the exoccipital run obliquely back-
ward and inward and almost, but not quite, meet at the bottom of the notch between
the exoccipitals. These sutures are very complex and sharply and finely interdigitating.

No trace of a basioccipital or basisphenoid could be made out on the specimen, on
either the upper or the lower side of the floor of the brain- case, or by a careful examination
of the edges of broken pieces in the process of preparation.

The pterygoids. — There are two strong rami of the pterygoids visible on the lower
surface. The anterior runs forward and outward to join the palatines and the trans-
verse (?) or maxillary. There is a slight median depression on this ramus, and near
the anterior end there is a small prominence extending slightly downward and outward.
The second ramus runs directly outward to the quadrate; its lower surface is rounded.

The quadrates. — The articular surface is concave from side to side and convex in its
longest diameter, which runs obliquely from without, inward and forward. The inner
and outer edges are raised into slender, sharp ridges. The suture between the quadrate
and the quadratojugal is not clearly marked, but it was evidently at the extremity of a
process extending inward from the quadratojugal. Anteriorly the quadrate sends a
process forward and inward on the anterior face of the anterior rising process of the
pterygoid. Posteriorly the quadrate is overlapped by the conjoined outer ends of the
rising processes of the pterygoid. Between the quadrate and the quadratojugal there
is a relatively large quadrate foramen. As shown in figure 2 b, there is apparently
a small, distinct element present between the anterior rising process of the pterygoid
and the inner process of the quadrate. This may be a deceptive appearance, but it is
present on both sides and the lines between the elements are distinct and filled with
matrix. Its meaning is unknown; certainly it is not a part of a descending process from
the skull-roof, as the lower side of the roofing bones at this point is perfectly smooth.

The exoccipitals carry the large, distinct condyles, which are separated by a deep
notch extending forward to the parasphenoid bone. There is a good-sized foramen on
the lower side of each, which probably transmitted the X and XI nerves and the jugular
vein. On the middle of the outer side is the foramen for the XII (?) nerve. At the
base of the rising process which articulates with the dcrmsupraoccipital and the tabulare
there is a large foramen on the inner side which transmitted the X nerve.

The posterior jace of the skull (fig. 1 c, and plate 2, fig. a). — The dermsupraoccipitals
are visible in the median line and are separated by a distinct suture. They join the
tsbulan by straight vertical sutures. The space between the dermsupraoccipitals was
filled in large part by a mass of cartilage in the absence of a supraoccipital.

CASE

PLATE 2

H

A. Posterior surface of skull of Buettneria perfecta. X0.3.

B. Ekdfllged view of port inn of the palatine and maxillary of same, showing the

teeth opposite the internal nares. The labyrinthine structure is
clearly seen. X2.

C. Metopoaourws(T) joneri, No. 3814, U. of Mich. X0.3. Ventral surface of

clavicles and interclavicles.

THE UPPER TRIASSIC OF WESTERN TEXAS. 17

The exoccipitals meet the tabulare and the dermsupraoccipitals by strong, finely
interdigitating sutures. On the left side there is preserved a portion of a thin process
extending inward, which, with its fellow of the opposite side, forms a more or less com-
plete shelf dividing the foramen magnum from the space left by the disappearance of
the supraoccipital cartilage. The articular surfaces of the condyles face backward and
inward.

On either side of the exoccipitals the posterior rising processes of the pterygoids
extend outward to the quadrates. The processes rise toward the upper edge of the
squamosal, but do not reach to it in the specimen, ending in a thin and broken edge on
both sides. Anterior to this process the posterior face of the anterior rising process
may be seen through the great otic vacuity. Between the two rising processes is received
the descending process of the squamosal. The otic vacuity lies between the exoccipital,
the tabulare, the squamosal, and the posterior rising process of the pterygoid (fig. 1 c).

The quadrate foramen is relatively larger than in Anaschisma.1

The ophisthotic is completely covered by the descending process of the dermsupraoc-
cipital and the tabulare. The general similarity between the posterior surfaces of the
skulls in the two forms, Anaschisma and Buettneria, is obvious from a comparison of the
figures, but it is equally obvious that there is one great difference: in the latter form
there are no post-temporal fenestras; the place of each fenestra is taken by a deep pit
with an imperforate bottom. The resemblance is even greater to Metoposaurus diag-
nosticus as figured by Watson.2 It is probable that the part which he has called the
epipterygoid (E. Pt. ?) is a part of the descending process of the squamosal. In corre-
spondence, Doctor Watson suggests that the space between the two ascending processes
of the pterygoid was filled with a "persistent pterygo-quadrate cartilage that is essen-
tially an epipterygoid," and suggests that the part described in this paper is an ossifica-
tion of that cartilage; but it is clearly a process from the squamosal, as the broken edges
in the specimen were clean and fitted perfectly. His figure shows but a slight develop-
ment of the posterior rising process of the pterygoid, which in Anaschisma and Buettneria
reaches nearly to the upper edge of the squamosal. In the place of a large post-temporal
fenestra he shows a small foramen on the dermsupraoccipital-tabulare suture, a condition
approaching very closely to that found in Buettneria.

The upper (inner) surface of the basicranial bones. — The condition of the specimen
was such that it was possible to remove the matrix completely from the brain-case and
the inner side of all the bones, leaving them as clean and intelligible as the outer side
(fig. 2 a).

The anterior portion of the parasphenoid is marked by a deep, flat-bottomed groove
with thin, abrupt borders; opposite the anterior edges of the palatine vacuities this
groove is nearly equal to the width of the bone (23 mm.), but it contracts gradually and'
regularly to the rear, and opposite the posterior edges of the palatine vacuities it is
not more than 5 mm. in breadth; at the same time the groove becomes much shallower,
due to the gradual thickening of the bone. The edges become more rounded and lower
as they thicken to the rear, and finally disappear about opposite to the center of the
broad posterior portion of the parasphenoid. On either side of the center of ossification
and a little to the rear there are openings which extend obliquely outward and backward
and are inclosed by a low arch of bone, so that they do not lie within the substance of
the parasphenoid, but rather upon its surface, covered by the arch. The outer end of
the arch is slightly wider than the inner and terminates in an irregular edge, as if it had
been covered with cartilage during life. The length of the arch is 32 mm. The meaning

1 Branson, E. B Journal of Geology, vol. xm, No. 7, fig. 3, 1905.

2 Watson, D. M. S., Transactions Royal Society of London, vol. 209, Series B, fig. 10. 1919.

18 NEW KEPTILES AND STEGOCEPHALIANS FROM

of these peculiar structures is not certain; a careful study of both the upper and lower
surfaces and the broken edges revealed no indication of separate basioccipital or basi-
sphenoid ossification. It is probable that the basisphenoid cartilage became much reduced
and the internal carotid arteries were forced down upon the parasphenoid and passed
for a short distance through the arches. This suggestion is supported by the position
of the arches and the fact that the internal carotid arteries entered the brain-case through
the posterior part of the palatine vacuities and then turned backward. Watson1
identified markings in the specimen of Laccocephalus in this position as the grooves of
the internal carotids, though he interpreted the method of the arterial entrance into the
skull in a different manner. On the surface of each arch, nearer to the inner than the
outer end, there is a small foramen penetrating through the wall to the canal below.
Just anterior to the inner ends of the arches there are slightly elongated rugosities;
these are in the direct line of the continuation of the ridges on the upper side of the
parasphenoid and of the ridges on the inner edges of the pterygoids. Anterior to these
rugosities there are small foramina entering the skull from before backward; they are
probably no more than openings for nutrient vessels. The sutures between the para-
sphenoid and the pterygoids can be easily traced; they run irregularly backward from a
point at the center of the posterior edges of the palatine vacuities and pass beneath the
outer edges of the arches just described, i. e., the arches are entirely upon the parasphe-
noid. Posterior to the arches the sutures appear again and continue backward, convex
outwardly, and then converge to end very close together at the base of the deep notch
between the exoccipitals.

The pterygoids. — The upper face of the anterior ramus of the pterygoid is marked
by a wide, shallow groove which becomes more distinct posteriorly to a point about
opposite the posterior edge of the palatine vacuity; the outer edge of the groove is distinct
for its whole length, but the inner becomes definite only in its posterior portion. The
outer edge of the ramus is raised into a slight prominence just at the origin of the anterior
rising process of the quadrate ramus of the pterygoid. The pterygoid joins the para-
sphenoid, as is indicated in figure 1 b, and connects with the exoccipital just posterior to
the origin of the quadrate ramus. Just anterior to the junction of the pterygoid and the
exoccipital there is a foramen near the inner edge of the pterygoid; on the left side the
bone is broken around the foramen and shows that it enters into a considerable cavity
in the body of the bone.

The upper surface of the quadrate ramus is most interesting. The perfection of
the structures preserved has rendered it a little difficult to reconcile the conditions found
with the descriptions of less perfect specimens, but, in the light cast by the study of this
specimen, the previous descriptions can be brought into more or less harmony. There
are two distinct rising processes, an anterior and a posterior (see figs. 2 a and 2 b). These
are directly continuous with the body of the bone and extend for nearly the full length
of the ramus; there is no trace of a suture in the specimen, where every suture is revealed
with almost diagrammatic clarity. The processes are inclined backward as they rise
and are separated by a considerable space at their inner ends, but are very closely approxi-
mated, if not actually united, at their outer ends. The anterior process rises to the roof
of the skull; its upper edge was thin, with the inner half attached to the lower surface
of the parietal and squamosal by cartilage; this is demonstrated by the thinning and
ii -regularity of the tipper edge, by the line of low rugosities on the roof-bones mentioned,
and by the fact that the bones were found in this position with a thin line of matrix

between them (fig. 2 b).

•Watson, D. M. S., Transact ion! PhfloMphicol Society Of London, vol. 209, Scries B, pi, 2, 1919.

THE UPPER TRIASSIC OF WESTERN TEXAS.

19

The inner end of the anterior process rises from a triangular base. From the
center of the posterior end of the groove on the palatine ramus of the pterygoid a thin
wall rises obliquely backward; beneath this there is a deep pit, with its mouth just
anterior to the outer end of the arch described on the parasphenoid. This pit extends
obliquely backward and its apex is perforated by two small foramina on the left side;
the pit on the right side is perforated by a single foramen. Posterior to this pit the
base of the process is thickened (fig. 2 b) just opposite the anterior edge of the outer

Fig. 2. — Buettneria perfecta.

A. Upper (inner) view of the bones

of the basicranial region.
d.sq., descending process of
the squamosal; i.e., open-
ing of cavity for the inter-
nal carotid artery; x.
outlet for the tenth cranial
nerve. Other lettering as
in figure 1.

B. Upper view of the left quadrate

ramus of the pterygoid.
Lettering as in previous
figures.

opening of the arch on the parasphenoid; beneath it a second pit passes forward and
slightly inward and is imperforate. The apices of the two pits are separated by a
thin wall. The third part of the triangular base is formed by the lower part of the anterior
rising process. The upper part of this rising process is somewhat crumpled; it is impos-
sible to interpret the appearance exactly, for the distortion gives the appearance of
reduplication, as if the bone were made up of superimposed laminae or, which seems very
improbable, of separate thin bones.

On the posterior edge of the quadrate ramus is the posterior rising process. Its
inner end originates just posterior and external to the second pit described above; it
reaches nearly as great a height as the anterior one, but there is no indication that it
was ever attached to the roof above; in the specimen there is quite a space between the
two. This process is shown by Quenstedt' as reaching up to the squamosal in Cyclo-
tosaurus (Mastodonsaurus) robustus. Fraas,2 however, shows the plate of much less
height in Cyclotosaurus posthumus. Watson, in his figure after Fraas, draws it much
higher.3

The outer ends of the two processes unite apparently without suture and clasp the
inner and part of the posterior face of the quadrate. Into the deep cleft between the
two processes descends the process from the squamosal. The relations of these processes

1 Quenatedt, F. A., Die Mnstodonsaurier im griinen Keuper.sandjsU'ine Wurtemberg's sind Batrachier, Tubingen,

1850, pi. .3, fig. 16.
■ Fraas, E., Neue Lahvrinthodonti-n aus der Schwabischen Trias, Paleontographica, Bd. lx, pi. xvm, fig. 2,

1913.
• Watson, D. M. S., loc. cit., fig. 30 c, p. 54.

20 NEW REPTILES AND STEGOCEPHALIANS FROM

is von- puzzling in places; towards the outer ends the three seem to be indistinguishably
fused; this may be natural or may be due to the very close approximation of the elements,
partly due to a very slight compression in fossilization. It is possible that the descending
process was surrounded by cartilage and that the partial ossification of this cartilage
may have produced the appearance of fusion of the bones. The thin outer edge formed
by the united or approximated ends of the processes descends obliquely forward and
disappears behind the quadrate; there is a small foramen between the bones at this point.

The quadrate.— The upper surface of the quadrate was seen in the specimen during
preparation; it sends a shorter and heavier process backward and inward to articulate
with the posterior process of the pterygoid, and a longer and more slender process for-
ward which lies on the anterior face of the anterior process of the pterygoid. This
anterior process is expanded vertically and in life was either applied closely to the
pterygoid or attached to it loosely by cartilage. In the specimen it is separated from
the pterygoid on both sides by a space of a millimeter or two, now filled with matrix.

The exoccipitals. — These join the pterygoids and the parasphenoids by finely inter-
digitating sutures. The edges of the parts of the exoccipitals, which form the walls of the
brain-case, are broken away, but they did not rise to and connect with the roof, as is
proven by the absolutely smooth surface of the under side of the roof-bones. Anterior
to the rising process there is a foramen; a little farther forward and somewhat laterally
there are two small foramina on the left side and a single foramen on the right side. At
the base of the rising process on the posterior outer angle there is a small foramen. On
the right side there is a second foramen on the posterior inner corner, but this is lacking
on the left side.

The lower surface of the roof of the cranial area (fig. 3). — The dermsupraoccipitals
and the tabulare send down processes which unite with the rising processes of the exoc-
cipitals. On the inner side of these processes there is a thin plate of bone, evidently
the very rudimentary opisthotic. On each dermsupraoccipital there is a small, elongate
area with a rough surface and slightly raised edges; that of the left side reaches to and
even crosses the dermsupraoccipital-squamosal suture; that of the right side is shorter.
These are evidently points of cartilaginous attachment for some element of the brain-
wall, probably the prootic, but no trace of such an element was preserved. The tabulare
has small presentation on the lower surface and the outline is nearly square. About
the center of the bone there is a deep pit with the inner edges slightly elevated. These
pits end blindly and do not penetrate more than halfway through the bone. Outside
of the pits the remainder of the surface is rough and evidently afforded cartilaginous
attachment; the rough area terminates sharply at the tabulare-squamosal suture, but
the adjacent portions of the squamosal are marked with fine grooves, as if by a plexus
of blood-vessels. There is no element which can be supposed to have attached to this
area, unless it be the opisthotic, which has nearly disappeared in the genus; this sugges-
tion seems the more unlikely, as, if it were attached, the opisthotic would appear on the
posterior face of the skull, which it does not do in the Metoposauridse. On the other
hand, Watson1 shows an ascending plate on the inner edge of the exoccipital in Capito-
gaurus. The thin plate on the inner side of the rising process of the exoccipital of
Buettneria may possibly be such a process, though it is apparently free from the exoc-
cipital and has been interpreted as the opisthotic. The meaning of the large pits on the
tabulare is not understood.

Attached to the inner edge of the anterior rising process of the pterygoid there was
found, on the left side, a small plate of bone, somewhat constricted in the middle. This
lay vertically in the matrix and extended horizontally inward, across the posterior end

1 Watson, D. M. S., loc. cit., fig. 11 a.

THE UPPER TRIASSIC OF WESTERN TEXAS.

21

of the groove on the upper face of the posterior end of the pterygoid. It is perhaps an
ossification of the alispenoid or orbitosphenoid cartilage; no corresponding element was
found on the other side.

One of the remarkable features of the skull is the evidently large amount of cartilage
in the walls of the brain-case. The stapes were lost and the otic opening is very large,
the bones around it terminating in smooth edges. There seems no escape from the
conclusion that the whole of the area of the organ of hearing was cartilaginous. The
whole of the lower surface of the bones in the area occupied by the brain is perfectly
smooth, with the exceptions noted. There was no attachment of exoccipital, epiotic,
or opisthotic to the upper wall, and the same condition in front shows that there
was no ossification of the ethmoidal elements.

FlG. 3. — Buettneria perfecta.

View of lower surface of roof-bones of cranial
region.

x', rugosity showing place of attachment of
the prootic; x, rugosity showing
place of attachment of the anterior
rising process of the pterygoid;
opf, opisthotic. Other lettering as
in figure 1.

The teeth are imperfect, the tops having been broken away and the stubs worn
as the skull was dragged, palatal side down, over the bottom. In even the smaller
teeth the broken surfaces show the labyrinthine structure (plate 2 b).

It is evident that this genus has its nearest relations in Metoposaurus and belongs
in the family Metoposauridae of Watson,1 but it is very doubtful whether Anaschisma
can be retained in the same family. From Metoposaurus the genus Buettneria differs in
the narrower processus cullriformis and the shape of the posterior portion of the para-
sphenoid, in the smaller (or absent) transverse, in the shorter quadrate ramus of the
pterygoid, in the loss of all trace of the post-temporal foramina, in the presence of a
series of small teeth on the internal nares, etc.

Measurements. — Total length of skull, 443.2 mm. Breadth of skull at posterior
end, 344.2 mm.

For this new form I take great pleasure in proposing the name Buettneria perfecta,
in recognition of the interest, patience, and skill of Mr. William H. Buettner, who has
prepared the skull and mounted it in all of its perfection.

A second skull of the same genus, but far less perfect, showing only the upper
surface, was found in the same region. Many other isolated bones were found, indicating

1 Watson, D. M. 8., loc. tit., p. 07. See also v. Huene, Gonioglyptus, ein alttriasicher stegocephale aus Indien,
Acta Zoologica, p. 456, 1920. In this paper v. Huene arrived, independently, at almost exactly the same
arrangement of the families of the Stegocephalia as did Watson a year earlier.

22

NEW REPTILES AND STEGOCEPHALIANS FKOM

other parts of the skeleton, but of these the majority are vertebras and bones of the pec-
toral girdle; remains of limb-bones are scarcely represented in the collection. Some of
the isolated bones are described below.

A nearly perfect left ramus of a lower jaw, No. 7469, University of Michigan col-
lection, was found by the author in the breaks of Sand Creek, just south of Cedar Moun-
tain. The jaw lacks the extreme anterior end and the posterior end is somewhat crushed,
but the greater portion is in good condition. The length of the jaw is, allowing for the
missing anterior end, 43 mm., somewhat shorter than the jaw of Buettneria as indicated
by the skull. The jaw as preserved shows too great a curvature to be associated with
the skull in the same genus. The jaw is illustrated in figure 4, a and b. On the inner

3"t

SP ^f

EP ang

Fig. 4.

A. Outer surface of loft ramus of jaw of a large

Stegocephalian, No. 7469, U. of Mich.

X 0.25.
U. Inner surface of the same jaw shown in A.
tli rit., dentary; cor., coronoid; sur. ang., suran-

gular; art., articular; sp., splenial;

pt. sp., postsplenial; ang., angular;

/)/-. art., prearticular.

side there is a relatively large supra-Meckalian foramen and below this a surprisingly
large lower opening. The splenial is short and confined to the anterior fourth of the
ramus; it appears on both the inner and the outer sides of the lower edge and evidently
took part in the symphysis. The sutures marking the outlines of this bone are clear
and distinct. Posterior to the splenial there is a postsplenial which forms the middle
portion of the inner surface of the jaw; it appears on both the inner and outer sides of
the lower edge, but on the inner side rises in a smooth plate to form the greater part of
the inner surface. It is pierced by two good-sized foramina. The posterior edge of
the postsplenial extends back to the middle of the lower edge of the lower Meckalian
opening, where the suture is clear and distinct. From the middle of the upper edge
there is apparently a suture of irregular outline extending downward, but it can be traced
for only a short distance; the exact nature of this apparent suture can not be made out;
it may indicate the presence of an intercoronoid bone, but no other indications of either
an intercoronoid or a precoronoid can be detected. Posterior to the lower Meckalian
opening the inner surface of the bone is badly crushed and covered with a multitude of
small fracture lines which render it impossible to determine the course of any sutures.
It is altogether probable that the greater part of this surface is formed by the pre-
articular. The anterior edge of the supra-Meckalian opening is formed by the coronoid
and the prearticular; it is probable that the angular appears on the lower edge of the
inner surface, but the suture can not be made out. The coronoid extends forward in
a long process between the postsplenial and the inner edge of the dentary. The dentary
ends posteriorly in a sharp process between the coronoid and the angular or surangular.
Anteriorly it extends forward, forming the whole of the alveolar edge. The teeth are
small at the posterior end, but rapidly increase in size forward until they reach the
maximum size. I'nder the binocular microscope the structure of the teeth is seen to
be labyrinthine at the base, becoming much more simple near the apex. No teeth
appear upon the coronoid.

CASE

PLATE 3

X.^

pq

©
X

o
£>

00

CO
CO

"3 n

4

CO

9
S

3

c

>

a.

k ©'
| X

JA

CO

O

THE UPPER TRIASSIC OF WESTERN TEXAS.

23

The posterior part of the outer surface of the jaw is largely formed by the angular;
it is marked in its posterior portion by a very coarse sculpture, which changes to a
coarse radiating pattern anteriorly and gradually dies out near the anterior third of
the jaw. No suture can be made out between the angular and the surangular, and it
is not certain that the surangular appears on the outer side, as it is figured in Eryops
by Broom. The sutures between the angular, dentary, postsplenial, and splenial are
all clear and distinct. Near the posterior end of the jaw there is a deep linear depression
devoid of sculpture; this may mark the position of the suture between the angular and
surangular, but no traces of the suture could be made out in or near the depression.

Fig. 5. — Outlines of intcrclavicles, all, except D, from University of Michigan.

A. No. 7368. B. No. 7448. C. No. 3814. D. Metoposaurusfrassi Lucas. E. No. 7266.
F. Posterior portion, No. 7367. G. Posterior portion, No. 7366. H. Posterior
portion, No. 7374.

Viewed from above, the posterior end of the jaw shows two distinct halves which
seem to meet as distinct bones at the posterior end; these may be the angular and the
surangular or the surangular and the prearticular. Just posterior to the elevation of
t he surangular on the posterior edge of the supra-Meckalian opening there is apparently
a distinct element represented by an egg-like mass of bone between the two separate
halves described above. This may be the articular, as no other trace of that bone can
be found.

A second jaw (No. 7503, University of Michigan), also from Sand Creek, is com-
plete, but badly rotted by gypsum. The sutures can not be traced, but the Meckalian

L'l \i:\\ REPTILES AM) BTBQOCBPHALIAN8 FROM

openings are the same as in No. 7409 and there is the same meeting of two distinct
hones at the posterior edge and the same distinct articular element.

Eight interclavicles were found in the same region as the skull of Buettneria. Five
of these are nearly complete; three are accompanied by more or less complete clavicles;
three are represented by the posterior portion only. These bones are shown in figure 5,
and plates 2 c, 3, and 4. It is evident that none of these forms is specifically identical
with any of the described forms from the Upper Triassic of Europe, and it is probable
that the difference is of generic value. Figure 6, copied from Fraas,1 shows the char-
acteristic form of the interclavicles in Metoposaurus, Cyclolosaurus, and Masiodonsaurus.
In comparing these with the interclavicles from the Texas Triassic, it is evident that
there is a great difference in the proportions; in the American forms the posterior pro-
longation is much less in relation to the anterior. In the specimen previously described
as Metoposaurus joncsi, No. 3814 of the University of Michigan collection, the form
most closely approaches that figured by Fraas as Metoposaurus in the shortness of the
anterior process and the breadth as compared with the length, but the posterior process
is much sharper and the sculpture is very different. In all of Fraas's figures there is
shown a strong posterior prolongation of the clavicles in the region of the articulation
with the shoulder girdle, which is apparently absent in the Texas forms, though this is
not absolutely certain, as all of the clavicles, except No. 3814, are imperfect in this
region.

In;. 0. — Outlines of interclavicles and clavicles.

A. Mehtpetaunu, after FraM. B. Cyclotosaurus, after Fraas. ('. Maslodonxnurus,

after Fraas.

