I

^igm,^>^wm^'^wm^'nw^p'^^m^i^^fi-v^^^

L I E) R.ARY

OF THE

U N IVER5ITY

or ILLINOIS

bbO.5
v.o-G>

Cop. Q
REIV \ '«

GEOLOGICAL SERIES

OF

FIELD MUSEUM OF NATURAL HISTORY

Volume VI Chicago, March 24, 1939 No. 24

NEW PANTODONTA AND DINOCERATA FROM
THE UPPER PALEOCENE OF WESTERN COLORADO

By Bryan Patterson
Assistant Curator of Paleontology

The upper Paleocene Plateau Valley beds have been famous
hitherto chiefly for having yielded skeletons of the large pantodont
Barylamhda faberi. The 1932 and 1933 expeditions of Field Museum
collected excellent material of this species, together with some
fragmentary remains of smaller mammals which suggested that the
horizon was equivalent in age to the Tiffany of southwestern Colorado
(Patterson, 1936, pp. 397-398). In addition, a few teeth and bone
fragments were obtained that were sufficient to indicate the presence
of other large mammals in the formation, but too incomplete to
provide any positive information about them. In 1935 Mr. Edwin B.
Faber of Grand Junction, Colorado, to whose interest and initiative is
due the discovery of the Plateau Valley horizon, collected the right
Mu of a large and specialized uintathere belonging to an unknown
genus. This specimen has been mentioned briefly by me (1936, p.
397) and by Scott (1937, p. 475). In the early spring of 1937 Mr.
Alfred A. Look, also of Grand Junction, and his son found a partial
skull of an entirely new type of pantodont. The skull was on the
surface and near-by a number of bones were exposed. Upper cheek
teeth of this new form were among the problematical specimens
found in 1933.

In the summw of 1937, thanks to the generosity of Mr. Stanley
Field, a third expedition was sent into the Plateau Valley area. The
personnel consisted of Mr. James H. Quinn and myself, reinforced
for parts of the season by Mr. Elmer S. Riggs, Mr. Clayton A.
Quinn, and Mr. Theodore Burdosh. Mr. Look's prospect was inves-
tigated and yielded a nearly complete skeleton. Several good
specimens of the large uintathere, first discovered by Mr. Faber,
were recovered, as well as remains of a smaller form of the same group.
Mr. Quinn was so fortunate as to unearth a partial skeleton of a

No. 441 351

JN»:uial History Survey
Lilrarv

352 Field Museum of Natural History — Geology, Vol. VI

small pantodont related to Barylambda. In addition to the outstand-
ing discoveries of the larger forms, a small collection of medium-sized
mammals was made. No true microfauna has yet been found, a
lack which is in striking contrast to other upper Paleocene localities.

Detailed description of this new material has been delayed by
pressure of other work, particularly by research on parts of the
South American collections brought together by the Marshall Field
expeditions. In view of this situation and because of the importance
of the new pantodont and uintathere material^it is safe to state
that it profoundly modifies current concepts of these two groups —
it has been decided to bring out this preliminary paper.

Grateful acknowledgments are due to my companions in the
field, and to those residents of Mesa County who aided in collecting
specimens or who facilitated our work in other ways. In particular,
I may mention Mr. Faber and Mr. Look, Mr. J. Elvin Harris and
Mr. Douglas Harris and their families, of Mesa, and Mr. Hatton
Edgerley of DeBeque. The preparation of the material has been
carried out with great skill by Mr. Quinn. The photographs are
the work of Mr. C. H. Carpenter, the retouching is by Mr. Carl F.
Gronemann, and the drawings are by Mr. Frank Gulizia.

Order Pantodonta^ Cope
Family Coryphodontidae Marsh 1876
Sparactolambda^ gen. nov.
Type species. — S. looki sp. nov.

Distribution.- — Upper Paleocene, Plateau Valley beds, Colorado.

Diagnosis. — Dentition l| Cy Pf M|. Upper canine enlarged,

tusk-like; P^"^ with antero-posteriorly compressed protocones,

' Simpson (1937a, pp. 265-269) has discarded the old name Amblypoda
altogether, and has employed Pantodonta for the natural group composed of
Coryphodon, Pantolambda, and their allies. My own preference would have been
to restrict the term Amblypoda to this group, and a fairly sound case could have
been made for such a procedure. In his original description of Amblypoda (1875,
p. 28) Cope stated that he proposed ". . . to regard Bathmodon [ = Coryphodon] as
the type of a distinct order of Mammalia, which will also include the genera Loxolo-
phodon and Uintatherium in a distinct subdivision." This would indicate that he
had the coryphodonts primarily in mind when he proposed the order, and is some-
what at variance with Simpson's statement that Amblypoda was based "...
about equally on the coryphodonts and the uintatheres." However, as Simpson
has pointed out, "Amblypoda" has so long been employed to cover an unnatural
association of three distinct groups that continued usage of the term, even in a
restricted sense, would probably be confusing, and for this reason it is perhaps
desirable to eliminate it entirely.

2 In allusion to its probable pulling-feeding adaptation (see below), and to its
tooth structure and relationship with the other -lambda genera.

FI

^-"^ '^ Paleocene Pantodonta and Dinocerata 353

cingula running internally from bases of paracones and metacones
not connected to apices of protocones; upper molars broader antero-
posteriorly and narrower transversely than in Barylamhda, propor-
tions of M- to M- to M- about as in Pantolamhda and Coryphodon,
paracones notably more internal than metacones. Lower incisors
broader and less conical than in Barylamhda and Pantolamhda, ap-
proaching those of Coryphodon; lower canine very large and long
antero-posteriorly, crown consisting of an anterior, pointed, slightly
recurved apex and a long, blunt edge or blade sloping downward and
backward, root very stout, buttressed antero-internally and tapering
posteriorly — altogether a unique tooth; Py with paraconid and
metaconid well differentiated, Pif with large centrally situated
metaconid; lower molars without metastylids, lower crowned and
with trigonids and talonids of Mx_tj farther apart than in Barylamhda,

Head rather smaller in proportion to body size than in Coryphodon,
larger in proportion than in Barylamhda and Pantolamhda. Skull
with anterior narial opening terminal; muzzle flaring; premaxillaries
very short, ascending rami confined to narial opening and not appear-
ing on side of face; nasals narrow, as in Coryphodon, not greatly
expanded posteriorly, in contact with maxillaries for nearly all their
lengths; zygomatic arches slighter than in Coryphodon, but flared
outward as in that form; sagittal crest high, long; occiput decidedly
triangular, condyles not extending posteriorly beyond it. Mandible
with horizontal ramus increasing in depth anteriorly; chin deep,
abrupt, and strong; condyle nearly on level with cheek teeth;
coronoid process comparatively low; angle not extending posteriorly
beyond condyle.

Neural arches of cervicals greatly strengthened; centra of dorsals
and lumbars considerably longer than in Barylamhda; sacrum low;
anterior caudals flattened dorso-ventrally, tending to unite with
sacrum.

Clavicle thick, very deep. Head of humerus wider antero-
posteriorly than in Coryphodon, with anterior groove as in that genus,
articular surface extending on to great trochanter; deltoid crest high,
thinner transversely than in Barylamhda; entepicondylar foramen
present, situated high on entepicondyle. Centrale completely
united with cuneiform, forming a thick wedge between lunar and
trapezoid; magnum not articulating with Mc. II; ungual phalanges
deep, short claws.

Pelvis approaching that of Coryphodon, ilium not as expanded.
Tarsus very Coryphodon-like, astragalus without neck; metatarsals

354 Field Museum of Natural History — Geology, Vol. VI

stouter than in Barylamhda, Mt. V with very large, robust, peg-hke
lateral process.

Sparactolambda looki^ sp. nov.

Holotype.—FM. No. P15520, partial skeleton.

Horizon arid locality. — Plateau Valley beds, about 80 feet above
the base, three miles west of DeBeque (one mile west of the Finley
Ranch House), Mesa County, Colorado?. found by Alfred A. Look.

