OCT 6 1^*^/ MEMOIKS OF THE T, I I CAENEGIE MUSEUM. VOL. IV. NO. 5 DESCRIPTION OF NEW CARNIVORES FROM THE MIOCENE OF WESTERN NEBRASKA. By 0. A. Peterson. While prospecting in the hiyer of fossil bones, which is exposed around "Car- negie Hill" and "University Hill" in the Miocene formation on the upper Niobrara River in Sioux County, Nebraska, with a view to opening additional quar- ries/ the writer, assisted by Messrs. T. F. Olcott and A. A. Dodd, was extremely fortunate in hndini^ in what is now called Quarry No. 3 (see Fig. 1, 3) various Fig. 1. Agate Spring Fossil Quarries. 1-2, quarries Nos. 1 and 2, on Carnegie Hill ; 3, quarry No. 3 ; AM, American Museum quarry on Carnegie Hill ; NU, Nebraska State University quarry on Uuivereity Hill ; A, quarry A. (From a photograph by the author.) remains of carnivora, among wliich is a nearly complete skeleton of a canid. The material having been entirely freed from the matrix and prepared for mounting by the skillful work of Mr. S. Agostini, the writer was entrusted by Director W. J. Holland with the task of describing and illustrating the new material. 1 No quarries except quarry A had at that time (1904) been opened in these hills. 205 206 MEMOIRS OF THE CARNEGIE MUSEUM In the quarry designated as No. 3 in the accompanying illustration, there were found from five to nine individuals of Daphmnodon superhus, of which No. 1589 is practically a complete skeleton. The latter was found imbedded together with another individual of the same species not fully adult. The soft sandstone in Avhich the two were found was taken out in a large block and transported to the Museum. On examination it was found that the adult individual was partly articulated, especially the vertebral column and portions of the feet. The tail was found attached to the sacrum and thrown backward in a graceful curve so that the tip was lying close to the pelvis and abdominal region. The anterior nine vertebrae and the anterior portion of the tenth were preserved in position. At this point the caudal region had been exposed and the vertebree had dropped out. By carefully measuring the natural curve, in which the tail was undoubtedly placed, its length was ascertained and the missing vertebrse were either restored in plaster, or by other vertebrse which were found in the talus below the place where the skeletons were dug out. The ribs of No. 1589 are not well preserved while those of the younger individual, No. lo89a, are quite complete. The posterior portion of the sternum is not present. One hind limb and foot of No. 1589 was found articulated. The posterior portion of the pelvis is weathered away, as is also the extreme anterior portion, leaving only the middle region. When material other than the type is used in the following description it is only for the purpose of elucidation and the catalog number will be mentioned in connection with the specimens referred to in order to prevent confusion. Scattered remains of a number of individuals, apparently of the same species, were found in all of the Agate Spring Fossil Quarries, but in Quarry No. 3 were found the most abundant remains of Daphcenodon superhus. Besides the type of Nothocyon annectens Peterson and another small carnivore described later in this paper there was little else found in Quarry No. 3 except remains of Daphcenodon supcrhm. This is interesting from the fact that hardly more than a stone's throw to the north and practically on the same level, the very rich and classic quarries, No. 1, No. 2, and the "University Quarry" are located (see Fig. 1). The lithological characters of the bone layer of Quarry No. 3 are similar to those in the largo quar- ries and it was observed that the same barren pinkish bedrock found in the main quarries is also present in this quarry. The irregularity of the fossil-bearing stratum was observed to be quite like that in the quarries immediately to the north. As has been stated elsewhere, the stratum containing fossil bones in the Agate Spring Fossil quarries was most probably laid down by a stream full of shallow channels which shifted from time to time during its existence. \ 1 PETERSON : NEW CARNIVORES FROM MIOCENE OF WESTERN NEBRASKA 207 Much field-work has been done and a considerable amount of material repre- senting the carnivora has been found in the lower and middle Miocene of South Dakota, western Nebraska, eastern Colorado and Wyoming during the past few years by tlie different museums of North America.' Since further explorations of this region are likely to occur in the near future it has seemed best to postpone a more detailed and systematic work upon the carnivora of these localities and for the present only to publish the new forms recently found by the museum parties and especially to accurately describe the skeleton of Daphtvnodon superhus. It is believed that the study of the lower Miocene Canidte will be facilitated by the publication of an account of the nearly complete skeleton of one individual. In the following description the exceptionally well preserved and almost com- plete skeleton of Daphmmts felinus Scott (No. 492, Carnegie Museum Catalog of Vertebrate Fossils) which Mr. J. B. Hatcher described (Memoirs Carnegie Museum, Volume I, pp. 65-108) will be used for comparison. In the first place it seems quite certain that Daphccnus is in the direct ancestral line of Daphosnodon, and secondly, the skeleton from the Oligocene now in the Carnegie Museum, represents the most complete individual of that genus as yet discovered. Valuable aid in this work was rendered by the authorities of the American Museum of Natural History and of Princeton University Museum, who courteously sent to the Carnegie Museum a number of types for comparison. The illustrations are from drawings made by Mr. Sydney Prentice and photographs by Mr. A. S. Coggeshall and the author. Family CANID.E. Subfamily Ami'hicyonin^: Daphoenodon superbus (Peterson). Amphicyon superhus Peterson, Annals Carnegie Museum, Vol. IV, pp. 51-53, PL XVIII, 1908. Daphomodon superhus (Peterson), Science, N. S., Vol. XXIX, pp. 620-621. Type: Skeleton nearly complete, No. 1589, Carnegie Museum Catalog of Vertebrate Fossils. Horizon : IMiocene (Lower Harrison Beds). Locality : Agate Spring Fossil Quarries (Quarry No. 3), Sioux County, Nebraska. Distinctive cliaracters, ~ In my previous notes in Science 1 g^\Q the following distinctive characters : Cranium comparatively short, hroad, and low; muzzle large; ^ Professor F. B. Loomis, of Amlierst, recently discovered two miles east of the Agate Spring Fossil Quarries tlie greater portion of a fine skeleton which he regards as belonging to the same genus described below. 208 MEMOIRS OF THE CARNEGIE MUSEUM sagittal crest prominent ; hrain-case small ; incisors heavy and short; canines compara- tively small and oval in cross-section; P* with antero-internal cusp of moderately large size and located relatively close to the main body of the teeth; 3P and 3P large andhroad; M^ present, though small, practically one-rooted and aligned with the internal border of AP and IP. To the above cited characters should be added : 1% Ci P| Tl/f; the symphysis of the iriferior ramus separated or slightly coossified ; the tubercles of all the molar-premolar teeth blunt; the heel of the lower carnassial tooth large: metatarsals not appressed at their proximcd ends ; feet sub-digitigrade ; tail long and heavy. rn The Skull. (Plates LXXIV-LXXVII.) In the articles cited above attention was paid to the cranial and dental charac- ters of Daphwnodon superbus. In tlie present paper, therefore, T will only give a short review^ of my previous description and add cliaracters which I regard as of interest and importance to the student. The cranium of this new genus represents an animal very nearly as large as a fully adult gray wolf, but the skull is proportionally shorter and broader, the brain- case smaller, the occipital condyles smaller, the basioccipital and basisphenoid of greater transverse diameter, the paroccipital process further separated from tlic tympanic bulla and directed more backward, the mastoid larger, the tympanic bulla smaller, and the postglenoid process heavier. When compared with Dapheenus felinus the skull is distinctly shorter and broader but the brain-case is evidently not much enlarged and there is a large frontal sinus; the muzzle is heavier; the pterygoids are shorter, and the region back of the pterygoids is apparently also somewhat shortened. The postglenoid and paroccipital processes are of the same relative size and position while the mastoid process is a truncated rounded knob, and is of proportionally greater size. In fact the entire mastoid region of Daphamodon seems to have increased, which is a distinct step away from the line of Canis. The space between the paroccipital process and the condyles is shorter, so that the occipital condyles arc more sessile in the specimen under description than in the Oligocene genus.^ The tympanic buUui are present on both sides in the type specimen ; they are of moderately large size and triangular in general outline. The tubular process, which encloses the external auditory meatus is quite prominent and is protected posteriorly by the heavy mastoid, though not so completely coossified with the latter bone as in the recent dogs. On the postero-internal angle of the bulla there is a sharp vertical ridge which unites with a similar though nmch ^In Daphmms fdinus (No. 492, C. M. Cat. Vert. Foss. ) the base of the skull iu the region of the condyles ia some- what injured and the separation of the condyle from the paroccipital process may, io part, be due to this fact. PETERSON : NEW CARNIVORES FROM MIOCENE OF WESTERN NEBRASKA 209 heavier ridge on the antero-internal angle of the paroccipital process, thus separating the Large pit for the tympanohyal externally, the jugular foramen, and the foramen laceruQi posterius internally. The bulla seems to be fairly well fused to the tym- panic region while in the Oligocene genus it apparently always slipped off From a study of the tympanic bulla in Daphienodon superhus it is quite evident that though quite sohdly attached along the antero-external angle and along the border of the basioccipital, the bone is not nearly so well ankylosed by the paroccipital process and the mastoid as for instance in Canis, In Daphcenus felinus the bulla appears to be more loosely attached to these processes than in the type under description. This fact apparently supports Mr. Hatcher's contention that the ankylosed auditory bulla in recent dogs should be regarded as a specialized character.