The figures showing the outline of the interclavicles have all been reduced to the
same scale on the single line which is comparable in all of the specimens, i. e., the
breadth just at the posterior edge of the articular faces for the clavicles. The dis-
proportion between the breadth of the interclavicles and the length of the posterior
process posterior to this line is sufficiently striking in the different specimens to need no
comment. Working out the proportion between these lines in all of the specimens
results in an almost perfect series, with no break where a line can be drawn between
groups. The varying outlines warrant the belief that there are several distinct species,
but upon such incomplete material the author does not feel that new genera or even
species should be founded. Future discoveries will undoubtedly associate tlie different
interclavicles with more characteristic parts of the skeleton; and if the differences in
outline and proportions are shown to be more than developmental or individual vari-
ations, new names may be based on such characters.

1 Fraai", K., Neue Labjrrinthodonteii mm dor Bchwabischen Ti-i.-i^. Paleontographica, Bd. uc, p. ~'N6, 1913.

v\

CASE

PLATE 4

Buettneria perfecta, ventral surfaces of posterior halves of interclavicles.
A. No. 7.367, U. of Mich. X0.34. B. No. 7366, U. of Mich. X0.48.
C. No. 7364, U. of Mich. X0.53.

THE UPPER TRIASSIC OF WESTERN TEXAS.

25

As the figures are reduced to the same scale, it is necessary to take the measure-
ments into account to realize the difference in size. The measurements are given in
the following table:

Specimen.

1

Total
length.

2

Breadth at

posterior edge of

face for clavicle.

3

Length posterior

to line indicated

in 2.

4

Proportion
between 2 and 3.

No. 7449, U. of M

No. 7367, U. of M

No. 7448, U. of M

No. 3814, U. of M

No. 7366, U. of M

mm.
355.6

381.0
330.2

428.8
279.4

508.4

mm.
279.4
381.0
273.2
279.4
254.0
309.0
172.0
189.6
355.6

mm.
101.6
149.4
120.8
117.6
114.4
150.0
89.6
95.4
225.6

0.36
.39
.44
.45
.45
.48
.50
.50
.63

No. 7368, U. of M

No. 7364, U. of M

No. 7265, U. of M

From the shape of the interclavicles it is believed by the author that only one
specimen, No. 3814, can be placed in the genus Metoposaurus; it has been described as
M . jonesi.1 All but one of the others correspond in a general way, and as they were found
in the same region as the perfect skull of Buettneria and a second, imperfect skull of the
same genus, they may be regarded, provisionally, as belonging to that genus. Specimen
No. 7364 is decidedly different in form, proportions, and sculpture and may be distinct;
the imperfect state of preservation renders this somewhat doubtful.

The clavicles are not well preserved, except in the specimen No. 3814, but the
general form and sculpture do not differ markedly in the portions preserved. In
specimen No. 7367 the clavicles are long in correlation with the long anterior process
which the interclavicle must have possessed.

Few other bones of Stegocephalia were found. A single humerus of small size
and with poorly ossified articular ends was found associated with the small interclavicle,
No. 7368, and probably belonged to the same individual. Numerous vertebral centra
of typically stereospondylus form were found scattered in the beds. One with a double
articular face is undoubtedly the anterior one of the series. No neural spines or other
parts of the skeleton were found.

1 Case, E. C, Occasional Papers of the Museum of Zoology, University of Michigan, No. 82, 1920.

26 NEW REPTILES AND STEGOCEPHALIANS FROM

DESCRIPTION OF DESMATOSUCHUS SPURENSIS AND THE NEW

SUBORDER DESMATOSUCHIA.

The remains of this specimen were found near the east bank of the Blanco or
Catfish River, about a half mile east of the crossing of the old mail-road from Spur to
(rosbyton, in Crosby County, Texas. The site of the discovery is a patch of sand and
sandy clay occupying not more than a few hundred square feet, about a half mile north
of the road. This patch lies within a half mile of one of the areas of exposure of sands,
clays, and conglomerates which mark the occurrence of old river-washes at rare inter-
vals, among the barren clays of the Upper Triassic. The particular patch of sand and
clay in which the specimen occurred was evidently accumulated in some isolated hole
distinct from the nearest large area of river deposit. No other bones were found within
half a mile of the spot, and this important point, bearing upon the association of the
bones found in one specimen, was checked during three separate visits to the locality.
In working the specimen out of the ground a few scattered bones were found in the
sandy clay of the upper part of the deposit, but the greater part of the material was
found in single mass near the bottom, where the matrix was a light-colored clay with
included grains of sand, abundant traces of badly decayed vegetation, and lumps of
charcoal. This part of the clay was filled with gypsum, which had in some places
badly rotted the bones and in other places formed a protecting coat which preserved
the surface in perfect condition. Next to the bones there was, in may places, a thin
layer of pyrite, which rendered the cleaning very difficult, especially the rugose surface
of the skull. The pyrite frequently penetrated and filled the small cavities and
foramina, making it especially difficult to work out some of the finer details. Some
parts of the matrix were very hard and colored a deep black ; when this sort of material
was in contact with the bone the preparation was even more complicated.

It seems evident that the remains of the animal were washed into a hole with a
considerable amount of vegetable material and then the hole was filled with a cleaner
sand. There can be no question that the major portion of the bones found together
at the bottom of the pit belong to one specimen, but there is also no question that
some bones of a second individual found their way into the hole, as there is an excessive
number of dorsal vertebrae and there is a second axis. Only a very few of the char-
acteristic dorsal plates of Desmatosuchus were found among the abundant remains of
Phytosaurs in the adjacent regions of the county, suggesting the comparative rarity of
the form and perhaps its restriction to a certain habitat. The peculiar condition under
which the specimen was found and the fact that no other bones were found nearer to
it than a half mile must again be emphasized by the author as justification for associating
the bones in a mount.

A preliminary description of this form has been published in the Journal of
Geology1.

The material collected consists of the skull lacking the lower jaw and the anterior
end of the nose, the vertebral column, which is apparently complete to the sacrum and
a few caudal vertebras, the dorsal armor of the body and a portion of the tail, an im-
perfect right scapula, and a portion of the light side of the pelvis. A fragment of one
of the long bones is the single bit of evidence of the skeleton of the limbs. One small,
irregular plate was found with the mass of vertebrae and may be taken as evidence of
the presence of isolated plates upon the sides or abdomen.

The greater part of the material was found in an irregular heap, with no continuity
except in the cervical region and a few of the dorsals. A few of the plates seem to

'Case, E. C, Preliminary Description of a New Suborder of PhytOHHUiao Reptiles, with a Description of a
New Specie* of Phylosaurus, Journal of Geology, vol. xxvin, No. 6, p. 624, 1930.

THE UPPER TRIASSIC OF WESTERN TEXAS.

27

have retained their normal position. The restoration of the skeleton has been accom-
plished by placing the bones in their most evident position, with such checks as their
original position afforded.

The skull. — In many of its characters the skull resembles that of the Phytosauria,
but differs so radically in others that it must be placed at least in a separate suborder.
The close occlusion of the sutures and the condition of preservation has made the
determination of the most of the sutures impossible. The surface of the skull has been
most carefully cleaned with a needle and has been treated with acid to remove the
pyrite and the other matrix. It has been repeatedly examined under a binocular
microscope, so that the author feels certain that all possible traces of the sutures have
been made out.

Pie. 7. — Skull of DespuUofuehut tpurenti*.

A. Upper surface. / 0.:i. B. Lateral surface. X 0..3. C. Posterior surface. X 0.3. D. Lower surface of pos-
terior portion, x 0.8. E. Lateral surface of wall of brain-ease. X 0.6.

bs., basispheiiiml; /«., parasplienoitl rostrum; oL, otic channel; als., alisphenoid; pro., prootic; o(., otic opening;
Op., opistliotie; I l-XI I , outlets of cranial nerves.

2S NEW REPTILES AND STEGOCEPHALIANS FROM

The general shape, with the exception of the anterior end of the nose, is shown in
figure 7 and plate 5. The small size relative to the body is a striking difference from
the condition found in the Phyiosauria. It is, of course, uncertain just what length
was attained by the nose, but from the position of the nares, the shape oi the nose,
and the approximation of the maxillaries, it is apparent that not a great deal is missing.
The lack of a second temporal fenestra is the most striking character of the skull. This
might be considered as a primitive character were it not for the associated characters
oi the lack of a pineal foramen and the small size of the post-temporal openings, which
amount to no more than foramina. The nearly complete closure of the posterior
surface of the skull, with the single temporal opening and the prolongation of the jugal
forward and downward, led the author to repeated attempts to trace some resemblance
between this form and those from South Africa, especially the Deinocephalia — attempts
that were just as repeatedly demonstrated to be useless; but superficial suggestions
recalled the South African forms many times. The presence of the deep notch below
the orbit, with the projecting edges of the jugal and the maxillary, led to the suggestion
that the temporal opening was the upper one and that the lower part of the temporal
region was broken away and lost. This suggestion was also made to the author
independently, by both Doctor von Huene and Doctor Watson, after their consideration
of the figures published in the preliminary description of Desmatosuchus. The ideas
of these friends are indicated in figure 8. This is, however, impossible, as the lower
edges of the bones in this region are paper-thin and complete; moreover, the position,
character, and form of the bones on the two sides are practically identical, a condition
that is almost inconceivable if part had been accidentally lost.

The lateral aspect of the skull (fig. 7 a and plate 5 b). — As mentioned above, the
anterior end is missing, from about the anterior third of the narial opening forward.
The maxillaries are elongate, with only a limited exposure on the sides of the skull;
the alveolar edge is slightly convex antero-posteriorly. At the anterior end, below the
anterior half of the nares, there is the appearance, on both sides, of a wedge-like insertion
of a posterior prong of some bone in front. If this is a distinct element it can only be
a part of the premaxillary. Between the narial and the antorbital openings the upper
edge of the maxillary rises in a process which is reflected at its upper end and joins the
lachrymals and the nasals; the suture is here clear and distinct. The line of juncture
between the maxillary and the lachrymal can not be made out on either side, but a
break on the left side of the skull, which possibly may be along the line of the suture,
indicates that the suture was perhaps convex downward, permitting the maxillary to
form the posterio-inferior border of the antorbital opening. The upper edge of the
maxillary forms the anterior half of the lower border of the orbit; the suture between
it and the jugal is inclined backward and downward. From a point at the posterior
edge of the dental series the lower edge of the maxillary is inclined somewhat abruptly
backward and downward, terminating in a blunt point of thin bone. The posterior
edge of this process is nearly straight and forms the anterior edge of a nearly rectangular
notch which lies below the orbit and between the maxillary and the jugal. A posterior
narrow prong of the maxillary which articulates with the jugal above forms the upper
border of the notch. The complete edges of the bones outlining this notch and the
almost exact similarity of the two sides show that this is a notch and not the upper part
of a second temporal opening. It is a structure which has not previously been noted,
so far as the author knows, in any fossil skull. There arc 12 clearly defined sockets
on the alveolar edge of the maxillary of the left side and 13 on the right; these indicate
teeth of moderate size with nearly cylindrical roots. There is no indication of enlarged
teeth at the anterior end and there is no suggestion of antero-posteriorly elongate teeth

v8~

H~

CASE

PLATE 5

Desmatosuchu-s tpwreruis, skull, No. 7476, U. of Mich. All figures X0.36.
A. Upper surface. B. Lateral surface. C. Lower surface. D. Posterior surface.

THE UPPER TRIASSIC OF WESTERN TEXAS. 29

such as occur in the posterior part of the jaws in the Phytosauria. The sockets of the
median part of the series seem a little larger than those of the anterior or posterior ends ;
this may be due to the condition of the specimen. Only a single, badly rotted tooth
was found associated with the specimen and from this nothing can be made out.

The nasals are elongate bones lying in the normal position, forming the upper
edge of the narial openings and separated from the maxillaries and the lachrymals by
distinct sutures. The posterior edge can not be made out exactly, but a suture is
traceable for a short distance from a point a little in front of the posterior edge of the
antorbital opening toward the median line.

No trace of a septo-maxillary can be made out.

The lachrymal forms the upper border of the antorbital opening and is separated
from the nasal and the prefrontal by a distinct suture. It forms the broad bar between
the orbit and the antorbital opening and joins the maxillary below, but the suture, as
stated above, can not be made out definitely. The portions of the lachrymal and the
maxillary adjacent to the upper part of the antorbital opening are smooth and evidently
afforded attachment to a strong muscle which helped to close the jaws. The upper
portion is marked by a deep but relatively small sculpture. A foramen enters the
bone on the inner rim of the orbit.

The jugal occupies the normal position, forming the posterior half of the lower
rim and the lower half of the posterior rim of the orbit. The union with the maxillary
is sharply defined. The posterior border of the notch beneath the orbit is formed by
the jugal. The lower border rises sharply from the posterior lower corner of the notch
to the articular region of the skull. The suture between the jugal and the postorbital
is very distinct : the two bones meet at about the middle of the slender postorbital bar.
The jugal narrows towards the rear and forms the major part of the bar closing the
temporal opening below. Neither the position nor the outline of the jugal-quadrato-
jugal suture can be made out.

The limits of the quadratojugal can not be made out, but it is evident that it is
a small element. The posterior lower angle of the skull is narrow and the bar behind
the temporal opening is even narrower. The whole region is more slender than appears
at first sight, for the distal end of the opisthotic appears on the lateral surface and
increases t lie apparent size of the angle when seen in profile.

The sutures outlining the prefrontal, poslfrontal, and frontal can not be followed;
the whole region is very rugose and the most careful inspection has failed to reveal more
than short traces of the lines of separation.

The postorbital. — Just posterior to the heavy rugose knob which marks the upper
posterior angle of the orbit, a suture can be traced for a short distance upon the smooth
surface of the skull on both sides; this evidently marks the posterior edge of the post-
orbital.

The squamosal begins at the suture just mentioned; the anterior part lies in the
smooth area above the temporal foramen; the posterior portion is slightly but abruptly
elevated and forms a protecting point on the back of the skull. The edge of the abruptly
elevated portion is continued downward and backward and forms the posterior edge of
the lateral temporal opening until it meets the quadratojugal at some undetermined
point.

The upper surface of the skull (fig. 7 A and plate 5 a). — The skull is narrow in front
and gradually widens to the posterior edge. The nasals form a little more than the
anterior half; their surface is sculptured a little more deeply on the median than on the
outer parts. The median suture is traceable to the posterior edge of the nasals, but
from there back is indistinguishable; it is entirely obliterated in the parietal region,

30 NEW REPTILES AND STEGOCEPHALIANS FROM

where the smooth area permits the most careful examination. This smooth area is
the surface of a shallow depression which runs directly across the top of the skull, from
a point opposite the anterior half of the upper edge of the temporal opening to the same
point on the other side; its posterior border is somewhat sharply elevated by the
rugosity of the surface of the squamosal and parietal bones, and it is divided by a low
ridge on the median line of the skull. The most of the upper surface is so deeply
sculptured that the sutures are not visible, but it appears from the arrangement of the
sculpture that the frontals, paired or unpaired, occupy a heart-shaped area from about
the posterior line of the orbits as far forward as the nasals. There is, on each side, a
deep groove or elongate pit just within the very rugose upper border of the orbit. At
the posterior inner edge of this groove there is the center of rugosity which may mark
the anterior inner corner of the postorbital.

The posterior face of the skull (fig. 7 c and plate 5 d). — This face of the skull was
readily freed from the matrix and presents a smooth surface of uninjured bones, upon
which no marks or sutures could be missed. The whole face is semicircular, with the
prominent occipital condyle in the center of the horizontal lower edge. The surface
is completely closed, with the exception of the post-temporal openings (which are so
small as to be reckoned as little more than foramina) and the large foramen magnum.
In the median line above there is a thin but prominent ridge flanked by vertically
elongate pits. The whole surface slants backward at a slight angle and is concave
from side to side above the line of the opisthotics; below this line the distal ends of the
opisthotics slant more acutely backward, increasing the concavity. The line of division
between the opisthotics and the plate above is traceable inward for a short distance
beyond the post-temporal openings; the sutures between the exoccipitals and the ad-
jacent elements are completely obliterated. Just above the foramen magnum there are
small prominences which suggest the presence, in the complete skeleton, of a proatlas.
The opisthotics turn sharply outward from their junction with the exoccipital portion;
their posterior faces are smooth, with a low ridge which runs outward and slightly
downward on the inner half. The distal portion is slightly expanded and the fairly
broad, somewhat triangular distal ends are visible from the rear.

Between the distal end of the triangular terminal face of the opisthotic, which
points forward, and the quadratojugal there is visible a wedge-like surface of bone
which must be a portion of the nearly concealed quadrate. There is no trace of a quad-
ratojugal foramen. The occipital condyle is nearly spherical and is supported upon
a well-defined neck; there is no trace of a pit marking the forward extension of the
notochord. A distinct line marks the contact of the posterior plate of the skull with
the squamosal-quadratojugal; the plate sends a long process to fill the space between
the distal end of the opisthotic and squamosal-quadratojugal.

The lower surface of the skull (fig. 7 d and plate 5 c). — The pterygoids and the bones
of the palatal region are, most unfortunately, absent, but the bones of the brain-case
are so perfectly preserved that all the foramina can be made out; the sutures are all
closed and can not be traced. The phytosauroid characters are seen in the prominent
basioccipital-basisphenoid mass lying well below the level of the occipital condyle, in
the deep pit which lies in the usual position of the openings of the Eustachian canals,
in the strong basipterygoid processes of the basisphenoid, and in the strong alisphenoids
rising to the roof of the skull and firmly attached to it. The primitive character is
shown in the evident loose attachment of the pterygoids and the prominent, free,
parasphenoid rostrum.

The basioccipilal.— Anterior to the condyle the lower surface of the bone bends
sharply downward and forms, or approaches close to, the posterior edge of the deep

THE UPPER TRIASSIC OF WESTERN TEXAS. 31

pit in the basisphenoid. The suture between it and the basisphenoid is not distinct,
but on either side of the pit there is a deep groove which marks the position of the
suture; it is possible that the sutural region was occupied in life by a considerable mass
of cartilage, which in its decay has left the groove. The upper part of the anterior
edge is marked by a sharp, thin ridge; just anterior to this is a large foramen through
which passed the jugular vein and the IX, X, and XI nerves. Near the posterior edge
there are small foramina in the exoccipital portion of the complex, which permitted
the escape of the XII nerves.

The basisphenoid. — The lower surface is marked by the presence of a remarkably
large and deep conical pit. This is in the position of the openings of the Eustachian
canals in the Crocodilia, but the pit was thoroughly cleaned out and no traces of such
openings could be detected. On either side of the pit the bone is thickened and passes
forward to form the prominent basipterygoid processes. Posterior to the processes
the sides of the bone are marked by a groove and then swell backward and outward
to the edge of the groove which marks the line between the basisphenoid and the
basioccipital. Between the basipterygoid processes there is a thin parasphenoid rostrum
of considerable vertical extent. The anterior portion has been injured by decay, but
it is apparent that not a great deal has been lost. The upper part of the posterior end
of this process is excavated to receive the lower end of the infundibulum and the edges
of this excavation are continuous with the alisphenoids above. The anterior face of
the basisphenoid is rather broad and slightly inclined backward to the posterior face
of the cavity for the infundibulum. Near the middle of this face there is a pair of
foramina for the entrance of the internal carotid arteries. The point of juncture of
the basisphenoid and the bones above is not determinable. From the bottom of the
pit on the upper surface of the posterior portion of the parasphenoid rostrum foramina
run through the sides of the basisphenoid to open into the lower part of the groove
described as lying posterior to the basipterygoid processes; the function of these foramina
is not known.

The walls of the brain-case (fig. 7 e) are formed by the alisphenoids, the prootics,
the epiotics (?), and the opisthotics, which form a solid lateral mass without distinguish-
ing sutures. The outer surface of the brain-case is marked by two distinct ridges.
The first starts below the origin of the bar between the orbit and the temporal opening
and descends almost vertically, becoming continuous with the anterior edge of the
alisphenoid at a point about one-third of its height above its origin from the parasphenoid
rostrum. This ridge probably marks the line of contact between the prootic and the
alisphenoid.

The lower ends of the alisphenoids are closely approximated, forming a narrow,
opening through which the hypophysis extends into the infundibular cavity below;
above this the anterior edges of the bones rise nearly vertically, with a slight notch
which permitted the escape of the III and IV nerves; at about one-half of their height
the edges incline gently forward and contract into a second notch which marks the
escape of the II pair of nerves; beyond this they incline more sharply forward and
diverge to their contact with the upper wall of the skull.

Posterior to the first ridge, the walls of the brain-case are concave and perforated
by the large foramina for the V nerves; the bulk of this portion of the wall is formed by
the prootic. The second, or posterior, ridge on the wall of the brain-case comes sharply
forward from the rear. Its posterior portion is formed by the lower edge of the opisthotic ;
just above the otic opening it bends sharply downward and forward and very nearly
reaches the lower end of the first ridge. There are no sutures visible in the space between
these ridges, but there is a short, low elevation extending forward from just below the

32

\1W REPTILES AND STEGOCKl'HALIAXS I'HiiM

upper posterior portion of the temporal opening, which carries slight rugosities and
may indicate the location of the suture between the epiotic and prootic. The epiotic,
if ever a distinct element, is fused with the opisthotic in the specimen. The edge of
the anterior half of the posterior ridge is elevated and behind it is a groove which is
continuous with the groove on the side of the basisphenoid. On the inner end of the
opisthotic there are two openings — an upper, the otic opening, and a lower, the jugular
foramen.

The position and form of the quadrate is one of the puzzling features of the skull.
Between the distal end of the opisthotic and the quadratojugal is a deep pit, partly
cut off in front by a ridge which runs from the inner side of the posterior end of the
jugal-quadratojugal bar to the opisthotic. The quadrate apparently occupies the bottom
of this pit, and the articulation with the lower jaw must have been formed by the con-
cave surface, but no distinct articular surface is visible. If this interpretation is correct,
the quadrate fills a small triangular area between the distal end of the opisthotic and
quadratojugal, forms the bottom of the pit described, and runs forward for a short
distance on the inner side of the opisthotic and for an uncertain, but not great, distance
on the cranial wall. At best it must have been a relatively small and inconspicuous
bone.

opo

pt.f?

mx

_J Fie. 8. — Restoration of skull of Deematosuckut.

After ideas suggested by Watson and von
Huene. Dotted portions are supposed

to have ljeen lost.
opo., distal end of the Opisthotic. Other lettering
as usual.

In comparing the skull with that of other forms, it is at once apparent that its
closest relations lie with the Parasuchia, but the differences are so fundamental that
it can not be placed in that order. The most notable differences are the single temporal
foramen, the lateral and anterior position of the nares, and the extremely reduced
condition of the quadrate bone. The last is the most fundamental. The lateral
position of the nares is found in the Pseudosuchians and in some primitive Parasuchians.
The development of the upper temporal opening is so irregular, though constantly
present, in the two groups that its complete disappearance would be no great step in
specialization; or its absence might be regarded as of primitive character, since it has
become apparent that the presence of two temporal openings and arches is not the
primitive condition of the Reptilia. The great reduction of the quadrate is a character
which does not appear outside of the South African phyla, which can hardly be com-
pared with this form, for in those phyla the reduction of the quadrate is accomplished
in an entirely different manner. In most of the Reptilia the quadrate has a greater or
less vertical extent upon the side of the skull and is not always closely associated with
the bones of the brain-case, its main associations being with the bones of the suspensorial
region. Only in the Crocodilia (sens, lot.) does the quadrate extend far inward, having
relations with the wall of the brain-case and the bones of the otic region. It is not
possible in the specimen to determine whether the quadrate has the relations figured
by Huene for the I'arasuchia (Geolog. u. Paleont. Abhdlg., vol. x, N. F., Hft. 1, p. 31,
fig. 20), permitting the squamosal, epiotic, and prootic to appear in the posterior lateral

THE UPPER TRIASSIC OF WESTERN TEXAS. 33

wall of the brain-case, but it seems probable that all of these, except possibly a distinct
epiotic, had such an arrangement and that the quadrate runs forward on the brain-case
nearly the full length of the opisthotic and joined the prootic in front and perhaps the
squamosal above. The quadratojugal did not send any process forward on the anterior
or lower edge of the quadrate, as in the Crocodilia.

In figure 8 is given a sketch of the condition of the skull as originally imagined by
the author and as suggested to him by both Doctor F. v. Huene and Doctor D. M. S.
Watson. Reasons have been given above why this suggestion must be rejected, but
there is much weight in the objection offered by Doctor Watson that in the phytosauroid
reptiles the basicranial region is never so far below the quadrate. Future discoveries
may show that the alternative restoration is correct and that Desmatosuchus must be
placed among the Parasuchia. At present the restoration and explanation here given
seem to the author the only ones possible.