Paratype.—F.M. No. PI 5523, incomplete skull, sacrum, right
OS innominatum. Found by Alfred A. Look one-quarter mile west
of the holotype.

Diagnosis. — As for the genus; for measurements see pages 360-361.

Discussion. — This animal is of extraordinary interest from two
points of view. In the first place, it shows profound structural
modifications in response to the acquisition of an unusual habitus,
one hitherto unknown among the Pantodonta. In the second, it is,
apart from these modifications, intermediate in many respects
between Pantolamhda and Coryphodon, and therefore of great impor-
tance in pantodont phylogeny. The habitus adaptation may be
briefly discussed first in order to clear the way for consideration of
the more important although somewhat less spectacular characters
of taxonomic significance.

Sparactolambda presents the following combination of unusual
features: large, unique lower canine consisting of anterior hook and
posterior blade, heavy chin (associated with the canine develop-
ment), low coronoid and condyle, small angle (in comparison with
other pantodonts), greatly strengthened cervicals, strong clavicle,
and clawed manus. In addition, the ulnar articulation on the
humerus is very extensive, the upper lip of the sigmoid notch is
slight, thereby permitting great mobility, and there is a tendency
toward the development of grooves and rugosities above the distal
articular surfaces of the metacarpals, seen particularly on Mc. II,
as in the chalicotheres and certain carnivores. This array may
most logically be interpreted as an adaptation to a diet consisting
in large part of roots and underground tubers. These presumably
were laid bare by the powerful claws and then caught up by the
hooks of the lower canines. They were then further uprooted by
pulling and jerking the head and neck — the strong dorsal neck
muscles imparting great power to these movements^ — and sliced

1 Named for Mr. Alfred A. Look, the discoverer of the type specimens.

2 Bohlin believes, correctly I think, that dorsal thickening of ;;halicothere
cervicals is indicative of similar habits (1936, p. 325).

o

a

CD

O

•?:

II

V o

V.

CO S.
>»

Sw

fe

355

356 Field Museum of Natural History — Geology, Vol. VI

off by the upper canines acting against the blades of the lower.
The reduced coronoids and angles and the low condyle permitted a
gape sufficiently wide for the mandible to clear the upper canines.

Certain heritage characters, such as the clavicle and the promi-
nent deltoid crest of the humerus, which did not come into existence
pari passu with the acquisition of this adaptation, were neverthe-
less of great use to it and probably underwent additional responsive
modification. The upper canine probably comes under this cate-
gory. If Sparactolamhda, as I believe, is to be regarded as an offshoot
of a stock that was approaching Coryphodon from a Pantolamhda-
like ancestry, then an enlarged upper canine was very likely a part
of its heritage. Such a structure would have been of advantage in
the development of the feeding mechanism suggested above, in all
probability playing a large part in conditioning the peculiar shape
of the lower canine. Incorporation of the upper canine into this
mechanism may in part explain those features — more vertical
implantation and more compressed form — in which it differs from
that of Coryphodon.

Turning now to the taxonomic bearing of this new form, those
characters, apart from the adaptive modifications just mentioned,
in which it compares or contrasts with Pantolamhda^ and Coryphodon
may be reviewed briefly.

Dentition. — In the upper canine and lower incisors Sparacto-
lamhda approaches closer to Coryphodon than does any other early
pantodont now known. The premolars, as indicated in the diagnosis,
display some distinctive characters of their own; the lower molars
have strong paraconids, somewhat more prominent than those of
Pantolamhda. The upper molars resemble those of the two genera
under comparison both in intramolar and in length-width propor-
tions, but seem to be aberrant in that the paracone is much more
internal in position than the metacone. At first sight this condition
might be interpreted as supporting Matthew's view of the homologies
of the coryphodont molar (1928, p. 970; see Simpson, 1929, fig. 5B,
for graphic representation). The structural series of Pantolamhda
and Coryphodon milk and permanent molars presented by Simpson
(1929, pp. 5, 6, fig. 6), however, strongly suggests that the paracone
migrated outward rather than inward in the coryphodont phylum.
For the present at least it seems preferable to regard the internal
paracone as a peculiarity of the Sparactolamhda phylum.

1 Comparisons made with P. bathmodon (Matthew, 1937, pp. 167-180, figs.
38-42, pis. 40-50). I suspect that P. caviridus may represent another genus.

Paleocene Pantodonta and Dinocerata 357

Skull. — Apart from the notably shortened premaxillaries, a
character probably correlated with the feeding mechanism, the skull
is almost ideally transitional between Pantolambda and Coryphodon,
and has on the whole a decidedly Coryphodon-like appearance. The
muzzle is rather more expanded than that of the former genus and
is about as much so as that of the latter. The arches are comparable
in slenderness to those of the Torrejon genus but are bowed out-
ward to nearly the same extent as in Coryphodon. The sagittal
crest does not extend as far forward as in Pantolambda and would
require but little widening to bring it to the condition seen in C.
wortmani (Osborn, 1898, fig. 18B, p. 194). The basicranial region
is not as compressed antero-posteriorly as that of Coryphodon,
being about as long as in Pantolambda.

Fig. 101. Sparactolambda looki gen. et sp. nov. Right lower dentition,
crown view. F.M. No. P15520, holotype. X i-

Vertebrae. — The formula cannot be determined at present. The
dorsals and lumbars agree with those of Pantolambda and Corypho-
don in having much longer centra than those of Barylambda. The
structure of the sacrum and anterior caudals of Sparactolambda
constitutes one of the most interesting and important resemblances
to Coryphodon in the entire skeleton. The sacrum in both genera
is decidedly flattened, the posterior centrum being very thin dorso-
ventrally. One of the notable features of Coryphodon is the presence
of a long pseudosacrum composed of five or more vertebrae. This
was first reported by Osborn (Osborn and Wortman, 1892, p. 120)
who found it detached, and suspected that it might have occurred
in the distal half of the tail. The true structure has only recently
been revealed by specimens in the United States National Museum
and in the Museum of Comparative Zoology (Patterson, 1939).
The pseudosacrals are exceedingly flat and are united to each other
by the neural spines and transverse processes as well as by the
centra. Two incomplete pseudosacrals are preserved in the holotype
of Sparactolambda and these, as far as preserved, are almost identical
with those of Coryphodon.

358 Field Museum of Natural History — Geology, Vol. VI

Shoulder girdle and. fore limb. — The scapulae are unfortunately
incomplete; the parts preserved indicate, however, that they were
somewhat wider than in Coryphodon and hence probably closer in pro-
portions to those of Pantolamhda; the coracoid process is extremely
long and robust. On the other hand the humerus resembles that
of the lower Eocene rather than that of the Torrejon form, especially,
as noted in the diagnosis, in the structure of the head; the anconeal
fossa is much shallower than in Coryphodon. Since an entepicondylar
foramen is now known to occur in certain specimens of Coryphodon
(Patterson, 1939), the presence of this element in Sparactolamhda

Fig. 102. Sparactolambda hoki gen. et sp. nov.
P15520, holotype. X h

Left manus. F.M. No.

does not constitute a difference between the two forms. The ulna
is close to that of Coryphodon in massiveness, in the posterior bowing
of the shaft, and in the inclination of the olecranon, but differs in
the less produced upper border of the sigmoid notch, as noted above,
and in the much less rounded lateral and medial borders of the shaft.
The radius is closer to that of Coryphodon both in shape and pro-
portions than it is to that of Pantolamhda.

The manus combines features of both genera in a most instructive
manner. The lunar is more nearly as in Pantolamhda and not as
wide as in Coryphodon.^ The scaphoid and centrale are completely

^ The various species of Coryphodon exhibit some variation in the development
of these features.