* It is interesting to note that the tympanic bulla in Gynodesmus brachypus (Cope) is, in the first place, larger; and, secondly, much better ankylosed with the sur- rounding bones than in the form under description, and more nearly approaches that of Ca7iis. Professor Cope ' and Dr. Matthew ' have observed that the paroccipi- tal process in Cynodesrnus hrachypus projects " backwards nearly as far as the pos- terior face of the occipital condyles." After a careful study of this region in the type of Cynodesmus hrachypus, now before nie, it would seem that the backward pro- jection of the paroccipital process, regarded as a primitive character by Matthew, is in reality due very largely to crushing, there being no part of the base of the skull which is left undisturbed. Besides, if the present position of the paroccipital process of the skull of G. hrachypus is a natural one it would interfere with the free move- ment of the occipital condyle upon the atlas. Unfortunately the paroccipital proc- ess of the left side of G. hrachypus has been broken off, which otherwise might have furnished good verification. From Dr. Matthew's statement and illustrations {L c.) it is very evident that the paroccipital process in Cynodesmus ihomsoni is closely connected to the tympanic bulla and points almost directly downward at the tip, \vhi\Q in Gynodcsmus thooides Scott' the process is heavy and projects faore back- ward. It is quite likely that in a more perfect specimen of G. hrachypus the condi- tion of the paroccipital process would be found to be more like that in C. thomsoni and certainly not more backwardly directed than in G thooides. The postglenoid process of G. hrachypus has proportionally a less transverse diameter than in Daphmndon superhus and is more nearly similar to that of Canis, while the foramen opticum occupies a similar position, i. e., more in advance of the exit of the foramen rotuiidum than in recent dogs. *" Oligocene Cauidsc/'^MEMOiBS Caknegie Muskum, Vol. I, pp. 72-73, 1902. ^BuU. U. S. Geological and Geographical Survey of the Territories, Vol. VI, p. 389, 1881. "Bull. Am. Mus. Nat. ITistory, Vol. XXIII, p- 186, 1907. 'Trans. Am. Philos. Society, Vol. XVII, p. 6(5, PI. I, fig. 1893. 210 MEMOIRS OF THE CARNEGIE MUSEUM In the type under description the greater number of the foramina are Avell pre- served. With the exception of the more internal position of the stylomastoid fora- men ^ and the sliglitly more anteriorly placed foramen opticum, the size and position of the foramina are quite similar to those of Canis, which is well illustrated in Plate LXXV. The more internal position of the stylomastoid foramen is due to the large ^^^ r development of the mastoid process and limited backward extension of the tympanic bulla, while the position of the optic foramen may be due to the more advanced positicm of tlie oi'bit in Daj)hwnodon than in Canis. The alisphenoid canal is appar- ently larger in proportion than in the Oligocene genus. Fig. 2. Skull of young indiviJnal of Daphcemdon superhus. j nat. size. No. 1589a, C. M. Cat. Vert. Fose. Per, mastoid portion of periotic ; Gp, posfcglenoid process ; J, jugal ; PI, palatine ; S(i, sq^lamosaI ; Pa, parietal ; Fr, frontal ; As, alisphenoid ; Os, orbitosphenoitl ; I, lachrymal with foramen ; Mx, maxillary ; to, infraorbital foramen ; op, optic foramen ; fr, foramen rotnndum ; s/, sphenoidal fissure- (Note: The occipital plate, the jugal, the nasals and the preraaxillaries are missing and a portion of the zygomatic arch is removed in order to give a better view of the dif- ferent elements of the skull in this region.) In Fig. 2 is given a side view of the skull of the younger individual (1589a) which was found together with the type. The premaxillary, nasal, malar, supra- occipital, and the base of the skull are slipped off and lost, but what remains furnishes an admirable idea of each element and their relative proportions in com- parison with those in Canis. It is seen that the parietal and frontal are longer, lower, and less convex ; the alisphenoid of apparently the same proportion ; the orbitosphenoid as long, but of less vertical diameter ; the lachrymal larger ; and the maxillary shorter than in the recent form. From the type specimen the process of the premaxillary, which extends upward between the maxillary and the nasal, is much shorter and weaker than in Canis and in this respect more nearly similar to what is observed in Daphccmts. 3 The stylomastoid foramen is not nearly so well separated from the large pit for the tympanohyal as is generally the case in recent dogs. PETERSON: NEW CARNIVORES FROM MIOCENE OF WESTERN NEBRASKA 211 Lower jauK 212 MEMOIRS OF THE CARNEGIE MUSEUM has a simple crown ; it is separated from the canine and P~ by short diastemata. The latter tooth is also isolated, but the diastema back of it is very short. This tooth has two roots ; its ciown is comparatively simple, there being no anterior or posterior basal cusps. P- is simple-crowned, like the anterior premolar, and other- -J wise has about the same relative proportions as in Daphcenus felinus. Its position in the alveolar border is oblique, which makes it appear of small antero-posterior diameter in the side view of the skull. In P^ (carnassial) a slight change from what is observed in the Oligocene genus has taken place. The deutorocone is slightly reduced in size and is closer to the main body of the tooth than in either Daphcenus or any of the allied John Day forms, and in this respect is more like the corre- sponding tooth in A'mphicyon major of Europe. The deuterocone is, however, sup- ported by a much heavier root than eitlier in Am^phkyon or Canis and its oblique position in the alveolar border is totally unlike the latter genera. The smaller J antero-posterior diameter and otherwise primitive features of the tooth places the genus rather closer to the Oligocene form and reiKioves it further from Cards. M- and M- are larger than in any of the allied John Day forms {Temnocyon, Mesocyon, Para- daphxrms) and liavc also undergone some changes from the corresponding teeth in the Oligocene genus, while M^ is apparently the same. This change of the molars is principally due to the greater development of the postero-internal angle, which gives these teeth a slightly greater antero-posterior diameter internally than is usually seen in the Oligocene species, find the tooth is otherwise apparently inter- mediate between that of these early forms and Canis occidentalis, which has a very prominent, sharp, and crescentic postero-internal tubercle, especially on M-. The development of the postero-internal angle is unusually like what we find in Am- phicyon major and it is quite curious to find the conical structure of the different tubercles so nearly alike in the two genera. M- and M- of the European genus are, however, proportionally larger, the posterior intermediate cusp (metaconulc), especially on M2, is more strongly developed and the antero-external cusp (para- cone) is the larger, while in Daphwnodon para- and metacones are subequal and of the primitive type met with in Baphwnus ; in fact the tubercles are more depressed than in the latter genus, as is seen by referring to the illustrations, PL LXXV. The internal border of the grinding surface of M- extends slightly below that of the preceding tooth. This character, which is well shown on PL LXXIV, may, in part at least, be due to slight displacement. M^ is, as stated, practically a duplication of that tooth in Daplmnns, i. c, a small tooth with transversely oval and very low crown with a very short root, the alveolus which is posteriorly formed by a thin and very delicate border in the maxillary. r m PETERSON : ]S'EW CARNIVORES FROM MIOCENE OF WESTERN NEBRASKA 213 rn 1 he lower incisors are laterally compressed. ly is quite small in comparison with that of Canis, while I^- and I3, though small, are in better proportion. The canine is robust and resembles that of Canis occidentalis very closely, the cross-section of the crown being somewhat more regularly oval and less excavated on the antero- internal angle. Py is a small single-rooted tooth with the crown extended well forward from the base, the apex being nearly directly over the anterior face of the root. The tooth is isolated by diastemata, of which the one in front of it is the longer. The next three premolars are also well separated in the jaw. P^ and P-3 have simple conical crowns with larger posterior than anterior bases and their shape and relative diameters are quite similar to those in Daphwmis, while Py is somewhat enlarged and has a strong posterior cusp and prominent cingulum similar to some species from the Oligocene, Daphojn.us dodgei Scott and Temnocyon altigenis Cope from the John Day formation of Oregon. In My (carnassial) is observed the greatest change from what appears in the Oligocene form in the direction of the more recent type. It is at once seen that this tooth has increased in its antero-posterior diameter, the metacouid is situated lower down,^ and the hypoconid is bluntly trenchant and is greatly developed, while the entoconid is rather small. The tooth is not, however, modified into the perfect shearing blade met with in Canis, as it still retains to a greater or less extent the conical structure of the tubercles so characteristic of Bapha^nus and its ancestors, the Miacid^. M^ is quite large ; the trigonid is raised only very slightly above the heel ; the proto-, para-, and metaconids are fused into a more or less solid and obtuse mass, the protoconid being the largest of the three, while the paraconid is represented only as a basal heel on the antero-internal angle. The heel is quite large and is occupied principally b}^ an antero-posteriorly placed ridge (hypoconid) which is low and located nearer the external than the internal border. The entoconid is represented by a low and oblique ledge on the postero-internal angle of the crown. Mg is single- rooted and is reduced to a small oval-sliaped crown with a slight elevation on the anterior half, while postero-in tern ally there is a shallow valley produced, to some extent by trituration. The size of the tooth with reference to the one preceding it ft IS proportionally the same as in Daphcenus felinus here used for comparison. ^It is interesting to note that Cynodesmits hrachypus from the Miocene of the Laramie Peak locality ia apparently inrther advanced towards the recent dogs in having the premolars larger, M^ reduced and the lower carnassial of less transverse diameter. 