Description of the endocranial cast (plate 6, figs. A, B, c). — A description and
discussion of the endocranial cast of Desmatosuchus has already been published1 and
only a portion of it is repeated here.

"The cranial cavity of Desmatosuchus was completely cleared out, leaving the surface of the
hone in good condition with all the pits and foramina clearly marked. As may be seen in the fig-
ures, the cast as finally secured shows the general form and proportions of the brain cavity and the
positions of the main outlets. The anterior wall of the cavity was entirely cartilaginous or mem-
branous, and as this jxjrtion was not preserved in the fossil the opening was stopped with plastic
clay which is easily detected in the figures. For this reason, the form of the olfactory tract and of
the pituitary body is not completely shown.

"The olfactory tract was evidently large, as in most of the primitive forms, and extended well
forward directly beneath the upper wall of the skull. The cerebral portion was relatively small,
scarcely any swelling Ix-ing revealed in this part of the cast except at the posterior end of the
prosencephalic region. The anterior-lower part of this region was enclosed by the ali-, orbitosphenoid
bones, ami the approximation of the hones of the two sides forms notches in two places which indi-
cate the points of escape of the nerves which supplied the eye. No indication of the origin of the
II, III, or IV pair of nerves is shown on the cast, and no outlets except the notches mentioned.

"It is impossible in the cast to distinguish between the diencephalic and the mesencephalic
regions of the brain, but the area occupied by these two is marked by a slight but distinct depres-
sion, which is outlined by definite elevations, the posterior one amounting to a low, sharp ridge.
From this depression rise the processes, above and below, which may be referred to in general
terms as the epiphysis and the pituitary body.

"The upper process is complex and is, perhaps, composed of two parts. Just posterior to the
edge of the prosencephalic portion there are two protuberances which mark the position of a pair
of deep pits in the upper wall of the brain-case. In cleaning the skull it was impossible to be cer-
tain that the bottom of these cavities had been reached, but it seemed probable that it had. With
the aid of a dentist's mouth-mirror and fine curved awls the pyrite filling was picked out until it
seemed that the bottom had been reached, but because of the inaccessibility of the cavities and their
small diameter it is possible that the cavities may have been deeper and even that they may be
the entrances to foramina. Cope, in describing the endocranial casts of a Phytosaur, Belodon,
and of a cotylosaurian reptile, Diadectes, speaks of the 'lateral processes of the epiphysis' and in
describing the skull of Belodon sjH'aks of the process as lying in a 'large canal which enters the
posterior part of the orbit.' To this canal he gave the name of the orbitopineal process on the
casts. The function of this canal he was unable to determine, but suggests that it carried a nerve
or blood-vessel.2 In his earlier papers he was inclined to the belief that Diadectes was blind because
he could find no outlet for the optic nerve and because the structure of the animal suggested that

'('use. El C, On :m Endocranial ( ':ist from u Reptile, Deamatoguchu* rpurentu, from the Upper Triassic of
western Texas, .Jounce! of Comparative Neurology, vol. 33, No. 2, p. 133, 1921.

2 Dr. H. L Moodie, in conversation with the author, baa suggested that these processes may indicate a portion
of the eOUne of the ductus eucloh luphuticus.

34 NEW REPTILES AND STEGOCEPHALIANS FROM

it was burrowing in habit; as the parietal foramen is exceptionally large in this form, he was in-
clined to believe thai the orbitopineal canal might have carried a nerve from the large, probably
functional eye which occupied the parietal foramen, which in part supplied the necessary vision.
It is impossible to tell whether such a canal existed in Desmatosuchus, but if it was present it was
very small, and the author of this paper is inclined to believe that it did not exist. Moreover,
there is a decided difference in the endocranial casts in this region. In Belodon and Diadectes the
processes are large and rise from the sides of the epiphysis, in Desmatosuchus they are small and are
entirely anterior to the epiphysis. It is possible that if a true cast of the brain could be obtained
the origin of the processes might l>e found to be the same in all, but as the casts were all made from
empty cavities a similai origin should be apparent.

"The epiphysis is very different in form. In Belodon and Diadectes there is a strong posterior
process, and ( 'o|>e describes the epiphysis of the former as 'subquadrate.' The orbitopineal piocess
extends either directly outward from the side, Diadectes, or outward and forward, Belodon. In
Desmatosuchus the epiphysis is erect and narrow antero-posteriorly with no posterior process. In
both Belodon and Desmatosuchus the processes referred to as the epiphysis are casts of a deep pit
on the under side of the skull in the exact position of the pineal foramen in other reptiles, but in
neither of these is the roof perforated. In looking up this matter the author has found that much
uncertainty exists as to the exact character of this process in the brain; it is known that both the
epiphysis and the paraphysis may reach large size and that either one may terminate in a func-
tional eye; at least, either one may carry organs which possess the histological structures of the
retina and the crystalline lens. In some forms there has also been found a third ex-agination of the
brain, posterior to the epiphysis, called the pineal organ, which has a similar histological structure.
Wilder, in his History of the Human Body, states that it is the paraphysis which was developed in t he
extinct Stenocephalia and filled the parietal foramen, and the epiphysis which was developed in
the reptiles, birds, and mammals. On the other hand, it is known that the epiphysis is not devel-
oped in the modern alligator. The term epiphysis is used in thjs paper only in a general sense and
without knowledge of its true nature.

"On the lower side of the diencephalic region of the cast is the second process; this represents
the combined infundibulum and the saccus vasculosus, or the pituitary body. Only the posterior
and lower borders are shown, for the anterior part was enclosed by the cartilaginous anterior wall
of the skull which was lost in fossilization. The process extended directly downward by a narrow
neck which passed through a narrow notch formed by the approximation of the alisphenoid bones
at their lower borders. Its posterior face lay against the basisphenoid bone, not penetrating it,
and its lower surface in an excavation on the upper surface of the posterior part of the parasphenoid.
The lower part of the process was enlarged and the posterior face, at least, extended backward at
a sharp angle. The lower end terminates in a bifurcate extension which is formed by the casts of
the begirmings of two foramina which open outwardly and downward in an excavation on the upper
surface of the parasphenoid. These foramina continue and terminate in deep grooves on the side
of the basisphenoid. On the posterior face of the process are two small prominences which mark
the position of two foramina on the lower face of the basisphenoid, evidently the openings for the
internal carotid arteries.

"The posterior part of the depressed area mentioned above must also include the mesen-
cephalic portion of the brain, but there is nothing to mark the presence of either optic lobes or optic
thalatni. This does not, however, suggest either their absence or relatively small size, for if a cast
were made of the cranial cavity of Sphenodon or of an alligator no evidence of these structures
would appear, though they are of large size.

"Posterior to the depressed area tht whole cast is curved sharply downward and then straight-
ened out horizontally in the metencephalic region. On the lower edges of the anterior part of this
region there are large prominences which mark the position of the large foramina for the passage
of the V pair of nerves. There is no indication in the cast of the division of this nerve into its
parts; this must have taken place external to the cranial wall. Within and a little posterior to
these prominences is indicated the position of a pair of small foramina in the floor of the skull,
evidently the outlets for the VI pair of nerves. Posterior to the Y and at about the middle of the
posterior part of the cast, there are a pair of processes on each side, one almost directly above the
other. The upper pair are the caste of the otic cavities and mark approximately the position of
the \ III. and probably, also, the VII nerves, for these two nerves escape from the skull of the Croco-

CASE

PLATE 6

atostichtu tpunturu, \o. 7476, U. of Mich. Endocranial cast. X about 0.6.
A. Lateral view. B. Upper view. ('. Lower view.

I). Dorsal armor, side view of lateral plate of dorsal region. X 0.23.

!•:. Dorsal armor, top view of median plate of dorsal region. X 0.23.

1'. Dorsal armor, side view of lateral plate of caudal region. X 0.23.
G. Dorsal armor, top of fifth cervical series. X0.32.
H. Dorsal armor, posterior view of fifth cervical series. X0.32.

THE UPPER TRIASSIC OF WESTERN TEXAS. 35

dilia in almost the same place. The otic cavities were injured in fossilization both by pleasure
and by the crystallization of the gypsum and pyrite which filled the cavities of the skull. It is
apparent that there was a thin wall between the otic cavity and the brain cavity, but this has
been so injured that it is impossible to determine the original form of the otic cavity or the form
and position of the semicircular canals.

"Below are the large cylindrical projections which filled large foramina carrying the IX, X,
and XI nerves and the jugular vein. All of these must have escaped through a common opening,
as the walls of the brain-ease are very perfect in this place and no other openings are present.

"Near the posterior end of the east are slender processes which mark the position of the XII
nerves. Above these processes there are small prominences which filled pits in the inner walls of
the exoccipitaJ hones. These pits were entirely cleared, and it is certain that they were not the
beginnings of foramina, their meaning is unknown.

"The whole meteneephalic portion of the cast is rather high and narrow. It is possible that
this is due in some degree to crushing, but there is no indication of such crushing in the skull, and it
is probable that it is the true form. The whole cast is very small relative to the size of the animal,
and even assuming that the brain occupied the whole cavity its size would be remarkable, though
after all it is not much smaller, relatively, than the brain of Sphenodon or of an alligator.

"It is difficult to make any satisfactory comparison of this endocranial cast with the one made
by Cope from the specimen of Bdodon buceroe because of the unsatisfactory nature of his figures,
but some points can be made out. As can lie seen by the figures in his paper, the whole shape is
different, the east from Bdodon docs not show the sharp downward curve posterior to the middle
region. The cast of Desntalosurhus is thinner for its height and does not have so long a meten-
eephalic portion.

"The epiphysis lacks the posterior prolongation, certainly it is not 'subquadrate' in form,
and the lateral processes rise in front of. not at the sides of, the epiphysis. The olfactory tracts
were much broader. The optic nerves did not escape through distinct foramina. Only the origin
of the pituitary body is shown in Cope's figures, but he describes it as small and occupying a fossa
in the base of the cranial cavity. These characters support the evidence afforded by the bones of
the skull that Desinatosiulntx must l>c placed, at least, in a distinct suborder from the Phytosauria."

The vertebral column (see plates 6 to 10). — The exact number of presacral vertebra?
has never been determined in the Phytosauria. McGregor says that it is "practically
certain that there were not less than 25 or more than 27" in the presacral series. In
the mount and restoration of Desmatosuchus, 28 vertebrae have been included in the
presacral series; while this number may not be exactly correct, every consideration
of the condition of the specimen and the character of the vertebra; indicates this
number; of all the material found, three cervicals, which are obviously duplicates, and
one dorsal have been excluded in making the mount. It is possible that the single
dorsal should have been included, but it is apparently a duplicate and its inclusion
would have made the presacral series seem unduly long. This restoration is subject to
correction by future discoveries.

The atttu and axis (fig. 9 a to d). — The first two vertebra; are closely united but
not eoossified. The shape of the two is very similar to that figured by McGregor1
for Mystriottichw. The atlantal ring is complete, with but slight rugosities indicating
the position of the suture between the intereentrum and the neural arch. The arch is
completed above by a thin plate which passes obliquely backward beneath the anterior
end of the neural spine of the axis. On either side the centrum bears projections from
the anterior edge near the lower level of the neural canal; these are marked off by a deep
notch below and project sharply from the anterior edge of the neural arch above. The
one on the right side is smooth, as if there were an articulation for a proatlas. This
may be deceptive, however, as the corresponding projection of the other side does not
have this appearance. There are strong projections from the rear of the neural arch

McGregor, J. H., American Museum Natural History, Memoir ix, pt. n, p. 62.

36

\i:\V REPTILES AND STEGOCEPHALIANS FROM

which extend back to the middle of the axis and carry distinct facets for articulation
with the anterior zygapophyses of the axis. The lower face of the centrum is broad
and flat or even slightly concave. On the posterior external corners there are small
I nit prominent processes for the capitals of the atlantal ribs; smaller processes, almost
destroyed by decay, lie on the sides of the arch near the level of the base of the neural
canal. The anterior face is deeply concave below, forming a socket for the nearly
spherical occipital condyle. The upper edges of the face are drawn in sharply below
the anterior processes and form the sides of the deep notch which receives the anterior
end of the odontoid process. The diameter of the neural canal is considerable.

Fig. 9. — Desmatosuehus spurensis.

A. t'pper surface of atlas and axis. B. Anterior surface of same. C. Anterior surface of a
second axis, No. 7504, U. of Mich. D. Lateral surface of same. All figures X 0.5.

The axis is much larger than the atlas, and though the sutures are closed it is
evident that the vertebra is composed of the elements suggested by McGregor, an
intercentrum and the axis proper. The neural spine is elongate and thin, except at
the anterior end, where it is expanded and heavy; this portion lies upon the down-
wardly and backwardly sloping arches of the atlas. The anterior zygapophyses are
covered, but are of good size, with the faces looking almost directly upward. The
posterior zygapophyses are nearly normal in form ; they overhang the posterior edge
of the centrum for some distance. The lower face of the centrum is flat from side to
side and concave antero-posteriorly on the posterior or true axial portion. The process
for the capitulum is larger than for the tuberculum and a little in advance of it. The
neural canal is still large, but smaller than in the atlas; the transverse diameter is the
larger. The posterior face of the centrum is deeply concave.

A second axis found with the remains of Desmalosuchus, and evidently belonging
to an individual of the same genus and species, is even more perfect. In this (fig. 9,
c and d) it is seen that the odontoid process is quite broad; the posterior zygapophyses
show a peculiarity which is continued backward for some distance in the cervical series;
the posterior end of the zygapophyses extends beyond the articular face and on its
upper surface there is a decided prominence.

The next two vertebra; (fig. 10) were found isolated in the matrix of the specimen
and are regarded as the third and fourth because of their size and shape. They are in
poor condition, but show the main characters. The neural spine of the third is partly
rotted away, but the base of the spine indicates that it was thin laterally and somewhat
elongate. It teas evidently low. The anterior and posterior zygapophyses are on
nearly the same level and are normal in form and position. The transverse process is
elongate and extends nearly straight downward, but a little to the rear. The sides of
the centrum are concave and the lower end of the transverse process is free for some

36

CASE

PLATE 7

Vertebra of Dcxmatosuchus spurensis
All figures X0.45.

A. I'il'thC.') cervical.

B. Tenth cervical.

( '. Twelfth vertebra.
D. Anterior dorsal.
E to H. Mid-dorsals.
I and J. Posterior dorsals.
K and L. Anterior dorsals.

THE UPPER TRIASSIC OF WESTERN TEXAS.

37

distance. The capitular facet stands on a short process (fig. 10) rising from the anterior
outer angle of the lower face of the centrum; this process is inclined slightly backward.
The anterior face of the centrum is nearly circular; the posterior face is oval, with the
long axis horizontal. The neural canal is still large.

The fourth vertebra has lost the upper part of the neural arch and the zygapophyses.
The transverse process (fig. 11 b) is curved outward and downward; the capitular process
is similar to that of the third. The lower face of the centrum is concave antero-
posteriorly, but nearly flat transversely. The anterior and posterior faces are trans-
versely oval.

Fk,. 10, lit sitiiiUisnchus spurenxis.

A. Lateral view of first tea vertebra;; the last six as they were found connected.

B. Upper view of third to ninth cervical vertebra). X 0.3.

X0.3.

The next six vertebra; (figs. 10 and 11) were found in close association, the
first five in position and the sixth slightly distant and turned to one side. The last is
thought, from its shape, to be the twelfth of the whole series and will be described in
that position. In the first five the faces of the centra are transversely oval and the
lower surface is fiat. These characters change slowly toward the rear until in the
tenth vertebra (fig. 11) the posterior face is nearly circular and the lower surface of
the centrum is much narrower and longer than in those anterior to it. The zygapophyses
of all are low and flat and closely overlapping. The posterior zygapophyses of the
fifth vertebra show clearly a structure which is only indicated in the less perfectly
preserved ones adjoining it, but is very prominent on the second axis. From a point
just above the articular facet there is a strong spinous process which extends backward
and outward for at least a centimeter. This character is confined to the anterior
cervicalfl, as it is not noticeable on the eighth vertebra, where the zygapophyses are
thin and broad.

38

NEW REPTILES AND STEGOCEPHALIANS FROM

The neural spine of the fifth is low, thin, and broad, but posterior to this one the
spines of the vertebra1 increase in height, become thinner at the base, transversely, and
elongate antero-posteriorly, and develop a knob or expansion at the apex. The
transverse processes rise progressively toward a horizontal position and increase in
length and thickness. On the sixth vertebra is seen the beginning of a ridge running
from the upper outer angle of the posterior face of the centrum to the base of the
transverse process; this continues to increase in importance until on the eighth it is
continued on to the lower face of the process as a thin supporting ridge inclined backward,
and the cross-section of the process is that of a T-beam. In the same manner the
capitular process increases in thickness and weight, but not in length. On the ninth
vertebra this process is over a centimeter in its two diameters, and a low, sharp ridge
runs from its upper face upward near the anterior face of the centrum, until it joins t he
ridge on the lower face of the transverse process at the level of the base of the neural
canal.

Flo. 11. — Desmatomichus spurensis.

A to G. Anterior views of the third to ninth cervical vertebrae. The figure of the fifth vertebra is slightly

restored. X 0.8.
1 1 Posterior view of a duplicate tenth (?) vertebra, No. 7504, U. of Mich. X 0.6.

In all these vertebra; the neural arch is deeply excavated between the zygapophyses
l>oth anteriorly and posteriorly, so that a considerable portion of the neural canal is
left uncovered between the vertebrae.

The vertebrae posterior to the ninth were all found isolated, with the exception of
the one reckoned as the twelfth and a few in the mid-dorsal region. They have been
placed according to their size and form. There may be some error, but it can hardly
be more than one position.

In the vertebra reckoned as the tenth the zygapophysial processes are broad and
thin and the faces are nearly horizontal. The transverse process is broad above; its
anterior edge is nearly on a line with the center of the face of the anterior zygapophyses.
The T-section of the process is well developed. On this vertebra a ridge extends back-

THE UPPER TRIASSIC OF WESTERN TEXAS. 39

ward from the anterior upper angle of the centrum and ends on the anterior face of the
posterior ridge as it turns outward on the lower surface of the transverse process. The
neural spine is elongate antero-posteriorly. The expanded upper end is wedge-shaped,
with the faces of the wedge directed laterally; the spaces between the faces of the wedge
are roughened. The lower part of the anterior face of the spine is rendered deeply
concave by the development of ridges on each side, which continue downward and
outward to end upon the surface of the anterior zygapophyses. The space between
these ridges and the zygapophyses is floored by a thin plate of bone which covers the
neural canal. On the posterior face of the spine there is a similar development, but the
excavation of the face of the spine is much deeper. The faces of the centrum are still
transversely oval and the lower surface flat, but the faces are approaching a circular
form and the lower surface is becoming rounder.

The eleventh vertebra has the transverse processes broken away, but the bases
are still preserved. The capitular face is now elongate vertically, and a strong ridge
runs from its upper origin, above the level of the base of the neural canal, backward
and upward to join the posterior ridge which forms the lower arm of the T-shaped
transverse process.

The neural spine is elongate antero-posteriorly and thin transversely. The upper
end has a triangular expansion on each side, but the ridges on the edges of the anterior
and posterior faces run to the apex. The depression between these ridges is similar to
that in the preceding vertebra?; the zygapophyses are closer together. In the tenth and
eleventh vertebrae is seen for the first time the deep depression in the floor of the neural
canal. The anterior face of the centrum is transversely oval and much larger than the
posterior face. This is in part due to the development of the capitular process with its
vertical extension.

In the twelfth vertebra the capitular process rises from a point above the lower
edge of the centrum, which is now rounded, and is elongated vertically, the face being
nearly twice as high as broad. The ridge upon its upper face, noted as beginning in the
eleventh vertebra, is now developed as a strong process which rises as a wall and
is united with the lower part of the transverse process at about the middle of the
length of the latter. The ridge from the posterior corner of the centrum is still
strong at its origin, but has become obsolete on the transverse process itself; the
transverse process is still T-shaped, but the lower flange is now formed by the anterior
ridge instead of the posterior one. In this vertebra the transverse plate between the
zygapophyses has a different form. From the inner edge of each of the posterior zyga-
pophyses a thin plate runs inward to form the transverse plate and there is a vertical
extension downward in the form of a very thin septum which reaches to the upper edge
of the neural canal, which is not more than half as large as in the cervical vertebra?.
There is also a tendency for a similar plate with a vertical septum to form between
the anterior zygapophyses.

In the thirteenth vertebra the transverse process stands out directly from the side
of the vertebra and rises slightly in its course, so that the distal end is higher than the
origin of the process. The capitular face has now left the centrum entirely and is attached
to the transverse process at about its middle portion. This face is supported by three
ridges — one very slender, running from the lower edge of the face to the anterior
upper angle of the centrum, the second running from the lower edge of the face upward
and inward to the lower surface of the base of the anterior zygapophyses; the third
ridge is the anterior half of the posterior ridge. Its posterior half, so strong in the
preceding vertebrae, is now obsolescent. Its position marks the dividing-line between
the capitular and the tubercular portions of the transverse process. This sudden change
of position in the capitular face is characteristic of the Crocodilia and the Phytosauria;

40

NEW REPTILES AND STEGOOEPHALIANS FROM

with it, in the present specimen, begins the elevation of the sides of the neural arch which
is so notable in Desmatosuchus. The neural spine is lost, but the space between the
gygapophysea is much narrower. The centrum is narrower and more elongate, with
the base rounder and the faces more nearly round.

From the fourteenth back to the twenty-fourth (?) the vertebrae are of the true dorsal
series (fig. 12). The transverse processes increase to their maximum length and are
inclined slightly backward, the capitular faces gradually come to lie on the anterior

A. Anterior view of a mid-dorsal vertebra

H. Posterior view of A.

('. Lateral view, left ride of A.

I). Upper view of A.

Fi<;. 12. — Desmatosuchus .iparensis. All figures X 0.3

E

Lateral view df a posterior dorsal vertebra with the

rib of left side attached,
l'\ Posterior view of E.
(i. Upper view of a posterior dorsal. The broad apex is

broken off.

side of the transverse processes, and the supporting ridges become obsolete on the bice
of the elevated neural arch. The neural spines become more elongate antero-posteriorly
and the apex La expanded abruptly into a rounded rugose table. The centra become
elongate antero-posteriorly, and the edges of the faces flare out so that the lower face is
concave antero-posteriorly and convex from side to side.

THE UPPER TRIASSIC OF WESTERN TEXAS.

41

The posterior dorsal vertebrae (fig. 12 e and f) become heavier in all of their propor-
tions; the spines remain thin transversely, but the apex becomes heavier; the transverse
processes become shorter by the abbreviation of the distal end, so that the capitular
face approaches the tubercular. Even in the posterior dorsals the sides of the neural
arch remain elevated and the transverse processes take their origin entirely from them.
The edges of the faces of the centra are expanded so much that a section of the middle
of the centrum is not more than half as large as that of the faces.

Fig. 13. — Desmatosuchus spurensis. All figures X 0.3.

A. Posterior view of the first true caudal vertebra.

B. Lateral view, left side of A.

< ' Lateral view of a median caudal vertebra with the incomplete chevron, left side.

D. Lower view of C.

E. Posterior (upper) view of the chevron shown in C.

The posterior presacral vertebrae resemble those of the dorsal series in all major
particulars; the last three or four become decidedly heavier, with larger centra, and the
transverse processes become broader antero-posteriorly, with the tubercular and capitu-
lar faces at nearly the same distance from the origin of the transverse process. The
ribs are lost, but there is evidence that even to the last the ribs were articulated with
the transverse process and not reduced and coossified with them; in this sense there are
no true lumbars.

A single very poorly preserved vertebra which seems to have a very heavy neural
spine and a short, heavy transverse process, or origin of a sacral rib, may be regarded as
one of the sacrals. This region is in poor condition; only a portion of a very poorly
preserved half of the pelvis, as noted below, is preserved.

42

NEW REPTILES AND STEGOCEPHALIANS FROM

There are six caudate preserved, all chevron-bearing (fig. 13). It is probable
that there were two or three pygal vertebrae present. In the caudals preserved the trans-
verse process springs from the side of the neural arch and curves outward and down-
ward, becoming thin and more horizontal toward the distal end. The zygapophyses
are well developed, with the faces nearly flat ; the characteristic recess continues at
the base of the neural spine, but with the decrease in height of the neural arches the
median descending vertical plate between the zygapophyses disappears. The first
three, which are evidently anterior caudals, have heavy spines with thickened apices.
The lower face of the centrum is marked by a median depression which separates two
strong ridges on the sides; these terminate posteriorly in the faces for the chevrons,
which look more downward than backward.

