MEMOIRS OF THE CARNEGIE MUSEUM. VOL. IV. NO. 3. A REVISION OF THE ENTELODONTIDyE.' By O. A. Peterson. Introductory Remarks. Since A. Aymard (1, pp. 227-242)- and S. A. Pomel (73, p. 307 ; 74, p. 1083) described the genus Entclodon from the Tertiary deposits of France, much material has been found, which represents this unique family of mammals, especially in the Oligoceneand Miocene formations of the North American Tertiary. The object of the present Memoir is first to give a systematic review of the known genera and species of this family ; and secondly to describe and illustrate in detail the type specimen of Dinohyus hoUandi, which was discovered in the Agate Spring Fossil (Quarries, Sioux County, Nebraska, by the Carnegie Museum Expedition of 1 905, and briefly described in Science (78, pp. 211-212) and the Annals of the Carnegie Museum (81, pp. 49-51). At the very outset of his work the writer became fully aware of the fact that generic and specific determinations in this family have sometimes been based on rather inadequate types, which present few and unsatisfactory characters. Frag- 'Pomel's (lesoription (73, 74) of Elotherium did probably appear before that of Aymard on Fnlelodon, but, iiias- luuch as the type of the former was rather inadequate, no illustrations were published, and the type has been since lost (see page 43), the present writer is of the opinion that the latter name should be used, as both text and figures are clear. ' For the references in parentheses, see the Bibliography appended. ( Publislied May, 1909.) 41 42 MEMOIRS OF THE CARNEGIE MUSEUM mentary types, which very often are most exasperating to the student of paleontol- ogy, cannot be regarded as finally determining genera and species, and in the pres- ent case we must still await the slow process of discovery before a number of questions can be satisfactorily determined. The paper here presented can only claim to be a report of progress upon our knowledge'of this most interesting family, which must be regarded as very distantly related to the Suidie. The types of the different genera and species of the family, though sometimes very fragmentary, clearly indicate diverging lines. These lines do not possess the radical variability met with for instance in the Merycoidodonts (Oreodonts) and the Cameloidea, but they are nevertheless of importance, when the conservative charac- ters of the family as a whole throughout the lower Oligocene are considered. This paper has l)een considerably delayed by the fact that the writer in the early part of the year 1908 was suddenly detailed to take charge of the field-work in the Agate Spring Fossil Quarries. The labors in these quarries were rewarded with unusually good success, as has been the case every season since they were first opened by the Carnegie Museum. Among other important material were found a numberof portions of individuals of the genus Dinohijus which are of great value, as they help to clear up certain anatomical features of that genus which otherwise would have been left in obscurity in the present puljlication. For the privilege of study and for aid otherwise given me, I am indebted to Dr. W. J. Holland, Director of the Carnegie Museum, at whose request the present work was undertaken. To Professors Osborn, Scott, Schuchert, and Lull I am indebted for free access to the splendid material of the American Museum of Natural History, the Princeton Museum, and the collection of the late Professor Marsh of Yale. To Dr. W. D. Matthew of the American Museum, Dr. Witmer Stone of the Academy of Natural Sciences, Philadelphia, Mr. J. W. Gidley of the National Museum, Dr. M. S. Farr of Princeton University, Professor John C. Merriam, and Mr. E. L. Furlong of the California University, and Miss Lucy P. Bush recently of Yale Museum, I here wish to extend my thanks and appreciation for important letters and information. The excellent series of illustrations in the present paper are mainly due to the skilful work of Mr. Sydney Prentice. These illustrations explain many features which, without them, would be quite unintelligible. Most of the photographs except when otherwise stated, were made by Mr. Arthur S. Coggeshall. Last, but not least, hearty acknowledgment should be expressed to Mr. Andrew ( Jarnegie, the founder of the Listitute, whose well-known generosity has made it possible to carry on the work done both in the field and the laboratory. PETERSON : A REVISION OF THE ENTELODONTID.T; 43 Family ENTELODONTID.E Lydekker.