214 MEMOIKS OF THE CARNEGIE MUSEUM Measurements of DENTiTroN, Mm. Length of superior premolar series ^^ Length of superior molar series ^^ Autero-posfcerior diameter of siiperior canine at base 13 Transverse diameter of superior canine at base 10 Antero-pOrtterior diameter of P^ ^ Transverse diameter of P-^ ^ Antero-posterior diameter of P^- > ^^ Transverse diameter oEP-^ ^ Antero-posterior diameter of P^ : 1^ Transverse diameter of P^ ^ Antero-posterior diameter of P+ ^^ Transverse diameter of P\ anteriorly ^^ Antero-posterior diameter of M^ 1^ Transverse diameter of M^ '^^ Antero-posterior diameter of M^ 1^ Transverse diameter of M^ 1^ Antero*posterior diameter of M- ^ Transverse diameter of M'^- I'ETERSON : NEW CARNIVOKES FROM MIOCENE OF WESTERN NEBRASKA 215 The remains indicate an animal of slightly larger size when fully adult (see Figs. 2 and '3). Fig. 3. Lower Jaw of Yonng ludividnal of DapJiisnodon supcrhiis. f nat. size. No. 1589a. In section 17 of Quarry No. 1 was found a fragment of a mandibular ramus with the deciduous sectorial in position and the permanent p:^ buried in the jaw. The latter tooth is only partially developed. The differ- ent cusps of the deciduous tooth are quite sharp and the heel is of considerable size and completely basin- sluiped. The fangs are excessively spread in the an- tero-posterior direction, due to the constant pressure from below by the permanent tooth, and furnish a sure indication of the juvenility of this tooth. Except for its slightly smaller size, the tooth most nearly resembles that of "Cams" anceps Scott and is here figured for the purpose of warning the student. The specimen may or may not represent Daphwnodon siq^erhus. Fig. 4. Daphccnodon snperhus? Frajjinent of lower jaw of young individual with deciduous sectorial in pliice nud p^ buried iu the jaw (No. 2201). Nat. size. The Vertebral Column. The formula of the vertebral column is as follows : cervicals seven, dorsals thir- teen, lumbars seven, sacrals three, and caudals twenty-eight? + > With the exception of the missing region of the caudals, already referred to in, the introduction, the vertebral column of the type specimen is in excellent preservation. Atlas. 216 MEMOIRS OP THE CARNEGIE MUSEUM reveal conditions more nearly identical with those found in Baphamus. Thus the posterior exit of the vertebrarterial canal is not superior as in the dogs, but more nearly posterior as in the cats;^'' the anterior cotyles are small; the transverse r. Fig. 5. Atliis of D. superbus. ^ nat. size. 1, dorsal view ; 2, lateral view ; 3, veutral view. diameter of the articulation for the axis is proportionally less in size ; and tlie bone, with the exception of the transverse process, is, as a whole, rather heavier than in Canis. The atlanteo-diapophysical notch is deep and as in the dogs is not covered with a bony bridge to form a foramen. This character was observed in Daphosnus , by Professor Scott," while in some of the genera of the Alia- ^^^^^ \ <^'>^dx (Oodectes) this notch was already converted into a fora- '^-'^^Zf^^^^ men.^'^ The foramen in the arterial groove, which pierces the wall of the neural arch on the ventral face of the trans- FiG. 6. Atlas of Cmodes- . i-i tii-t'i,t mn. hrachypusco^e. No. 8140, ^ersc proccss, IS exceedingly small, which IS also the case m Cope Collection, Am. Maseum the atlas of Dapliwnus here used for comparison. In Canis Natural History, i Dat. size. . i PETERSON : NEW CARNIVORES FROM MIOCENE OF WESTERN NEBRASKA 217 Fig. 7. Lateral and Posterior Views of Axis of * 218 MEMOIRS OF THE CARNEGIE MUSEUM Measurements of Axis, Mm. Anteroposterior diameter of the centrum, odontoid process included 54 Greatest transverse diameter of centrum, anteriorly 36 Greatest traasverae diameter of centrumj posteriorly 22 Antero-posterior diameter of spine 53 Transverse diameter across postzygapophyses 34 Third Cervical Vertebra. PETERSON ; NEW CARNIVORES FROM MIOCENE OP WESTERN NEBRASKA 219 Measurements of Pourth and Fifth Cervical Vertebra. C. 4 Mm. Antero-posterior diameter of centra 31 Transverse diameter of centra, posteriorly 29 Vertical diameter of centra, posteriorly 15 Transverse diameter across transverse processes, posteriorly 55 Transverse diameter across prezy^apopbyses, posteriorly 43 Transverse diameter across postzygapophyses 41 C. 5 Mm. 30 22 16 59 43 39 Fig. 9. Sixth Cervical Ver- tebra of D. siqyfrbiis. ^ nat. size. Sixth Cervical Vertebra. > 220 MEMOIRS OF THE CARNEGIE MUSEUM In the sixth cervical vertebra of Cynodesmus hrachypus the inferior branch of the transverse process is divided, but the anterior portion is much larger proportionally than that in Daphwnodon. In all other respects the bone is more nearly like that in Canis than Daphmnodon. In fact the entire cervical series in C. hrachypus is strikingly similar to that of the recent form except that the posterior exit of the vertebraterial canal of the atlas is further back, the neural spine of the axis more produced posteriorly, the inferior lamella of the transverse process of the sixth cervical is divided, and the neural spine of the seventh cervical is more robust and higher. Measubements of Sixth Cjebvical Veetebba. Mm. Antero-poaterior diameter of thecentrura 29 Transverse diameter of the centrum posteriorly 23 Vertical diameter of the centrum posteriorly 27 Transverse diameter at transverse processes 59 Transverse diameter of prezygapopbyses 43 Transverse diameter of postzygapophyses ^"^ Seventh Cervical Vertebra. PETERSON : NEW CARNIVORES FROM MIOCENE OF WESTERN NEBRASKA 221 1 First Dorsal Vertebra.222 MEMOIES OF THE CARNEGIE MUSEUM cal with that which has been described in the preceding vertebra, and is well shown in the illustration, Fig. 13- MEASURE:\rENTS OF SECOND DORSAL VERTEBRA. Mm- AnterO'posfcerior diameter of the centrum "^ Transverse diameter of the centrum, posteriorly 27 Veitical diameter of the centrum, posfceriorly 1" Transverse diameter at the transverse processes ^0 r Transverse diameter at the prezygapophyses ^5 Transverse diameter at the postzygapophyses 22 Height of neural spine ^' TJdrd and Fourth Vertehrx.^The third and fourth dorsals liave their centra similar to that of the second. The transverse processes are lighter, the prezygapophyses are less complicated, facing directly upward,and the neural spines are less expanded at the top than in the preceding vertebra ; there are otherwise no characters which require a separate description. Fig. 14. Lateral Views of Third and Fourth Dorsals of D. superhus. ^ nat, size- Measurements of Third and Fouiith Dorsal Vertebra. D. 3. D 4. Mm. Mm. Antero-posterior diameter of centra *^ ^^ Transverse diameter of centra '^^ ^^ Vertical diameter of centra ^^ ^" Transverse diameter at transverse processes '*' ^^ Transverse diameter at prezygapophyses "^ ^1 Transverse diameter at postzygapophyses ^^ Height of neural spine approximately ::^^->^;^:-:^>:j-;<,::=^^---^lir7m:^^^ PETERSON : NEAV CARNIVORES FROM MIOCENE OF WESTERN" NEBRASKA 223 processes decrease but very slightly from the fifth to the ninth, the articular sur- faces for the ribs are extensive on the free and expanded ends, but the excessory processes on the superior faces, which are so prominent in the recent dogs, are very inconspicuously developed. The neural spines in the whole series are heavy, but gradually reduced in height in proceeding backward, the ninth being quite thin transversely though broad antero-posteriorly. Measurements of Fifth and Ninth Dorsal Vertkbrte. D. i D. 9 Mm_ Mm, Antero-posterior diameter of centra 24 25 Transverse diameter of centra 24 26 Vertical diameter of centra ]6 17 Transverse diameter at transverse processes 50 47 Height of neural spines 60 43 Tenth Dorsal Vertebra. 224 MEMOIKS OF THE CARNEGIE MUSEUM Eleventh Dorsal Vertehra. PETEKSON : NEW CARNIVORES FROM MIOCENE OE WESTERN NEBRASKA 225 and tlie zygapopliyses and neural spines, especially of tlie thirteenth dorsal, are typically lumbar in character. t I Measurements of Twelfth and Thirteenth Dorsal VEETEBEiE. n 12 Mm. Autero-posterior diameter of ceutra 28 Transverse diameter of centra posteriorly 26 Vertical diameter of centra posteriorly 17 Transverse diameter at transverse processes 31 Height of neural spine D. 13 Mm. 29 27 19 33 25 First Lumbar Vertebra. 226 MEMOIRS OF THE CAniNEGIE MUSEUM is a curioxis feature of this genus which seems to have departed from Daphcenus, Temnocyony^'^ and also from the recent canids generally, while, according to observa- tions by Dr. Matthew {I. c, p. 367) upon Miacis, its aflfinity with regard to the short transverse processes would seem to recall the earlier tertiary types. As in Daphcenus the metapophyses, anapophyses, and neural spines are proportionally more robust than those in the recent dogs. The meta- and anapophyses are heaviest on the first lumbar and gradually decrease in size backwards, the anapophysis of the fifth lumbar being quite small. Except on the first, second, and seventh lumbars the neural spines are broken off superiorly, but enough remains to ascertain their great an tero- posterior diameters, which are not unlike those in canids generally. Measurements of Second and FiiTii Lumuar Vertedr^. L. 2 L. 5 Mm. Mm. Antero-posterior diameter of centra 36 40 Transverse diameter of centra posteriorly 27 29 Vertical diameter of centra posteriorly 19 19 Transverse diameter of transverse processes 50 72 Length of transverse processes from base (posterior) to tip end 14 24 Height of neural spine 18 ? r .