Fig. 14. — Desmalosuchus spurensis. All figures X 0.3.

A. Upper surface of the seventh (?) rib of the left side.

B. Lower view of A.

C. Lower view of tenth rib, left side.

D. Upper view of B.

E. Upper view of eleventh rib, right side.

F. Lower view of E.

The last two caudals (fig. 13 c and d) are in poor condition, due to decay, but are
more elongated than the anterior ones, and there is a decided tendency for the posterior
faces to descend lower than the anterior ones.

The chevron bones (fig. 13 A and e) are widely separated at the proximal end, but
meet in a strong terminal portion. The exact length is not shown in any one.

The ribs. — There were free ribs on all of the vertebrae of the presacral series, but
the posterior ones show a strong tendency to a close anchylosis with the transverse
process. Those of the cervical series were small and turned back close to the side of
the vertebrae, as is evidenced by the small amount of space between the vertebrae and
the sides of the dorsal armor. The first rib preserved was, in all probability, attached
to the seventh vertebra, as it was found close to it and fits the articulations quite closely.
This rib (fig. 14 a and b) is still short, and the capitular and articular processes are
nearly equal in length. The dorsal surface of the rib is expanded into a broad, thin
surface, which is supported by a thin, high ridge running the length of the rib on the
under side; this ridge has its origin on the tubercular process.

CASE

PLATE 8

5S

o

a.
a

oj

& A

£

03
g

«3

a

a

-^
i

o

a

s

=

SP £ ~£

co w

cu 0->

"H -C >S
o ^s „g

e *

a

9

it

3

s

E

9

N

CO

3

oo

. C
~3 O

CD V?

o3 ev
.S ?
00 to

§2

.2 0

T3

4)

09

C

I

—

W

o '—

. o

o O

J- > > —

S3 — -

0

*

>

3

—

THE UPPER TRIASSIC OF WESTERN TEXAS.

43

The second well-preserved rib is assigned to the tenth vertebra (fig. 14 c and d).
It is much longer than the seventh and the tubercular process is much shorter than the
capitular. There is still some expansion of the upper surface, though much less than
in the ribs anterior to it. Traces of the supporting ridge on the lower side are still
present.

The third well-preserved rib is assigned to the eleventh vertebra (fig. 14 e and f).
This closely resembles the rib assigned to the tenth, but is even longer, and the expan-
sion of the upper surface is more closely confined to the proximal end. The capitular
and tubercular faces are close together, corresponding to the approximation of the facets
upon the vertebra?.

A fourth well-preserved rib is assigned to the
twelfth vertebra (fig. 15). This rib is much stronger
than the preceding ones and longer. The whole rib is
flat and rather thin. The capitular process and face
are stronger than on the anterior ribs and the tuber-
cular face is located as a notch on the body of the
rib. This is the first rib that shows any decided down-
ward curvature of the shaft.

In the following ribs, which are only incompletely
preserved, the tubercular and capitular faces are on
the same plane; the ribs are relatively broad in the
dorsal scries, with a considerable outward extent, indi-
cating a well-rounded barrel in the animal. In the pos-
terior dorsal and lumbar (?) regions the ribs become
shorter and more slender; the capitular and tubercular
faces gradually approximate until they are nearly on
a line.

The dermal armor (plates 8 and 9). — The back
of DexHtatoxucliii.s was covered by a complete armor
consisting of four rows of plates. The presence of any
lateral or abdominal armor is uncertain. Aside from the presumption that such plates
occurred, there is a single small, irregular plate which was found intimately associated
with the rest of the armor and skeleton.

The dorsal armor consists of four rows of plates arranged symmetrically in succes-
sive transverse series. The two inner rows are composed of flat plates with a small
subcentral knob; these plates increase in breadth toward the dorsal region, at the same
time decreasing in their antero-posterior extent. The outer rows are composed of plates
which have the two sides of the bases bent almost at right angles to each other, so that
a portion lies over the back and a portion over the sides; they bear spines which are of
varying height in the different parts of the body. It is probable that the median rows
disappeared in the posterior-caudal region, but of this there is no direct evidence. It
will be best to describe the plates according to the transverse series. The general
arrangement La shown in the plates and figures. The transverse series of dorsal plates
begins immediately behind the skull, but the anterior ones are not arranged in relation
to the vertebrae below; the first five series of plates cover the entire cervical series of
vertebra;.

The plates of the first series are represented by the lateral one from the right
side; it is incomplete, but shows a decided spine.

The second series is represented by the median plate of the left side and the two
lateral plates. In common with all the plates of the cervical region, the median plates
were quite thick on the inner and outer edges and thin on the anterior and posterior

Fig. 15. — Desmatosuchus spurensis.

A. Lower view of the twelfth rib, left

side. X 0.3.

B. Upper view of A.

44

\i:\V REPTILES AND STEGOCEPHALIANS FROM

edges. The low knob is nearer to
anterior edge is the broader and the
terior edge. The left plate of the
inner edge is thickened to form a firm
and slopes backward and outward t
on the outer half of the plate and
anterior surface of the spine is drawn
to the middle of the plate.

the posterior corner of the plate (fig. 16 a). The
outer edge slopes backward and inward to the pos-
outer row of this series is nearly complete. The
union with the inner plate; the outer edge is thinner
;o the broader posterior edge. The spine is located
is inclined outward and slightly backward. The
out into a narrow, sharp edge which slopes forward

Fig. 16. — Desmatosuchus spurensis, dorsal armor of. All figures X 0.3.

A. Upper view of median plate of second series, right side.

B. fjpper view of median plate of the third series, left side.

C. Upper view of the two median plates of the fifth series.

D. Anterior view of the right side of C.

E. Inner (upper) view of the lateral plate of the third series, left side.

F. Posterior view of E.

The third series has the plates (fig. 16 b) nearly complete. The median plates
are similar to the second, except that they are proportionately larger. It is clear in this
series that the plates overlap to the rear, and this is continuous throughout the series.
There is little, or a very fine, sculpture. The outer plates have high spines located
near the posterior edge, with a thin flange extending down the anterior side. This is
shown in figure 16 a and b. The two parts of each plate lie nearly at right angles
to each other. The inner edge is thickened for articulation with the median plate,
while the outer part is thin and larger. The anterior edge is narrower, but the outer
edge slopes backward and outward to meet the broader posterior edge. The outer side
shows the beginning of a rough sculpture of relatively small pits and ridges. When
in position the spines point outward and a little upward.

The fourth series is represented by the conjoined plates of the left side and an
imperfect median plate of the right side. The left plates are complete, except for the
extremity of the spine. The median plates are similar to those preceding, but are
larger. The outer plate carries a heavy spine, of which the lower half only is preserved.
The inner side of the spine is nearly flat, but there is a thin, strong ridge on the anterior
side. As in the preceding plates, the anterior edge is shorter than the posterior, and the
outer edge slopes backward and outward. Both the inner and the outer plates show an
increase in the size of the sculpture, which becomes so heavy and rough in the dorsal
region.

CASE

PLATE 9

THE UPPER TRIASSIC OF WESTERN TEXAS.

45

The fifth series is represented by two complete median plates (fig. 16 c and d), the
complete plate of the outer row of the left side and an incomplete spine of the right side.
The median plates are similar to the preceding plates. The outer plates present the most
remarkable feature of the armor. Each one bears an enormous spine, over 45 centimeters
in length ; this extends almost directly outward, but a little upward, and from its middle

Fio. 17. — Desmalosuchiu spurensis.

A. Lower (outer) view of the lateral plate of the fifth series in the dorsal armor,

left side. X 0.3.

B. Posterior view of A.

point curves backward as far as the middle of the seventh series of plates (fig. 17). The
base of the spine is flattened on its inner and anterior side, but is rounded on the outer
side. The angle between the inner and anterior faces is continued to about the middle
of the spine and then disappears, the section then becoming circular, which continues
to the acute distal extremity. The base of the plate is actually larger than the preceding
ones, but relative to the spine much smaller; the inner edge is thickened for attachment
to the median plates; the outer edge is thin and the anterior and posterior edges are
of about equal width.

The accurate fit of the plates of this row permits a determination of the outline
of the armor at this point, and this is checked by the conjoined plates of the fourth ring
(see plate 8 e to h). This is most fortunate, as the position and attitude of the spine
would hardly be credible otherwise. The position of the great spine curving outward
and backward could hardly have been imagined, and its meaning is still a puzzle.

Posterior to the fifth series the armor becomes flat and broad and descends but a
short distance on the sides. The plates of the median row rapidly contract antero-

46

NEW REPTILES AND STEGOCEPHALIANS FROM

posteriorly and broaden transversely through the sixth, seventh, and eighth rows, and
by the ninth they have reached the normal form of the dorsal series, which continues
back nearly to the sacral region. The low spine or knob is located a little outside of
the center and is directed backward and inward. The plates are covered by a rough
sculpture, shown in figure 18 c, which extends to the thin edges except for about 1 or 1.5
centimeters on (he anterior edge. This smooth area on the edge of the plate indicates an
overlapping, but it must have been very slight in the dorsal region. Examination of the
plates in the armor of several species of alligators, crocodiles, and caimans shows that there
is no, or very slight . overlapping of the plates in any of these forms, and that it occurs only
in an overflexion of the back. The same smooth area appears on some of the plates
in these living forms and seems to be a region of attachment of tough skin or tendinous
material. The outer plates have low, sharp spines, which are broader antero-posteriorly
in the sixth and seventh rows, but of nearly equal diameters in the ninth and following
ones. The spines were inclined outward and a little backward. The inner half of the
base is shorter than the outer. The inner half is nearly horizontal and the outer half is
nearly vertical, but inclined a little outward. There is evidence from one side or the
other of 17 plates of this kind, making at least 9 rows in the mid-dorsal series.

Fhj. lK—DesHHitosuchiisi spurensis.

A. Median plate of a mid-dona] scries, x 0.3.

B. Outer view of a lateral plate of a caudal series, left side.
( '. Posterior view of B.

X 0.3,

Posterior to the plates of the dorsal series there are no representatives of the median
plates, but there is a considerable number of plates from the outer series. They differ
notably from those of the prepelvic region (fig. 18 b and c). The outer half of the base is
much longer than the inner, and is in the same plane with the outer sides of the spines.
This character becomes more noticeable as the pjates become smaller, that is, in the pos-
terior part of the caudal series. The spines are relatively higher than in the dorsal series.
The sculpture becomes less rough in the posterior plates. It is believed that in the
ilistal portion of the tail the median row disappeared, as it does in the Crocodilia.

The srapula-coracoid. — With the skeleton was found the distal end of a scapula-
coracoid .so radically different from the usual form found in the Phytosauria that it is
difficult to reconcile it with the vertebral column, but it is no more peculiar than the skull.
It does, however, resemble very closely the scapula-coracoid of Slagonolepis. The
cotylus is large and cleanly formed and is somewhat obliquely placed in the bone. The
upper edge overhangs strongly; the lower edge is broken away. A ridge interrupting
the smooth face of the articular surface is perhaps due to pressure, but seems normal
•lust anterior to the edge of the cotylus is a pit which seems to belong to the bone and
not due to the action of decay and gypsum. The coracoid foramen is at about the level
of the lower edge of the cotylus and perforates the bone upward and inward; on the
posterior side the opening of the foramen is directly upward. On the anterior edge

CASE

PLATE 10

«f'

a
o

1

o

—

©

g S

§ °
o

* O

03
E
-Q
0)

t, a

¥ o

> --S

05 o

I n

» .3
I? "C

a. "2

m

e

« a

5 .2
<=> £

a.

3

c

-3

THE UPPER TRIASSIC OF WESTERN TEXAS.

47

opposite the upper edge of the cotylus there is a decided prominence; below this the
edge of the bone is quite thin. The lower and anterior edges of the coracoid have been
injured by decay, but could not have had a very much greater extent.

Measurements of Desmatosuchus.

Length of skull on right side 27.4

Width of skull at posterior end 16.4

Length of axis and atlas at base 7.6

Length, anteiior end of skull to first sacral 198.6

Length of large spine along outside curve 39.4

Same with extreme tip, now lost, added 40.9

Extent of long spine from mid-line 41.2

Average length of mid-dorsal plate 16.8

Average length of same, antero-posterior 8.4

The restoration (fig. 20). — In the attempted restoration consideration has been
given to several factors. It is obvious that such an animal as Desmatosuchus, with its
spiny armor, would have considerable difficulty in making progress through tangled
vegetation, either aquatic or terrestrial, and it is in accord with this character and the
suggestion of the sediments that sparse vegetation and open water are indicated in the
environment. It must be remembered
in considering the nature of the animal
that its remains are very limited in
quantity in a region where the remains
of Phytosaurs are relatively abundant.
It would appear that Desmatosuchus
inhabited a region somewhat different
from that occupied by Belodon and
Mystriosuchus, either more remote from
the pools which formed the favorable
environment of more common forms or
a region geographically remote from
the locality where the collection was
made. It is exceedingly unfortunate
that no teeth were found with the
specimen, and so no more than a guess
can be made of the nature of its food;
but as the sockets are simp'e and
rounded, it is safe to assume the teeth
were single cones and that the animal
was carnivorous. Inhabiting an area of
open land, it is probable that the animal must have had relatively long limbs, a prob-
ability borne out by the size of the pelvis and the shape of the acetabulum, and further
by the facts that no short limb-bones were found in the locality and that the frag-
ment of the shaft of a limb-bone found with Desmatosuchus is evidently part of a long
femur or humerus. No trace of the feet or claws was found, and the suggestion of par-
tially webbed feet, implying a dominantly aquatic habitat, may be erroneous. The
weight of the great carapace would, perhaps, also be an objection to the assumption of an
aquatic life. For the same reason — the possibility that the habitat was not dominantly
aquatic — it is possible that the tail has been made too long and too much of a swimming
organ.

Critical study of the specimen in the course of mounting the skeleton has led to
the conclusion that the plates of the dorsal armor did not overlap when the animal was
in a normal position, but overextension, as by a strong upward curvature of the body,
would have been possible only by the overlapping of the plates. This is the condition
found in the modern Crocodilia.

FlO. 19. — Desmatosuchus spurensis.

Anterior view of the distal portion of the scapula-coracoid'

right side. X 0.3.
Outer view of A. X 0.3.

IS

NEW REPTILES AM) STi:<;oci:iMIAUA\S KHOM

The relationships of Desmatosuchus. — To the author it seems apparent that Des-
matosuchus can bo considered only as a member of a branch of the Parasuchian stem
which has developed a high degree of specialization. The character of the vertebral
column and the carapace indicates close relationships to the Parasuchia; all the remarkable
features of the carapace are simple developments of possibilities clearly indicated in the
more conservative line.

4?sS ^ >?

"J1'

'JJ

FlO. 20. — Restoration of Desmatosuchus spurensis.

It is in the skull that the morphological divergence from the line of the Parasuchia
is clearly evident. The author, as indicated in the body of the description, is unable to
accept the suggestion made by Doctors Huene and Watson that the single large temporal
opening is the upper and that the bones outlining the lower opening have been lost.
Reasons for this opinion have been given in detail in a preceding portion of this article.
The angle temporal opening and the amazing condition of the quadrate region clearly
place the animal in a separate suborder or even order. The basicranial region, the
antorbital opening, and the lack of a pineal foramen, together with the character of the
vertebral column and the carapace, just as clearly indicate affinities with the Parasuchia.
I 'nt il future discoveries shall prove the author's interpretation of the skull to be erro-
neous, or shall reveal further structures which will make more evident- the affinities of
the animal, DatmaUmichtu must be considered as a member of a distinct order or suborder
of phytosauroid reptiles, highly specialized in its morphology and confined to the Upper
Triassic of North America.

CASE

PLATE 11

f«'

THE UPPER TRIASSIC OF WESTERN TEXAS. 49

A NEW PARASUCHIAN, PROMYSTRIOSUCHUS EHLERSI.

The specimen (No. 7487, University of Michigan) described below was discovered
by the author in a bed of yellowish sandy clay near the head of Holmes Creek, Crosby
County, Texas, in the summer of 1921. The patch of sandy clay is but a minor phase
of the larger deposit in which so many remains of reptiles and amphibians occur in this
locality. Near the specimen, and in the same kind of matrix, occur large numbers of
fragmentary bones, teeth, and coprolites. This small patch is evidently a bit of an old
sand-bar, or an accumulation of sandy clay in some small depression which has been
well leached by percolating waters. The relatively small percentage of clay rendered
the matrix easily permeable by the waters which converted the iron into limonite and in
spots removed it so completely that the matrix is pure white; in other spots the matrix
is still purple from the unchanged iron of the original deposit. The limonite was con-
centrated upon the bones, covering them with a thin scale, and was deposited in the
cracks between the pieces of the broken bones. Movements in the ground crushed the
skull and broke it into many small pieces which are slightly displaced. The cleaning
and restoration were rendered difficult by the distortion and breaking of the bones, but
as the fragments were only slightly displaced it was possible to harden the matrix or
replace it by plaster cement and work out the original form of the skull. The anterior
half of the long rostrum was weathered out, but it has been possible to fit the pieces into
place and determine very accurately the length and form of the nose. From the narial
opening forward, the two halves of the skull were separated slightly; the palate was
crushed and pushed somewhat toward the left side. The posterior part of the lower jaw
of the left side was recovered, but the anterior end was destroyed by decay before
fossilization.

Because of the fractured and compressed condition of the skull, it is intelligible
only after considerable study; the figures presented are restorations which have been
made with as great accuracy as possible, being the result of six different attempts, made
as checks, and are, in the opinion of the author, very satisfactory representations of the
original condition. The location of the sutures was rendered difficult by the innumerable
cracks and the condition of the specimen; only those that have been determined with
certainty are represented in solid line.

The peculiar characteristics of the specimen are the slight extension of the squamosal
region behind the occipital condyle, the elevated position of the parieto-squamosal arch,
the absence or small size of the post-temporal opening, the very long parasphenoid
process, and the large interpterygoid space. The lack of any great posterior extension
of the squamosal region, the great length of the parasphenoid, and the large interpterygoid
space are primitive characters recalling the condition in Mesorhinus, but the development
of the external process of the pterygoids, the character of the transverse, the shape and
condition of the palatines, and the elongate rostrum are features belonging to the more
highly specialized of the Phytosaurs of the Upper Triassic. The suggestion of immaturity
conveyed by the small size of the skull is not borne out by the condition of the bones
and the sutures. It is evident that we have to do with a fully mature Phytosaur of the
Mystriosuchid group, of small size and distinct in its characters from any previously
described.

The upper surface of the skull (plate 11, fig. a). — The openings in the skull: The
supratemporal openings are elongate oval and lie entirely upon the upper surface of the
skull ; their boundaries are shown in the figures. The orbits are somewhat oval in outline
and lie almost entirely upon the surface; the lateral presentation is very slight. The
narial opening is entirely upon the upper surface and is surrounded by an elevated
rim; it is divided somewhat deeply below the rim by a pair of small bones which are

50 NEW REPTILES AND STEGOCKIII AI.1ANS FROM

dcocribed below. The lateral temporal openings are elongate and inclined downward
and forward, reaching as far forward as the center of the orbit. The antorbital openings
are elongate and rather narrow vertically; they reach from just anterior to the orbit
to a point below the center of the narial opening. Both the lateral temporal openings
and the antorbital openings are but partially visible from above.

The parietal bones are small, each with a slender process extending outward and
backward and forming the upper edges of the deep notch at the posterior end of the
skull. These processes form the inner borders of the upper temporal openings and ter-
minate near its posterior end. The anterior portion of each parietal meets its fellow of
the opposite side in the median line; the extent forward is not accurately made out, but
it is. in all probability, not great. This portion of the parietal forms the anterior edge
of the upper temporal opening. There is no trace of a parietal foramen.

The squamosals are cruciform in outline. The anterior arm unites with the postor-
bital to form the bar between the two temporal openings. The inner arm is very narrow
and unites with the parietal; the outer arm is much wider and is overlapped by the upper
end of the quadratojugal. On the lower surface of this arm there is a concave surface,
elongate laterally for the reception of the upper end of the quadrate; this is clearly shown
on the left side, where the quadrate is missing, having been lost by maceration before
the fossilization of the skull. The posterior arm is short; its upper surface is marked
by prominent rugose ridges extending in the direction of the greatest length of the bone.
The posterior end overhangs the distal end of the opisthotic.

The postorbital forms a small part of the posterior edge of the orbit and articulates
with the jugal below. It sends backward and slightly outward a broad arm which unites
with the squamosal. Just anterior to the suture there is a broadening of the outer side
which forms a constriction in the lateral temporal foramen.

The outlines of the frontals, postfrontals, prefrontals, and nasals can not be made
out clearly. This portion of the skull is rugose and is badly broken and cracked, the
cracks being fdled with limonite, so that the sutures can not be followed. The probable
course of the various sutures is indicated in dotted lines in the figures. It seems alto-
gether probable that the nasals had considerable extent posterior to the narial opening.
Dividing the narial opening some distance below the upper edge of the rim, there is a
slender rod formed by two narrow, elongate bars which meet in a close, relatively broad
symphysis, as shown in figure 21. This septum is, in all probability, formed by the
anterior ends of the nasals.

The seplomaxillaries. — No trace of these elements could be found, but it may well
hi' that the limiting sutures are obscured in the specimen.

The premaxillaries form the greater part of the long, slender rostrum anterior to
the narial opening. The origin of the maxillary-premaxillary suture lies about 3 centi-
meters in advance of the opening; the suture is t rateable backward between the two bones
and then is apparently continued backward between the premaxillaries and the nasals,
allowing the premaxillaries to take a small part in the anterior portion of the rim of the
nates: tin- last point is. however, uncertain. The anterior ends of the premaxillaries
are only very slightly thickened and expanded to accommodate the alveoli of the anterior
teeth. The rugosity of the upper surface terminates with the narial opening and the
surface of the maxillaries and premaxillaries is relatively smooth. The quadratojugal, jugal,
maxiUory, and lachrymal are only obliquely and imperfectly exposed in the upper view.

The side of the skull (plate 11, fig. «)• — The small extent of the projection of the
squamosal beyond the quadrate and occipital condyle is very noticeable in this view.
The posterior process of the squamosal is excavated below, so that the whole projection
18 thin and without a terminal descending portion. I'1 the notch below can be seen the
distal extremity of the opisthotic.

THE UPPER TRIASSIC OP WESTERN TEXAS. 51

The quadratojugal is high behind, forms the posterior edge of the lateral face of
the skull, and conceals the quadrate. Its upper anterior edge forms the posterior half
of the lower edge of the lateral temporal opening. Its lower edge forms the extreme
posterior part of the lower edge of the skull, but is soon covered by the jugal; the suture
between these two bones passes obliquely upward and forward.

The jugal forms the lower edge of the skull as far forward as the anterior end of the
lateral temporal opening; it is then covered by the maxillary and the suture between
them runs forward and upward, terminating, apparently, on the lower edge of the antor-
bital opening, about one-third of the length of the edge from the posterior end. The
jugal is deeply notched posteriorly by the anterior end of the lateral temporal opening
and forms both the upper and lower edges of this part of the opening. The suture
between the jugal and the postorbital lies at the posterior end of the orbit and the jugal
forms much of the outer orbital rim. The suture between the jugal and the lachrymal
can not be made out.

The lachrymal can not be delimited with certainty; it probably occupies the same
position as in most of the Phytosauria and forms the posterior border of the antorbital
opening.

The maxillary starts from a point below the anterior edge of the lateral temporal
opening and forms the lower edge of the skull as far forward as about 3 centimeters in
front of the narial opening. Its upper edge forms the anterior two-thirds of the lower
border of the antorbital opening. A thin groove, which apparently marks the position
of a suture, rises from the middle of the upper border of the antorbital opening and runs
forward close to the edge; it is traceable out upon the face of the skull for some distance;
it apparently joins the suture between the maxillary and the premaxillary. It is believed
that this groove marks the position of the suture between the maxillary and the nasal,
ns the form of the maxillary thus outlined is found in several genera of the Phytosaurs.

The nasals can not be exactly outlined. If the suture described above is correctly
determined, the nasals form most of the sides of the skull above the antorbital opening,
form a part of its upper border, and join the lachrymals and prefrontals posteriorly.
Anteriorly the nasals probably form the borders of the narial opening, except, possibly,
the median portion of the anterior edge, and send slender processes forward between
the maxillaries and the premaxillaries. The septum in the narial opening is in all proba-
bility formed by the nasals.