^ The characters of this group of mammals from the Oligocene and Miocene for- mations of both the Old and New Worlds amply justified Lydekker (53b), Marsh (64, p. 408), and Scott (87, p. 322) in placing them in a separate family. The com- bination of the primitive bunodont dental structure with the highly modified limbs and feet gives to this group an unusual and quite unique appearance. Family Characters : Teeth bunodont. Muzzle long ; cranium short. Limbs elongated ; feet didactyl. Genus ENTELODON Aymard. There has been much confusion both in Europe and America regarding the pri- ority of the names Entehdon Aymard and Elolher'mm Pomel. Quite recently Miss Lucy P. Bush, who was one of Professor Marsh's assistants, wrote an article (4, pp. 97-98) which stimulated inquiry as to tliese two names. After carefully looking over the literature, it is quite plain that Aymard's paper on Entehdon, though per- haps written in 1846, was not published until late in 1847 or during the year 1848. The most conclusive evidence of this fact is found on page 247 of Vol. XII, of the Annales de la Societe d' Agriculture, des Sciences, et de (Jommerce du Puy for 1842- 1846, or on p. 23 in the reprint of this article. On these pages of Aymard's paper is a foot-note correctly referring to page 385 of the Bulletin de la Societe Geologique de France, Vol. IV, which was published late in 1847 ; consequently Aymard's paper could not have been published in 1846. Pomel's description of Elothermm on the other hand was apparently published in 1847 ; Ijut, in the first place, the type was rather inadequate, judging from Pomel's article ; secondly, there is not in Pomel's paper, or elsewhere, any illustration of the specimen ; and, thirdly, I am informed that the type is lost.^ These facts collectively should, in the mind of the writer, exclude the use of the name " Elothermm" especially since Aymard's article on Entelodon is accompanied with good illustrations of satisfactory types. There was apparently not a great lapse of time between the appearance of the two publi- cations, and it may yet possibly be established that Pomel's description was also delayed and not published l)efore Aymard's paper actually appeared. Principal Generic Uiaractcrs of Entelodon : Upper and lower premolai-s relatively 'The family name EnModonlidir was apparently first used by Lydeljlver in 1883 (53b, p. 146), while Edward Richard Alston in the Zoological Record for 1876, p. 18, refers to ParahyuK mgans, Ueloliyus, and " Elolherium " under the caption Elne associated with jaws and skulls from the type locality is of extreme impor- tance in connection with the study of this subgenus. Pelonax bathrodon (Marsh). Tijpe : M3 of right side. Horizon: Upper 01 igocene. Protoceras Sandstones? "In a letter of reply from Dr. Matthew, of the American Museum of Natural History, New York, dated Feb. 2C, 1908, I am assured that Coin's statement regarding the single-rooted premolars is correct. " No true Entelodonts from the Oligocene of America or Europe have as yet been_found with four digits on all the feet, as in hippopotamus. 58 MEMOIRS OF THE CARNEGIE MUSEUM Locality : Bad-lands of South Dakota. Locality of Type : Collection of Yale Museum, No. 1 2030. The tooth figured in 1893 by Professor Marsh (63, PI. IX, fig. 4), and referred to the genus Ammodon, though inadequate as a type, fortunately shows characters which at once separate it from Ammodon leidyanum, not only specifically, but also generically (see fig. 13). In Marsh's original description in 1874 (59, p. 534), he gives the principal characters of M3 of his proposed species •' Elotherium " bathrodon, which constitute the chief differences between that tooth of the latter species and the genus ^mmo(7o7i. Marsh says : "This molar differs essentially from the same Fig. 13. Last Lower Molar of P. bathrodiin Marsh. \ nat size. (After Marsh.) tooth in the other known species of this genus [Elotherium'], especially in having the anterior pair of tubercles much larger than the posterior pair, and elevated high above them." '* On comparing Marsh's figures reproduced in this connection (fig. 13) it is at once seen that the tubercles are more distinctly separated by narrow longitudinal valleys, and that the cross-valley between the anterior and posterior tubercles is open and very broad, which is due to the small development of the pos- terior tubercles ; a character common to the older American types of the family Entelodontidse. In Ammodon and Dinohyus the development of the posterior tuber- cles (ento- and hypoconids) is relatively greater and the cross-valley is very much reduced in the anteroposterior direction. The distinct development of the fifth cone (hypoconid) in Ammodon leidyanum seems to have progressed in the same gen- eral ratio as that of the posterior tubercles. In the Yale Museum is a specimen of an Entelodont, which was collected in 1890 by Mr. J. Brown and bears the catalogue number 10U08. This specimen is a skull of a large individual apparently from the Protoceras sandstones of the Oligo- cene. The front and back of this skull is broken off so that its length cannot be ascertained. The size of the molars corresponds quite well with that of the tooth described by Marsh as the type of Pelonax bathrodon, and, from the fact that it was '* This elevated position of the anterior portion of the crown seems to be greater than in Pelonax potens. I'ETERSON : A REVISION OF THE ENTELODONTIDyE 59 found