^ Sixth lAtmhar Vertebra.' '228 MEMOIRS OF TEIE CA.ENEGIE MUSEUM resented b}^ a comparatively delicate roughened ridge, while the third has a promi- nent process which extends laterally and slightly back of the posterior face of the centrum. In the recent forms the transverse j)roccsses of the sacrum are more strongly developed, the first and second sacrals, especially, taking a more prominent part in the support of the ilium. In comparison with the heavy centra the neural canal is rather small, thus agreeing with what is observed in the Oligocene genus and differing from recent forms. The neural spine of the first sacral is relatively smaller than in the recent dogs or cats, while that of the last sacral is much heavier, the spine on the median vertebra consisting only of a sharp low ridge, which uninter- ruptedly connects with the spines in front and beliind. In Cjjnodesmus hrachypus the sacrum is shorter and broader, the second sacral helps to support the ilium more prominently and the anterior neural spines were apparently proportionally more robust than in Daphwnodon. The spinous process on the last sacral in C. brachypits however, is, quite prominent, and is more com- pletely separated from the spine in front of it than in either Baphwnodon or the recent dogs. Measurements of Sacrum. Mm, Antero-posterior diameter of centra 77 Greatest transverse diameter across the pleurapophyses 59 Greatest transverse diameter at transverse processes of last sacral 49 Caudal Vertehrsc. \ PETERSON : NEW CARNIVORES FROM MIOCENE OF WESTERN NEBRASKA 229 that of the first, while back of this vertebra the increase in length is more noticeable. The transverse processes are single up to the sixth, and the seventh caudal still bears the pre- and postzygapophyses witli perfect articular facets in front and behind, while on the eighth and ninth the postzygapophyses are more or less imperfect, but the neural canal is complete and extends throughout nearly tlie entire length of the vertebrae. The last mentioned vertebra is the longest in the series which was at- tached to the sacrum when the specimen was found, but there is reason for believ- ing that the vertebra continued to increase in lengtli up to the eleventh or twelfth caudal. There are seven chevron bones present whicli were found more or less in position in the series of caudals here described. At the anterior portion of the tenth caudal the skeleton was exposed and a section of the tail dropped out. As mentioned in the introduction, the natui-al curve in which the tail undoubtedly was lying would allow room for six or possibly seven vcrtebrie. Working on this hypothesis the next vertebra which was found in position in the skeleton would be the seventeenth or eighteenth. This vertebra has the pre- and postzygapophysial prominences, anterior transverse processes, and ventral tubercles for chevrons well represented, and the centrum is of considerable length and robustness. The ten succeeding vertebrsi gradually decrease in length and robustness, the last being only seven millimeters long. In tlietypeof Ci/nodcsmus b rack i/pus the anterior four caudal vertebrfe are pres- ent together with various others which represent different parts of the tail. The transverse process on the first is yery long and robust, while on the second it is much shorter and indicates a greater decrease backward than is seen in Daphwnodon. The centra are less depressed than in the latter genus and approach more closely those in the dogs. It is quite likely that the tail in C. hrachyjms when completely known will be found to be proportionally shorter than in Daphcenodon sitp^ , 230 MEMOIRS OF THE CARNEGIE MUSEUM rior region of the thorax the ribs have largo caxjitular and tubercular articulations and rather flat shafts as in Canis, while pos- teriorly they are raore cylindrical; the elev- enth, twelfth, and thirteenth having no tuber- cular articulations. Sternum. "I .--'. 232 MEMOIRS OF THE CARNEdlE MUSEUM before the suprascapular border is reached, and the acromion process on the spine is well developed, while the postscapular fossa is unequally divided by the prominent and sharp glenoid border, identically as in the bears. In the fox the scapula has a small area developed for the teres major muscle at the superior portion of the glenoid border, but it is quite small in comparison with that in the present genus, in which the scapukx as a whole is proportionally broader transversely. The spine is high and terminates in an overhanging acromion, which, however, does not project below the glenoid cavity. Judging from the heavy acromion process there was undoubtedly a clavicle of perhaps considerable development. The glenoid cavity is not so large proportionally as in the large cats, and in this respect is ratlier more like that in Cards. The coracoid process also has about the same proportionate development as in Canis and is relatively much shorter than in the tiger. Professor Scott states {l. c, p. 381) that in Cynodiciis "the coracoid process is unusually large ..." that the "prominent acromion descends below the level of the glenoid cavity ..." and that in general character the scapula " is rather viver- ine or raccoon-like than canine," which is not true of the scapula in Daj^limnodon. In Temnocyon ferox, according to Eyerman's description '''"' the greatest vertical length [of the scapula] is from the coracoid to the posterior upper edge and not along the spine, as in C. familiaris. This is not the case in the present genus, nor is the spine " faUing short of the supra-scapular border," as in the John Day genus, but in this respect it is more nearly like the dogs. It is interesting to note that the area for the teres major muscle on the upper part of the glenoid border of the scapula in Iloplo'plioncus is also quite large, and is quite similar to that in Daphcei'iodon. Measurements of Scapula. Mm. Greatest height from glenoid cavity to siipraFcapular herder 167 Greatest antero-posterior diameter 104 Antero-posterior diameter of glenoid cavity, including the coracoid process 39 Transverse diameter of glenoid cavity 33 Humerus (Plate LXXIX).-. 234 MEMOIRS OF THE CARNEGIE MUSEUM The olecranon process of the uhia is very short, truncated, and more inwardly produced than in the Oligocene genus. The tendinal sulcus, though well defined and wide, is not so deep as in Canis lupus, and consequently much less developed than in the cats. The principal feature of the greater sigmoid notch is its heavy coronoid process, which character is feline rather than canine. The shaft is very heavy and not much ' curved. The tuberosity for the flexor profundus digitorum muscle on the external border of the shaft is very heavy, as is well shown in Figs. 5 and 6, PI. LXXIX. Distally the shaft is more trihedral in cross-section due to the sharp ridge which separates the internal and posterior surfaces as in the tiger. The articulation for the radius on the distal end is a heavy rounded knob much as in the large cats and entirely unlike that in the dogs. The styloid process ex- tends well down and has a decided constriction above the articulation for the cuneiform. In Cj/nodesmus hracliypus both ulnse are represented. The olecranon process is short and truncated^ the tendinal groove is deep as in CaniSj but the coronoid process is larger in proportion than in the recent dogs^ as is also the shaft- The distal end^ however, lacks the heavy and broad development of the rounded tubercle which articulates with the radius and is so conspicuous in the genus under descrip- tion. The feature of this reduction of the ulna in C- brachypus is apparently a decided step towards conditions found in Canis. Measurements of Radius and Ulna. Mm, RadiuBj greatest length 184 Eadius, antero-posterior diameter ot proximal end 16 EadiuSj transverse diameter of proximal end 23 EadiuSj transverse diameter of shaft medially 15 Eadius, antero-posterior diameter of shaft medially , 12 Eadius, greatest antero-posterior diameter of distal end 23 Eadius, greatest transverse diameter of distal end 34 Ulna, greatest length 293 Ulna, antero-posterior diameter at coronoid process 32 Ulna, transverse diameter at coronoid process 25 Ulna, transverse diameter of distal end at tubercle 20 Ulna, antero-posterior diameter of distal end at tubercle 11 lite Manus (PL LXXX).- PETERSON : NEW CARNIVORES FROM MIOCENE OF WESTERN NEBRASKA 235 that the similarity is remarkabl}^ close. The dorsal face of the bone in the present genus has a less vertical diameter than in the Oligocene form, which is due to the fact that the articulation for the radius is carried further downward on the anterior face than in the older type, in fact lower than in the lion and tiger, consequently far removed from conditions in the recent dogs, in Avhich the dorsal face of the scapho-lunar is very large, the bone being much higher throughout. The articular facet for the radius also extends far downwards on the palmar face near the ulnar angle, as in the lion and tiger, which indicates that there was even more power of flexure between these two bones (radius and scapho-lunar) in the present genus than in DajyliceniLS, and is a condition strikingly similar to what is seen in the large cats. The radio- palmar process is well developed, but has not the vertical diameter seen in Cams, and is more nearly like that in the cats. It is, how- ever, more distinctly separated, b}^ a constric- tion, from the main body of the bone than in the felines, which is due to the greater devel- opment of the palmar face of the scapho- lunar in the latter. The ulnar face has no facet for the cuneiform, which is identical witli that in Daphmnus, in which the cuneiform and scapho-lunar did not come in touch with one another sufficiently to establish facets. Distally there are well defined facets for the unciform, magnum, trapezoid, and trapezium respectively. The facet for the unciform is deeply excavated as in Daphcenus, while that for the magnum is extended downward somewhat less near the dorsal border, the beak of the lunar being less produced downward in the present genus. The facet for the trapezoid is a large, irregularly shaped, and obliquely placed surface, indicating a trapezoid of considerable size. The facet for the trapezium is located on the extreme radial angle, but is poorly demarcated from the facet for the trapezoid. Fig. 29. Scaplio-luBar (left sides) of D .super- bus. Nat. size. 1, dorsal view ; 3, palmar view; 3, radial view ; 4, ulnar view ; 5, inferior view ; 6, sux:)erior view. Measueemtents of Scapho-lunar. Mm. Greatest antero-posterior diameter, iucladiug the radio-palmar process 29 Greatest transverse diameter 29 Greatest vertical diameter near dorsal face 10 Cuneiform. 236 MEMOIRS OP THE CARNEGIE MUSEUM face of the bone, and furnishes muscular attachment for Mc.V, but docs not articu- late with the latter as in the dogs. The facets for the ulna and pisiform on the pi'oximal face are sub- equal in size and are separated by a promhient ridge, the former being more concave antero-posteriorly than the latter. In the dogs the facet for the ulna is difiPcrentj being more sinuously curved, with an antero- posterior cavity on the ulnar angle and an antero- posterior convexity on the proximo-dorsal angle of the bone. The distal fiice is taken up by a large facet which is concave in all directions and articu- lates with the unciform. The bone as a whole is not unlike the cuneiform of the tiger. Cuneiform (left aide) of Z>. superhus. Nat, size. 1, dorso-uinar Tiew ; 2, dorsal view ; 3, inferior view ; 4, superior view ; 5, palmar view ; 6, pal mo-radial view. Measueementh of Cunbiform, Mm. Greatest autero-posterior diameter 14 Greatest transverse diameter 20 Greatest vertical diameter of dorsal face 8 Pisiform. .i5»3?