The premaxillaries. — The outline of the posterior end of the premaxillaries is uncer-
tain, but they apparently send long processes backward and upward which terminate in
the anterior median portion of the edge of the narial opening and join the nasals and
maxillaries. Anteriorly they form the major portion of the long, slender rostrum. The
anterior end is but slightly decurved, and there is a very shallow constriction just
posterior to it.

The lower surface of the skull (plate 11, fig. c). — The bones of the palatal surface
have been largely preserved, but have been badly broken, slightly displaced to the left
side, and crushed against the lower surface of the bones of the roof of the skull.

The batioccipital. — The condyle is nearly hemispherical and has a rather long neck,
with the sides flattened. Low ridges appear near the posterior end of the lower face
and rise rapidly forward until just anterior to the suture between the basioccipital and
the basisphcnoid they form high processes on the sides of a deep notch.

The I V/.s /.sphenoid is separated from the basioccipital by a distinct suture. On
either side of the notch described above there are prominent processes, the tubera basioc-
cipitalia, rounded externally and with flat, oval faces looking posteriorly. These pro-
cesses contract anteriorly, disappearing just posterior to the origin of the basipterygoid
processes. The lower face of the bone is flat and there is no deep pit; the bone is broken

52 NEW REPTILES AND STEGOCEPHALIANS FROM

in this place and is filled with matrix, but an examination of the depression shows that
the edges of the pieces are sharply fractured and that there is no surface leading down
into a pit. The basipterygoid processes are prominent, extend slightly forward as well
as outward, and terminate in slightly convex faces. The parasphenoid process is of
remarkable length, extending forward for 75 millimeters. It gradually contracts ante-
riorly to a very thin lower edge. The sides of the process are flat, about 20 millimeters
high at the posterior end, and contract to not over 5 millimeters at the anterior end.
There is no indication of a division into upper (presphenoid) and lower (parasphenoid)
portions.

The iipisthotics are indistinguishable from the exoccipitals; they extend outward
and backward. The proximal portion of the complex is thick, but the distal portion
contracts to a thin oval, with the major axis placed vertically. To the outer, i. e.,
toward the quadrate, side of the extremity of the right opisthotic there is attached a
small fragment of bone of irregular form. This was at first regarded as a displaced
fragment of no significance, but the extremity of the opisthotic of the opposite side shows
a sutural surface in the same position. The fragment is in exactly the position where
MoCiregor found a small element in Mystriosuchus1 which he considered to be calcified
cartilage and identified as a hyoid element.

The stapes. — On the lower side of the right opisthotic there is preserved a portion
of a very slender stapes apparently in position. The portion preserved is about 1.5
centimeters in length, with a diameter of less than 1 millimeter. The outer extremity
is lost, but the inner end is evidently in place in the fenestra ovalis; the crushed condition
of the specimen in this region renders it impossible to determine the exact relations of
the inner end.

The quadratojugal and the jugal are represented on the lower surface by their thin
lower edges only.

The pterygoids. — These bones are badly broken, but the parts are sufficiently well
preserved on the right side to permit a restoration of a major portion of the bones. The
articulation with the basipterygoid process of the basisphenoid is accomplished by a
well-formed, slightly concave facet on the inner edge of the bone, which is slightly
thickened at this point. The quadrate process extends backward, with a narrow,
concave face presented downward, which rapidly contracts toward its distal end.
A- this face contracts, the body of the process assumes a vertical position. Union with
the quadrate is accomplished by a definite suture with the slender pterygoid process,
which is of considerable vertical extent. The external process of the pterygoid extends
outward almost at a right angle from a point a little in advance of the basisphenoid
proce>-. The external process is not greatly thickened, but the posterior end descends
somewhat abruptly from the palatal surface, and from this point the bone slants gently
forward and upward. The posterior edge is thin and the outer corner is nearly a right
angle; the outer border descends rather sharply beneath the transverse. The anterior
process is broad posteriorly and contracts anteriorly as its inner edge is cut out by the
development of the interpterygoid vacuity. This portion of the bone is very thin and
much has been crushed and lost ; the outline of the interpterygoid vacuity as represented
is uncertain, but enough can be made out to show that it was elongate, broader posteriorly,
and uradually contracting in front until the bones of the two sides met near the anterior
end of the parasphenoid process. The suture with the palatine is obscured, as the latter
hone- arc -omrwliat overthrust on both sides.

The transverse is thin throughout; its posterior end is expanded and fits closely over
the lower face of the outer part of the external process of the pterygoid; the inner edge is

1 HcQngpr, .1 EL, Memoirs Aiiicricim Musi-urn Natural History, vol. i.\, pt. II, 190(1, pi. vii.

THE LPPER TRIASSIC OF WESTERN TEXAS. 53

marked by a slight ridge, and below this is a groove which is continuous with a foramen
which excavates the pterygoid and connects with the small palatine vacuity which is
not visible from below. The bone is narrowed in the median portion and then expands
to articulate with the maxillary, the palatine, and the jugal. The outer face is bent
sharply down and forms a nearly vertical face; its distal end is reflected upon the inner
sides of the maxillary and jugal.

The palatines are elongate bones lying between the maxillaries and the pterygoid.
On the right side the palatine is separated by a fracture from the maxillary, but on the
left side the bones are in contact. The surface of the palatine is convex downwardly
in the middle of its length, but the anterior and posterior ends are natter. The greatest
height of the convexity is near the inner edge of the bone and the whole surface of the
middle portion is slightly rugose.

The anterior termination of the pterygoids and the palatines is uncertain. They
apparently join the vomers near the median line, but as the premaxillaries of the two
sides meet in the median line as far back as a point just below the posterior edge of
the external narial opening, the palatines, pterygoids, and vomers must have risen
somewhat in the skull at this point. Just posterior to the point where the premaxillaries
terminate, the inner edges of the median bones of the palate are slightly excavated and
rounded, indicating the position of the very narrow internal nares, but whether the
boundaries of these openings are formed by the vomers alone or by the vomers and
pterygoids can not be made out. There is no suggestion of a median septum, but the
opening was probably paired. In the light cast by the study of the skull of Leptosuchus,
this interpretation may be erroneous.

The maxillaries appear on the lower side of the skull only as the alveolar edges;
anterior to the external narial opening they join the premaxillaries by oblique sutures
running backward and inward, so that the posterior ends of the premaxillaries extend
backward as long processes between the maxillaries.

The premaxiUariet extend forward with nearly straight outer sides; the lower
surface is divided between the alveolar edges and smooth convex surfaces; the two bones
meet in a broad median symphysis. This symphysis is divided into an upper and a lower
part; between the two articulating surfaces there is a smooth, somewhat concave area,
indicating the presence of a narrow median cavity running nearly the full length of the
rostrum. On either Bide of this groove the palatal surface of the premaxillaries is raised
into convex (longitudinally) ridges, forming, with similar process on the dentaries, a
buttress which prevented injury to the teetli when the jaws were snapped together.

The teeth are typically phytosaurian in shape and arrangement. The posterior
teeth are broader and lower, with sharp anterior and posterior serrate cutting edges.
The inner side of each tooth is nearly flat, except at the tip, where the convexity becomes
greater; the outer side is decidedly convex. The cutting edges are sharply crenulate
from the apex nearly to the base. The greater part of the teeth are lost, but certainly
at the middle of the premaxillaries the teeth were elongate and slender. The inner side
of the few anterior ones preserved is less convex than the outer and the cutting edges
are lower; the crenulations do not appear on the single tooth, which is well preserved,
but they were probably present to some extent. The extremity of the snout was occupied
by four large teeth, which on the right side are represented only by the base of the inner
one and the apex of the just erupting outer one. The apex of the outer tooth shows
the same disparity between the convexity of the inner and outer sides that appears in
the other teeth of the series. There are 47 teeth and alveoli on each side in the upper
jaws. The separation of the two sides of the skull at and anterior to the external narial
opening permits of some observations upon this region. As shown in figure 21, there
is a pair of bones, at the level of the surface of the skull posterior to the narial opening,

;.t

\I.\V HKI'TILKS AND STK(i()('KI'IIAUANS FKOM

Fig. 21. — Promyatrio-
suchun ehlersi.

Median surface of the
narial region, left side.
no., nasal. X 0.6.

which divides the nares some distance below the level of the rim. These bones are thin,
but nearly a centimeter in height on the broad inner surface, which was closely applied
to the bone of the opposite side. These are probably extensions of the nasals. On
the left side, in what is apparently a direct continuum with the median palatal element,
probably the palatine (vomer ?), there is a slender bar of bone distinct from both the
nasal and the premaxillary. As this bar is above the premaxillaries in the region where
they meet in the median line, it can not be a bar separating the internal nares; rather it
seems to be a bar vertically placed beneath the nasals and defining an opening which
permits the nasal to open laterally into passages which lead back to the somewhat
posteriorly placed internal nares. The separate character and
distinct identity of this bar is proven by the fact that it was
surrounded by matrix, which has penetrated between it and the
adjacent bones, and by its rounded edges revealed under the
binocular microscope.

The posterior face of the skull (plate 11, fig. d). — The crushing
of the skull has almost completely closed the foramen magnum, and
in the restoration allowance has been made for this depression as
accurately as possible. The sutures between the elements forming
the edges of the foramen magnum can not be made out.

The basioccipital carries a good-sized, nearly hemispherical
condyle; the sides of the neck run forward and upward to join
the exoccipital-opisthotics. It has been impossible to determine
the position of any of the nerve outlets.

The exoccipilal and the opisthotic are closely fused; the opis-
thotic portion extends outward and backward at a fairly sharp angle; the outer portion
becomes thinned laterally, but retains its vertical extent; the outer end terminates freely
below the distal end of the squamosal.

The supraoccipital is a rather broad plate inclined sharply forward in the median
line, but becoming more nearly vertical toward the sides; it is closely attached to the
slender parietals above, which are visible only as thin edges from the rear. The post-
temporal openings are entirely closed, if they were present, which is quite probable.
An enlargement visible on the upper edge of the opisthotic in the usual position of the
openings may be due to crushing. Small post-temporal openings have been shown in
the restoration.

The squamosal has a relatively small presentation on the posterior face of the skull ;
it -ends a short prong inward between the distal end of the opisthotic and the supraoc-
cipital and a second one above the supraoccipital to unite with the parietal. Just above
the distal end of the opisthotic is the prominent posterior termination of the squamosal,
and below this is the small fragment of bone on the outer side of the opisthotic which
McGregor identified as a hyoid element in Mystriosuchus. Externally the squamosal
sends a strong process forward and outward to join the quadratojugal; the lower surface
i- excavated for the reception of the upper end of the quadrate.

The quadrate has a nearly quadrangular posterior surface. This surface inclines
inward and forward beneath the opisthotic; the upper half of the inner edge forms the
articulation for the pterygoid. The outer edge is covered by the quadratojugal, which
-end- a short process around the outer edge on to the posterior face near the outer lower
corner. About midway up the outer edge there is a good-sized quadrate foramen
surrounded by the quadrate and quadratojugal. There is no indication that the quadrate
process of the pterygoid appeared on the outer face of the skull, as described by Huene'
in Mystriosuchus pleiningeri.

' Bocae, P. %•., Geoio«. u. Paleoatolog. AbhdJg., N. F., Ud.x, lift, l, s. 81, 1911.

THE UPPER TRIASSIC OF WESTERN TEXAS. 55

A review of the known skulls of the Parasuchia reveals the fact that they may be
divided into two more or less clearly marked groups — those with long, slender, depressed
rostra and those with elevated rostra. To the first group, the Mystriosuchid group,
belong Mystriosuchus, Rhylidiodon, Paleorhinus, and Promystriosuchus; to the second
group belongs Phytosaurus (Belodon). Intermediate between the two are the forms
which have been described by Cope as Belodon buceros and by Mehl as Machceroprosopus
validus. In the opinion of some workers, Belodon buceros Cope and Machoeroprosopus
validus Mehl should be placed in the genus Phytosaurus. The two specimens described
in this volume as Leptosuchus help to further bridge over the gap between the two groups.

The form described as Promystriosuchus belongs in the Mystriosuchid group and
is nearest to Angistorhinus. Because of its retention of certain primitive characters,
it will be best to compare it first with Mesorhinus, the earliest and most primitive of the
Parasuchia. Mesorhinus was described by Jaekel1 from the middle Bunter sandstone
of Bernburg.

Comparison with Mesorhinus.

The anterior portion of the rostrum of the specimen was not recovered, and
Huene considers that it has been made too long in the restoration by Jaekel. Granting
the correctness of Jaekel's restoration, which would be the maximum possible length
of the rostrum, there is still a great disparity between the relative lengths of the prenarial
and postnarial portions of the skull and the same measurements in Mystriosuchids
from the Upper Triassic Measuring from the anterior end of the narial opening, Jaekel's
figure (as copied by Huene) gives a relative proportion of lengths between the prenarial
and postnarial portions of the skull as 5 : 6; the same measurements on the figure given
by Abel (St annuo der Wirbelthiere, p. 516) give proportions of 4 : 5. In Promystriosuchus
the rostrum has become relatively much elongated, a distinctive character of the special-
ized Parasuchians of the Upper Triassic, the proportional lengths being as 3 : 2. In
Mesorhinus the narial opening is elongate and clearly divided by a septum at the level
of the surface of the skull, formed by an anterior extension of the nasals; the antorbital
vacuity reaches either to the middle of the narial opening or only a little way anterior
td the posterior edge, according to figures of the lateral and the upper view of the skull
as given by Jaekel (copied by Huene); the lateral temporal opening is nearly circular;
the parieto-squaniosal bar defining the inner border of the supratemporal foramen is
complete and at the level of the roof of the skull; the supratemporal foramen itself is
entirely on the top of the skull; the squamosals do not extend far posterior to the occip-
ital condyle. In Promystriosuchus the narial opening is round, is surrounded by a high
rim, and the septum is far down within the cavity of the nares; the antorbital opening is
elongate and reaches as far forward as the middle of the narial opening; the lateral tem-
poral opening has assumed, more or less perfectly, the characteristic parallelogrammic
form of the Parasuchia; the parieto-squaniosal bar is at the level of the roof of the skull,
and the supratemporal opening is entirely upon the upper surface; the squamosals do
not extend far posterior to the occipital condyle.

On the posterior surface of the skull the shortness of the opisthotic process in
Mesorhinus is a striking difference from the condition found in the younger Parasuchia.

On the lower surface the palate of Mesorhinus, as figured by Jaekel, is radically
different from that of the majority of the Parasuchia. The pterygoids lack the strong
cxt final process; the transverse articulates with the outer edge of the pterygoid instead
of underlying an external process; the palatine vacuity is large, lying between the pala-
tines, pterygoids, maxillaries, and transverse. The peculiar articulation between the

'.Jaekel, ().. Sitz. Her. d. Cesell. N'.iturf. I-'reuwle. Berlin, No. 5, p. 197, 1910; see also Huene, F. v., Geol. u.
Paleont Abhdlg., X. K, Bd. x, lift. I, s. 50, 1911.

56

\K\V REPTILES AM) 8TEGOCEPHALIANS FROM

pterygoid and the palatines i> not found in any other form of the Paraauehia. In
Promyttrioiuchtu there is a strong external process of the pterygoids which is underlain
by the posterior end of the transverse; the palatine vacuity is represented by a small
foramen between the palatine and the transverse, which appears on the lower surface
near the posterior end of the transverse.

The primitive characters of Promyatriosuekw appear in the elevated parieto-squamo-
sal arch, the relatively anterior position of the narial opening,1 and the elongation of the
parasphenoid process. The other characters cited are those of the more specialized
Parasuckia, and it is evident that the form here described must be placed in the Mystrio-
suchid group of the Phytosauridae. The primitive characters are conservative features
which are found in at least one other member of the same group and family.

Comparison with Ammstorhinus.

Angistorhimis- is in many ways the nearest known form to Promystriosuchus.
A consideration of the comparative characters will make more evident the similarities

V io. 22. — Anffistorhinus, lateral view of skull, after Mehl.

A. Lateral view of skull.

B. Upper view of skull.
('. Lower view of skull.

D. Posterior view of skull, somewhat larger.
ix., septomaxillary. other lettering as usual.

and differences between J'romyslrioHUchus and the other Mystriosuchids. The relative
proportions in length of the prenarial and postnarial portions of the skull of .1 ngistorhinue
are as 10 : ti.4, or approximately 5 : 3; the antorbital vacuity is elongate and extends as
far forward as the anterior edge of the narial opening; the opening is surrounded by a

1 Queue, in the QeoL u. Paleont. Abhdlg., N. I'., Bd. x, lift, l, a .'>:i, has suggested that the position of the
nanal opening ami tin- elongation of the rostrum .-in' two distinct things .- < i > • I that tin1 retreat of the narei is oonnecti 'l
with tie- shortening of the posterior pari of the skull. The relation between the antorbital vacuity and the nans is
regarded by him as a phylogenetie character; as the nares retreated the antorbital opening would appear more and
nan anteriorly in the skull.

: Mehl, II <;.. .Journal of Geology, vol. 21, No. '-', p. 186, 1913; /'•<.'.. vol. 23, No. 2, p. 129, mi:,.

THE UPPER TRIASSIC OF WESTERN TEXAS. 57

high rim, but the septum is placed deeply within the cavity; the parieto-squamosal arch
is complete and on a level with the roof of the skull; the interpterygoid vacuity is
relatively large, and the parasphenoid process is long, reaching to the anterior end of the
vacuity. The two forms differ in that in Angistorhinus the squamosals are much larger
and extend well beyond the occipital condyle, so that it is not visible from above; there
is a much sharper depression of the rostrum anterior to the external nares; the anterior
end of the rostrum is much more down-turned and is wider; the pterygoid is represented
by Mehl as having no external process and the transverse joins it laterally, not by
underlying an external process; the posterior arm of the parietal is short, the greater part
of the upper edge of the posterior part of the skull being formed by the squamosal; the
opisthotic process extends out to the squamosal, but is surrounded by it both above and
externally; the post-temporal foramen is relatively large.

Comparison with Paleorhinus.1

If Huene's suggestion of the phylogentic value of the relative position of the externa
nares and the antorbital vacuity is correct, Paleorhinus retains the primitive character
in a marked degree, for the antorbital vacuity is entirely posterior to the nares and has a
rounded anterior outline, indicating a very different form of the jaw-muscle; the squa-
mosal has a very large extension behind the occipital condyle. The proportion between
the relative lengths of the prenarial and postnarial portions of the skull is as 8 : 8.8,
or nearly as 1 : 1, approaching in this the probable condition in Mesorhinus. On the
other hand, the specialized character of the skull appears in the depression of the
parieto-squamosal arch, the shortness of the posterior bar of the parietal, the elevated
rim of the external nares, the position of the internal nares posterior to the external,
the small interpterygoid vacuity and the short parasphenoid, the strong external process
of the pterygoid with the transverse articulating with its lower surface, and the position
of the palatine vacuity anterior to the middle of the transverse. The characters de-
scribed by Lees2 of the elongate vomers reaching back to take part in the anterior edge
of the interpterygoid vacuity, the otic foramen between the squamosal and the quad-
ratojugal, and the position of the quadrate foramen on the lateral surface of the skull
have all been questioned by Mehl and Huene3 and are so unlike the normal characters
of the other Paratuckia that they can not be accepted or discussed until the skull is
reexamined. The improbability of the existence of an otic foramen in the position de-
scribed by Lees is further emphasized by the discovery of the stapes in the normal position
of the lower side of the opisthotic in Promystriosuchus.

Comparison with Mystriosuchus.

This comparison is based on the published descriptions of Mystriosuchus planirostris
and M. pleiningeri by McGregor and Huene.4

The lengths of the prenarial and postnarial portions of the skull of Mystriosuchus
are very closely as 5 : 2; the antorbital openings are elongate and extend forward either
to the anterior end of the nares (planirostris) or slightly anterior to them (pleiningeri);
the nasal septum reaches the level of the skull; the parietals have no posterior bar
defining the inner side of the supratemporal opening; the parietals descend to the top

1 Williston. s. \v., Journal of Geology, vol. xa, p. 696, 1904; Lees, J. 11., Journal of Geology, vol. xv, No. 2,

p. 121, 1907.
- Lees, J. II., The Skull of Paleorhinus, Journal of Geology, vol. xv, NO. 2, p. 133, 1907.
M' M. M G., Journal of Geology, vol. xxm, No. 2, p. 157, 191.5; Huene, F. v., Neuee Jahrb. f. Min. Geol.

u. Pal., No. 19, p. 587, 1909.
' McGregor, J. 11., Memoin American Museum Natural History, vol. ix, pt. n, 1906; Huene, F. v., Geolog-

ische u. Paleontolojrische Abhaodhingen, N. F, Bd. x, Hft. 1, p. 68, 1911.

58

M.W REPTILES AND STEGOCEPHALIANS FROM

of the opisthotie and the openings look as much backward as upward; the posterior face
of the skull is very similar, except for the depressed paricto-squamosal bar; the post-
temporal opening is small, and the opisthotics extend out to the squamosal and are

li... •.>:{.
Paleorhinws, after Lee*. Lettering as usual.

A. Lateral view of skull. H. Upper view of skull.

imuchus, after McGregor,
I). I '|>per view of skull. E. i/ower view of skull.

C. Lower view of skull.

F. Lateral view of skull.

overlapped by it. In the specimen of M. planirostris, described by McGregor, the small
element on the outer .side of the distal end of the opisthotie is apparently calcified car-
tilage; in Promyttrumtckut the fragment in this position is true hone; the opisthotie
proces- standi a little higher above the condyle in Promydriomchus. On the lower

THE UPPER TRIASSIC OF WESTERN TEXAS.

59

face of the skull the relation of the transverse to the pterygoid is very different in the
two forms. In M. planirostris and pleiningeri the transverse lies external to the process
of the pterygoid and articulates with it laterally instead of underlying it; the outer
corner of the dependent process is formed entirely by the transverse; the palatine foramen
is small, but lies anterior to the middle of the transverse; the quadrate process of the
pterygoid is shorter and articulates with an elongate process of the quadrate; the suture,
as figured by McGregor, lies, at least on the lower face, near to the basisphenoid process of
the pterygoid; there is but a small interpterygoid space and the parasphenoid process is
short or in large part covered by the pterygoids; the internal nares are directly beneath
the external nares; the inner edges of the palatines are elevated, but do not show the
rugosity which appears on these bones in Promystriosuchus. It is apparently this ele-
vated edge of the palatines which Huene1 refers to as the median edge of the pterygoids.

pf JL

I •'!>;. 24. — Phytosaunis (Machcerojirosopus) , after Mehl.

A. Lateral view of skull.

B. Upper view of skull.

' '. Posterior new of skull.

1). 1'oslcrior view of skull of Miistriosuchux, after McGregor.
E. Posterior view of skull of I'hi/tosnurwi kappfi.
Lettering M usual.

Comparison with Rhytidiodon.8

Most of the material representing this genus was recovered from the Triassic
coal-beds of Egypt, Chatham County, North Carolina. Most of the skulls, as reported
by McGregor, tire in a very fragmentary condition; a single skull, lacking the anterior
end. permits of some comparison. The antorbital opening extends beyond the anterior
edge of the external nares; the orbits have a considerable lateral presentation; the
parieto-squamosa] bar is depressed; the squamosal region extends well posterior to the
occipital condyle; the palatine vacuities arc at the extreme anterior end of the transverse.

' Huene, 1". v., loo. tit., p. 17. * McGregor, J. H., loc. cit, p. 58.

60 M :\\ KI1TILKS AND STEGOCEPHALIANS FROM

The figures show this specimen as having no interpterygoid vacuity and no parasphenoid;
this is, in all probability, the result of the condition of the specimen. There is an elevated
rim to the external nares, and the septum is at the level of the top of the skull. By
utilizing a specimen in the New York State Museum, which he assumes to have been of
the same size as the one in the American Museum, McGregor determined the prenarial
and postnarial proportion to be as 3 : 2.

( 'OMI'AKISON WITH PhYTOSAURUS (MaCHjEROPROSOPUs).

The specimen described as Machwroprosopus by Mehl1 and regarded by him as
congeneric with lielodon buceros ( lope is in many ways intermediate between the Mystrio-
suchid and the Phytosaurid groups of the Parasuchia. The elevation of the posterior
half of the rostrum, the posterior position of the nares, and the extreme depression of
the parieto-squamosal arch are all characters of the Phytosaurid group, while the long,
slender anterior half of the rostrum is distinctly Mystriosuchid-like. The prenarial
portion of the rostrum is relatively short, its proportional length to the postnarial portion
being as 7.8 : 8.3, or a little less than 1:1; the antorbital opening is elongate and reaches
almost to the anterior end of the extern] nares; the parietals have no posterior bar, and
the squamosal region is much extended, reaching far behind the condyle; the supratem-
poral openings look almost entirely to the rear; Mehl2 has drawn in outline a rather
broad bar separating the supratemporal from the post-temporal openings on the posterior
surface of the skull. This is probably correct, as its width is indicated by the broken
edges, but, if so, it leaves a much larger post-temporal opening than occurs in any other
of the Phytosaurid group.