Paleocene Pantodonta and Dinocbrata 359

fused, as noted in the diagnosis, the centralar portion being nearly
as large as the separate bone in Pantolambda. Matthew long ago
suspected (1897, p. 321) that the pantodont centrale disappeared
by absorption and not by fusion. If the condition in Sparactolamhda
is a reliable criterion, it would seem that fusion preceded absorption,
which would account satisfactorily for the small, thin scaphoidal
wedge extending between the lunar and trapezoid of Coryphodon.
In the distal carpal row the trapezium > magnum > trapezoid,^
as in Coryphodon and in contrast to Pantolambda (Osborn, 1898,
p. 187). On the other hand the unciform does not cover more of
the proximal end of Mc. Ill than does that of Pantolambda,^ and
likewise the magnum does not articulate with Mc. II. The meta-
carpals are relatively a little larger than in Coryphodon but are notably
broader than in Pantolambda. The proximal phalanges are shorter
than in Pantolambda, but longer than in Coryphodon; the intermedi-
ate ones are about as long as in the former. They are robust, and
have very deep, grooved articular surfaces for the claws.

Pelvis and hind limb. — The pelvis has advanced far beyond the
condition seen in Pantolambda. The ilium has lost the trihedral rod-
like shape, and has expanded into a broad flat plate resembling, but
not as large as, that of Coryphodon. The antero-external angle has
begun to be carried backward, resulting in a rounded supra-iliac
border. The flange along the medial border described by Matthew
in Pantolambda is absent. The ischium and pubis are shortened to
almost the same extent as in Coryphodon. The proportions of the;
post-acetabular lengths, measured from a point opposite the center
of the acetabulum, to the total pelvic lengths in the three genera is
as follows: Pantolambda^ 47 per cent, Sparactolambda 41 per cent,
Coryphodon^ 38-40 per cent. Matthew (1937, p. 177) suspected
that the differences between the ilia of Pantolambda and Coryphodon
might be due largely to size, but stated, nevertheless, that the shape
of the ilium "... forms a striking difference between taligrades and
amblypods as at present known, and emphasizes the relationship
between taligrades and condylarths." I believe that the Sparacto-
lambda ilium satisfactorily disposes of this argument. The only
femur preserved is too crushed for accurate comparison; it may
be noted, however, that the third trochanter is considerably higher
than in either Pantolambda or Coryphodon. The cnemial crest of the

* I.e. in area of anterior faces, to which Osborn was presumably referring.

* See footnote, page 358.

» Measured from the published figures of Matthew, Cope, and Osborn.

360 Field Museum of Natural History — Geology, Vol. VI

tibia does not extend as far distally as in the Torrejon genus. In
common with the manus, the pes combines features of both genera.
The astragalus, the foot bone which has been most heavily
emphasized taxonomically, agrees in every essential with that
of Coryphodon. The calcaneum and navicular are closer to those of
Pantolambda, the former having a smaller fibular facet and a more
slender tuber calcis than the calcaneum of Coryphodon. The distal

Fig. 103. Sparactolambda looki gen. et sp. nov.
PI 5520, holotype. X h

Right pes. F.M. No.

tarsals are arranged as in Pantolambda and are not displaced on the
metatarsals as they are in Coryphodon. The metatarsals, on the other
hand, are much shorter than in Pantolambda.

MEASUREMENTS
{In millimeters)

P15520
A-p diam.*.. . .•.
Tr. diam.t

C

... 32.4
... 14.0

pi pa pi
12.3 11.0 11.8
8.9 23.4

M-L
14.5
22.3

M^

17.8

28.5

M^
19.5
32.4

Paleocene Pantodonta and Dinocerata 361

P15523 M^ M^ M^

A-pdiam.* 15.0 18.4

Tr. diam.t 23.3 28.2 31.6

P15520 It I^ I7 C Py P^ P5 P5 Mr M, M^

A-pdiam.* 10.0 13.9 13.8 43.0 12.8 11.0 13.3 14.7 19.0 23.3 29.7

Tr. diam. trigonidt 6.0 7.8 8.2 22.3 8.0 12.0 13.8 15.0 14.4 16.0 17.9

Tr.diam.talonidt 13.4 15.2 15.8

* Measured from mid-points of anterior and posterior faces.
t Maximum.

Skull (P15523)

Length from posterior border of canine to condyle 325

Depth of facial region at anterior border of orbit 85

Width across zygomatic arches 232

Width of occiput 123

Depth of occiput to ventral rim of foramen magnum 92

Mandible (P15520)
Length from anterior border of symphysis to posterior border of condyle . 290
Depth of ramus at anterior border of Mx 59

To sum up, knowledge of Sparactolambda enables us to unite
Pantolambda and Coryphodon much more closely than has been
possible hitherto. These three genera seem to constitute a natural
group distinguished in many ways from Barylamhda and its newly
discovered ally, Haplolamhda (diagnosed below). In view of this,
the old classification that placed BaryUxmbda and Pantolambda in
one family and Coryphodon in another no longer seems to be in accord
with the facts. As pointed out in greater detail below (pp. 369-373),
it now seems necessary to suppress the family Pantolambdidae and
to elevate the following group from subfamily to family status.

Family Barylambdidae Patterson, 1937 (new usage)
Barylambda Patterson.

Barylambda Patterson 1937, Field Mus. Nat. Hist., Geol. Ser., 6, pp. 229-231.

A detailed account of this form was nearly ready for publication
at the time the 1937 expedition left for the field, but was immediately
rendered out of date by the discoveries made in that season. It
was then decided to defer full publication until a complete account
of all the upper Paleocene Pantodonta could be prepared. To
facilitate comparisons with the newly defined genera I give here the
emended diagnosis.

I-^- Ct Pt M|. First and second upper incisors small, conical,
I^ sometimes absent; I^ large, caniniform. Canine moderately
enlarged. Upper molars fully selenodont, narrower antero-posteriorly
and wider transversely than in Coryphodontidae and M- relatively
larger. Lower incisors conical, increasing in size from ly to I7.
Lower canine medium sized. Lower premolars with paraconid

362 Field Museum of Natural History— Geology, Vol. VI

wings directed well anteriorly, especially on P^-^. My larger,
M^ smaller relative to Mtj than in Coryphodontidae; metastylid
rudimentary on My_^, developed on M^, metastylid ridges running
downward and backward well into the talonid valleys.

Skull with anterior nares terminal; muzzle not expanded; facial
region short, wide, deep; nasals very wide, greatly expanded poste-
riorly; premaxillaries weak, not sutured -medianly, ascending rami
barely appearing on side of face and not reaching nasals; skull roof
extremely wide and somewhat flattened across orbits; temporal
ridges strong; zygomatic arches but little bowed outwardly; cranium
low; occiput semicircular, condyles extending well beyond it poste-
riorly. Symphysis of mandible long, sloping; coronoid process high;
angle large.

Cervicals with exceedingly weak neural arches; dorsals and
lumbars very short in comparison with those of Coryphodontidae;
sacrals with high spines, transverse processes and ribs greatly
expanded and encroaching upon gluteal face of ilium; tail long,
massive, provided with chevrons, deeper than wide anteriorly,
spines of anterior caudals high, centra notably longer than those
of dorso-lumbars, neural arches complete on anterior half of series.

Clavicle present, well developed. Scapula almost as wide as long;
suprascapular border extensive, gently arched; acromion process
long, robust. Head of humerus large, very deep antero-posteriorly,
without anterior groove, articular surface not extending on to great
trochanter; deltoid crest very strong, wide; entepicondylar foramen
present. Manus with large separate centrale; magnum articulating
with Mc. II; phalanges shortened; unguals broad, flat, those of
digits II-IV deeply fissured.

Pelvis exceedingly broad; ilium shorter antero-posteriorly than
in Sparactolamhda or Coryphodon, antero-external angle considerably
more anterior in position than in the genera just mentioned and
bearing a plate-like process,^ ilium notched medially by gluteal
process of sacrum. Femur broad, flattened antero-posteriorly, no
pit for ligamentum teres in head. Tibia and fibula much more
massive than in Coryphodontidae. Pes with heavy tarsus, relatively
slender metatarsus and phalanges; astragalus without neck; cal-
caneum with very robust tuber calcis, fibular facet smalP or absent;

1 By a lapsv^ calami in an earlier paper (Patterson, 1935, p. 154) this process
was stated to be on "the antero-internal corner."

