Arretotherium, to be described later, and part of the skull of a quite large rhinoceros.
12 DOUGLASS FOSSIL MAMMALIA
The mandible is somewhat smaller than that of S. pansus of the Loup Fork of New
Mexico. It is less robust, not being so deep or thick. The masseteric area does not extend
so far forward. There is some difference in the enamel pattern of the teeth, but just how
much is due to wear I cannot say. That of the premolar in the present species is more
complex, being crenulate or minutely folded on the inner lakes and inflections. This
tooth is smaller especially at the top, but much wear would bring this surface nearer to
the area of that of S. pansus. The two anterior molars are broader and longer. The
teeth do not appear so high from the outside, but are higher above the alveolar border on
the inside. The outer enamel inflections do not extend so far down on the outside of the
teeth, are more open and incline more forward and the outer lobes are more angulate.
There are small extra enamel islets en M T and M-J-. There is no sharp antero-internal
angle on the premolar. These comparisons are made with No. 10575 of the Princeton
collection labeled Castor pansus, Camp creek, Oregon, which evidently does not differ
at all from Cope's type from New Mexico (1877, p. 297, PI. LXIX, Figs. 4 and 5).
In none of the teeth are the principal inner and outer enamel inflections converted
At its preserved stage of wear the grinding surface of P T has three divisions, an
anterior outer, an anterior inner and a posterior. The anterior outer division is three-
lobed, or is a crescent with a long limb extending inward and somewhat backward from
near the posterior horn. The anterior horn is at the antero-external angle of the tooth.
From this the crescent sweeps backward and inward, the outer border forming the outer
margin of this part of the premolar. The inner limb extends transversely inward to the
inner border. The anterior inner crescent of the tooth begins near the anterior border of
the preceding one, the outer margin forming the anterior and anterior inner border of the
tooth. On the posterior part of this crescent near the antero-external horn is a rounded
lobe. Much wear would make the two anterior crescents confluent. The posterior part
of the tooth is a transverse, pointed oval with the smaller end outward and slightly
deflected forward. Its enamel lake has nearly the same form but with sinuous margin
Much wear would perhaps make the enamel pattern nearer like that of S. pansus.
The enamel folds and lakes are inclined to be more curved and the enamel borders
more sinuous than in S. pansus. There is an extra islet in the two anterior molars. In
M T it is posterior to the anterior lake and in M^ it is anterior.
The anterior surface of the incisors is more convex than in the beaver, Castor
densis, and the antero-posterior diameter is proportionally greater.
Length of molar-premolar series 017
Length of Pj at base QQJ
OF THE WHITE RIVER BEDS OF MONTANA. 13
Width of P at basr 0045
Length of M T 004
Width of M T " 0045
Length of M z 004
Width of Mj 0042
Length of M j 004
Width of Mj 0036
Depth of ramus under middle of Pj 012
Depth of ramus under posterior of M T Oil
Steneofiber complexus, sp. nov.
The type of this species is part of a skull and mandible with complete dentition (No.
42). It was found in a gray sandy layer in a ravine cutting through high bluffs of sup-
posed White River age on the west side of the Madison river, in the Lower Madison
valley, about nine or ten miles south of Three Forks. As this specimen and some limb
bones of a rhinoceros-like animal are the only fossils of importance obtained here, and
as the exact relations with the lower White River fossil-bearing beds northwest of Three
Forks, about fourteen miles distant, have not been made out, it is impossible to say to just
what horizon these beds belong.
This is a young individual. Only one premolar the right upper one has been shed.
All of the molars are fully protruded and considerably worn. The upper permanent
premolars are farther advanced than the lower, which evidently had only begun to grow.
This species seems to be nearest like Cope's S. (Castor) penlnsulatus (1883, p. 840,
PI. LXIII, Figs. 18-21). The skull is damaged, yet some points of interest can be made
out. There is a marked difference in the mandibles of the two species. In the present
one the masseteric area does not extend so far forward, the anterior margin of the coro-
noid process rises opposite the back part of the third cheek tooth (M T ), and is nearer to
the molars. This process has an entirely different form in the present species. It is
high ; the anterior border is straight and rises steeply. The angle is inflected inward
and is rounded, not angulate anteriorly and posteriorly as in S. viciacensis (see Filhol,
1891, PL 5).