1 Mehl, M. G., Quarterly Bulletin University of Oklahoma, n. s., No. 103, 1916.
' Mehl, M. G., loc. cit., fig. 3.

THE UPPER TRIASSIC OF WESTERN TEXAS. 61

NEW PARASUCHIANS, LEPTOSUCHUS CROSBIENSIS AND LEPTO-
SUCHUS IMPERFECTA, FROM CROSBY COUNTY, TEXAS.

The first skull (Leptosuchus crosbiensis, No. 7522, University of Michigan) is
singularly perfect. All of the bones except those of the posterior portion of the upper
surface are preserved in place and the loose ones were found close to the specimen.
One half of the lower jaw was found overlying the skull, and the posterior portion of the
other half was found not far distant. Total length of the skull, 87.5 cm.

The second skull (No. 7523, University of Michigan) is much larger than the
first and was found entirely washed out and lying on the surface as a mass of fragments.
Some parte had disappeared and the skull as mounted is in part restored. Total length
as restored 112 cm.

Skull No. 7522 is so singularly perfect that it warrants a rather full description.
It was completely covered when found, the removal of the half of the lower jaw over-
lying it leading to its discovery. The matrix is a fine yellowish clay broken into
innumerable small pieces, between which are sheets of calcite not over 0.1 mm. thick.
The skull is slightly compressed, the right side is a little displaced upward, and the bones
of the posterior portion are a little distorted. The extremely complex fracturing of the
matrix compelled a fracturing of the bone which is only slightly less in degree, but the
fragments are all in position. The skull was collected in a block of matrix and has been
prepared and mounted by the painstaking and skilful work of Mr. William H. Buettner.

Particularly important is the preservation, in the natural position, of the extremely
thin bones of the palatal region. This permits a determination of the complete osteology
of the skull.

The upper surface of the skull. — The best idea of the position and relation of the
bones may be gained from figure 25 a. The narrow upper surface of the squamosals is in
striking contrast with most other forms of the Phytosaurs.

The lateral $urfaee »/ the skull. — Seen from the side, the premaxiliartee are similar
in form and relations to the same hones in the majority of the forms of the Mystriosuchid
group. The anterior end is sharply down-curved, with a decided notch behind the
prominent anterior tusks for the reception of the tusks of the lower jaw. Near the
middle of the bone there is a low but very evident prominence on the upper surface.
This seems to be a common and normal structure in most, if not all, of the Mystriosuchids.
It occurs in all the specimens preserved in the University of Michigan collection and is
figured and described by v. Huene in M. pleiningeri. Huene attributed the prominence
to an accident, but it is clear that it may be a normal feature. The peculiar condition
found in thi< region in specimen No. 7523, University of Michigan, is described below,
but is not yet understood.

The articulation with the maxillary starts on the alveolar edge, just posterior to
the twenty-third tooth. The surface of the premaxillary is devoid of rough sculpture,
unless it be at the extreme posterior end, where it approaches the nasal prominence just
anterior to the external nan-.

The maxillary has the usual position and relationships. It forms the lower border
and the anterior border of the antorbital vacuity and the anterior part of the upper
border of the same opening. The portion below the vacuity is smooth and the alveolar
edge is convex downward. There are 21 tooth-sockets in the maxillary, which are
rounded in the anterior portion, but by the eighth or tenth from the anterior end become
decidedly oval in outline. The upper portion of the maxillary is more rugose. The
position of the suture between it and the nasals is indicated by the sudden increase in
the coarsenesi of the rugosities, as well as the change in the course of the bone-fibers.

62

NEW REPTILES AND STEGOCKI'llALIANS FKO.M

The exact outline of the nazals is not determinable. The anterior end appears to
be opposite a point halfway between the anterior end of the maxillarics and the anterior
end of the antorbital vacuity. The bone is marked by a deep radiating sculpture which
apparently starts from the region around the external nares, but is more pronounced
on the sides of the bone. The eminence of the nose is 2 or 3 centimeters anterior to the

Fig. 25.— Leptosuchus crosbiensis, No. 7522, U. of Mich. X 0.16.
A. Lateral view of skull. B. Upper view of skull. C. Lower view of skull.

Lettering as usual.

external nares, and from this point the front slopes gently down to join the premaxillaries.
The lower edge of the nasals is excluded from the upper edge of the antorbital vacuity
by the juncture of the maxillary and the lachrymal.

It is impossible to determine the exact outline of the septomaxillaries. It is evident
from the direction of the bone-fibers that a distinct element is present in front of the
nares, perhaps taking part in their anterior edge, but the sutures have not been made
out with certainty.

The jugate articulate with the maxillaries by a broad sutiiral surface which underlies
the posterior end of the maxillary. A broad vertical process rises to articulate with the
po-t orbital and a relatively slender bar extends backward to articulate with the quad-
ratojugal and the quadrate. The articulation with the quadratojugal is as described
previously by the author.1 The quadratojugal is deeply cleft on the lower edge and

1 Cane, E. C, Journal of Geology, vol. xxvni, p. 530-531, 1020.

THE UPPER TRIASSIC OF WESTERN TEXAS. 63

the jugal is deeply dovetailed into it. This method of articulation is confirmed by its
presence in two specimens in the University of Michigan collection. The jugal is visible
along the lower edge of the skull to the posterior end of the quadrate.

The quadratojugal is a nearly triangular plate which covers the outer side of the
quadrate almost entirely and takes slight part in the formation of the posterior edge of
the lateral temporal fenestra.

The quadrate appears on the side of the skull only at the posterior lower corner and
as a thin edge at the posterior line of the skull.

The squamosal has a very large lateral presentation and a correspondingly small
presentation on the upper surface, apparently less than in any previously described form.
The lateral surface is nearly vertical; its posterior portion extends 52 mm. beyond the
end of the opisthotic, forming a very decided projection at the posterior end of the skull.
From the lower edge of this projecting portion there is a strong descending process,
the anterior edge of which forms the posterior border of the otic notch. Within the
notch the upper part of the quadrate and a portion of the distal end of the opisthotic
can be seen from the side.

The posterior face of the skull. — This face of the skull is more distorted than any
other, due to the upward movement of the right side in the compression accompanying
fossilization. The occipital condyle is forced somewhat to the left and the bones above
the condyle are somewhat broken and displaced. It is, however, easy to restore the
original condition.

The parietals descend very abruptly and lie upon the upper edge of the opisthotic.
The post-temporal opening was either entirely occluded or was represented by a small
opening not greater than a large foramen. The outer edge of the parietal articulates
with the anterior edge of the process described on the inner side of the squamosal.

The supraoccipital is a small triangular bone which in the specimen is separate
from the exoccipital-opisthotic on the left side, so that the outline is plain.

The exoccipitals are separated from the basioccipital by distinct sutures, but are
still in position. The lower portions meet in the median line, excluding the basioccipital
from any part in the floor of the foramen mangum, at least its posterior part. There
is no suture visible between the exoccipitals and the opisthotics.

The opisthotics are expanded at the inner end, where they contain a portion of
the cavity of the inner ear. Beyond the exoccipital portion they contract rapidly, and
the outer half is spatulate in form, with the outer end reaching to the extreme outer
edge of the squamosal, but, in the specimen, not visible in a lateral view of the skull.
The lower edge of the inner third is marked by a distinct groove, which evidently
protected the stapes in part. This groove becomes deeper toward the inner end and
is continued on the lower edge of the prootic.

The squamosals. — The inner side of the squamosal is best seen from the posterior
surface. This face is nearly vertical, but is divided by a horizontal ridge which originates
near the posterior end and, growing more prominent as it advances, terminates near the
middle of the length of the squamosal. To the anterior end of this ridge is articulated
the distal end of the parietal. Immediately below the ridge lies the distal end of the
opisthotic, which is here a narrow, oval plate with its greatest diameter nearly vertical.
The extremity of the opisthotic reaches nearly to the posterior end of the descending
process from the squamosal, but is not visible from the outer side. The upper outer
corner of the quadrate articulates with the squamosal by a rounded head just anterior
to the end of the ridge described on the inner surface of the squamosal.

The quadrates. — These are much as described in other specimens. Each quadrate
stands upright in the skull, with a broad posterior face, which contracts toward its

i I \K\V KKl'TII.KS AND STEGOCEPHALIAN8 llfo.M

upper extremity. The articular face for the lower jaw is convex antero-posteriorly and
divided into two parts by a sharp median elevation which runs obliquely from within
outward and backward. The suture with the quadratojugal is vertical and nearly
straight, running from the lower extremity of the bone nearly to the top. There is
a large quadrate foramen at about the middle point of the suture, which runs slightly
downward and forward. On the inner face there is a slight concavity just above the
articular surface and then a strong spout-like process, concave downward, which extends
forward for a short distance, but not nearly so far as is indicated by v. Huene in M.
phhiingeri. The inner edge of this process rises gradually, shortening as it rises, and
disappears on the anterior face of the quadrate at about the middle of its height. At
t he origin of this process and above and outside it there is a fairly deep pit. The posterior
edge of the quadrate process of the pterygoid fits over this process on the lower and
inner side. The upper portion of the outer edge of the quadrate is inclined inward
slightly, and the upper extremity terminates in a smooth, nearly hemispherical face
which fits into a pit on the lower surface of the squamosal. There can be no doubt
that there was a certain amount of motion possible at this point, which would be per-
mitted by the sliding of the quadrate process of the pterygoid on the spout-like process
of the quadrate. This reveals a greater flexibility of the posterior portion of the skull
of the Parasuchia than has previously been suspected. Versluys1 regarded the skull of
the Parasucltin as akinetic, but there is a possibility of movement between the parietal
and supraoecipital and a possibility of movement of the pterygoids on the basipterygoid
processes of the basisphenoid ; these possibilities, taken with the evidence of the movable
quadrate and the smooth face between the pterygoid and the transverse, indicate that
the skull of the Parasuchia was certainly kinetic.

The lower surface of the skull. — In this portion of the discussion is included the
description of the walls of the brain-case and the narial region.

The basioccipital-basisphenoid. — On the lower surface the basioccipital differs
decidedly from specimens Nos. 7261 and 7505, described on a later page. In these
specimens the lower surface of the neck of the condyle is smooth and rounded, without
any ridges running forward to the tubera basioccipitalia. In No. 7522 the lower surface
of the neck is flat, with two prominent rounded ridges on the sides of the neck running
from the condyle to the tubera. This forms a decided concave area with a nearly flat
bottom, but a little more depressed on the sides. A similar condition is indicated in
the imperfect specimen, No. 7474.

The tubera are prominent and formed by the basioccipital and basisphenoid.
The basipterygoid processes are well formed and project outward, downward, and
forward; at the posterior end of each there is a sharp, low prominence; such prominences
are described in No. 7257, and their position is indicated in the other specimens by more
or less evident areas of slight rugosity. The space between the basipterygoid processes
is occupied by a low, sharp ridge which is continuous with the very sharp lower edge
of the parasphenoid rostrum. This condition is quite different from the other spec-
imens. In No. 7257 the space is occupied by three ridges; the middle one, somewhat
asymmetrically placed on the right side, was probably continuous with the lower edge
of the parasphenoid rostrum; in No. 7261 there are faint ridges on each side and no
median ridge; in No. 7505 there IS the faint beginning of a median ridge: in No. 7474
the region is not preserved.

The parasphenoid rostrum start- to rise almost directly upward in the skull; the
proximal portion of the lower edge is nearly at right angles to the line of the basicranial
•i x i ~. hut quickly assumes a horizontal position. The lower edge is very sharp. There

1 Veruluys, J., ZoologischtT Jahrbuch, tupplemenl xv, 2 Band, i. 847, 1012.

CASE

^

PLATE 12

4 "'J&

A. Lower surface of sacrum of a phytosaur, No. 7266, U. of Mich. X0.3.

H. Upper surface of the sacrum shown in A.

< '. Lateral view of the hasicranium of a phytosaur, No. 7261, U. of Mich. X0.72.

D. Posterior view of the hasicranium of a phytosaur, No. 7474. X0.75.

K. l"p|>er surface of a hasicranium of a phytosaur, No. 7505, U. of Mich. X0.67.

F. Outer surface of the ilium of a phytosaur, No. 7333, U. of Mich. X0.38.

('•. Lower surface of snout of a phytosaur, No. 7505, U. of Mich. X0.36.

THE UPPER TRIASSIC OF WESTERN TEXAS. 65

is no indication of a composite structure of the rostrum, i. e., a presphenoid and a
parasphenoid portion, as described by Huene in M. plieningeri.

The side of the basioccipital-basisphenoid.— The structure of the sides of these
bones coincides very closely with those described in other specimens. Between the
ridge connecting the side of the condyle with the tubera and the suture between the
basioccipital and basisphenoid there is a deep groove passing upward and terminating,
apparently, in the otic opening. Just posterior to the upper end of this groove and in
the exoccipital bones, which are in place, there are two foramina — a posterior, larger for
the exit of the XII nerve, and an anterior, smaller, which probably transmitted the
IX, X, and XI. Between the tubera and the basipterygoid process, on each side, are
the openings of the foramina for the internal carotid arteries, which open into the base
of the small hypophysial cavity.

The upper surface of the basioccipital-basisphenoid. — The exoccipitals come together
below, forming the floor of the foramen magnum; anterior to these can be seen the very
distinct suture between the basisoccipital and basisphenoid. On the right side this
suture terminates in a deep pit entirely upon the upper surface of the bones. The
left side is partly obscured by the prootic, but the pit is apparently present, showing
that its occurrence on the right side is normal. The presence of the pit can not be
confirmed from other specimens, but its presence is suggested in No. 7505. The sella
turcia is prominent just posterior to the hypophysial pit. On either side of the
pit the edges of the basisphenoid are thickened, forming a strong articulation with the
prootic.

The proolics. — The prootic of the left side is in place; that of the right side was
found detached, but it fits well into its position. The posterior portion extends back-
ward in a point for some distance, and articulates with the opisthotic both above and
below; an epiotic is not distinguishable as a separate element. The upper edge is
thickened for attachment to the parietal, and the lower and anterior edges are thickened
for attachment to the basisphenoid. Near the anterior end is the foramen for the
V nerve; this is a vertical oval with the greatest diameter nearly twice as long as the
horizontal diameter. The portion of the bone anterior to the large foramen is bent
inward nearly at right angles, to form the anterior wall of the brain-case. The two
bonea do not meet in the median line. On the anterior face of this portion of the bone,
at about the level of the lower edge of the large foramen (in the undistorted bone of
the right side), there is a small foramen which probably transmitted a branch of the V.
On the inner side of the prootic, posterior to the large foramen, is a large and deep pit
which sheltered a portion of the inner ear. When the bone is in position this pit comes
into close association with two pits on the anterior inner end of the opisthotic and the
large pit on the suture between the basioccipital and basisphenoid. On the lower edge
there is a groove continuous with the groove on the lower edge of the opisthotic.

The pterygoid* show the usual tripartite form, but are rather different in shape
from that common to most of the reptiles. The middle portion of the bone is thickened
and has a decided notch which fits, loosely, the basipterygoid process of the basisphenoid.
The quadrate process fa a thin plate standing nearly vertical in the skull, which extends
backward and slightly outward and articulates with the pterygoid process on the
inner side of the quadrate. The outer process is fairly heavy. Its anterior edge is
deeply concave, giving the whole process a distinctly crescentic form. The process
extends outward, downward, and backward. The distal end is broad and very thin
and lies upon the upper surface of the distal portion of the transverse, meeting it by
a very smooth, flat surface which undoubtedly permitted free movement. The
anterior process extends forward and quickly becomes a thin vertical plate. In the

t'»t> \i:\\ REPTILES A\l) STEGOCBPHALIANS PROM

specimen this plate lies directly in contact with the parasphenoid process and rises
considerably above it, but not so high as figured by v. Huene.1

The upper edge of the pterygoid rises slowly, forward, from the median portion
and panes above the upper edge of the parasphenoid process at about its middle point.
The anterior portions of the pterygoids are joined at their lower edges, forming a V, which
lies between the anterior portions of the palatines.

The ( pipU rygoid.- There is no evidence of an epipterygoid, either of a separate bone
or of an articular space upon the pterygoid. If such a bone were present, as figured by
Huene.- it is remarkable that no evidence remains in a specimen otherwise so complete.

The transverse. — This bone is radically different, both in position and form, from
the conception which has been derived from the study of crushed specimens. It
stands nearly vertical in the skull, slanting slightly backward as it descends. The upper
end is attached to the maxillary and the jugal. The upper portion of the shaft is a
rather thin oval in section. Below, the bone becomes broader and thicker, with a
prominent, heavy Mipporting ridge on the lower surface. The upper surface of the
distal end is smooth and flat, meeting a similar face on the pterygoid. The anterior
outer edge of the smooth face of the transverse is terminated by an elevated shoulder
which limited the direction of motion between the bones. The palatine was attached
to the posterior end of the inner side of the transverse by a close suture, forming a rigid
joint: this fixed the transverse in its position and allowed motion in the pterygoid only.
1 he transverse forms the greater portion of the descending process of the palatine
legion: the pterygoid is visible from the side only as the thin edge of the distal portion
of the descending process.

On the inner side of the jugal, or the jugal and the postorbital, there is a thin,
prominent extension of the bone which forms the anterior border of the lateral temporal
opening: this gives the bone an L-shaped section and contributes largely to its strength.
The lower portion of this extension becomes quite heavy just posterior to the last of the
maxillary teeth and is penetrated by a large foramen running antero-posteriorly for a
short distance. The transverse is attached just below this heavy extension. Just
anterior to its attachment to the distal end of the palatine the transverse is free for a
short distance, but the bones are united again, leaving a small palatine fenestra opposite
the middle of the shaft of the transverse.

The palatines are rather heavy posteriorly, where they unite with the transverse,
but soon become more thin and plate-like. The median portion of the palatine is thicker,
and this portion forms the characteristic ridge on the lower surface of the palate and the
outer edges of the choanae. On the inner side of the heavier portion is the shoulder
described and figured by v. Huene as supporting the lower edge of the pterygoid. The
palatine sends out two very thin extensions: an upper, which overlies the pterygoids
for a short distance, not nearly so much as is indicated by v. Huene, and an outer, which
unites with the palatine plate of the maxillary. The upper plate forms the posterior
and outer borders of the choanae. Near the middle of the plate there is a strong vertical
ridge which ends upon a prominence upon the upper edge. Just anterior to this ridge
the palatine divides into two vertical plates. The upper edge of the inner plate descends
rapidly and soon joins its fellow of the opposite side, leaving a shallow V-shaped channel
on the upper surface; the two together form the septum which divides the choanal.
'1 he outer plate runs forward and joins the premaxillaries. The lower, or horizontal,
plate of the palatine is deeply concave just below the heavier median portion, forming a
channel which runs as far forward as the anterior end of the choanae. This channel

1 lliniK-. I', v . Beitrftge zur Kenntau und BeurtheiluiiK der PuMuchiet, Geol. u. Paleont. Abhandiungen,

N. !•'. B«L x, fit' I I'll I.
• lliicni', !■'. v., lor cii.. Bfh 5 and 6.

THE UPPER TRIASSIC OF WESTERN TEXAS.

67

may he somewhat exaggerated by pressure in the specimen, but even in the normal
skull it was a prominent feature.

The miners do not appear upon the lower surface of the skull. This is very different
from the condition of the palate as figured by the various authors in the described forms.
If seems probable that the interpretation of the vomer as appearing on the lower surface
of the skull is due to the condition of the specimens. In the deep V formed by the approxi-
mation of the pterygoids in their anterior portions there is evidence of a pair of very
slender bones which have the same relations, i. e., they meet below to form a V; these
are apparently the much-reduced vomers. The position and reduced condition of the
veil ers. the articulation of the palatine with the anterior bones of the roof of the mouth,
in this case the promaxillaries, and the elevation of the pterygoids are very suggestive
of the condition in the Crocodilin.

Fro 26.
bde with outer baaet removed to ihow the structure of internal nares.

Lettering as usual.

X0.8.

The choaruB. — The openings of the internal nares are separated by a bar composed
of the palatines, pterygoids, and probably the vomers. There was a distinct space
between the internasal septum of the external nares and the septum between the choanse
below. Huene's description of this region is fairly accurate, but he has not recognized
the exact relations of the pterygoids, palatine, and vomers, and has not recognized
that the palatines extend forward to meet the premaxillaries. The bar between the
choana is formed by three bones, all paired. The lower pair is the palatine portion.
These bones gradually approximate; the posterior portion of the bar is V-shaped, the
anterior portion a complete oval bar with a complete upper edge. Between the arms of
the V formed by the palatines there is a second V formed by the anterior portion of the
pterygoids, which retain this form to the anterior end The lower edge of the pterygoid
V rests upon the upper edge of the palatine bar at the extreme anterior end. Between
the arms of the V formed by the pterygoids there is evidence of a third V, formed by
the vomers.

The maxiUaries. The alveolar edge of the mamillaries is supported on the inner
side by a strong, horizontal palatine plate. There are 21 sockets in the maxillary, the
anterior of which are rounded in section, but the posterior 10, or more, are oval. The
strong median prominences on the premaxillaries gradually die out upon the palatine
processes of the maxillary.

The premaxillariea. The anterior end of the premaxillary is but slightly expanded;
it accommodates two enlarged sockets for the tusk-like anterior teeth. Posterior to

tiN NEW REPTILES AND STEGOCEPHALIANS FROM

these sockets the premaxillary narrows, and there are three small sockets, rather on the
side of the bone than on its lower surface. There are 23 sockets in the premaxillary ; in
the anterior portion these are rather more on the side of the bone, but by the middle of
the series they come to lie entirely on the lower surface. The middle of the snout is
occupied by the two prominent rounded ridges which are considered to have acted as
buffers to proven! injury to the jaws when they were snapped violently together.

The second skull {Leptosxichus imperfecta, No. 7523, University of Michigan)
is much larger than the first, measuring approximately 112 cm. over all. There is a
break in the continuity of the snout, and it is impossible to determine the exact length,
though only a very small part, if any, can be missing. The whole skull seems propor-
tionately stouter than that of the smaller skull. As certain portions must be restored,
it is desirable to designate this as a new species only tentatively, until more detailed
study can be made. One very remarkable feature appears in this skull: There is an
elevation on the upper surface of the snout, somewhat farther back than in the others
in the collection. This elevation is decidedly rugose and its upper surface is excavated
by a nearly hemispherical pit with a smooth inner surface. The pit is 5 cm. long and
4 cm. wide, extending into the symphysis of the premaxillary bones. The condition is
so remarkable and unexpected that the pit was at first taken for the opening of the
external nares, but this is easily demonstrated to be impossible, v. Huene1 described
a somewhat similar condition in a specimen of M. pleiningeri. There is, in his speci-
men, a rough area near the anterior end, which he attributed to a fracture or injury.
In the center of this rough area he found a nodule of matrix. From the evidence cited
in this paper it is apparent that an elevation on the premaxillary portion of the snout of
members of the Mystriosuchid group of Phytosaurs is a normal thing, but in none of
the other specimens is there any evidence of a pit. The region of the snout surround-
ing the pit is rugose, but without any evidence of fracture or injury. The meaning
of this apparent abnormality is at present entirely conjectural.

Contrasting characters in the skulls of the known Phytosaurs,2
Paleorhinus:

1. Post-temporal arcade depressed.

2. Antorbital opening posterior to the external nares rounded.

3. length of the prcnarial portion of skull is to the postnarial portion as 1+ : 1.

4. PosHcni|x>ral fenestra small.

5. Posterior bar of parietal short.

6. External nares with elevated rim; internal nares posterior to external.
7 Squamosals not greatly extended behind.

8. Orbits look up and out.

9. Septomaxillary (?).

10. Premaxillary and maxillary teeth 3(5, rounded in sect inn.

1 1. Palatines with inner edge elevated into a ridge. Palatine vacuity anterior to the middle of the transverse.

12. Transverse overlaps the external pterygoid process below.

13. Vomers extend back to interptervgoid vacuity.

14. Parasphenoid process shori

15. Interptervgoid vacuity small.
Angislorhinus:

1. Post-temporal aroade at the level of the top of the skull.

2. Antorbital opining with the anterior end even with the anterior end of the external nares. Oval. Anterior

end acute, posterior end rounded.