in the upper Oligocene, it would seem quite reasonable to suppose that it belongs to the same species. The premolars of this skull are rather pointed, with wrinkled enamel, and with comparatively small antero-posterior diameter, approaching Dinohyus in. these char- acters, pi has a small paracone and a very strong cinguluin, which nearly surrounds the tooth ; the external face is quite smooth, except the posterior and anterior angles. The molars have distinctly separated tubercles and rather broad or open cross- valleys. Opposite the cross-valley on the internal faces of M^ and M^ the heavy rounded cingula, or swellings, which are so i)rominent in Dinohijus hollandi, are indicated. The anterior tubercles of M^ are well separated, while the posterior tubercles are fused and elevated posteriorly, so as to form a basin of the cross-valley ; the tooth as a whole is relatively small when compared with Dinoliyus. The frontals are much less inflated than in the type of Dinohyus. Judging from the crushed condition of the skull, I am inclined to think that the anterior border of the orbit is opposite the posterior part of M^, or perhaps somewhat further back. The anterior border of the posterior nares is opposite the posterior part of M-. The dependent process of the jugal is very broad and extends well down. A second specimen in the Yale Museum, which may be referred to Pelonax bathrodon (No. 10009), is a skull and lower jaws of a rather young individual. This has an unusual development of the posterior basal tubercles (hypoconulids) of the lower molars. The ento- and hypoconidsof My and Mt,- are also" relatively high in comparison with most Entelodonts from the Oligocene, while those of M3 are low as in the type of Pelonax bathrodon and the cross-valley is also broad in the same manner. Measurements. Skull No. 10008, Yale Museum. Mm. Total length of the skull fragment about 540 Breadth of .skull at M^ I44 Antero-posterior diameter of P^ 32 Transverse " " pa jg "Pi 28 Antero-posterior " " pa 4j 5 "PA ....'..^.^i 31 Transverse " " Pt 34 "Ml ""!''^""!'!!""!'""' 40 Antero-posterior " "Mi 38 " " " M" 39..5 Transverse " " M^ 43 "Ml ' \ 37 Antero-posterior " " M^ 533 60 MEMOIRS OF THE CARNE(iIE MUSEUM Pelonax potens (Marsh). Type : Greater portion of left ramus and symphysis. Horizon : Not indicated ; most probably upper Oligocene. Locality : Oligocene of Colorado. Locality of Type: Collection of Yale.Museum, No. 12042. The most conspicuous feature of this type is the gi-eat development of the ante- rior dependent process of the inferior face of the lower jaw, which, as in Pelonax ramosum, gives to the chin an unusually short and rather broad appearance. The alveolus for Vj indicates that the roots were welP'' coalesced near the crown, a character usually found in types from the upper Oligocene, and which should be regarded as of subgeneric value. The antero-posterior diameter of Y^ must have Fig. 14. Type of /". poiens Marsh. About J nat. size. (Yale Museum, No. 12042. ) been quite great, judging from the space which the alveolus occupies. The molars indicate a smaller individual than Pelonax hathrodon, as stated by Marsh (63, p. ^410), but possessing the same characters as the latter, /. e., the anterior- tubercles of the molars are higher than the posterior, and the latter separated from the former by broad, transverse valleys. ^Vhen more complete material from tlie Oligocene of Col- orado is found, the validity of the type of Pelonax potens will be more satisfactorily established. "In Pelonnx ramosum Py and P., are apparently single-rooted. PETERSON : A REVISION OF THE ENTELODONTID^ 61 Subgenus (?) BOOCHCERUS Cope. Boochoerus humerosus Cope. Type: A portion of a skeleton without the skull or teeth Horizon: (Upper?) John Day For- mation. Locality : John Day River, Oregon. Locality of Type : The American Mu- seum of Natural History, Cope Collection, No. 7380. The genus Bobchwrus was proposed by Cope and the type was fully described in 1879 (10, pp. 59-66). This John Day form has been regarded as belonging to Archxotherium by authors who have occa- sionally referred to it, and it is even now difficult to correctly place the remains generically, as there are unfortunately no remains of the skull or teeth with the type. The present arrangement must, therefore, be regarded as only provisional, pending the discovery of more material in the typi- cal locality. After restudying the type, and after comparing it with the splendid skeleton of Dinohyus hollandl from the Agate Spring Fossil Quarries, as well as with the Prince- ton specimen of Arclucoilierium ingens from the Titanotherium beds, which was de- scribed by Professor Scott, it appears that there are certain characters of generic value. When the skeleton of Dicodon is thoroughly known, Boikhccrus may pf)S- sibly have to be referred to that genus. Principal characters of Boochosrus Im- merosxis : On comparing BoocJuenis humeroms with I)inohyu.'