The Dentition. The anterior face of the lower incisor is not so convex as in S. hes-
IKTUS just described. The posterior angle is not so acute but is broadly rounded. The
outer part of the cutting edge is rounded, not angulate, and the inner worn, beveled
surface is long and longitudinally concave. Below this the transverse is just a trifle
greater than the antero-posterior diameter.
The upper cheek teeth are very complex on account of the number and the curving
of the enamel lakes and inflections and the sinuosity of the enamel. This might be simpli-
fied on further wear. The inner loops extend forward as much as outward, and some of
14 DOUGLASS FOSSIL MAMMALIA
the lakes are nearly antero-posterior. In the lower teeth also the enamel of the lakes
and inner enamel inflections is much lobed.
I will defer a detailed description of these teeth for a paper which has been partly
written, describing several new species of Mylagaulida? and throwing considerable light
on their development.
From anterior of incisor to posterior of MA
Prom posterior of incisor to anterior of P 1
Transverse diameter of incisor
Antero-posterior diameter of incisor
Length of molar-premolar series . .
Length of temporary P* 0033
Width of temporary P 0038
Length of Mi 0032
Width of MI 0035
Length of M* 0032
Width of M* -0034
Length of MA 0082
Width of MA 0032
Anterior tip of incisor to posterior of Mj 033
Length of molar-premolar series 0146
Length of P ? 0038
Width of P, 003
Length of M T 003
Width of M T -004
Length of M j -0032
Width of M, 004
LengthofM s -0032
Width of M, -0033
In the Pipestone beds quite a number of jaws and teeth of Palwolagus were found.
They apparently belong to three species. One has the third column on the lower
molars as in P. triplex. The three portions of mandibles preserved are somewhat smaller
than the one described by Cope (1883, p. 881, PL LXVII, Fig. 28), and may belong
to a different species.
Several upper and lower jaws I cannot distinguish from P. turgidus.
Palaeolagus temnodon, sp. nov.
There are three series of upper teeth (Nos. 43, 44 and 45) which differ in some
respects from previously described species. They vary somewhat in size. P^ the first
OF THE WHITE RIVER BEDS OF MONTANA. 15
cheek tooth has two anterior grooves instead of one. It lops backward, its upper pos-
terior part resting against P 4 . On M 3 of the largest one (No. 42) there is a rudiment of
a posterior lobe. From Pipestone creek.
No. 43. No. 44.
Length of molar-premolar series 0123 .0141
Length of molar series 0051 .0061
Width of MA 0004 .0045
Cylindrodon fontis, gen. et sp. nov. (Plate IX, Figs. 9, 9a.)
Of this rodent I have two portions of right mandibular rami (Nos. 38 and 39). One
has all the cheek teeth and the greater part anterior to the ascending ramus ; the other
has the three posterior teeth and part of the ascending ramus, but not the angle, coronoid
process or condyle. They were obtained from the Pipestone beds, associated with Palceo-
lagus, Ischyromis, etc. The Pipestone springs, near where the fossils were found, sug-
gested the specific name.
The diastema was short ; the teeth are cylindrical with a central enamel islet and
an outer enamel inflection. The lower border of the ramus is convex fore-and-aft, and
is shallow but thick under M^. The anterior angle of the masseteric area extends for-
ward to M T and is near the alveolar border. The anterior border of the ascending
ramus rises opposite JVtg- and leaves a broad space between the two. The surface of the
masseteric area is nearly flat, with a convexity considerably behind and above its ante-
rior angle. On the inside, back of the teeth, from the horizontal ramus a large con-
vexity passes backward, outward and upward on the ascending ramus. Above and below
this the bone is thin. The mental foramen is small and is situated above the middle of
the jaw, a little in advance of P T .
The teeth are very characteristic, yet they are simple. As seen from above they
look like a large comma, with a short tail directed outward. Their order of size, begin-
ning with the smallest, is M-j, P T , then M T and M 7 , which are nearly equal in size. The
teeth are evidently quite long vertically, except the last which is short, as it is so nearly
approached by the posterior portion of the canine. The incisors are thicker antero-
posteriorly than transversely.
Length of molar-premolar series 0076
Depth of ramus under P ? 0064
Depth of ramus just back of M T 0046
Thickness of ramus under P T 003
16 DOUGLASS FOSSIL MAMMALIA
Thickness of ramus at M r 004
Length of P T 003
Width of P 0021
Length of M T 0021
Width of M r 0022
Length of M, 0021
Width of Mj 0022
Length of M 3 0014
Width of M T 0015
Sciurus jeffersoni, sp. nov.