3. length of the prcnarial portion of the skull is to the postnarial as ."> : .'{.

4. Post-temporal fenestra largest of any known form.

6. Posterior liar of the parietal very short.

ti. External nares with elevated rim, low septum; internal nares a little posterior to the external.

7. Squamosal eon-iderably extended backward and with a strong descending hook; overhangs the bones Ik'Iow.
s Orbit* look upward more than outward.

1 Iluene I ■'. v., Heitrage zur K'enntniM und Bcurtheilung der Parasuchier, Gcolog. u. Paleontlg. Abhand-

lungen. N. I'.. Bd. x, s. -,, fig. 2, 1911.
'The same character is described under the same numlicr under each genus. The characters given are as
ribed by the authors. Some of these must !«■ modified by further study. The probable errors arc

in the description of the palatal region.

THE UPPER TRIASSIC OF WESTERN TEXAS. 69

9. Septomaxillary present.

10. Preraaxillary and maxillary teeth, 41 ±, round and oval, with serrate edges.

11. Palatine with inner edge elevated into a ridge. The palatine vacuity between the palatine, pterygoid, and

the anterior half of the transverse.

12. Transverse meets the external edge of the pterygoid. No external process of the pterygoid.

13. Vomers do not appear behind the internal nares.

14. Parasphenoid process long, covered anteriorly by the pterygoids.

15. Interpterygoid vacuity short, narrow.
Phytosaurus (Machceroprosopus):

1. Post-temporal arcade depressed; not visible from above.

2. Anterior end of the antorbital vacuity anterior to the external nares; elongate oval.

3. Length of the prenarial portion of the skull is to the postnarial as 3 : 2.

4. Post-temporal fenestra large.

.5. No posterior bar on the parietal; the postorbital takes the place of the parietal on the posterior edge of
the skull.

6. External nares without elevated rim, septum at the level of the top.

7. Squamosal region greatly extended posteriorly, with a strong hook.

8. Orbits look outward more than upward; slightly forward.

9. Septomaxillaries large, united in front and form a prominence, "if correctly determined" (Mehl).

10. Premaxillary and maxillary teeth 36-38; both rounded and oval in section.

11. Palatal surface of the skull unknown.
Mystriosuchus:

1. Post-temporal arcade depressed.

2. Anterior end of antorbital opening anterior to the nares; nares over the middle of the opening.

3. Length of the prenarial portion of the skull is to the postnarial as 5 : 2.

4. Post-temporal fenestra small.

5. Posterior bar of the parietal depressed, lying on the opisthotic, long.

6. External nares with elevated rim, septum at upper level; internal nares directly under the external nares.

7. Squamosal region extended posteriorly.

8. Orbits look outward more than upward; upward in .V. j>leiningeri.

9. Septomaxillary present.

10. Premaxillary and maxillary teeth 47-50, lioth rounded and oval in section.

11. Inner edge of palatine elevated; palatine vacuity opposite the middle of the transverse.

12. Transverse underlies the external process of the pterygoid.

13. Vomers do not extend posterior to the internal Mm, or very little.

14. Parasphenoid process fairly long, anterior end covered.

15. Interpterygoid vacuity small.
Promyxtriosiichus:

1. Post-temporal arcade at the level of the skull.

2. Anterior end of tin- antorbital opening opposite middle of external nares; elongate oval.

3. Length of prenarial portion of the skull is to the postnarial portion as 3 : 2.

4. Post-temporal foramen small.

5. Posterior bar of the parietal long.

6. Rim of external nares elevated, septum low; position of the internal nares posterior to external.

7. Squamosal region not greatly extended posteriorly.

8. Orbits look upward, perhaps in part due to enisliing.

9. Septomaxillary (?).

10. Premaxillary and maxillary teeth 47; both oval and rounded in section.

11. Inner edge of palatine raised and rugose. Palatine vacuity small, near middle of transverse.

12. Transverse underlies the outer process of the pterygoid.
18. Vomers (?).

14. Parasphenoid process long; anterior end probably covered by the pterygoids.

15. Interpterygoid vacuity long.
Lej>tosuchus:

1. Post-temporal arcade depressed.

2. Antorbital opening with the anterior end even with the anterior end of the nares; oval in outline.

3. Length of the prenarial portion of the skull is to postnarial portion as 3 : 2.

4. Post-temporal fenestra small.

.".. Posterior bar of the parietal long, lying on the opisthotic.

ti. External nares without rim, septum at, level of the skull. Internal nans direetb below the external.

7. Squamosal extended posteriorly with a strong descending process; overhanging the bones below.

8. Septomaxillary present.

9. Orbits look upward and outward.

10. Premaxillary and maxillary teeth 44; both rounded and oval in section.

11. Inner edge of palatine strong, a low ridge; palatine foramen small, opposite middle of transverse.

12. Transverse underlies the pterygoid.

13. Vomers do not appear on the palatal turface.

14. Parasphenoid proem long, covered anteriorly by the pterygoids.

15. Interpterygoid space short and narrow.

7(1 \i:\V REPTILES AND BTEGGCEPHALIANS FROM

DESCRIPTION OF ISOLATED BONES OF PAKASICHIAXS.

Many isolated hones were found in the beds in Crosby County, Texas, and
a few in the Bad Lands west of San Jon, New Mexico. Some of these are described
and figured below:

An iiiterclavicle (No. 7112, University of Michigan) found by Mr. Paul Miller
in the Bad hands "> miles west of San .Jon, New Mexico, resembles in general the inter-
clavicle <>f Myatriosuehua pUrnvroatris, figured by McGregor.1 The blade is stouter and

the articular faces for the clavicles are deeper and larger. The anterior portion is heavily

rugose on the lower side, and the dividing septum between the two articular faces h
elevated and strong. The distal end is incomplete, but enough is preserved to show t hat
the whole bone was shorter and wider than in .1/. planirostrU. The upper face of the
blade is gently convex and covered by fine radiating lines in the posterior half. The
lower surface of the hone is gently convex in front, but the posterior half is nearly flat.
This portion is marked by rough, radiating rugosities imposed upon a sculpture of very
line lines. The whole bone is 216 nun. long as preserved, and could not have been a very
great deal longer when complete (fig. 27 a I.

A second iiiterclavicle (No. 7313, University of Michigan) was found in the same
region; it is very imperfect, but indicates the presence of a very large animal, in which
the posterior extension of the interclavicle was much greater than in the one described
and figured above.

There are five ilia and partially complete pelves (Nos. 7244, 7266, 7322, 7333,
7470), all in the University of Michigan collection. Four of these are so decidedly
different that they indicate the presence of different forms of at least specific value.
They are described separately below.

No. 7322 is a perfect ilium from the right side, discovered in the breaks of Sand

k, Crosby County, Texas. It is shown in figure 27 b; plate 13, figure a. The
anterior extension of the upper edge is somewhat elongated and down-curved: it extends
as far forward as the articulation of the lower anterior edge of the ilium with the pubis.
This hook-like extension is rather broad above, hut not nearly as broad as in some of the
other ilia. The lower edge of the posterior extension of the ilium extends forward at a
gentle angle to form the articulation with the ischium. The extreme posterior edge
of this slanting surface is curved inward, giving a broad posterior face inclined backward
and outward, 'the cotylusis very deep and the faces for the articulation with the ischium
and pubis meet at a gentle angle or on the surface of a broad curve. The upper edge
of the blade is moderately broad, anil the anterior face of the blade rises almost directly
from the upper edge of the cotvlus. The inner face i- somewhat water-worn and the
articulations for the sacral ribs are in part destroyed, hut it h apparent that they stood
very much lower upon the hone than in other specimens and that the whole blade of
the bone was more slender.

No. 7211. also from Sand Creek, Crosby County, Texas, is very different in form
from the preceding (fig. 27 c; plate 13, fig. b). The anterior process of the blade of
the ilium i- very short, not reaching halfway to the articulation of the ilium with
the pubis. The extremity does not turn downward at all, hut is directed slightly upward
through its whole length. The upper edge of the blade is relatively thin and slightly
rugose on it > anterior face. 'I he whole blade is more than twice as thick as in the
preceding specimen, for on the inner side of the bone there i> a concave area looking
inward and upward before the articular faces for the sacral vertebra' begin. The posterior
process extends almost directly backward) from the articulation witli the ischium the

1 McGregor, l II Memoin American Museum Natural History, vol. ix. pari n, fin. 9, 1906.

CASE

A. Outer surface of ilium of a phytosaur, No. 7322, U. of Mich. X0.54.

B. Inner surface of ilium of a phylosaur. No. 7244, U. of Mich. X0.45.

( '. A plate of donal armor of a phytoaaur of the Mystriosucbid group, No. 7247, U. of Mich. X0.8.
I). The cranial region of skull of a small dinosaur, Cedophysia sp., No. 7478, V. of Mich. X0.9.

E. Lower view of I). I'. Lateral view of I).

("/., ravity for iiifurxlil niliiin ; ni., otic region; bpt., baaipterygoid process of baaisphenoid; etmd., occipital oondyle.

THE UPPER TRIASSIC OF WESTERN TEXAS.

71

lower edge runs upward and forward for a short distance and then turns directly back-
ward. The posterior end is broken away and lost; so the total length of the process is
uncertain, but it was evidently quite long. The lower edge of the process is turned
inward, as in No. 7322, but in this specimen it looks almost directly downward rather
than down and back. The cotylus is not so deep and the faces for the ischium and pubis
meet in a sharp angle. On the inner face (see plate 13, fig. b) the upper part of the blade
of the ilium is deeply concave; the lower edge of the concavity marks the upper edge of
the articular surface for the sacral ribs. Near the anterior end there are two deep,

Fig. 27.

A. Lower surface of interclavicle, No. 7442, U. of Mich. X 0.3.

B. Right ilium, outer side, No. 7322, U. of Mich. X 0.3.

C. Left ilium, outer aide, No. 7244, D. of Mich. X 0.3.

D. Left ilium, outer ode, No. 7333, l". of Mich. X 0.3.
I. Left ilium, outer side, No. 7266, IT. of Mich. X 0.3.

rugose pits below this edge, marking the place of insertion of the distal ends of the sacral
ribs. Below the posterior pit there is a wide rugose surface which received the posterior
half of the distal end of the second rib. Above this surface the lower edge of the concave
surface extends back as far as the posterior end of the broken bone; it evidently extended
to the extremity in the perfect specimen. The whole bone is much heavier than in
No. 7322, the anterior edge just below the anterior process of the upper edge being excep-
tionally thick.

No. 7333 differs from either of the previously described forms in that the height
is much less relative to the length (fig. 27 d; plate 12, fig. f). The length of the anterior

72 NEW REPTILES AND STEGOCEPHALIANS FROM

process of the crest is intermediate in length between that of Nos. 7322 and 7244, reaching
about halfway to the articulation with the pubis. This process is much thinner on the
upper edge even than No. 7322, as the outer side is beveled by a rough surface. The
posterior process is relatively much longer than in either of the others and extends
backward with much the same slant as in No. 7322. The posterior extremity is thickened
by the inward curvature of the lower edge, the face thus formed looking almost directly
downward. The cotylus is exceptionally deep, as there is a strong buttress developed
near its anterior edge which maintains its width to the upper edge of the bone. The
upper edge of the cotylus is extended forward on the anterior portion of the bone, forming
a sharp ridge and giving the whole region, and the articular face for the pubis, a triangular
section. The articular faces for the ischium and the pubis are exceptionally heavy.
On the inner side there are no deep cavities for the articulation of the distal ends of the
sacral ribs. The upper portion of the blade is smooth and is free above the sacrum for
some distance; this surface is only gently concave. The upper edge of the ilium is
thickened at the anterior and posterior edges, but the median portion is quite thin and
sharp. This is decidedly different from the two other forms described, where the upper
edge is thick and heavy throughout its length, the median part being nearly as thick as
either of the ends.

No. 7266, discovered near the head of Holmes Creek, in Crosby County, Texas,
consists of the two ilia and the sacral vertebrae complete, except for the neural spines.
The specimen has been slightly crushed from above downward, but not enough to materi-
ally distort the specimen. The ilia are similar in form to No. 7333, but nearly twice
as large; they probably belong to an animal of the same species, but of larger size. The
anterior process of the blade of the ilium (fig. 27 e) is relatively slender, reaching a little
more than halfway to the articulation with the pubis; its extremity is very slightly
down-curved. The posterior process is long, and the lower edge extends backward,
almost directly from the articulation with the ischium, at a gentle angle. The cotylus
is deep, especially near the anterior end, where there is a strong buttress reaching down
from the upper edge of the bone to the upper edge of the cotylus. The articular edges
for the ischium and pubis meet in a large angle. The upper edge of the blade is relatively
thin, with slight thickenings at the anterior and posterior ends. The articular portions
of these edges outside of the cotylus are very heavy. The slight depression of the
pelvis has thrown the lower edges of the ilia inward and the upper edges outward, slightly,
BO there is an apparent flattening of the upper surface. The blade of the ilium on each
side rises free from the upper surfaces of the sacral vertebrae, and there is a decided
concave area on each bone on the inner upper side. The articulation of the distal ends
of the sacral ribs is close and firm, the anterior one being received in deep rugose pits
and the posterior one articulating along the length of a rather heavy ridge (fig. 28 a;
plate 12, figs, a and b).

The sacrum is formed of two heavy vertebrae; the neural spines have been destroyed
and the neural arches are pressed downward, almost closing the neural canal, but they
were not high even in the undistorted specimen. The t wo vertebrae were not anchylosed,
even in this evidently adult specimen, and the zygapophyses between the two sacral
vertebrae are still well formed and distinct, though somewhat injured. The sacral ribs are
very wide and strong, giving a powerful attachment of the pelvis to the vertebral column.
The ribs originate high upon the sides of the neural arch, almost at the level of the zyga-
pophyBee; there is no indication of sutures at the proximal end.-. The anterior ribs
originate along the full length of the side of the vertebrae, the anterior edges being as far
forward M the anterior ext remit v of the prezygapophysis, and the posterior edges originat-
ing relatively as far back. This rib is very heavy at its proximal end and becomes more
so as it extends outward in a fan-shape; the distal end is at least 3 centimeters thick

THE UPPER TRIASSIC OF WESTERN TEXAS.

73

where it fits into the deep pits on the inner side of the ilium. There is a notch on the
posterior side near the proximal end which is in opposition to a similar notch on the
anterior side of the proximal end of the second rib, leaving an opening for the escape
of the nerves between the two vertebrae. The anterior rib slightly overlaps the posterior
rib at the distal end.

The second rib is nearly as large at its origin as the anterior one, but the neck is
somewhat narrower, due to the presence of notches on the anterior and posterior edges.
Its anterior edge slants somewhat forward, but not at so great an angle as the posterior edge,
which extends back to the extremity of the posterior process of the blade of the ilium.
The distal end is not so much thickened as in the anterior rib, and the posterior portion
of the distal end attaches to the ridge on the inner side of the ilium by an overlap.

Specimen No. 7470 consists of the nearly entire right half of the pelvis, with the
two sacral vertebrae and several associated dorsals. It was found by the author near

Fig. 28.

A. Upper surface of pelvis and sacrum, No. 7266, U. of Mich.

B. RiKhl side of pelvis, No. 7470, l". of Mich. X 0.25.

X 0.25.

the head of Holmes Creek, Crosby County, Texas. The posterior half of the ilium has
been slightly crushed down and out, and the distal portion of the pubis is lacking, but
the ischial portion of the symphysis is complete and a part of the ischium of the left
side is preserved.

The form of the ilium (fig. 28 b) is most like that of No. 7322. The anterior process
is long, reaching as far forward as the articulation with the pubis, and the distal end is
slightly but definitely down-turned. There is no buttress over the anterior portion
of the upper edge of the cotylus. The posterior process is somewhat distorted by the
crushing described above, but it is apparent that its lower edge rose rather abruptly,
much more so than in No. 7322.

The upper edge of the blade of the ilium is decidedly thickened, more at the anterior
and posterior ends, but also in the median portion. The top of the anterior process is
wide and only slightly convex. Near the extremity of the posterior end there is a very
decided thickening of the bone, due to the development on the outer face of a triangular
prominence resembling the buttress which is described upon specimens 7333 and 7266,
but at the opposite end of the blade of the ilium. The cotylus is very deep in the anterior

74 NEW REPTILES AND STEGOCEPHALIANS FROM

upper portion, but otherwise it is more shallow than in any other of the specimens;
this may be in pari due to the distortion of the specimen. On the inner side there is a
much wider area above the place of the attachment of the distal ends of the sacral ribs;
tlie anterior rib was received into a deep rugose pit, as in specimen 7244, and the posterior
rib was attached along a ridge which ran backward from the pit. The anterior face
of the lower part of the ilium descends toward the articulation with the pubis much
more directly than in any other of the specimens.

The ischium and pubis are closely anchylosed with the ilium, so that it is difficult
to follow the sutures. 'The anterior edge of the proximal portion of the pubis continues
almost directly downward from the articulation with the ilium; just below the artic-
ulation there is a large pit near the edge. A little below this, and in the lower part
of the cotylus. is the opening of the pubic foramen. The distal portion of the pubis
is lost.

The posterior edge of the ischium runs downward and then turns sharply forward,
forming the upper edge of a narrow but strong process. The symphysis is preserved
forward almost to the juncture of the ischium with the pubis; the union of the bones of
the two sides is strong, as they meet in a wide articular face and are firmly fastened
together in tiie specimen. The line of the symphysis, so far as preserved, is sigmoid.

The sacrum of this specimen is complete and resembles that of No. 7266. The
anterior rib rises from the level of the zygapophyses and is very heavy; the rib stands
with its greatest diameter vertically, which is perhaps the normal position. The distal
end was received in a pit on the inner side of the ilium. The two vertebrae are separate
and the zygapophyses between them are distinct and well formed. The neural spines
of the vertebra; are complete and show an enlargement of the upper end, forming heavy
flat tables which overhang the sides of the neural spine. The upper surface of the tabular
expansion is marked by a shallow depression which runs antero-posteriorly.

In comparison with the pelvis of Rhylidiodon carolinensis, as figured by McGregor,

7333 and 7266 are apparently of the same type and are probably Mystriosuchids,
hut Nob. 7470, 7244, and 7322 are decidedly different. The ischium of No. 7470 resembles
that of R. carolinensis, but the posterior (upper) edge is perhaps a little more curved,
the symphysis is stronger, and the lower edge of the symphysis is sigmoid in outline.
There is no suggestion of a space between the ischium and pubis, filled during life with
cartilage: the two halves of the pelvis meet at a decided angle, and there could have
been little opportunity for a median vacuity. The pubis is largely lost, but the anterior
edge is decidedly convex in its upper portion, suggesting a very different contour from
that of R. carolinensis. These differences, combined with the shape of the ilium, suggest
that this pelvis belonged to some member of the Phytosaurid group, a suggestion borne
out by the few plates found with the specimen which are typically phytosaurian. The
peculiar outline of the upper pail of the anterior edge of the pubis, suggesting a more
vertical position of the anterior face of the pubis, recalls the peculiar pelvis described by
Mehl,1 from the Triassic of New Mexico, as Acompsosaurus wingalensis, and even more
the condition found in the Pseudosuchians Ornithosucftm and Euparkeria. It is not
suppo»ei| that the resemblance implies genetic relationships, but so little is as yet known
of the Triassic reptiles in North America that note must be taken of suggested resem-
blances until their value may be correctly estimated. In v. Meyer's paper upon the
Reptilia of the Steuben Sandstone of the Upper Keuper- there are figured three phytosau-
rian ilia. Figure 1 has much the same form as Xos. 7333 and 7266, described and figured
above, while (inure ."> resembles more closely Xo. 7244. Von Meyer did not assign the
ilia figured by him to any definite genus or species.

1 Mehl, M. <;., Quarterly Bulletin of the University of Oklahoma, dew series, No. 103, p. 33, 1916.
■ Meyer, II. v.. I'aleontographica, Bd. 7, Tat. 41, figs. 1, 3, 5.

THE UPPER TRIASSIC OF WESTERN TEXAS.

75

There is a single left scapula (No. 7472, University of Michigan), in a somewhat
imperfect condition, the proximal end being slightly injured by decay. It is impossible
to give the exact outline of the anterior edge (see fig. 29 a), but there was evidently a much
greater extension of the region than in Rhytiodon carolinensis. The wide extension of
the anterior lower border recalls that of the Permian Pelycosauria and some of the
South African Upper Paleozoic reptiles. The articular face for the coracoid is thick and
heavy. The distal end of the blade was broken and the parts somewhat overthrust,
but making allowance for this the total length was 277 mm., slightly larger than R.
carolinensis.

Via. 29.

A. Right scapula. No. 7472, l". of Mich. X 0.15.

B. Right femur. No. 3305, U. of -Mich. X 0.15.

C. Right humerus, Xo. 7312, IT. of Mich. X 0.15.

D. Right radius, No. 7312, U. of Mich. X 0.15.

E. Right ulna, No. 7312, U. of Mich. X 0. 1 .",.

A single right coracoid (No. 7315, University of Michigan) is too imperfect for
figuring or description, but shows the presence of a large notch on the anterior edge.

A very imperfect scapula-coracoid (No. 7471, University of Michigan) shows only
the proximal portions and the outlines of the cotylus. Evidently there was in this
specimen the same large expansion of the anterior lower portion of the scapula as in
Xo. 7472, but the size of the preserved portions indicates an animal at least one-half
larger.

76 NEW REPTILES AND STEGOCEPHALIANS FROM

There are two femora in the collection (Nos. 7330 and 3395, University of Michigan).
The first is a very small femur 73.5 mm. long, apparently of an immature individual.
The articular ends are imperfect, as if incomplete ossification had permitted their rapid
decay. The trochanter is large compared to that in larger specimens and is well formed.
The second femur is from the right side of a fairly large individual. It corresponds in
form very closely with that of R. carolinensis (fig. 29 b).

In one small spot north of the Spur-Crosbyton road the remains of two or three
large Phytosaurs were found in close proximity. Most of the bones were badly weathered,
but a humerus, radius, and ulna were collected in close association and in good condition.
It has been assumed that these bones belonged to one individual, as they are of the
proper size and of the same side. They have been numbered 7312 in the University
of Michigan collection.

The humerus (fig. 29 c) has the two articular ends of nearly equal breadth, with the
long axes inclined at a slight angle to each other. The deltoid ridge is broken away in
part, but stood nearly at a right angle to the proximal end of the bone. The shaft is
nearly cylindrical. The lower end has a well-developed, nearly hemispherical head for
the radius and a distinct articular surface for the ulna. There is a distinct ectepicondylar
process which is not complete in the specimen. There is no entepicondylar notch
or foramen. The total length of the humerus is 28.35 cm.; the breadth of the upper
end is 14.2 cm., and the breadth of the lower end 12.7. cm.

The radius (fig. 29 d) has a nearly straight cylindrical shaft; the upper end has a
well-formed concave articular surface; the lower end is but slightly expanded. The
total length is 24.6 cm.

The ulna (fig. 29 e) is very broad proximally and gradually contracts in breadth to
the lower end. The shaft is compressed, so that it is much thinner than broad; this
may be, in part, due to compression. The articular face is shallow, and there is a very
small olecranon process. The out corner of the lower end was broken away and lost
and has been restored in plaster. The total length of the ulna is 30 cm.

All of these limb-bones, as well as the femur described above, and the remains of
less perfect limb-bones in the collection indicate that the Phytosaurs of the region were
relatively long-limbed forms and must have been capable of raising the body from the
ground for a time at least, and also that they were capable of relatively rapid progression
upon the land.

There are in the University of Michigan collection four basi-cranii of Phytosaurs,
Nos. 7257, 7261, 7474, and 7505. The first three show the condyle complete, the last
has the condyle broken away and lost. These are all typically phytosaurian in form and
correspond closely with Huene's description of P. kappfi as opposed to his P. rutemeyeri.
The general form is shown in figure 30, and plate 12, figure c, of No. 7261. All show
clearly the pit in the occipital condyle for the anterior end of the notochord.

No: 7474 is the smallest, and is interesting in that it shows the lower portions of
the exoccipitals in position, meeting in the median line and excluding the basioccipital
from any part in the foramen magnum. (Plate 12, fig. d.)

No. 7505 is the next largest. The exoccipitals are lost, and it shows clearly the
corrugated articular surfaces on the basioccipital for the exoccipital. The floor of the
brain-case i- eroded and the suture between the basioccipital and the basisphenoid is
clearly traceable all round the bone. The edges of the shallow hypophysial cavity
are broken away and the entrance of the foramina for the internal carotid arteries into
the cavity is very clear. (Plate 12, fig. e.)