' Stated to have been entirely wanting (Patterson, 1935, p. 158), but present
on a specimen prepared subsequently.

c«

-I

0) g

I'i

(1)

> w

ea

c3

„ 3
(WO

CW S

^ Qi

SO-

CQTa

13

2^

363

364 Field Museum of Natural History — Geology, Vol. VI

distal tarsals not displaced on metatarsals; phalanges shorter than
in Pantolambda, longer than in Coryphodon; unguals small.

Fig. 105. Barylambda faberi Patterson. Posterior view of slceleton. F.M.
No. P14945. Epiphyses of pelvis, left lower leg and pes largely restored from
other individuals. Approximately X to-

Paleocene Pantodonta and Dinocerata 365

Barylambda faberi Patterson.^

Titanoides faberi Patterson 1933, Amer. Joum. Sci., (5), 25, pp. 415-425;

1934, Proc. Amer. Phil. Soc, 73, pp.71-101; 1935, ibid., 75, pp. 143-162.
Barylambda faberi Patterson 1937, Field Mus. Nat. Hist., Geol. Ser., 6, pp.

229-231.

The largest known pantodont;^ for preliminary description and
discussion see the papers cited.^ A left lateral view and a posterior
view of the mounted skeleton are given here. These supplement a
right lateral view given by Scott (1937, fig. 302, p. 480) and an
oblique front view published in the Field Museum Annual Report
forl937(pl. 21).

Haplolambda* gen. nov.

Type species. — H. quinni sp. nov.

Distribution. — Upper Paleocene, Plateau Valley beds, Colorado.

Diagnosis. — Generally similar to Barylambda in the known parts,
differing as follows: canines very small; M| reduced, metacone of
M- vestigial, talonids of My-^ narrower. Skull with much weaker
temporal ridges. Mandibular ramus more slender, symphysis shorter.
Shaft of radius more slender, proximal and distal ends notably wider.
Trapezium and Mc. I completely fused, the compound bone having
a double articulation with Mc. II, the additional articular surface
situated below, slightly behind and almost at a right angle to the
usual one; Mc. V wider, nearly twice as wide as the other metacarpals,
distal articular surface with median constriction.

Haplolambda quinni^ sp. nov.

Holotype. — F.M. No. P15542, anterior half of skeleton.

1 The parentheses surrounding the author's name have been omitted. See
W. H. Osgood, "An Outworn Nomenclatural Practice," Science, n.s., 89, pp.
9-11, 1939.

* I know of no Coryphodon, certainly of no mounted specimen, that exceeds
Barylambda in size.

2 1 may mention here that the isolated tibia and fibula (F.M. No. P14904)
referred to this species in 1934 (pp. 88-91, figs. 10-11) appear, in comparison with
material obtained later, to be too small for this species. Among the undetermined
specimens collected in 1937 are some jaw fragments of a pantodont similar to
Barylambda and Haplolambda, and intermediate in size between them. It is
possible that the bones in question belong to this form.

* In allusion to its less spectacular appearance when compared with Bary-
lambda and Sparactolambda, and to its tooth structure and relationship with the
other -lambda genera.

' Named for Mr. James H. Quinn, the discoverer of the holotype specimen,
without whose diligent and successful collecting the Plateau Valley faunal list
would be much smaller than it is.

X

o

Z

^
fe

V

.1

c

fa,

o

£

366

Paleocene Pantodonta and Dinocerata 367

Horizon and locality. — Plateau Valley beds, fifteen feet above
the base. One and one-half miles north of Mesa and about 75 yards
east of the DeBeque-Mesa road.

Diagnosis. — As for the genus; known parts comparable in size
to Pantolambda caviridus; for measurements see below.

The entepicondyle of Haplolambda appears to be considerably
reduced in comparison with that of Barylambda, but, since the
humeri were in fragments and the entepicondylar contacts not exact,
I have refrained from mentioning this probable additional character
in the diagnosis. The discovery of this new Paleocene genus, which
is more closely related to Barylambda than to Pantolambda and

Fig. 107. Haplolambda quinni gen. et sp. no v. Left upper cheek teeth, crown
view. F.M. No. P15542, holotype. X h

Sparactolambda, is, for pantodont phylogeny, an event second only

in importance to the finding of Sparactolambda itself. It is greatly

to be hoped that future field work will bring to light the posterior

half of the skeleton, and reveal how far this medium-sized animal

had progressed toward the acquisition of graviportal characters in

the pelvis and how close the structure of the sacrals and caudals is to

that of Barylambda.

MEASUREMENTS
{In millimeters)

pi pa pa pi M^ M^ M^

A-pdiam.* 9.7 9.8 11.6 15.9 13.8 11.9

Tr. diam.f 8.7 21.0 24.0 24.5 27.8 29.4 25.2

It I^ U C Pt Ptt P5 Mj M, M^

A-pdiam.* 6.5 8.2 10.0 8.7 13.0 16.1 15.9 17.3 18.0 18.9

Tr.diam.trigonidt 5.2 5.8 7.4 7.3 8.2 13.1 14.3 .... 15.0 12.2

Tr.diam.talonidt 11.6 10.0

* Measured from mid-points of anterior and posterior faces, f Maximum.

Skull

Length premaxillary to condyle 282

Width across zygomatic arches 152

Depth of face at anterior border of orbit 65

Width of occiput 101

Depth of occiput from apex to ventral rim of foramen magnum 52

Mandible

Length anterior border of symphysis to posterior border of condyle 222

Depth of ramus beneath anterior border of Mr 30

368

Paleocene Pantodonta and Dinocerata 369

THE TAXONOMY OF THE PANTODONTA
Until very recently it was generally supposed, largely as a result
of Osborn's work (1898), that Coryphodon was closely related to
the uintatheres, and that Pantolambda, although showing some
resemblances to Coryphodon, was a near ally of the periptychids.
This view found extreme expression in Matthew's great monograph
on the Puerco-Torrejon faunas (1937) in which, following Gregory
(1910), Pantolambda and Coryphodon were separated ordinally.
Although Matthew tended in general to favor the "horizontal"
method of classification it would appear from his text that he was
not entirely satisfied with it in this case. Thus in one place (p. 172)
he says of Pantolambda that "it is a very primitive type, and its
amblypod affinities have been unduly stressed"; in another (p. 163),
"Coryphodon should be derived from some as yet undiscovered genus
of Pantolambdidae . . ."; and again (p. 183), "Nevertheless, it [P.
cavirictus] may serve to show that a great part of the difference
between the two genera [Pantolambda and Coryphodon] is merely
due to the difference in size and does not indicate any wide diversity
of origin." The discovery of Barylambda, a graviportal form com-
bining resemblances to Pantolambda with structural approaches to
Coryphodon, occurred subsequent to Matthew's death but prior to
the appearance of his monograph. The combination of characters
exhibited by this animal indicated that Matthew's last quoted
suggestion was correct and that Pantolambda and Coryphodon were
members of a natural group — the assemblage now known as the
order Pantodonta. Recognition of this fact supported ordinal
separation of the uintatheres (a separation that I believe to be firmly
established by the new material described below), and at the same
time cast doubt on the reality of the supposed Pantolambda-
periptychid relationships (Patterson, 1934, pp. 95-97). This latter
question has been settled recently by Simpson (1937a, pp. 216-224)
who, after a thorough re-examination of all evidence, has concluded
that the Periptychidae should be replaced in the Condylarthra.
With this transfer accomplished the last of the obstructions that
interfered with a proper conception of the three orders concerned
has been removed and the way cleared toward a better understanding
of them. The scheme of classification to be presented below is about
as far removed as possible from Matthew's placing of Pantolambda
and Coryphodon in different orders. The evidence herein reported
may make it appear that he was unduly conservative but it must
be remembered that all the new material bearing on pantodont

370 Field Museum of Natural History — Geology, Vol. VI

phylogeny has only been made available during the past five years.
Pantolamhda is a generalized form so much more primitive than
Coryphodon that, lacking knowledge of the intermediates, no cautious
investigator would have been justified in connecting the two except
in the broadest way.