- i^5si^ir!rs-^ 1 PETERSON : NEW CARNIVORES FROM MIOCENE OF WESTERN NEBRASKA m process for attachment of muscles. On the ulnar face is a large facet for the trapezoid, which extends well back, cov- ering the base of the palmar process, referred to above. On the proximal radial angle there are Avell formed facets for the scapholunar ; the one on the ra- dial angle is scarcely separable from the one on the palmar process, Distally there is a large obliquel}^ placed facet 237 Fig. 32. Trapezium (right side) of D. supcrbiis. Nat. size. 1, dorsal view; 2, radial view; 3, superior view; 4, posterior view; 5, uln^r view; 6, inferior view. for Mo. I which is concave from before backwards. Measurements of Trapezium. Mm. Greatest autero-posterior diameter 238 MEMOIRS OF THE CARNEGIE MUSEUM Mc. Ill and the smaller for Mc. IV. In IJaphcenus the latter facets are divided in about the same proportions as in the present genus. Measurements of Magnum. Mm, Greatest antero-posterior diameter 20 Transverse diameter of dorsal face 11 Greatest vertical diameter 16 Unciform. PETERSON" : IS^EW CARNIVORES FROM MIOCENE OF WESTERN" NEBRASKA 239 \ and the articulation for the trapezium is broader. In comparative size the bone is very little smaller than in Daphccnus and consequently much heavier and longer than in Cards. On the ulnar face the rugosity for the ligamentary attachment to Mc. II is quite prominent and the tuberosity on the radial side of tlie head is as large proportionally as in recent cats, though the bone as a whole however, is, lighter and longer than Mc. I of the pollex in the latter. The distal end has a well formed facet which is separated palmarly by a prominent keel. The shaft is sliglitly curved, so that the distal end points a little in the ulnar direction. Measurements of Metacarpal I. Fro. 35. Dorsal and Ra- dial Views of Left Metacar- pal I of D. superbus. Nat. size. Mm, Greatest length 32 Greatest antero-posterior diameter of head 10 Greatest trausverse diameter of head.PETERSON : NEW CARNIVORES FROM MIOCENE OF WESTERN NEBRASKA 241 Measukements of Metacarpal IV- Mm. Greatest length 60 Greatest antero-posterior diameter of head 15 Greatest transverse diameter of head 12 ,242 MEMOIRS OP THE CARNEGIE MUSEUM X !* fl The phalanges of the second row are also proportionally shorter than tliose in the older genus, but display the same -_'K- ^PETERSON : NEW CARNIVOKES FROM MIOCENE OF WESTERN NEBRASKA 257 Restoration of Daphcenodon superhus. (Plates LXXXIII and LXXXIV.) The unusually well preserved and symmetrical bones of the type of Dajjhwno- don superhus have caused comparatively little trouble in correctly articulating the different parts of the skeleton. The restoration has been veryAvell executed by Mr. Serafino Agostini. The mounted skeleton furnishes many additional features of the animal which only in this way can be fully appreciated. The position of the head is as high in relation to the horizontal line of the vertebral column as the animal ordinarily held it in life. The long neural spines on the cervical vertebrae farther indicate that the animal probably carried the head rather low. The curvature of the neck at the junction of the thorax is gentle, and the position of each dorsal vertebra increases in height until the lumbar region is reached, where a sudden downward curvature again takes place, not unlike that in the recent tiger. The twelfth and thirteenth dorsals have no facets for a tuberculum of the ribs as in Felis figris and it is quite doubtful whether the eleventh dorsal of the present form carried a tubercular facet for the rib as the transverse process is + poorly developed. The thorax was on the whole quite small, which is also charac- teristic of the Oligocence genus. The heads of the ribs of the left side (except the eleventh or twelfth) are all represented, while the shafts are restored after the more completely preserved ribs of No. 1589a. On the left side the ribs are poorly pre- served, there being present only the heads and portions of the shafts of seven. The fore limbs display a close similarity to those in Daphcenus. The deltoid and supinator ridges of the humerus are prominent, the former low down on the shaft, indicating the downward extent of the heavy muscles, while the development of tlio latter indicates the power of supination of the manus. Though the forearm is proportionally longer than in the latter genus the feet are remarkably little advanced, being short and broad, the pollex little reduced, and the position of the phalanges sharply angled as in the felines, but proportionally somewhat shorter than in Daphmnus. The power of retractility of the unguals was nearly as great as in the older form and no doubt still served the purpose of assisting in catching and holding the prey. As has already been stated the posterior portion of the pelvis, portions of the femora, and the proximal ends of the tibia? were unfortunately weathered out and disintegrated when the specimen was found, but from other material the compara- tive measm-ements Avere made, and the proportions, as they are in the skeleton repre- sented on Pis. LXXXIII-LXXXIV, are thought to be approximately correct. The n y \ 1 li 258 MEMOIRS OF THE CARNEGIE MUSEUM : hind feet are also quite broad, but somewhat longer than the fore feet^ while the ungual phalanges of the former had apparently a power of retractility equals or nearly equals to that of the latter. Altogether the skeleton shows a long and com- paratively slender body^ long tail, and short and heavy limbs with broad and short feet Measurements of Skeleton- Cm. Greatest length of vertebral column, the skull and all curvatures includerl 210 Length of vertebral column^ atlas to posterior end of sacrum 93 Length of tail, approximately 88 Length of lumbar series 26,5 Length of dorsal series 33 Length of cervical series 21 4 Height of skeleton at first dorsal 57 Height of skeleton at first lumbar 61 Height of skeleton at supra-iliac border, approximately 58 r Notes on Individual Variation of Daphcenodon. The superior dentition of the material at hand sliows very httle variation in general character, except in the young individual (No. 1589a) found with the type of Daphcenodon superhus and described on pages 210 and 215. The mandibular symphysis in some individuals however, is, longer, py has been dropped out and the alveole closed up, while m-g is sometimes inserted by two strong fangs. In a series of cervicals (No. 1589b), the fourth, fifth, sixth, and seventh, which were found on the surface and in the talus at Quarry No. 3 (Agate Spring Fossil Quarries), are seen some differences from those described of the type of Daphwnodon superhus. These differences are most likely due to individual variation and sexual distinction, and will bo thus treated in this connection pending the discovery of more material. This cervical series is in the first place considerably heavier and pertains to a larger individual than the type specimen. Secondly, the ventral keels appear to terminate posteriorly in a more condensed rugose tubercle approacliing more nearly that of Canis in its characters than to that of the type of Daphcenodon. Thirdly, the r anterior projection of the transverse process of the sixth cervical vertebra in the series under consideration is much smaller than the corresponding portion in this vertebra of the type specimen. One of the more significant differences from the latter is seen in the sixth and seventh cervicals. The base of the neural spine of the seventh cervical is more nearly oval in cross-section, with no such excavations in front or back of the spine as is characteristic of the type specimen described above. The base of the spine of the sixth cervical is also without these excavations, so that Ei^s PETERSON : NEW CARNIVORES FROM MIOCENE OF WESTERN NEBRASKA 259 it is quite safe to say that this individual did not possess this character of the verte- brae. Another important feature is seen in the prominent antero-posterior ridge in the floor of the neural canal, which is pierced by a transverse venous foramen, while in the type specimen this ridge is rather poorly developed and the area in the middle region, where the foramen is located in the larger specimen, is much greater and was not always bounded above by a bony bridge so as to form a foramen. The transverse process is proportionally much shorter, which would indicate a narrower neck in the larger individual. In the Amherst specimen, which is about equal in size to the one represented by the cervical vertebrae described above, there are some lumbar vertebra preserved. One of these has the transverse processes preserved. These processes are propor- tionally longer than those in the corresponding region of the lumbars in the type, and more in accordance with the conditions found in Daphwrms, Tcmnoci/on, and the recent forms. The os penis of the Amherst specimen of Daphcenodon is greatly different from that in No. 1589a (Carn. Mus. Cat. Vert. Foss.) and is described on pages 230 and 231. Relationships of Daphcenodon. That Daphwnodon had its true ancestor in Daj^liccnus of the Oligocene forma- tion can hardly be doubted, from the study which has resulted in the above com- parative description. Daphccnus in turn is generally regarded as a descendant of the Miacidse of the earlier Tertiary of Nortli America. In the Oligocene and early Miocene there were a number of diverging lines, Avhich as a whole were neverthe- less quite homogeneous in their general structure, though apparently sufficiently diversified to show with more or less clearness their destiny in later times. These points have already been ably treated by Scott, Boule, Schlosser, I\Iatthew, Merriam, Hatcher, Wortman, and others, and will not be discussed in the present paper, except in so far as to assign the phyletic position of Daphomodon. It would appear that Daphcenodon is not yet represented by any very closely allied species ^^ in the intervening formations from the lower Harrison beds down to the lower portion of the Oligocene where Daphcenus, its predecessor, is found. The genera from the John Day are already too flir advanced in different ways to be seriously considered as intermediate forms. Thus Temnocyoih and Mesocyon have m- absent and differ in many other important respects, some species^^ of the latter genus having a well developed entoconid on the lower sectorial, and being possibly ^^3Usocyon robusius Matthew from the lower Miocene of Soath Dakota bears geueral similarities, but m.^ is apparently too much reduced aud more trenchant, and m^ is probably sometimes absent. ^^Mesocyon Josephi {?). See Merriam, Univ. of Cal. Publications, Bull. Dept. Geol., Vol. V, pp. 19-20, 1906. 260 MEMOIIiS OF THE CARNEGCE MUSEUM more nearly in the ancestral line of Cynodesmus (C. hrachypus). 