Type No. 40.
This species is larger than 8. relictus (Cope, 1883, p. 817, PL LXV, Fig. 35).
The teeth do not increase regularly in size backward. The mental foramen is higher
and nearer to the incisor. The diastema was shorter.
In this species the anterior cheek tooth is the smallest and the posterior one the
largest, but the two between are nearly equal in size. In all of the teeth the anterior
inner tubercle is the higher and larger and the posterior inner one the smaller. The
latter does not appear as a separate tubercle on M.-g-, as a continuous wall extends from the
anterior inner to the posterior outer one, enclosing the basin internally and posteriorly.
A small ridge, which may represent the posterior inner tubercle, extends from the wall
backward and inward in the posterior part of the basin. On the anterior three teeth
there are minute tubercles between the two internal ones. There are also minute median
tubercles on the outer margins of all the teeth.
From White River beds, Pipestone creek.
Length of lower molar-premolar series 012
Length of P, 0026
Width of Pj 0029
Width of molars, each 003
Length of M T 0029
Length of M T 003
Length of M 7 004
Eumys minor, sp. nov.
The type of this species, No. 37, is part of a right ramus of a mandible from the
Pipestone beds on Pipestone creek, in Jefferson county. The anterior part of the incisor
is gone and all of the mandible back of the second molar. The two anterior molars are
OF THE WHITE RIVER BEDS OF MONTANA. 17
Tliis species is much smaller than E. elegans Leidy. The mental foramen is near
the upper border of the diastema and apparently much in advance of the middle. The
anterior angle of the masseteric area is under the anterior molar. Both teeth are some-
what worn on their grinding surfaces. Their antero-posterior diameters are nearly
eijual, hut the second is broader transversely than the first. The anterior part of the
lirst molar (cheek tooth) is not plainly divided into two lobes, but is subconical with n
pit in the middle of the truncated apex. Its transverse is greater than its antefo-
]>nsterior diameter. It is not much worn. It is smaller than the posterior portion of the
tooth, in which there is a posterior outer tubercle, from which three slender lobes extend
inward. The anterior lobe is the longest and the posterior the shortest. The second
tooth has the two rounded outer tubercles and five slender inner lobes. It looks as if
each outer tubercle sent in three inner lobes, but the contiguous two unite to form one
median one. On looking at the grinding surface, it is not easy to make out the four
principal lobes as it is in Leidy's figure of a tooth of E. elegans (1869, PI. XXVI,
Fig. 13). On examining the tooth from the inside it is seen that, beginning anteriorly,
the second lobe forms the large anterior inner tubercle and the fourth the posterior one.
The first and fifth are accessory, and the third is in the median valley between the two
tubercles. It seems that a little further wear would unite the two anterior lobes.
Depth of jaw at mental foramen 0031
Depth under second molar 0039
Thickness under second molar 0025
Length of first molar 0015
Width of first molar 0013
Length of second molar 0015
Width of second molar 0017
Hycenodon montanus, sp. nov.
Type No. 46.
A portion of a skull and mandible with all the teeth represented except the lower
incisors. Found northeast of Toston, a village on the Missouri river, southeast of
Helena. This was the only fossil secured from this stratum ; but teeth of two species
of Colodon were found in lower strata. The skulls of Oreodon robustum, 0. culberi-
soni (?), and Eucrotaphus hclence were found farther to the northeast, but I am not sure
whether the beds are higher or lower.
Larger than H. crucians; crown of upper canine nearly straight; crown of P 2 high
and without posterior cusp; talons on P^ and P , but not on P 7 .
18 DOUGLASS FOSSIL MAMMALIA
The canine is long and slender but not curved backward toward the apex. P^ has
a high crown as in H. cruentus and H. horridus. It has no heel and is therefore nearly
like that of the former species. The anterior border is only slightly convex longitu-
dinally. The posterior is concave as in that species. There is a small tritocone on P^,
and the rudiment of a deuterocone. supported by an inner root. In P^ the tritocone,
the deuterocone and its supporting root are all larger. There is no antero-external basal
The lower canine curves backward. P T is small and low, P-j bends backward. It
is higher than P^. P T is much larger and higher than P^.