Prior to the discovery of the new genera here described, the
American Pantodonta^ seemed to fall naturally into two divisions —
the Coryphodontidae and the Pantolambdidae. Reference of
Barylamhda to the latter family was the only course then possible,
although it was recognized that the animal was not a descendant of
Pantolamhda. The selenodont dentition, small size of the head in
comparison with that of the skeleton, presence of a centrale, tail
size, and various other osteological features of Barylamhda were
regarded as family characters of the Pantolambdidae (Patterson,
1934, 1935).

It has been pointed out above that Haplolamhda is, in its known
parts, essentially a small edition of Barylamhda and that Sparacto-
lamhda is structurally intermediate between Coryphodon and Panto-
lamhda. The characters in which these two groups of genera
contrast with each other may now be reviewed:

Coryphodon group Barylamhda group

Dentition Dentition

Molars increasing in size posteriorly ;" M- largest, M- large, M^ compara-

upper molars wider antero-posteriorly tively small; Mx relatively larger, Mg
along ectolophs and narrower trans- relatively smaller than in the Corypho-
versely; lower molars usually without don group; upper molars wide trans-
metastylids. versely, narrow along ectolophs; meta-

stylids rudimentary on Mx-2. developed

on M75.

Skull Skull

Muzzle and arches flared; nasals Muzzle and arches not flared; nasals

comparatively narrow, moderately ex- very wide, greatly expanded posteriorly;

panded posteriorly; facial region but facial region notably deeper than

little deeper than cranial; occiput cranial; condyles extending posteriorly

extending posteriorly beyond condyles, beyond occiput.

Vertebrae Vertebrae

Dorsals and lumbars comparatively Dorsals and lumbars much shortened;
long; sacrum flattened dorso-ventrally, sacrum with high spines, transverse pro-
spines low; tail originally long but cesses and ribs forming a crescent-
reduced in later genera, without chev- shaped mass that encroaches upon
rons, anterior caudals much flattened gluteal face of ilium; tail very large,
in later genera and tending to unite provided with chevrons, deeper than
with sacrum. wide anteriorly.

^ The Mongolian Pantolambdodontidae are not discussed here.

2 A tendency rather than a hard and fast rule. Certain of the Coryphodon
specimens figured by Cope have M^ as large as, perhaps slightly larger than M^.
M{ tend to be larger in Pantolamhda (cf. Matthew's figures) than in Coryphodon;
Sparactolambda is closer to the latter in this respect.

Paleocene Pantodonta and Dinocerata 371

Manus Manus

Centrale small, becoming fused with Centrale large (possibly enlarged

scaphoid and finally reduced to a ves- from the primitive condition),
tige.

Pelvis and hind leg Pelvis and hind leg

Ilium developing from a trihedral Ilium very broad, antero-extemal

rod-like form to a broad fiat plate on angle situated somewhat farther for-

which the antero-extemal angle is ward. Femur with flatter and broader

carried well backward. shaft, tibia and fibula shorter and

stouter than in the Coryphodon group.

I believe that the two generic groups just contrasted may be
regarded as natural assemblages. It would seem therefore that the
division of the American Pantodonta into Coryphodontidae and
Pantolambdidae was a horizontal arrangement, with the latter
family containing members of divergent stocks, rather than a
natural classification. In this light a re-appraisal of the resem-
blances between Pantolambda and Barylamhda strongly suggests
that the points of similarity are due to common retention of heritage
characters (probably ordinal) and do not necessarily indicate close
relationship. In fact the differences between the two genera now
appear to be more important than the resemblances, and no good
reason remains for referring them to the same family. The validity
of the Pantolambdidae depends therefore upon whether it can
properly be retained for a part of the Coryphodon group. Sparacto-
lamhda seems to me to prevent any such arrangement. Combining
as it does the molar structure of Pantolambda with many skeletal
characters hitherto regarded as typical of Coryphodon, it renders
family separation of the extremes of the Coryphodon group impracti-
cable if not impossible. Attempts to force Sparactolambda into
narrowly redefined Pantolambdidae or Coryphodontidae result in
a manifestly artificial and thoroughly unsatisfactory arrangement.
The new evidence seems to require that the Pantolambdidae be
suppressed, that the Coryphodontidae be broadly redefined so as to
include Pantolambda and Sparactolambda, and that the Barylambda
group be given family status.

The families may be defined as follows:

Coryphodontidae ;! Medium-sized to large animals. Molars tending to
increase in size posteriorly; earlier genera with fully selenodont upper
molars, later genera with crescents considerably modified; lower molars
usually without metastylids; canines becoming tusk-like in certain genera.

"^The Corjrphodontidae are a more inclusive group than the Barylambdidae
and it is conceivable that future discoveries may necessitate the erection of sub-
families. Pantolambda may eventually be separable as a persistent primitive and
Sparactolambda as an aberrant adaptive type, but such a course would be decidedly
premature at present.

372 Field Museum of Natural History — Geology, Vol. VI

Head increasing in size relative to size of skeleton; muzzle flaring; nasals
of moderate width; skull roof broad and flat posteriorly in late forms.
Dorsal and lumbar vertebrae comparatively long; sacrum with low spines,
becoming flattened; tail originally long but undergoing progressive reduc-
tion, without chevrons, anterior caudals flattened in later genera and
tending to fuse with sacrum. Centrale small in early forms, becoming
fused with scaphoid and largely absorbed in later. Ilium developing
from a trihedral rod-like structure to a broad flat plate with the antero-
external angle carried well backward in the terminal genera.

Barylambdidae: Medium-sized to large animals. Mi larger, Mf smaller
relative to size of Mf than in Coryphodontidae; upper molars remaining
fully selenodont; lower molars developing metastylids; canines small to
moderately enlarged. Head remaining small relative to size of skeleton;
muzzle not flared; nasals very wide, greatly expanded posteriorly. Dorsal
and lumbar vertebrae short; sacrum with high spines, transverse processes
and ribs forming a crescent-shaped mass that encroaches upon gluteal
face of ilium; tail very large, with chevrons, deeper than wide anteriorly.
Centrale large, remaining separate. Ilium very broad, antero-external
angle not carried as far backward as in later coryphodontids. Tibia and
fibula short, stout.

Recognition of these two divisions makes the position of Titanoides
a Httle more certain. The lower molars agree with those of the
Coryphodon group in their progressive increase in size and in the
absence of metastylids. On the basis of these characters (the only
ones definitely known) the genus falls into the Coryphodontidae
as here redefined, thus confirming Simpson's (1937b, p. 14) assign-
ment (to the "Pantolambdinae") and rejecting my tentative refer-
ence to the "Barylambdinae" (1937, p. 230). This is in accord with
the course followed by Jepsen (1930, pp. 506, 508) who placed
Titanoides in the Coryphodontidae,^ and stated that it ". . . seems
to be a pre-Coryphodon stage of amblypod evolution and may be
in the direct ancestral line of Coryphodon . . . ." It is very possible
that this suggestion will prove to be correct. If, as seems conceivable,
Titanoides resembled Sparactolambda in the possession of coryphodont
skeletal characters it may well have been near the direct line. The
normal lower canine (Gidley, 1917, p. 431) shows at least that the
genus did not pursue root-pulling evolutionary bypaths. Scott
(1937, p. 479) has recently erected the family Titanoideidae for
Barylambda and Titanoides. On present evidence, however, it
appears that the former genus cannot be included in such a group,
and that the latter alone is not entitled to familial distinction.

1 The soundness of this reference was inferentially questioned by my earlier
mistaken assignment of Barylambda faberi to Titanoides, an inference that I am
glad to be in a position to correct.

Paleocene Pantodonta and Dinocerata 373

From the new discoveries it is becoming apparent that the
pantodonts were the dominant large ungulates during part at least
of Paleocene time. In common with later dominant groups they
seem to have undergone a considerable adaptive radiation, of which
we probably know only a small part as yet. Knowledge of the
Pantodonta has more than doubled during the present decade and
in this increase the Plateau Valley has played a major part. It
is therefore a great pleasure to have been able to name the im-
portant forms obtained from this horizon in honor of the three men
whose discoveries have so notably enlarged our conception of the
order.