No. 7261 is the third largest and the most perfect. The tubera basioccipitalia
are seen to be formed by both the basioccipital and the basisphenoid. The tubera are

CASE

1t'

PLATEIM

?**

:v* '

Leptoauchus croabienais, 1'niversity of Michigan. All figures X0.15.

A. Upper surface of skull. D. Inner side of the right lower jaw.

B. Lower surface of skull. E, < Klter side of the right lower jaw.
('. Right side of skull.

THE UPPER TRIASSIC OF WESTERN TEXAS.

77

prominent and rugose, with a deep, irregular groove marking the position of the suture.
Between the tubera is a sharp angulation marking the line between junction of the
basioccipital and the basisphenoid, and along this the suture between the two bones is
easily traced. On the sides of the basioccipital, just posterior to the basisphenoid suture,
there is a shallow groove running upward and backward. Between the tubera and
basipterygoid processes on either side are the large openings of the foramina for the
internal carotid arteries. These foramina run upward and inward and open a little to
either side of the lower end of the hypophysial cavity. The cavity is triangular in
section above, but contracts rapidly below; it is surprisingly small and shallow. The
basipterygoid processes are well developed and point outward and downward, with a
smooth, flat, oval face for the pterygoid. The sphenoidal rostrum is broken away in
all of the specimens, but was evidently vertical, narrow, and of great extent. In none
of the specimens is there any suggestion of a composite structure, i. e., an upper (pre-
sphenoidal) and a lower (parasphenoidal) portion, such as described and figured by Huene.

Fig. 30.

A. Basicranial region, left side, No. 7261, U. of Mich.

B. Lower side of same

X0.6.

No. 7257 is the largest of the specimens, being nearly twice the size of No. 7474
and the least well preserved. It has lost the occipital condyle and parts of the tubera
and basipterygoid processes. The anterior lower part of the basioccipital is still in
position, and clean breaks show (he suture between it and the basisphenoid. The
whole basisphenoid is shorter and higher than in the other specimens. The basipterygoid
processes are shorter and less well developed; they pointed more directly backward,
and on the inner posterior corner of the base of each there is a prominent rugosity, which
on the left side has the form of a pit with elevated edges. The space between the basi-
pterygoid processes is marked by two prominent rugose ridges, which converge forward
and unite at the origin of the sphenoidal rostrum. It would appear that this space
between the processes was almost vertical instead of slanting rather steeply upward and
forward, as in the other specimens, but as this would make the sphenoidal rostrum rise
vertically in the skull it is probable that the position of the occipital condyle was different
from that it occupies in the other specimens. This specimen is so much larger and so
different from the others and from any described form that it indicates, in all probability,
a new species, but it seems undesirable to give a new name to such imperfect material,
especially as the large size suggests the possibility that the peculiar features may be
due to advanced age.

7* NEW REPTILES AND STEGOCEPHALIANS FROM

DESCRIPTION OF THE REMAINS OF DINOSAURS.

Certain specimens collected in Crosby County, in the same beds with the
Phytosaurs, indicate the presence of a small Dinosaur. Notable among these are
a posterior cranial region (No. 7473, University of Michigan) and a string of cer-
vical and anterior dorsal vertebra (No. 7507). Both of these specimens were
collected on the west side of the Blanco River, north of Cedar Mountain, the first
in the breaks to the south of the old Spur-Crosby ton mail-road and the second in
the breaks just north of that road.

The cranial region is well preserved in a hard matrix of clay, which has permitted
the working out of the smaller details and the foramina of the region. As shown in
plate 13 d to f, the basicranial floor is preserved as far forward as the hypophysis.
The distal portions of the exoccipital opisthotics are lost. The occipital condyle is nearly
hemispherical, with an extremely short neck. The portion of the basioccipital below
the condyle descends almost vertically and the posterior edges of the tubera are not in
advance of the anterior edge of the condyle. The vertical portion of the basioccipital
is deeply concave posteriorly and the lower face of the neck of the condyle is perforated
by two small foramina near the middle line, probably for nutrient vessels. The sutures
between the basioccipitals, exoccipitals, and supraoccipitals can not be made out.
The foramen magnum is depressed oval in outline, with the transverse axis nearly as
great as the width of the condyle. The supraoccipital is smooth, slightly concave, and
slants upward and forward to the rough, incomplete edge, which must have been very
near to the articular surface for the bones above. As described by Huene in Anchisaurus,
the exoccipital has three ridges; one runs inward and backward to the upper outer
edge of the condyle, one downward on the outer side to the tubera, and one outward
and backward to form the lower edge of the opisthotics. In the angle formed by the
last two there is the good-sized opening of the foramen for the XII nerve. This foramen
can be traced on the broken surface of the bone to its opening on the inner side, not
far within the edge of the foramen magnum.

The lateral view of the specimen (plate 13 e) displays the peculiar characters
of the skull most strikingly. The inner portion of the opisthotic and the prootic meet
in a smooth concave surface which is inclined to the posterior surface of the supraoccipital,
so that the two would meet at a sharp angle. The beginning of the opisthotic shows that
that bone extended outward and backward and slightly upward. On its lower side is
a deep groove which leads into the surprisingly large otic cavity. Be*low this cavity the
basisphenoid extends downward, terminating in the long and slender basipterygoid
processes. The distal ends of these processes are broken away, so that it is impossible to
determine the nature of the articulation with the pterygoids. The basipterygoid pro-
CCOBCB are rounded and heavier behind, but become very thin anteriorly; the anterior
edges are broken away, but it appears that the processes of the two sides came together,
or very nearly so, in the median line. Between and beneath the anterior portions of the
process is a large cavity extending forward as far as the groove on the side of the basi-
sphenoid described below. This cavity is apparently blind; no openings in its wall have
been detected, though an artificial opening was made by the needle in the thin anterior
port ion . The side of the basisphenoid is peculiar. In the broken condition of the speci-
men it appeared as if the bone were composed of two parts, an anterior overlapping below

00 to the posterior part, which in turn overlapped in the same way upon the basioccipital;

1 hi- appearance, however, is less apparent when the upper half of the brain-case is
put in place, for the upper pari of the apparent groove on the side of the basisphenoid
is resolved into a part of a foramen. The anterior portion of the basisphenoid has

THE UPPER TRIASSIC OF WESTERN TEXAS. 79

the sides slanting obliquely inward and downward until they meet in a thin edge
below; the extreme terminus of this edge is broken away. Posteriorly there is a break
between the edge and the anterior portion of the posterior part of the basisphenoid,
formed by the approximating edges of the basipterygoid processes. This break empha-
sizes the apparent bipartite character of the basisphenoid. On the sides of the basi-
sphenoid and marking the line between the two portions is a groove, which is very pro-
nounced below and becomes very shallow in its upper third. A little below the middle
of the course of the groove there is a deep extension forward of the cavity which runs a
little downward as well as forward. This cavity is so deep that exploration is difficult,
but apparently its anterior end communicates by a very small foramen with the hypo-
physial cavity below. Above this cavity the groove becomes very shallow and finally
terminates in a foramen which penetrates the wall of the brain-case. This foramen
is the common terminus of two leads, one the groove just described and the other a
foramen which in the specimen is first detected on the broken lower outer edge of the
opisthotic; the distal termination of the latter is not determinable. From its position
on the inner wall of the brain-case the common foramen is evidently the outlet for the
VII nerve. The deep vacuities at the middle of the course of the groove are probably
the inlets of the internal carotid arteries. Anterior to the grooves the sides of the
anterior portion of the basisphenoid contract rapidly into the thin ridge on the lower
edge; above this sharp edge and covered by it is a cavity, conical in form, with its posterior
end almost a point and the anterior part much larger in diameter. The anterior edge of
this opening is apparently coincident with an opening into the brain-cavity and so must
be the cavity for the pituitary body. On the sides of the bones, just about opposite
the cavity for the pituitary body, there are two small foramina on each side, probably
for blood-vessels. On the upper edge, just anterior to the openings of the VII nerves,
there is a large foramen on each side which can only be the opening for the V nerve.
The anterior portion of the basisphenoid described is in the position of the bone which
is called by Osborn the orbitosphenoid.

On the inner side of the brain-case there is visible a decided angulation between
the parts of the floor formed by the basioccipital and the basisphenoid; near the middle
of the flat floor formed by the basisphenoid, but nearer the anterior than the posterior
edge, there is a pair of small foramina in the position usually occupied by the VI nerve.
The posterior part of the inner side of the brain-case wall is marked by a decided swelling
which sheltered the inner ear. The exact interpretation of this region is impossible,
because of the slight injury to the bone on both sides. Beneath the swelling marking
the position of the canals of the ear there is an opening of considerable size on both sides,
which leads directly into the large cavities on the sides of the skull at the inner end of
the opisthotic. This opening in all probability was the outlet for the IX-XI nerves.

lust anterior to the swelling marking the position of the canals there is a shallow
notch with an elongate form, its greatest diameter vertical; it is apparent that the bottom
of this notch has not been reached in cleaning, and it seems very probable that there is
a small foramen leading from it into the region of the otic canals; if this is correct, this
notch marks the position of the escape of the VIII nerve; but a similar notch is shown by
Osborn, in his figure of Tyrannosaurus, just posterior to the otic region, which he regards
as marking an evagination of the dura mater into the cranial wall. Almost directly
above these notches and near the anterior edge of the preserved portion of the cranial
wall there is a second deep pit on each side which can not be traced through the bone
and appear to be blind; these are almost exactly in the position marked by Osborn, in
\i\- figure of Tyrannosaurus, as other evaginations of the dura mater.

This fragment of a skull is so radically different from that of the Phytosaurs found
in the same beds, and corresponds so well with the figures and descriptions, so far as they
have been given, of the same region in the primitive Theropodous Dinosaurs, that there

Ml NEW REPTILES AND STEGOCEPHALIANS FROM

can be very little doubt of its subordinal position. Little can be said of its generic
position. The only other Dinosaur remains that have been found in the Triassic of the
Southwest are the very incomplete remains of forms described by Cope as CaHophysU
from the Gallina Mountains in New Mexico. These forms are placed in the family
Ccelurida?, occurring in the Upper Keuper in Europe and in the Upper Triassic in North
America. For the present, this fragmentary skull and the few vertebra1 described below
may be provisionally placed in or near this genus, especially as the remains of Phytosaurs
and Stegocephalians which occur in the same beds indicate an Upper Triassic age.

In 18S7, Cope1 described the remains of a small genus of Dinosaur from New Mexico,
which he at first placed in the genus Ccelurus, later referred to Tanystrophceus, and
finally placed in a new genus, CcelophysisS Three species were described — longicollis,
Iniitri, and unllistoni. According to Huene, the type material of the last species has
been lost.

During the summer of 1921 the author collected a series of small Dinosaur vertebrae
(No. 7o()7, University of Michigan), a few miles north of Cedar Mountain, in Crosby
County, Texas, in the same beds with the Upper Triassic Phytosaurs and Stegocephalians
described in this paper. When found the vertebra? were in position, but were fully
exposed and badly broken as they lay on the crumbled surface of a light cream-colored
clay. The anterior and posterior members of the series and many small fragments
of ribs were loosened from the rest of the series and had slipped down the face of a rather
steep slope. Because of the crumbled condition of the matrix, it was impossible to
collect the specimen in a block; it could only be taken up with as much of the matrix
as possible, which was filled with the minute fragments which had been separated from
the vertebra?. Reassembling the specimen has been a very long and tedious piece of
work and has been only fairly successful as yet, but enough has been accomplished to
show the character of the vertebra?. The specimen is of importance, as it is the largest
series of Dinosaur vertebra? yet found in a free condition in the Triassic of North America.

The series contains 20 complete vertebra?, with the remains of two very imperfect
ones. Of the series, the first four are elongated and very similar to those figured by
Huene as characteristic of Anchisaurus ceelurus and Coelophysis. The fifth of the series
is shorter and shows for the first time in the series a well-developed transverse process;
it is probable that the first four are cervicals and the remainder dorsals. There is no
certain indication of the sacrals; the transverse processes of the last two are indicated
only by the bases, but these are well formed and not the bases of sacral ribs. In Anchi-
saurus there are 14 dorsals and probably 9 cervicals. According to Osborn3 there are
23 presacrals in Strulhiomimus, Allosaurus, and Tyrannosaurus. In Slruthiomimux,
with 10 cervicals, the first dorsal is indicated by the presence of the first free rib according
to Osborn. If we accept this as a criterion, the fifth of the series here described would
be the first dorsal and there would be 16 dorsals, a number that seems excessive, but can
be questioned only by considering the last two as sacrals, a proceeding that is not war-
ranted by the form of the vertebra?.

The first of the series (sixth (?) cervical) was found a little separated from the
series, washed down the bank, and the zygapophyses and the neural arch have not been
reassembled. It can be seen, however, that the zygapophyses were elongate, low, and
horizontal and that the neural spine was low. The centrum is elongate, the anterior
face concave and broader than the posterior; its borders have been injured by decay.
The lower face of the centrum is broader anteriorly and nearly flat at the anterior end;

■Cope, l. I'. boolean Naturalist, pp. 36Z-368, 1887.

' The literature of this gHMM il pm by Hay, Bulletin United States Geological Survey, No. 179, p. 493, 1902,

and Huene, GeoL u. Paleoat Abhdlg., N. F., Bd. vm, s. 118, 1906. In the latter paper figures are given

of some of the specimens and the ipeciM are in part redeserilx'd.
'Oshorn, H r\, Bulletin American Museum Natural History, vol. 35, art. 43, p. 735, 1917.

THE UPPER TRIASSIC OF WESTERN TEXAS.

81

posteriorly the lower face becomes rounded from side to side, with no trace of a keel
or lateral ridges or angles. The posterior portion of the centrum descends slightly,
maintaining its rounded outline. The posterior face has an oval outline, being higher
than wide, due to the development of a lip on the lower edge. It is concave, but less
so than the anterior face, and slants from above downward and backward. Length
of the base of the centrum, 52 mm.

The second of the series (seventh (?) cervical) is much more complete. It lacks
the anterior zygapophyses and the neural spine. In this vertebra the anterior face is
complete; it is deeply concave and is inclined parallel to the posterior face of the preceding
vertebra, i. e., from above downward and backward. On either side of the anterior
portion of the centrum there are two processes; the upper begins as a low ridge at
about the middle point of

the side of the centrum id 9c^a^f\

and extends forward to the
edge of the anterior face,
becoming more prominent
during its advance. The
second process, below the
first, is shorter, not over
12 mm., and heavier; it
continues out upon the
edges of the anterior face
of the centrum, forming a
narrn w triangular face
W h i c h looks downward
and outward and slightly
forward. These two pro-
cesses arc distinct, as
shown by several vertebrae,
but are very suggestive of
the complete arch figure by Marsh1 in Coelurus and described by Cope and Huene in
Cadophytia longicoUis. It is possible, even probable, that there was such a complete arch
in the anterior cervicals, but in the posterior portion of the cervical series the arch has
broken into distinct diapophysis and parapophysis. The posterior zygapophyses are
depressed and elongate, extending back nearly as far as the upper edge of the posterior
face of the centrum. The rest of the vertebra is much like the preceding one. Length
of the base of the centrum, 52 mm.

The third of the series (eighth (?) cervical) has the upper process on the side of the
centrum stronger and extending outward more than in the second; the process is a little
higher on the side of the centrum. The lower process is shorter and has a larger face.
The anterior zygapophyses are preserved; they are horizontal, heavy, and extend well
forward of the anterior face of the centrum. Length of the base of the centrum, 44 mm.

The fourth of the series (ninth (?) cervical) has the anterior zygapophyses locked
with the posterior ones of the third; they are still nearly horizontal, but are slightly inclined
upward. The upper process on the side of the centrum is now very strong and stands
well out from the vertebra; it is so high upon the centrum that its lower edge is but
slightly below the lower face of the neural canal. The anterior edge of the centrum
is injured, but it can be seen that the lower process is now but a triangular face on the
lower edge of the face. Length of the base of the centrum, 42 mm.

' Manh, 0. C, Dinosaurs of North America, Sixteenth Annual Report U. S. Geological Survey, plate vn,
Oft*. -', 2a, 1896.

Fig. 31.

( Vrvicul and dorsal vertebra- of a small Dinosaur, Codophysis, sp. No.

1'. of Mich. X 0.5.

7507,

82

NEW REPTILES AND STEGOCEPHALIANS FROM

This structure of the posterior cervicals is very similar to the condition found in
Stttotaunu hermanniamu Huene.1

The fifth of the series (first (?) dorsal) has the upper process on the side of the
neural arch; only its proximal portion is preserved, but it is apparent that it stood well
out from the side of the arch, forming a distinct transverse process. The lower process is
a triangular face on the lower edge of the anterior face of the centrum. In this vertebra
the zygapophyses and the neural spine have not been replaced. Due in part, but not
entirely, to compression, the centrum is notably shorter and heavier than the preceding
one and the posterior face of the centrum does not descend lower than the anterior.
The faces of the centrum are now vertical and larger than in the cervical series. Length
of the base of the centrum, 27 mm.

The sixth of the series has a well-developed transverse process which stands out at
right angles from the high neural arch. This process is supported by two ridges running
from the upper edges of the anterior and posterior faces of the centrum to its base;
between these ridges and on either side of the base of the processes are deep pits. The
lower process is still visible as a small facet on the lower edge of the anterior face of the
centrum. The neural arch is high and the anterior and posterior zygapophyses rise at
a considerable angle instead of lying horizontal. Length of base of centrum, 27 mm.

The seventh of the series has well-developed transverse processes; the extremities
are not replaced, but the base shows that they stood out at right angles to the neural
arch. The base is supported by four ridges running respectively from the anterior and
posterior zygapophyses.and from the upper edges of the anterior and posterior faces
of the centrum. Between these ridges, on the anterior and posterior faces of the base
of the transverse process, there are deep pits. Upon this vertebra appears the last
trace of the lower process. The zygapophyses have not been replaced, but it is evident
that they rose at a decided angle from the neural arch. The lower edge of the centrum
is injured, but its length is approximately 32 mm.

From this point in the series backward the vertebra; do not change radically.
The transverse processes increase slightly in length and then decrease; the longest are
upon the tenth to the thirteenth. The zygapophyses are very steeply inclined upward
on the ninth and tenth and are less steeply inclined in the posterior vertebrae. The
centra become gradually heavier toward the posterior end of the series, and the bases of
the transverse processes are shorter anterio-posteriorly, due to the atrophy of the support-
ing ridges. In none of the vertebrae is there any sign of a median keel on the centra
or of any lateral ridges. The length of the base of the centrum of each is as follows:

Kth

9th
lflth
11th

mm.
31
88

33
36

mm.

13th 37

14th (injured)

15th 86

16th 88

17th
18th
l'.lth
20th

mm.
39
38
36
35

12th

35

These measurements are as accurate as possible, but are not quite
because of slight compression, breakage, etc., in the specimen.

Fractures in the mid-line of several vertebrae show that the centrum
but not so thin-walled as in Coelurus. The ribs are uniformly single-headed;
to the fourteenth vertebra has a thin head 7 mm. wide; other rib-heads not
apparently belonging to more anterior vertebrae, are nearly twice as broad
of the rib in certain parts of the series was T-shaped in section. The ribs
broken :m<l the ends so injured by decay that only portions have as yet been

as in nature

was hollow,
one attached

located, but
. The shaft
are so badly
reassembled.

Bnaae, F. v., Neon Jafcrb. f. Geol. Min. u. Pal. JahrK., 1915, No. 1, TV. 111.

THE UPPER TRIASSIC OF WESTERN TEXAS.

83

These vertebrae must be tentatively assumed to belong to a small Dinosaur, such
as bore the skull, the posterior portion of which is described above, as the sizes of the
two specimens correspond very closely, and as they were found in the same beds and but
2 or 3 miles apart. Until further evidence is
found, the vertebrae may be assigned to the
genus Caelophysis.

A femur, No. 3396, University of
Michigan (fig. 32), is rather puzzling. It is
42.16 centimeters in length. Compared with
a typical phytosaurian femur (No. 3395), it
is much heavier at the articular ends and the
articular surfaces are better formed. The
relation of the long axes of the articular ends
is different in the two forms; in No. 3395
the axes are nearly at right angles, so that
when the head was in the socket of the pelvis
the anterior face of the lower end was pre-
sented almost directly forward; in the second
femur the axes are at an angle of 60°, so that
the anterior face of the lower end was pre-
sented inward as well as forward.

In the phytosaurian femur the tro-
chanter has the characteristic form of an oval
elevation with a depressed center, and the
lower end has a simple articular region ob-
scurely divided into two parts. In the other
femur the trochanter is simply a heavy rugose
area but little raised above the rest of the
bone; the lower end is heavy and divided
into two areas, but on the posterior side of the A- Wt """""J • Dmosa"r> front view> No. 3396.

., ' ... U. of Mich. X 0.25.

outer part there is a prominent process which u. Posterior view of same

is continued upward on the shaft of the bone

as a ridge for some distance. This femur has a very dinosaurian aspect, resembling

in many details the figures of the femora of Plateosaurus and Gressylosaurus from the

Triassic of Europe, and, so far as can be determined, the femur of Anchisaurus from the

Triassic of North America. No remains of Dinosaurs of the size indicated by this femur

have been found in the Triassic of the western portion of North America; Coelophysis

Cope was decidedly smaller. It is to be hoped that more material revealing the skeleton

of this Dinosaur will be found.

Fig. 32.

COPROLITES.

There are several types of coprolites contained in the collection. The number found
in the Holmes Creek region is very large and the condition of preservation is in most
cases very good.

The first type is that which must be assigned to the larger reptiles and stegocepha-
lians. These vary from 5 cm. in length to as much as 15 or 18 cms. The form is generally
regular, with smooth surfaces. In none that have been found has it been possible to
detect anything, such as comminuted bones or scales, that would indicate the nature
of the food. Microscopic sections have not been made of any of the coprolites.

St

\i:\V REPTILES AND STEGOCEPHALIANS.

The second typo is small, ranging from 1 cm. to as much as 7, in extreme cases,
in length. They show very perfectly the spiral nature of the coprolite; two very good
specimens are shown in figure 33 a and b. It seems probable that these are the coprolites

Fia. 33.

A and B. Two coprolites showing trace of the spiral valve. X 1.

(' and I). Two coprolites showing the longitudinal ridges. XL

E. Figure of a tooth of an unknown form, No. 7506, U. of Mich. XL

of the Dipnoan fish of the region, but in the light of the great number of the coprolites
it is strange that not more of the skeleton of the fish has been found; only two teeth and
no other parts of the skeleton have been found.1

The third type of coprolite is very different from the other two. These range from
2.5 to 7 cm. in length. They are all somewhat curved, sometimes approaching a cres-
cent ic form. In many, especially of the smaller coprolites of this type, the outer side of
the curve is marked 'by deep and regular grooves very evenly spaced (fig. 33 c). This
is so striking that when the first fragment of one was found it was suspected that it
represented the surface of a Cephalopod shell. In most of the larger coprolites of this
type the markings are very obscure or wanting, but the curved form is always character-
istic. In the larger forms there is frequently an additional part of the coprolite which
lit- over the end as a cap or cloak (fig. 33 d). It is impossible to associate this type of
coprolite with any definite group of animals; the only suggestion that can be made is
that in many amphibians the distal portion of the rectum is thrown into parallel linear
folds — just such an arrangement as would make linear markings upon the fcecal mass.

INCERTiE SEDIS.

A small fragment of a jaw contains a very singularly shaped tooth which the author
has been unable to identify, or even to determine its relationships. As shown in figure
33 e, the tooth is elongate oval in section, with the greatest diameter parallel to the
length of the jaw. In general outline it is reminiscent of the teeth of the Permo-
( iarboniferous reptile JJiadecles, but the upper edge is contracted, slightly concave antero-
poeteriorly, and has a narrow, flat surface, slightly inclined toward the outer (?) side.
The outer (?) side ot the tooth is slightly concave, and there is a swelling at the base
Bad then a sharp contraction. The tooth is without root and attached directly to the bone.
The base of a second tooth shows the rather spongy character of the base (fig. 33 e).

•Cav I I Occasional Papers of tin- Museum of Zoology, University of Michigan, No. 101, Apr. 9, 1921.

861
C37

Case, Ermine Cowles
New reptiles

P&ASci

PLEASE DO NOT REMOVE
CARDS OR SUPS FROM THIS POCKET

UNIVERSITY OF TORONTO LIBRARY