Length of upper series of teeth, exclusive of incisors 094
Length of upper molar-premolar series 083
Length of upper premolar series 053
Length of upper canine 012
Width of upper canine : 0083
Height of crown of upper canine 026
Length of Pi 009
Height of crown of Pi 007
Length of P* 0125
Height of P* 014
Length of P* 0135
Height of PA 010
Length of P 013
Height of P* Oil
Length of Mi 013
Height of Ml 010
Length of M* 017
Height of M* '. .011
Length of lower series of teeth, exclusive of incisors 092
Length of lower molar-premolar series 090
Length of lower premolar series 052
Length of lower molar series 039
Length of lower canine Oil
Width of lower canine 0095
Length of P T 008
Height of P! 004
Length of P z Oil
Height of Py 010
Length of Pj 0125
Height of P T .011
Length of Pj . .0133
OF THE WHITE RIVER BEDS OF MONTANA. 19
Height of P, ... 013
Length of M r 0095
Length of M 7 0105
Height of M j . .010
Length of M y 017
Height of skull above M- 1 - alum I 062
Hycenodon ininutus, sp. nov.
Type No. 47.
This consists of only a lower sectorial tooth, but it is so minute that it cannot be
confused with any known species, unless it be H. mustelinus, which is considerably
larger. It is the second molar of the right side. It shows considerable lateral wear on
the outside, showing that it was not a young animal. At the top of the cusp are two
beveled, worn surfaces, where the white dentine shows through the black enamel. The
dentine also is exposed on the anterior surface near the root. The tooth is thick in
proportion to its length. There are minute ridges on the anterior angles and a faint
median convexity. There is also a minute ridge in the posterior median surface.
Probably the most characteristic thing is the size.
Antero-posterior diameter 0075
Transverse diameter 0049
From Pipestone beds near Whitehall, Jefferson county, Montana.
Colodon cingulatus, sp. nov.
Type No. 62.
Part of maxillary with PI, P*, Ml and base of P.
A prominent cingulum entirely surrounds the last two premolars. The internal
cusps the deuterocone and tetartocone are just beginning to divide, as indicated by a
shallow furrow on the inner side. The protostyle, protocone and tritocone are all con-
vex on the outside. The protocone and tritocone are nearly equal in size, but the pro-
tostyle is not so large or high. The metaloph is higher than the protoloph, the latter
nearly dying out before it reaches the protocone. In other words, it ascends steeply as it
passes outward toward the protocone.
In M 1 - the paracone, parastyle, metacone and hypocone all appear as subconical cusps.
The parastyle is much smaller than the paracone, and is situated in front of it and does not
send backward a prominent cingulum on the outer face of the paracone as in C. proem-
20 DOUGLASS FOSSIL MAMMALIA
pidatus. The distance from the apex of the paracone to the apex of the metacone is
the same distance as from the latter to the apex of the hypocone.
There are at least two infraorbital foramina. They open in a concavity in the face
above the third and fourth premolars.
From Toston beds near the Missouri river, above Townsend.
Length of last three premolars and first molar 0407
Length of P* at base '. 009
Width of Pa at hase 012
Length of P^ 0115
Width of Pi 0173
Length of P 012
Width of PA 0182
Length of Ml 0145
Height of infraorbital foramina above alveolar border 018
Distance between the two iufraorbital foramina exposed 0065
A last upper molar in a fragment of the maxillary.
This is larger than C. dakotensis. The tooth is considerably worn. The parastyle
is as large as the paracone and is more convex on the outer surface. The metacone is
small. Prom Toston beds.
Length of M* 020
Width of MA 022
Bathygenys, gen. nov. (Plate IX, Figs. 7 and 8.)
Type Nos. 48 and 66.
Among the remains of small animals that were found in the Pipestone beds are two
portions of mandibles that are of much interest, as, disregarding size, they are so very much
like the corresponding parts of some of the Loup Fork specimens from Montana that I
have put in the genus Merycochcerus (1900 and 1901). I have not made a careful study
and comparison, but since seeing the fine specimens of Meri/coehasrus in the American
Museum of Natural History, I think it is very likely that the Montana forms above
referred to should be put in a new genus, so I will at present refer these to Merycocha-ru*
with a question mark, as : Merycochoems (?) laticeps, M. (?) altiramus, etc.
OF THE WHITE RIVER BEDS OF MONTANA. 21
The parts of the present specimens are small, being anterior fragments of mandi-
bles, yet these parts are so perfectly characteristic in Merycochcerus (?) that I cannot avoid
the belief that if the skull of this animal is found it will show a strong leaning toward
that genus, and I would not be surprised if it proved to be the White River ancestor.