One further point remains for consideration. Matthew (1937,
pp. 301, 303) reported four major types of foot structure among
Paleocene mammals, two of which — the creodont-condylarth and
the "taligrade" — are affected by the evidence reported here. The
former was regarded by Matthew (pp. 106-107, 303) as ancestral
to the Carnivora, Artiodactyla, Perissodactyla and Notoungulata,
the latter as ancestral to the "Amblypoda," Arsinoitheria, Pyro-
theria, Astrapotheria and Hyracoidea. Transfer of the Peripty-
chidae to the Condylarthra, however, has destroyed "Taligrada"
in the inclusive sense, while Sparactolambda appears to have removed
all possibility of revival of the group in the original sense. Since
the Pantodonta (at least as now known) can hardly be regarded
as ancestral to any other order, it follows, broadly accepting
Matthew's view, that the entire ungulate ancestry is to be sought
in the creodont-condylarth complex. This does not affect, other
than perhaps to render somewhat more compact, Matthew's tenta-
tive conception of an early placental dichotomy into "proto-creodont"
and "proto-insectivore" divisions (1937, fig. 57, p. 211; fig. 83, p. 302),
an important generalization that seems to be in accord with present
evidence.

Order Dinocerata Marsh, 1872

Family Uintatheriidae Flower, 1876

Bathyopsoides gen. nov.

Type species. — B. harrisorum sp. nov.

Distribution. — Upper Paleocene, Plateau Valley beds, Colorado.

Diagnosis. — 1| C^ P| M|. Metacone on P^ more distinct than
in Probathyopsis; M^ notably smaller than P^ and M^; cingula
less extensive than in Probathyopsis, more extensive than in Pro-
dinoceras. Lower incisors conical, with long crowns and external

374 Field Museum of Natural History — Geology, Vol. VI

(posterior) heels; lower canine considerably larger than incisors,
flattened transversely; molar metastylids long antero-posteriorly;
Mx relatively smaller, M^ relatively larger than in Bathyopsis;
posterior talonid crests of My-^ compressed antero-posteriorly,
entoconids subordinated in them; M^ with small entoconid. Skull
without indication of parietal horns; sagittal crest extending forward
to plane of glenoid cavities, occiput consequently triangular and
cranial table less extensive than in Bathyopsis. Diastemata between
canines and P| about as in Bathyopsis, much longer than in Proha-
thyopsis or Prodinoceras. Mandibular flanges very large, gener-
ally similar to those of Bathyopsis but differing in that they meet
ventrally beneath the symphysis in a wide 11 rather than in a
narrow A.

Bathyopsoides harrisorum^ sp. nov.

"... a rather large and advanced uintathere . . ." Patterson 1936, Proc.
Geol. Soc. Amer., 1935, p. 397.

Holotype. — F.M. No. P15546, lower jaws lacking anterior end
of symphysis, coronoid processes, and condyles; dentition nearly
complete but cheek teeth poorly preserved; various skeletal frag-
ments. Found by James H. Quinn.

Paratypes. — F.M. No. P15552, partial skull, lacking rostrum,
left side of cranial table, and left zygomatic arch; cheek teeth poorly
preserved. Found by James H. Quinn. F.M. No. P15574, incom-
plete right mandible with F^-M^ and isolated canine, various
skeletal fragments. Found by Clayton A. Quinn.

Horizon and localities. — Plateau Valley beds, about 150 feet
above the base of the horizon. Holotype collected in Plateau Valley,
one mile north of the Douglas Harris ranch house. Paratypes col-
lected four miles north of Mesa and about 200 yards east of the
DeBeque-Mesa road, P15552 occurring some thirty feet above
P15574.

Diagnosis. — As for the genus. Comparable in size to Bathyopsis
fissidens. For measurements see page 378.

Discussion. — The upper canine is not preserved in the type
material. Incomplete isolated specimens of this tooth which are
surely referable to Bathyopsoides show that it was fully as large,
proportionately, as in the Bridger uintatheres, somewhat flattened
transversely, and with a ridge on the lateral face. The only notable

1 Named for the Harris family of Mesa, Colorado, to whose friendly aid and
hospitality we owe much of our success in the field.

T3 <I>
C S
« S

II

Pl, t>
. a;

$1 to

.2.3
-^ O

So.S

w

a>

1^

Em u

375

376 Field Museum of Natural History — Geology, Vol. VI

feature of the upper cheek teeth is their very close resemblance to
those of all other genera of the order. The lower incisors are strik-
ingly different from those of the Bridger forms (cf. Osborn, 1881,
pi. 3, figs. 3-4).^ Nevertheless the presence of well-developed ex-
ternal heels does foreshadow the great enlargement of these elements
in the later members of the order. It is unfortunate that these teeth
are unknown in Bathyopsis, a genus in which they might be struc-
turally intermediate between the two extremes. The lower canine
appears to have had the same degree of inclination as the incisors;
it agrees with that of Bathyopsis in being larger than the incisors.

Fig. 110. Bathyopsoides harrisorum gen. et sp. nov. Left M^j-^, crown view.
F.M. No. PI 5546, holotype. X h

The metastylid of M^ is almost as large and prominent as the
metaconid; that on M^ is very small. The compressed posterior
talonid crest and the subordinated entoconid of MY_ir are of some
interest because in these features Bathyopsoides differs from Proha-
thyopsis and from Bathyopsis^ and approaches, to judge from the
figures given by Osborn (1881) and by Marsh (1886),^ the middle
Eocene uintatheres.

The structure of the skull, except for those differences men-
tioned in the diagnosis, is very similar to that of Bathyopsis (Osborn,
1913). The preorbital constriction is not quite as prominent as in
the lower Eocene genus, and the occiput does not appear to extend
posteriorly beyond the condyles. The postglenoid processes of
P15552 do not project as far ventrally, but this condition is perhaps
accentuated by distortion. In palatal and zygomatic structure and
in breadth of condyles the two forms are essentially identical. The

1 Marsh's figures of the incisors are inaccurate, as has been pointed out by
Matthew (1937, p. 170).

'^ Amer. Mus. No. 14803, a very fragmentary specimen identified as Bathyopsis
includes an M^ similar to that of Bathyopsoides in the posterior talonid crest.
Conceivably this species represents another genus.

^ Matthew (1937, p. 170) states that this monograph bears the date 1884 on the
title page. This must have been either a misprint or a lapsus on Matthew's part
because such copies as I have seen bear the date 1886, and Marsh himself relates
(p. 237) that the author's edition was printed in February, 1885.

Paleocene Pantodonta and Dinocerata 377

basicranial region of Bathyopsoides is so poorly preserved that no
structural details can be seen but the proportions are almost exactly
those of Bathyopsis.

The mandibular flange resembles that of Bathyopsis and differs
from that of Uintatherium^ in that it slopes evenly instead of abruptly
backward and upward into the ramus. The resemblance and differ-
ence is reversed, however, in the case of the Jl-shaped ventral union
of the flanges.

The more important associated skeletal remains include two
incomplete vertebrae and the proximal end of an ulna belonging to
P15546, the holotype, and an unciform found with P15574. In
addition, an isolated femur, P15548, which is almost certainly refer-
able to Bathyopsoides, was collected by Mr. Edgerley three miles
west of DeBeque.

The vertebrae are an anterior dorsal and a lumbar. The former
is, as far as preserved, almost exactly like those of Uintatherium;
the latter has an equally large neural canal but a somewhat longer
neural arch. The olecranon preserved on the ulna fragment has
the same degree of inclination as that of Uintatherium, but is not
expanded transversely; the articular surface is narrower than in the
Bridger genus. The unciform is not as deep in the anterior face as
that of Uintatherium, and the articular area for the lunar and for
the medial half of the cuneiform is much higher. In the latter
feature Bathyopsoides approaches the contemporary pantodonts,
differing from them, however, in the lack of any pronounced demar-
cation between the surfaces for the lunar and cuneiform.