1 PETKRSON : NEW CARNIVORES FROM MIOCENE OF AYESTERN NEBRASKA 261 From the study of the osteology of Daphccnodon superhus it is quite evident that this genus belongs to a phylum existing through the Ohgocene and Miocene periods, with rather conservative adaptations when compared with such forms as Temnocyon, Mesoojon^ Gynodesmus, Tephrocyori and in fact with many coexisting forms of Canidx. Although the genus should be, for the present at least, regarded as belonging to the subfamily Amphicyoninx, it is not altogether unlikely that, upon gaining more complete knowledge of all the different supposed aUied forms, it will be regarded as representing a separate subfamily {Daph(eninx) paralleling Temnocyon, Mesocyon, Cynadesmiis, and Tephrocyou on the one hand and the true Amphicyon on the other, but less differentiated from Daphcenus than either of the genera regarded as descendants from the Ohgocene forms. The phylum terminated, most probably, in some such forms as Dinocyon or Borophagus, or perhaps in some of the so-called Amphkyones in hiter formations of North America, Avliile the Awphi- cyoninx of Europe were paralleling the line. This would seem to lend color to Mr. Hatcher's contention -Hhat rivtcmnocyon and Proamphicyon are both valid genera, the former pointing to Temnocyon and Meso- cyon, while the descendants of the latter, if not 'Amphicyon'' as Hatcher believed, may not yet have been discovered. The reduction or apparent absence of M^, together with other features in Protemnocyon from the Oligocene, is suggestive of the later John Day form, while M^ in true Daphwnus still persists in an almost unchanged form in Daphccnodon from the Miocene. The bony structure of Dapihcenodon, though extremely cat-like in many respects, is on the whole more closely related to the Canidx, of which family it was an aber- rant line not continued to the present time. A fairly good attempt has been made by Mr. Theodore A. Mills to construct the soft parts of Daphamas felinus, from the skeleton which is now in the Carnegie Museum. And the curious combination of the characters of the cat and the dog are, if anything, even more striking when an allowance of flesh is represented upon the bony structure (see PL LXXXV). The downward extent of the heavy muscles of the limbs, the broad and short feet, with the semi-retractile claws, the long body and tail are especially cat-like, while the head is dog- like in every respect. The tail is represented more or less like that of Fells Goncolor, there being no perceptible means of knowing whether this appendage Was busliy, as in the dogs, or more slender as in cats generally. Altogether the model represented by the figure on PI. LXXXV is instructive, as it furnishes at least a conception of a primitive form ancestral to cats and dogs. Of this primitive stem Daphoenus and especially Daphmnodon on the whole, appear to be far less specialized i 12 W Oligoceue^Cauidic," I. c, p, 105. I :=:*- - = .i....!LLJ'JS."?J 262 MEMOIRS OF THE CARNEGIE MUSEUM than most of their contemporaneous forms and consequently more nearly like the old types of the Miacidse from the Eocene formations. Canid (Undetermined Species). A humerus (No. 2400) of a very large canid, the size of a lion, was found in the lower Harrison beds on Whistle Creek, Sioux County, Nebraska, about eight or ten Fig. 56. Anterior View of Humerus of Canid. Sp. indt. ^ nat. size. No. 2400. miles east of the Agate Spring Fossil Quarries. With the exception of its greater size and the greater distal extent of the deltoid ridge this humerus most nearly resembles that of Daphosnodon. The bone is very nearly the size of the humerus of ArnqMcyon major of Europe and not unlike the latter, so far as comparisons can be made. In the Amer- ican species, however, the deltoid ridge extends lower down and the anconeal fossa is not so high as in the European species. The humerus possibly rep- resents some genus closely related to such largo forms as Dinocyon from the later Tertiary of Montana de- scribed by Mr. E. Douglass ^^ or D. {Borophag7is) gidleyi from the Miocene of Texas described by Dr. W. D. Matthew.^* The quarry in which the hu- merus (fig. 56) was found contains a similar fiiuna to that of the Agate Spring Fossil Quarries. A fifth metacarpal (No. 1897) of a large carnivore, which may belong to this species, was found by the writer in 1904 among the surface fragments in the quarry, which was worked the following years by Professor Barbour (Agate Spring Fossil Fig. 57. Fibular and Anterior Views of Calcaneum of Canid sp. indct. ^ nat. size. No. 2211. " "New Vertebrates from tbe Montana Tertiary," Ann. Car. Mus., Vol. II, p. 192, 1903. ""A Skull of Dinocyon from tbe Miocene of Texas," Bull. Amer. Mus., Vol. XVI, p. 139, 1902. PETERSON : NEW CARNIVORES FROM MIOCENE OF WESTERN NEBRASKA 263 QuarrieS; '' University Hill'' Quarry), This bone is very heavy, shorty and much arched in the dorsal direction. A calcaneum of nearly an equally large carnivore was found in the same horizon north of the Niobrara River in this general locality and is illustrated on page 262, Fig. 57. The remains indicate the presence of these large carnivores in the fauna found in tliese quarries. Mgasueements. Mm. Greatest length of humerns 305 Greatest antero-posterior diameter of head 87 Greatest transverse diameter of head 68 Greatest length of Me. V 63 Greatest length of calcaueiim 90 Greatest trausverse diameter at susteutacular fiicet 45 Borocyon'^ robustum gen. et sp. nov. Type: (No. 1918, Carn. Mus. Cat. Vert. Foss.) Front of right ramus, caudal vertebroSj fragments of manus, part of hind hmb and foot, and various other fragments. Horizon : Miocene, Upper Harrison beds. Locality : Whistle Creek, Sioux County, Nebraska. Mr. T. F. Olcott, while engaged with the Carnegie Museum party in 1905, dis- covered and collected this specimen about four or five miles southeast of the Agate Spring Fossil Quarries. As indicated above it was found in tlie upper Harrison beds. r , Although the type is fragmentary it enables me to give some characters which may be regarded as of generic value. The type is perhaps more closely allied to such forms as Amphicyon sinapius Matthew from later horizons than to A. americanus, A. ursinus, or any of the European forms with which it has been compared. The jaw is undoubtedly quite short and heavy, and the premolars are of considerable size when compared Avith A. ursinus. The portion of the jaw preserved compares fairly well with jElurodon taxoides Hatcher, but the canine is heavier and P^ has no pos- terior cusp as in that species. A. meandrinus had the jaw proportionally longer in front of Py than the present genus. It is, of course, quite provisionally that this genug is placed in the Daphcenus-Daphvnodon phylum, but pending the discovery of more complete material it would seem most satisfactory from the study of the limb and foot structure to regard the genus as a later member of this aben'unt line. The material represents an animal nearly the size of a fully adult lion. Generic characters. Canine proj^ortionally large and oval in cross-section, P^ single- rooted, P2 low-crowned tuith protoconid well anteriorly, no posteriar cusp, and a large * y'ioptSf =r greedy ; /ci'wv^dog. 264 MEMOIRS OF THE CARNEGIE MUSEUM posterior base, M^f present, heavy and elongated tail, limbs and feet vjell proportioned and sub~dih(jenus and Daphasnodon, little or no retractility of the unguals. General Description and Specific Characters. The material represents a fully adult animal judging from the much worn canine. The latter is oval in cross-section and is proportionally larger than that tooth in Daphcenodon. Judging from the abraded anterior face of the canine it is quite evident that I^ was of large si;^e as in Daphcenodon. The postero-external angle of the canine is also worn by the contact with the upper canine. As in Fig. 58. Right Side of Lower Jaw-fragment and Crown View of Teeth of Borocyon rohustnm, Nat. size. No. 1918. Fig. 59. Caudal Vertebra; of Borocyon roluslnm. J nat.? size. No. 1918. 1, lateral and ventral views of caudal 7 ; 2, lateral and ventral views of caudal? 9. Daphcenodon there is a short diastema back of the canine. Px is single-rooted. Pa- is low-crowned and has the protoconid placed more anteriorly than in DaphcfModon, which resiilts in a less developed anterior and greater posterior base than on P^ in the latter genus. There is no posterior tubercle on P^ as in M/arodon, but the cingulum is well developed ; it surrounds the tooth and is smooth on the external side, while internally it is slightly raammilated. There is a large mental foramen directly under the posterior portion of Pg- The symphysis is quite heavy and extends well back as in Daphcenodon. The only parts present representing the vertebral column in the type are a number of caudal vertebrse from different regions of the tail. The centra of these vertebrge are long and heavy, indicating a tail of approximately the same propor- tions as that of Daphosnodon. The manus is represented by a complete trapezoid, Mc. I, and other fragments. The trapezoid is depressed and wide transversely, indicating a broad manus. The iJtff'! i PETKRSON : NEW CARNIVORES FROM MIOCENE OF WESTERN NEBRASKA 265 facet for the scapho-lunar is slightly convex from before backward near the dorso- radial angle, while further back and along the ulnar portion it is concave in the same direction. Radially there is a large facet for the trapezium, which takes up nearly the entire face. On the ulnar side the facet, for the lunar, meets that for the magnum at an acute angle so as to form a prominent ridge, which extends a short distance back from the dorsal border. Back of the ridge the ulnar face is taken up by a large rugose area for the at- tachment of muscles and ligaments. On the distal face there is a large subtriangular articulation for the second metacarpal. The free end of the pisiform indicates this bone to be quite like that in Dapliwnodon. Mc. I has the head less expanded transversely, the shaft less arched forward, and less curved outward at the distal end than that of Daphwnodon superbus ; there is otherwise little or no difference between this bone in the two forms. Other fragments of the metacarpals suggest heavy bones in this region. The phalanges are proportionally short when compared Fia. 60. Foot Bones of Borocyoii rohustum. I nat. size. No. 1918. 