This is made more probable by the fact that none of the Oreodontidce that have been
described can be considered as ancestral to Merycochoerus, either the true genus or the
doubtful one. I do not refer to the John Day forms, Promerycochasri, which have been
included in that genus. They are a very different animal.
The specimens are (48) the anterior part of a right rainus of a mandible, with part
of the alveolus of the canine, the root of P T , the alveolus of P T and the last two premo-
lars complete, and (66) a part of a right ramus with the last three preniolars and the
There are four ways in which it differs from Merycochcerus (?). It is only a frac-
tion of the size ; it has, like some species of Leptauchenia, two mental foramina ; the
premolar teeth are not crowded, and the teeth are not so high. The premolar series was
probably as long as the molar series.
In the form of the chin and symphysis, the depth of the jaw, the narrowness of the
space between the rami back of the symphysis, the evident reduction of the incisors, the
smallness of the canine, the lenticular section of P T , the forms of all the teeth preserved,
their narrowness in proportion to their length, are all like Merycochoerus (?). It may be
that these two fragments belong to different species, so to save confusion I will describe
Specimen 48- The anterior upper tip of the ramus is broken off. This shows ante-
riorly a transversely narrow broken surface, nearly in the middle of which is the canine
alveolus, but no hint of incisors. It is very doubtful if there was the full number. If
there was they must have been exceedingly small or placed anteriorly to the canine.
The anterior surface of the chin as far as shown is steep and straight along the sym-
physeal suture. It is convex transversely. The anterior mental foramen is beneath the
posterior part of P^, the posterior one a little behind the middle of Py. They are a
little above the longitudinal middle line of the ramus. The anterior is the larger. The
symphyseal suture is broadest below and narrows upward as in Merycochoerus (?). The
posterior of the symphysis is under the anterior part of P T .
The canine was evidently much smaller than P T . P T is lenticular in section with
nearly equal sides and rounded angles, the longest diameter being obliquely fore-and-aft,
and directed posteriorly outward and anteriorly inward as in Merycochoerus (?). P^- had
two roots near together. P^ has a proportionally large paraconid, which as the tooth is
viewed from the outside is seen to be separated above by a distinct notch from the pro-
22 DOUGLASS FOSSIL MAMMALIA
toconid. The deuteroconid is only represented by a narrow ridge passing part way down
the tooth inward and backward from the apex of the protoconid. The metaconid is
much better developed and is represented by a posterior tubercle, connected by a narrow
ridge with the apex of the protoconid.
In P T the paraconid is thicker and higher than in P^, and is inflected inward. The
other four elements the protoconid, metaconid, deuteroconid and tetartoconid* are all
well developed and surround a quadrangular cup-shaped depression. This is true also
of Oreodon and Eucrotaphus, and the form of the tooth differs very little from these.
In this particular tooth, however, the posterior depression would, on further wear, make
an islet, not an enamel loop. It is shallow and would entirely disappear if there should
be much wear. The great difference between this tooth and the corresponding ones of
the above-named genera is in the narrowness of the tooth as compared with its length.
P-g- also differs in this respect and they are, therefore, more like Merycochasrus (?). P^
differs from that of Oreodon culbertsoni in the simplicity of the inner ridges rudimen-
tary deuteroconid in this respect being much like the specimen I have examined of 0.
gracilis. It also differs in having the paraconid partly separated from the protoconid.
Again in these slight variations the differences are in the direction of Merycochcerus (?) .
In the type of M. (?) altiramus, which I have described from the Loup Fork of Mon-
tana, the teeth are little worn and give an excellent opportunity for comparison (1901, p.
73, Fig. 1). The present species shows almost no advance on this pattern. The para-
conid on the last two premolars is a little thicker, and this element is distinguishable on
P^, which has just begun to develop the posterior elements. The teeth have all increased
Specimen No. 66. In this specimen the mental foramina are farther apart. The
anterior one is under the anterior part of P^ and the posterior one under the anterior of
P . There is a thickening of the ramus an outer convexity between these two foramina
not seen in No. 48.
P-g- is shorter than Pg- and the roots are not close together as in the other specimen
P ? is about the same. P T has the metaconid and tetartoconid much lower and they are
not united at the posterior inner angle of the tooth, so that after considerable wear there