The isolated femur is very similar to that of Uintatherium in the
structure of the shaft, position of the second trochanter, and absence
of the third trochanter, the latter character sharply distinguishing
it from any pantodont femur. The head, however, is set more
obliquely to the long axis of the bone and is but little higher than
the great trochanter. The position of the head in the Bridger
uintatheres is an extreme graviportal adaptation comparable to
that occurring in the elephants, and it is therefore not surprising to
encounter a more primitive position for it in their Paleocene fore-
runners. In the structure of the distal end Bathyopsoides differs from
Uintatherium in the more upwardly extended patellar trochlea and
in the more divergent condyles.

1 Including Dinoceras, Tinoceras, etc. The synonymy of the Bridger uinta-
theres seems to be in an even worse state, if such be possible, than that of the
lower Eocene coryphodonts.

378 Field Museum of Natural History — Geology, Vol. VI

p4

Mi

M^

M^

7.2

16.4

22.6 22.5

L8.3

16.2

21.8 24.0

P15574

P15106*

16.7

11.8

10.9

16.4

11.4

9.7

21.0

20

4

15.0

15

7

12.8

13

8

MEASUREMENTS

(In millimeters)

PI 5552 P^ P^

A-pdiam 15.0 15.7

Tr. diam 11.5 14.8

P15546

Pj, a-p diam ....

P^^, tr. diam. trigonid .... ....

P^:, tr. diam. talonid ....

Mx, a-p diam ....

Mx, tr. diam. trigonid ....

Mx, tr. diam. talonid ....

M^, a-p diam 20 . 1

M^, tr. diam. trigonid 14.8

M^, tr. diam. talonid .... 13.6

Mu, a-p diam 25.0

Ms, tr. diam. trigonid .... 16.4

Ms, tr. diam. talonid 14 . 8

♦ The original specimen discovered by Mr. Faber.

Skull (P15552)

Length from posterior border of canine alveolus to posterior border of

condyle .' 316

Zygomatic width 220

Width across condyles 93

Depth from apex of occiput to ventral rim of foramen magnum 117

Mandible (PI 5546)

Length from posterior border of canine alveolus to posterior border of Ms ■ ■ 152

Width of symphysis behind canines 88

Depth of flange 91

Width across flanges, ventral 147

Femur (PI 5548)

Length 428

Proximal width 155

Distal width 126

Probathyopsis Simpson

Probathyopsis Simpson 1929, Amer. Mus. Nov., No. 387, p. 1.

Probathyopsis newbilli^ sp. nov.

Probathyopsis sp. Patterson 1936, Proc. Geol. Soc. Amer., 1935, p. 397.

Holotype. — F.M. No. P15549, incomplete lower jaws with partial
premolar-molar dentition; a young individual with M^ not yet
erupted.

Horizon and locality. — Plateau Valley beds, Plateau Valley,
Colorado, one mile north of the Douglas Harris ranch house. Found
by James H. Quinn some fifty yards away from the holotype of
Bathyopsoides harrisorum.

Diagnosis. — P:^, M-^, relatively larger than in P. praecursor,^
My relatively smaller; teeth larger and wider than those of P. prae-
cursor. For measurements see page 381.

^ Named for Mr. Thomas J. Newbill, Jr., my companion on the 1932 expedition.

* I am greatly indebted to Dr. Barnum Brown for the loan of the holotype of
this species.

Paleocene Pantodonta and Dinocerata

379

Discussion. — The cheek teeth, although of the usual, almost
stereot)T)ed uintathere pattern, exhibit several differences from
those of Bathyopsoides. The paraconid remnant is noticeably

Fig. 111. Probathyopsis newbilli sp. nov. Incomplete lower jaws and
dentition, crown view. F.M. No. P15549, holotype. X h

stronger on P^-M^ of P. newhilli, the talonid shelf on F^ is more
prominent and farther removed from the hypoconid, and the ento-
conid of Mij is very distinct. In addition, the metastylid of Mt^ is
smaller and that of M^ larger than in B. harrisorum. In contrast
to Bathyopsoides, Pj is present and there is little or no diastema

380 Field Museum of Natural History— Geology, Vol. VI

between it and the canine. The symphysial fragment indicates that
a small flange, comparable to that of P. praecursor, was present.^

In view of the fact that the tooth proportions of P. newbilli, its
diagnostic characters, are very similar to those of B. harrisorum,
the possibility that the former might be a female of the latter was
considered. However, the structural differences, all shared with P.
praecursor, can hardly be dismissed as ,sexual. Six specimens of
M^, three each of B. harrisorum and P. newbilli, show no inter-
gradation in entoconid and metastylid structure. It seems necessary,
therefore, to regard the characters of P. newbilli as having taxonomic
significance.

Four fragmentary specimens may be referred to the species under
discussion. These are a right M.-^, P15584, the trigonid of a right
M^, P15577, an incomplete left lower canine, P15578, and a
right Mi, P14939.2

The lower canine lacks the apex and part of the external surface.
It is considerably larger than the corresponding tooth of P. prae-
cursor and has a rather flatter lateral face, but agrees in the presence
of a posterior basal tubercle and of a prominent antero-internal
cingular ridge. In the last two features both canines differ from those
of jB. harrisorum.

Mi differs from that of P. praecursor in slightly smaller size
(significant in view of the small Mx of P. newbilli), stronger and
higher parastyle, metaloph and protoloph more nearly parallel and
slightly closer together, metaloph tending to run postero-internally,
and cingulum not encircling the protocone. From Prodinoceras it
differs in much smaller size, weak external cingulum, and in the
more posterior and less internal position of the hypocone. The
Bathyopsoides material does not contain an M^ sufficiently well
preserved for comparison.

^ Simpson's figure (1929, fig. 3) shows the flange as being rather more promi-
nent than it actually is; the ventral extremity merges more gradually backward
into the ramus than is shown. The total depth is 41 mm.

^ I referred to this specimen in 1936 (pp. 397-398) in stating that "Proba-
thyopsis sp. was found some 200 feet higher in the formation." In the 1932 and
1933 seasons all the material collected, except for this specimen, a fragmentary
Bathyopsoides upper canine (not then recognized as such), and an incomplete
chelonian, came from the lower 50-75 feet of the series. In 1937, however, it was
found that the fossils had a vertical distribution of some 300 feet. There is no
evidence as yet that more than one faunal unit is represented by the material
collected. Of the P. newbilli specimens the holotype, PI 5584, and P14939 were
found in the upper third of the fossiliferous zone, PI 5577 near the base, and
PI 5578 about midway between the two other occurrences.

Paleocene Pantodonta and Dinocerata

381

M-*-, a-p diam

M^, tr. diam

Lower C, a-p diam. . . .

Lower C, tr. diam

P5, a-p diam

P5, tr. diam. trigonid .
P5, tr. diam. talonid . .

Mx, a-p diam

Mt, tr. diam. trigonid
My, tr. diam. talonid.

M5, a-p diam

M^, tr. diam. trigonid
M^, tr. diam. talonid.

Ms, a-p diam

Ms, tr. diam. trigonid
M^, tr. diam. talonid .

MEASUREMENTS
(In millimeters)

P. newbilli

P14939
13.0
11

Holotype
P15578 P15549

16/

r

is.

11.

9.

13.

is!

13.

10.

P15584 P15577

19.2
13.5
10.8

P. praecursor

Holotype
A.M. 16786

13.3
8.5

13.4
9.2

8.7

11.9

9.7
21.0

i2!2

THE ORDINAL STATUS OF THE DINOCERATA

Since the appearance of Wood's paper (1923) pointing out the
fundamental differences between the uintathere and corjrphodont
molars, the view that the two groups represent distinct orders has
been steadily gaining ground (Simpson 1929, 1931; Patterson 1934;
Simpson 1937a). Matthew in his Puerco-Torrejon monograph (1937)
retained the old classification, but, as has been pointed out above, a
good part of the crucial evidence now available was unknown to
him. Scott (1937, pp. 477-478) admitted that ordinal separation of
uintatheres and pantodonts might be found to be desirable in the
future, but stated that "... it must ... be recognized that the
families now included in the fAmblypoda [Uintatheriidae, Cory-
phodontidae, Pantolambdidae, etc.] are more nearly related to one
another than they, or any of them, are to other groups." The bear-
ing of the Plateau Valley material on this question must be briefly
considered.