1, dorsal view o£ metacarpal I ; 2, dorsal view of a second pbalaux ; 3, dorsal view of a first phalanx ; 4, proximal view of trapezoid ; 5, dorsal view of trapezoid ; 6, radial view of trapezoid ; 7, uluar view of trapezoid. with Baplicenodon. Those of the proximal row are other- wise arched in the dorsal direction similar to the conditions found in -the latter genus, while the median phalanx is depressed and broad. There is no ungual pha- lanx present with the type. The hind limb is represented by the lower half of the tibia and fibula, the astragalus, navicular, ento-, meso-, and ectocuneiforms, and other fragments. With the exception of the proportionately somewhat greater antero-posterior diameter of the distal end, the tibia offers httle or no characteristic differences by which it may be distinguished from that of the smaller genus. The distal end of the fibula is more compressed laterally than in Daphcrnodon ; otherwise there is also little difference in this bone in the two forms. The astragalus has a slightly longer neck and the articular surface of the trochlea descends lower than in Daphcenodon, otherwise this bone does not differ in tlie two genera. The principal difference of the navicular from that of Daplmno- don is seen in the smaller development of the plantar tuberosity ; the rugose area for muscular attachment slanting more upward and backward, resulting in the com- paratively small vertical diameter of the tuberosity. The entocuneiform is quite large, and of good proportionate vertical diameter. The tibial plantar angleismore i> 266 MEMOIRS OF THE CARNEGIE MUSEUM concave in the vertical direction than in Daphosnodon, which results from the greater development of the ascending plantar tuberosity of the present form. The meso- and ectocuneiforms are similar to those in Daphcenodon. Metatarsal III has the articular surface for the ectocuneiform less oblique and the tibial and fibular faces of the head less exca- vated than in Daphcenodon. Mt. V is quite unlike that of the latter genus in respect to the articular facets of its head. The tibial face has a somewhat smaller facet for Mt. IV and the facet for the cuboid is entirely proximal, while in Daphmmis and Daphamodon the facet for the cuboid extends well in the fibular di- rectioUj encroaching on the base of the ascend- ing process so as to form a lateral as well as a proximal contact with the cuboid. Another important difference from the older forms is seen in the vertical position of the ascending process on the fibular face of the head. In the Oligocene and Miocene forms, here used for comparison, this process is very heavy and projects upward ai;id outward forming, with the excavation for the peroneus longus tendon on the cuboid, a transversely broad groove. In the present genus the ascend- ing process, though large, is directed more vertically and the space between it and the fibular face of the cuboid is comparatively narrow. Near the proximal ends the shafts of the meatarsals are slightly triangular in cross-section. The distal end of Mt. V? is somewhat less expanded transversely than in Daphosnodon, but the trochlea is quite similar^ being very convex dorsally and separated by a prominent cai'ina on the plantar aspect. Measurements. >3*-l -_^2.t /_ \,.--^jrjr-i^, ^'^-i. 268 MEMOIRS OF THE CARITEGIE MUSEUM Fig. 62. Pelvis and Feet ot Cynodesmus thomfioni Matthew. Nat. size. No. 1591. 1, anterior view of manus ; 2, external view of pelvis ; 3, anterior view of pes. When the skull of the present specimen (No. 1591) is compared with Matthew's description and figures {I. c, pp. 186-188) it is seen that in practically all the details of structure, except the appai'ently smaller brains in our specimen, which may be due to crushing, the two specimens agree quite closely. P^ in the Carnegie Mu- seum specimen has no posterior cusp, while Dr. Matthew's figure indicates one on the type. The present specimen represents an old individual, as the tubercular teeth are much worn. Tephrocyon (Canis) temerarius Leidy. i Proc. Acad. Nat. Sci. Phila., p. 21, 1858; Jour. Acad. Nat. ScL, Vol. VII, p. 29, PL I, fig. 12, 1869. A fragment of a lower jaw (No. 2404, Carn. Mus. Cat. Vert. Foss.) is provision- ally referred to this species as the measure- ments and general appearances are quite FIG. 63. Crown View of Teeth and External View ^^^^^ ^^ ^^^^^^ ^^^ ^j^^ ,^,j^g ^^^^^^^ of Jaw-fragment of Tephrocyon (Canis) temerarius Leidy, Kat. size. Specimen was fouiid on Whistle Creek, Sioux *-^^i. > PETERSON : NEW CARNIVORES FROM MIOCENE OF WESTERN NEBRASKA 269 L ^^ ^^ ^^ County, Nebraska ; it was lying on the surface and possibly pertains to a later Miocene or Pliocene deposit which apparently thinly covers the lower Miocene section in certain places in this valley. The accompanying figures are given in order to show the close similarity of this species to T. rurestris JMerriam and T. hip- pophagus Matthew and Cook.^* The only differences between the present form and , T. hippopJiugus so far as comparison can be made, is the smaller size and the slightly slenderer jaw of T. temerarius. MuSTELID.li:. Paroligobum's (Brachypsalis) simplicidens gen. nov. Annals Carnegie Museum Vol. IV, pp. 44-46, 1906. Tyjje: Right mandibular ramus (No. 1553, Carn. Mus. Cat. Vert. Foss.). Horizon: Lower Harrison beds (Miocene). Locality : Near Agate Spring Fossil Quarries, Sioux County, Nebraska. Plesiotype : Left mandibular ramus, greater portion of femur, fibula, and bones of fore and hind foot (No. 2389, Carn. Mus. Cat. Vert. Foss.). Horizon : Lower Harrison beds (Miocene). Locality : Agate Spring Fossil Quarries (quarry No. 3), Sioux County, Nebraska. Generic Characters. ~ Dentition, ^JJ:^- Canine heavy and oblong in cross-section. Premolars crowded, heavy, rather bluntly cro'wned, no accessory posterior cusps except on p^. (Jarnassial comparatively large, approaching condition in recent forms {Gido), but possesses a small entoconid which is located high up on the croivn. M^ much reduced in size. Femur longer than the tibia in the same proportion to that in Meles or Gulo. Feet broad and short; pes sub-plantigrade; unguals not retractile. General Description of the Plesiotyfe and Specific Characters. The lower jaw is only very slightly larger than that in the type of this species previously described and is much larger than Meles taxus. The mandible below p^ is very deep and the posterior mental foramen is located directly below the posterior fang of p;^. Judging from the bottom of the alveolus for the canine, that tooth was strongly developed and the premolars appear to have been crowded, as in the type. The inferior border of the horizontal ramus is gently convex in the fore-and-aft direction, as in the type. Opposite the vertical ramus the inferior border is only slightly raised as in Gulo luscus and unlike that in 3Ieles, in which the angle is higher. The posterior portion of the lower jaw as a whole is otherwise very like that of the latter genera ; possessing a large and deep tempoj-al fossa, a large ^'Matthew, W. D., and Cook, HaroldJ., Bull. Amer. Mas. Nat. Hist., Vol. XX Vl, p. 373, 1909. 270 MEMOIES OF THE CARNEGIE MUSEUM coronoid process, transversely broad and well rounded condylej heavy masseter ridge, and prominent liookdike angle. The condyle is on a horizontal line with the denti- tion and is slightly lower than in Mela taxus and more like that of Gulo luscus. The ramus is broken in the region of py but this tooth is present in the type'. P.j is Fig. 64. Lower Jaw of ParoUgobunis simplicidens. § uat. size. 1, internal view of jaw ; 2, external view of jaw, of considerably large size judging from the posterior root, which is still in position in the jaw. F^ is robust, and, as in the type, has a great transverse diameter with the protoconid placed well anteriorly. Internally there is a smoothly rounded cingulum, which continues from the antero-internal angle of the tooth, as in Gulo luscus, and is somewhat more developed than in the type. The crown of the tooth is much abraded showing that the animal was fully adult. Back of p-;; is a space of 11 mm. for P4. My is unfortunately broken anteriorly so that the blade cannot be described. The metaconid is preserved and is rather small and placed high up on the crown. The heel is of moderately large size with the hypoconid well developed, forming the external edge of the heel, while the entoconid is probably absent.^*"' Back of my there is a small alveolus, indicating a single-rooted tooth, which was even smaller than that tooth in the type. The contour of the lower jaw and especially the denti- tion are on the whole rather more suggestive of Gulo luscus than Meles taxus. ^^The abraded condition of the tooth does not admit of a positive statement as to the presence or absence of an entoconid. H- PETERSON: NEW CARNIVORES FROM MIOCENE OF WESTERN NEBRASKA 271 Measurements of Mandible. Mm. Total length of the rarans fragment 110 r A ntero -posterior Jianieter of P^- 11 Greatest transverse diameter of P3- 8 , Autero-posterior diameter of M^ approximately 18 Greatest transverse diameter of heel 8 Manns. 272 MEMOIRS OF THE CARNEGIE MUSEUM slants excessively downward in the ulnar direction, meeting the facet for Mc. II at an acute angle, so as to form a prominent and sharp ridge, which extends a short distance from the dorsal face backward almost identical with that in Meles and Lutra. Back of this ridge on the ulnar face is a rugose area for muscular attach- , ments, the bone terminating palmarly in a round obtuse point. Distally the sur- face for Mc. II is convex transversely, the radial side less strongly than the ulnar, while antero-posterioriy the articulation is very slightly concave. The radial face is taken up almost entirely by the large facet for the trapezium. The latter is undoubtedly of large size judging from the large facets on the trapezoid and on Mc. I. The first metacarpal is rather short when compared with that of Gulo, or even with that of the recent badger, and is more like that in the otter; it is also proportion- ally shorter than that in Megalictis ferox Matthew from the lower Miocene of South Dakota. The proportionate length of the metacarpals as a whole agree quite closely, however, with those in the latter genus. The facet for the trapezium is, as stated, of large size, and it is obliquely convex antero-posterioriy, extending well down on the shaft as in Meles, and is not unlike that of Megalictis feroxM-^Xi\\Q^N 2.\\d JEluro- cyon hrevifacies Peterson.^' The shaft has received some injury in the process of col- lecting, but enough is preserved to show that it was heavy, well arched forward, and quite cylindrical. The distal trochlea is fully developed and carried the usual two phalanges. The second metacarpal is quite heavy and in its general proportion it compares well with that in Megalictis. The head rises above that on Mc. Ill, so that the ulnar face has a facet for the magnum similar to that in 3Ielcs, while in Megalictis the head is apparently raised higher (see illustration, lig. 15, of Dr. Matthew's paper, I. c). The facet for the trapezoid is concave transversely, the ulnar angle being higher than the radial. The palmar tuberosity is a hemispherical knob, while radially the head has an uneven surface for muscular attachment. On Mc. II of the right side there is no articular facet for Mc. I, while the one on the left side r has a small facet near the dorsal angle. The origin of this facet may, in part, be due to injury received in this region, which is plainly shown by the rough surface of the proximal portion of the bone. The shaft is very heavy, slightly arched forward, and transversely oval in cross-section. The dorsal portion of the trochlea is not nearly as hemispherical as in Daphccnus, while plantarly it is divided by a prominent carina, the radial portion occupying a higher position than the ulnar, causing asymmetry of the trochlea. 27 See fig. 19, 5, Ann. Carn. Mus., Vol. IV, p. 71, 190G. PETERSON" : NEW CAENIVORES FROM MIOCENE OF WESTERN NEBRASKA 273 The tliird metacarpal is no heavier than the second, though longer. The articu- lar facet for the magnum slants inwardly, but the radial border is not raised as in Daj)hwnodon so as to form a transverse concavit}^, but instead it is very nearly plane transversely and evenly convex antero-posteriorly. On the radial side is a facet for Mc. II which is succeeded, posteriorly, by a deep sulcus, the palmar tuber- osity forming its posterior border. . On the ulnar side the proximal end is ex- cavated in order to accommodate the head of Mc. IV and the articular facet for the unciform is concave due chiefly to the development of the dorso-ulnar angle of the head. The shaft is very slighty arched and transversely oval, due to the divergence of the metacarpals directly from the proximal end as in themustelids generally. The troclilca is more symmetrical than that of Mc. II< Measurements. Mm, * Unciform, greatest antero-poaterior diameter 13 TJnciformj greatest transverse diameter 12 Unciform, greatest vertical diameter 15 Trapezoid, greatest antero-posterior diameter 10 Trapezoid, greatest transverse diameter - 9 Trapezoid^ greatest vertical diameter 6 Metacarpal I, greatest length 26 Met^iCiirpal II, greatest length 37 Metacarpal III, greatest length 41 Median phalanx, greatest length 17 4 A proximal plialanx of the pollcx or hallux has the proximal end much ex- panded, the shaft strongly arched and of considerable length, while the distal trochlea is confined principally to the plantar face. A median phalanx found with the specimen is much depressed, quite long, and does not indicate retractility of the ungual. The Hind Limb. The greater portion of the femur is preserved. The head is well rounded and sits on a long neck, as in 3Ieles, Lutra, and Gulo. The pit for the ligaraentuni teres is located on the posterior half near the lower border. Unfortunately the great tro- chanter, together with the upper portion of the shaft, is lost on its external side, while internally the shaft is preserved, showing a large lesser trochanter. On the lower half of the shaft the fibular border is produced into a very prominent ridge, which is even more pronounced than in 3Ieles faxus or Gulo luscus and terminates below in a rugose tubercle similar to that of Canis. The lateral development of the shaft gives it the characteristic transversely broad appearance seen in Gulo hiscusor Meles taxus. The rotular trochlea is shallow and broad, the condyles have great trans- 274 MEMOIKS or THE CARNEGIE MUSEUM verse diameter and are separated by a rather shallow groove, the external condyle being quite oblique to the long axis of the bone. The fibula is heavy and has the proximal and distal ends much expanded. The shaft sweeps gently towards the shaft of the tibia and is also gently arched in the backward direction. The facet for the astragalus is nearly plane and the pos- terior face has the tendinal groove and external malleolus fully as well developed as in the recent badger or otter. Fig. 67. Limb Bonea of Faroligohunis simplicidens. | uat. size. 1, anterior view o£ femur ; 2, posterior view of femur ; 3, fibular view of fibula ; 4, anterior view of fibula. Both astragali are represented. In comparison with the length of the bone the trochlea is somewhat narrower than that of Megalictis ferox. The trochlea is very shallow and is more oblique and imperfect than in either Meles or TAitra and rather more like that of Gulo luscus, the internal condyle being less developed, while the exter- nal is higher than in the two recent forms first mentioned. The tibial face of the inter- nal condyle is not interrupted by the sulcus, which is so characteristic of Daphcenodon, but is continuous as in Gulo and Meles. The pos- terior limit of the trochlea, as well as its down- ward extent on the neck of the anterior face, is approximately in the same propor- Fio. 68. Astragalus of ParoUgohunis simplici- dens. f nafc. size. 1, anterior view; 2, superior view ; 3, tibial view. PETERSON : NEW CARNIVORES FROM MIOCENE OF WESTERN NEBRASKA 275 tion to that in the recent forms. One of the more curious features of the bone is the deep and obhquely placed sulcus, which separates the groove for the flexor muscles from the astragalar trochlea. In this groove is located the astragalar fora- men (see fig. 68, 2), which pierces the bone and again appears in the groove be- tween the internal and external facets for the calcaneum on the plantar face, as in 3Ieles taxus. The neck is perhaps somewhat longer than in Megallctis, which again reveals a tendency towards Gulo hcscus. The head is also expanded and convex in the same way. A mesocuneiform which was found with the specimen and very likely belongs with itj is high and narrow when compared with Meles, and is apparently also slightly higher in proportion to that of Megalictis, but is otherwise quite like that of the latter, the tibial face possessing a large facet for the entocuneiform. Measurements. Mm, Femur, length from head to distal end I74 Femur, transverse diameter of distal end , 42 Femuij autero-posterior diameter of distal end 36 Fibula, greatest length 130 Astragalus, greatest length 33 Astragalus, greatest transverse diameter 2q Astragalus, greatest transverse diameter of trochlea 16 Mesocuneiform, greatest anteroposterior diameter \\ Mesocuneiform, greatest transverse diameter q Mesocuneiform, greatest vertical diameter g Affinity and Systematic Position of Paroligohunis. While some features of the limbs in Paroligohunis nmplicidens suggest tliat of Lutra ; i. e., the oblique and limited posterior extent of the internal condyle of the femur, the backward sweep of the shaft of the fibula, the proportionate length and arrangement of the metacarpals; tliere are more important characters which very strongly suggest that the animal was probably fossorial rather than aquatic in his habits. The femur is much longer than the fibula and the shaft of the former bone is straight and is transversly expanded near the distal end as in IJelcs. The fibula is on the whole more nearly like that of the latter genus than that of the otter, while the astragalus strongly suggests that of Gulo lusais and has the astragalar foramen, a characteristic feature in Meles. The John Day form, Oligohunis, w^as regarded as an aberrant member of the Ganidx by Professor Cope, while Dr. Matthew (Bull. Amer. Mus., Vol. XXIIT, p. 193, 1907) after the re-examination of the type, questions Cope's reference and transfers the genus to the Mustelidae. After a compar- ison and careful study of the newly acquired material now in the Carnegie Museum I 1 276 MEMOIRS OF THE CARNEGIE MUSEUM i }\ ;J h \ -^^^0r^^rf^^^:.:i^^r^^^^^:^<%^f^i^^^ *t PETERSON : NEW CARNIVORES FROM MIOCENE OF WESTERN NEBRASKA 277 Matthew and Gidley doubtfully referred this type to Potam.otherium. As this Ameri- can form now stands I cannot see that it is much nearer Potmnotherium than Sleno- gale as the jaw is deeper in proportion ; it is also somewhat shorter and more ro- bust, the symphysis is ligliter, the premolars are smaller and M^ larger than in the former genus, while the latter genus has the under border of the lower jaw very piuch more curved fore-and-aft, the premolars are too large in proportion to the car- nassial, the heel of the latter is entirely different in shape, and very much smaller, and the tubercles of the teeth are evidently sharper and higher than in the American specimen. The latter form should, I think, occupy a new generic position {B^m^ Stheni g^ somewhat near the O^o^nm-is phylum although the premolui-s are proportion- ally smaller than in the latter, especially in the transverse diameter. In Paroligo- hunis the jaw is proportionally shorter and heavier, the alveolar border is more curved from before backward, the premolars are more crowded, and the carnassial has apparently a shorter heel than in the new genus here proposed. Fig. 69. Alveolar Border Eind External View of Jaw of Brachijpsalis jxici/cephahis Cope. Nat size fConp col lection, No. 8544, Am. Mus. Natural History.) Brachnjpmlis pachycephahts Cope was proposed (/. c, p. 951) on a rather inade- quate type, a left mandibular ramus, from the same locality in which Slenoqale Tohusta was found. Professor Cope afterwards referred this type to Potmnotherium of Europe, while Matthew and Gidley (/. c, p. 254) accept the type as valid and distinct from Poiamoilieriiim. In a former paper ^" the present writer re<'-arded Paro- ligohunis {BrachypsaUs) dmplicidens as possibly an ancestral form to Brachypscilis, a view which is no longer tenable inasmuch as the latter species has M^ proportionally larger, and the carnassial apparently smaller in proportion to the premolars. (Compare figs. 64 and 65 with fig. 09.) r so Ami. Carn. Mu3-, Vol. IV, p. 45, 190(1. li (II n^. I I 278 MEMOIRS OF THE CARNEGIE MUSEUM JEliirocyon Peterson and MegalicMs Matthew are more modified forms which, however, are apparently related to this group of the Mustelidm. One of the chief generic differences between JiMurocyon, Megalidis, Oligohunis, and Paroliyohunis is the presence of the metaconid on My of the two latter and the absence of it in the case of the two former genera. In Paroliyohunis the metaconid is situated nearl^^ as high up as the summit of the protoconid, and quite closely fused to the latter. Some of the Siwahk mustelids (Melivoradon) are also closely allied to this general group. I fully agree with Dr. Matthew's"" statements that Mellivora and Gulo are the nearest living allies of jElurodon and Megalictis; I would in fact also include Oligohunis, Paroliyohunis, and possibly the proposed new genus B'i'ddh'Uuuh as n(*t so very distantly related especially to jElurocyon and Megalictis. 8' Bull. Amer. Mus. Nat. Hiat., Vol. XXIII, p. 204, 1907. Bthenictie I