The Plateau Valley appears to be somewhat older than the Clark
Fork, and it follows therefore that in Bathyopsoides harrisorum and
Prohathyopsis newbilli we are dealing with the earliest uintatheres
thus far found in North America, if not in the world. ^ In the case of
P. newbilli a phylum of the family already known from the Gray
Bull and Clark Fork is carried back a stage further. The new
species is fully as advanced as the Clark Fork P. prdecursor, per-

1 The Gashato of Mongolia which contains Prodinoceras cannot be exactly
correlated, but is almost certainly upper Paleocene.

382 Field Museum of Natural History— Geology, Vol. VI

haps slightly more so, and is probably not directly ancestral. This
is significant as to the stability of the phylum, but is not particularly
startling.

The discovery of Bathyopsoides, however, is an event of first-
rate importance in the furthering of our knowledge of uintathere
evolution. In this animal we have a form which in the upper cheek
teeth, in the upper canine, and in the development of diastemata
is comparable to the terminal North American members of the
group; a form as large as and more specialized in flange structure
than Bathyopsis from the upper part of the lower Eocene. The
known parts of its skeleton are clearly of uintatherid type. So
advanced a form strongly suggests a long period of independent
development for the group prior to upper Paleocene time. Neither
in Bathyopsoides nor in Probathyopsis can I detect any significant
approach either to the pantodonts or to the periptychid condylarths.
The Pantodonta do not seem to be closer to the Dinocerata than to
other groups such as the Creodonta and Condylarthra. The essen-
tially static cheek teeth of the known uintatheres (upper Paleocene
to upper Eocene) probably indicate that their characteristic pattern,
so different from that of the pantodonts, was acquired long before
the first appearance of the order in the record. For these various
reasons I feel compelled to regard the ordinal separation of the
Dinocerata as valid and necessary.

The structural succession Prohathyopsis-Bathyopsis-Elachoceras-
Uintatherium-Eohasileus admirably illustrates the trend of evolu-
tion in the later uintatheres. This succession has been regarded
as phyletic, but Bathyopsoides suggests that such a view may not be
entirely correct. Probathyopsis may well have been ancestral to
Bathyopsis but I am by no means certain that the latter is in turn
ancestral to the later members of the order. In at least one striking
character — flange structure — Bathyopsoides is closer to Uintatherium
than is Bathyopsis. Although we have no evidence at present that
the Dinocerata had as complex a radiation as the Pantodonta, it is
nevertheless likely that future discoveries will reveal that the genera
we know now are inadequate in number for a proper understanding
of the phylogeny.

REFERENCES

BOHLIN, B.

1936. Notes on Some Remains of Fossil Mammals from China and Mongolia.
Bull. Geol. Soc. China, 15, pp. 321-330, pi. 1.

Cope, E. D.
1875. Systematic Catalogue of Vertebrata of the Eocene of New Mexico,
Collected in 1874. Report to the Engineering Dept. U. S. Army; Washing-
ton, Government Printing Office, pp. 1-37, 1 fig.

1884. The Vertebrata of the Tertiary Formations of the West. Book I. U. S.
Geol. Surv. Terr., 3, pp. l-xxxiv, 1-1009, figs. 1-38, pis. l-75a.

GiDLEY, J. W.

1917. Notice of a New Paleocene Mammal, a Possible Relative of the Titano-
theres. Proc. U. S. Nat. Mus., 52, pp. 431-435, fig. 1, pi. 36.

Gregory, W. K.

1910. The Orders of Mammals. Bull. Amer. Mus. Nat. Hist., 27, pp. 1-524,
figs. 1-32.

Jepsen, G. L.

1930. Stratigraphy and Paleontology of the Paleocene of Northeastern Park
County, Wyoming. Proc. Amer. Phil. Soc, 69, pp. 463-528, figs. 1-4, pis.
1-10.

Marsh, O. C.

1886. Dinocerata. A Monograph of an Extinct Order of Gigantic Mammals.
Monog. U. S. Geol. Surv., 1(H pp. l-xvill, 1-243, figs. 1-200, pis. 1-56.

Matthew, W. D.

1897. A Revision of the Puerco Fauna. Bull. Amer, Mus. Nat. Hist., 9, pp.
259-323, figs. 1-20.

1928. The Evolution of the Mammals in the Eocene. Proc. Zool. Soc. Lond.,
1927, pp. 947-985, figs. 1-16.

1937. Paleocene Faunas of the San Juan Basin, New Mexico. Trans. Amer.
Phil. Soc, n.s., 30, pp. I-Vlii, 1-510, figs. 1-85, pis. 1-65.

OSBORN, H. F.

1881. A Memoir upon Loxolophodon and Uintntherium, Two Genera of the
Sub-order Dinocerata. Contr. E. M. Mus. Geol. Arch. Coll. New Jersey
[Princeton], 1, pp. 1-44, pis. 1-4.

1898. Evolution of the Amblypoda. Part I. Taligrada and Pantodonta. Bull.
Amer. Mus. Nat. Hist., 10, pp. 169-218, figs. 1-29.

1913. The Skull of Bathyopsis, Wind River Uintathere. Ibid., 32, pp. 417-420,
figs. 1-4, pis. 64-65.

OsBORN, H. F. and Wortman, J. L.

1892. Fossil Mammals of the Wahsatch and Wind River Beds. Collection of
1891. Bull. Amer. Mus. Nat. Hist., 4, pp. 81-147, figs. 1-19, pi. 4.

Patterson, B.

1933. A New Species of the Amblypod Titanoides from Western Colorado.
Amer. Jour. Sci., (5), 25, pp. 415-425, figs. 1-4.

1934. A Contribution to the Osteology of Titanoides and the Relationships of
the Amblypoda. Proc. Amer. Phil. Soc, 73, pp. 71-101, figs. 1-13, pis. 1-2.

1935. Second Contribution to the Osteology and Affinities of the Paleocene
Amblypod Titanoides. Ibid., 75, pp. 143-162, figs. 1-6.

383

384 Field Museum of Natural History — Geology, Vol. VI

1936. Mounted Skeleton of Titanoides with Notes on the Associated Fauna.
Proc. Geol. Soc. Amer., 1935, pp. 397-398.

1937. A New Genus, Barylambda, for Titanoides faberi, Faleocene Amblypod.
Field Mus. Nat. Hist., Geol. Sen, 6, pp. 229-231.

1939. A Skeleton of Coryphodon (in press).

Scott, W. B.

1937. A History of Land Mammals in the Western Hemisphere. Revised ed.
pp. l-xiv, 1-786, frontispiece, figs. 1-419. The Macmillan Co., New York.

Simpson, G. G.

1929. A New Paleocene Uintathere and Molar Evolution in the Amblypoda.
Amer. Mus. Nov., No. 387, pp. 1-9, figs. 1-9.

1931. A New Classification of Mammals. Bull. Amer. Mus. Nat. Hist., 59,
pp. 259-293.

1937a. The Fort Union of the Crazy Mountain Field, Montana, and Its Mam-
malian Faunas. Bull. U. S. Nat. Mus., 169, pp. l-x, 1-287, figs. 1-80, pis.
1-10.

1937b. Additions to the Upper Paleocene Fauna of the Crazy Mountain Field.
Amer. Mus. Nov., No. 940, pp. 1-15, figs. 1-4.

Wood, H. E.

1923. The Problem of the Uintatherium Molars. Bull. Amer. Mus. Nat.
Hist., 48, pp. 599-604, figs. 1-4.

i