< hollandl it is quite ap- parent that the humerus in the former is relatively long and the manus'" broad *" Although B. hmnerostis is heavier than /). holhindi, the feet of the former are considerably shorter and broader than in the latter. Fig. 15. Front and Inside View of Right Forefoot of Z(. AmiifrosKS Cope. About J nat. size. (Cope Collection, No. 7380.) 62 MEMOIRS OF THE CARNEGIE MUSEUM (21, p. 168) and short (see figs. 15 and 16). The absence of an articular facet for rat. IIP' on the lower tibial angle of the cuboid in the John Day form is also a noticeable feature. Boochwrus hwnerosus differs from Archfvothermm ingens by the absence of the trapezium and by the fact that the unciform and magnum do not Fig. 16. Lett Humerus of B. humeroms Cope. About 5 nat. size. (Cope Collection, No. 7380.) Fio. 17. Eight Hiud Foot of B. humerosus Cope. (Cope Collec- tion, No. 7380.) I nat. size. touch one another dorsally when in position in the carpus." From both Dinohyus hollandi ond Archxotherium ingens the John Day subgenus differs by a relatively greater transverse diameter of the distal end of the femur. Additional specimens from the John Day formation will undoubtedly give more complete information regarding this proposed genus. Cope states (10, p. 60) that the " remains indicate an animal of the size of the Rhinoceros indicus. The animal is characterized by the massive proportions of the humerus as compared with the femur, and by the short robust form of the metapodials." '' The head of mt. Ill in the type of Bnockcrrus is restored on the fibular angle, but it was evidently the same as in Archneoiherinm ingens, judging from the absence of the facet on the lower tibial angle of the cuboid. "^ Cope says (21, p. 171) that the lunar has penetrated so far as to almost divide the second row of carpals. PETERSON : A REVISION OF THE ENTELODONTIDJE 63 Genus D^ODON Cope. A description of this genus was given by Professor ( 'ope in a paper (9, p. 15) read before the American Philosophical Society, November 15, 1878, and published on December 3, of the same year. Owing to the very poor condition of the type, Cope Fia. 18. Type of Dxodon shoshonensis Cope. About J nat. aize. Cope Collection, No. 7387. Oblique View to show symphysis. was unable to give any characters, except the absence of the osseous tuberosities on the anterior under borders of the lower jaws, which differ from those of the Ameri- can Entelodonts generally. Cope placed Dscodon in the suborder Perissodactyla and associated it with Titanothermm and Chalicothermm. The genus has hitherto been regarded as belonging to the Perissodactyla and Zittel (99, p. 304) referred to it with a misprint under the name Duledon in connection with the .synonymy of Titanothermm. The fortunate discovery of additional material, which undoubtedly belongs to this genus, by the California University and published by Dr. W. J. Sinclair (89, pp. 132-134), furnishes much welcome light on the subject, and it now appears that Dasodon may confidently be removed from the Perissodactyla and placed in the family Entelodontidx. Principal Generic Characters of Dxodon : Chin without knob-like processes on the under border ; posterior mandibular protuberance small and hollow; a grad- ual backward slope of the dependent angle of the lower jaw; jugal process with a moderate downward extension ; fibula not'^'^ coossified with the tibia. " In a letter from Dr. E. L. Furlong, of the California University, Berkeley, California, he assures me that " there is no fusion of the shaft of the fibula with the tibia." 64 MEMOIRS OF THE OARNKOIE MUSEUM Dseodon shoshonensis Cope. Type : Symphysis of lower jaws. Horizon : (Upper?) John Day Formation. Locality : Bridge Creek, Wasco County, Oregon. Locality of Type: The American Museum of Natural History ((-ope Collec- tion, No. 7387). The principal characteristics of the type are the absence of the osseous bosses on the chin and the evenly rounded under surface of the symphysis. Only the roots of the teeth are preserved iti the alveolar border of the type. These show that there were three incisors, a canine, and two premolars. The roots of the median pair of incisors are laterally compressed and indicate that they were small as in Dinolvyus hollandi. 1^ and Ig- were evidently much increased in size, Px is close to the canine and Po^ is separated from Py by a short diastema as in Dinohyus. Tliat Professor Cope had compared the type of Dxodon with ArcJueotherium is evident from his statement (9, p. 15) that "the canine teeth are very robust, as in the species of Elotherium," but he associated Dxodon with Chalicotherium and Titaiiotherium, per- haps wholly on account of the absence of the large dependent processes of the chin. In this connection I may state that the symphysis of Moropus is much more pro- cumbent and is of a decidedly more delicate structvire than in the type of Dxodon. No Titanotheres have as yet been found in the John Day horizons, while Entelo- donts of the same size are sometimes discovered. Measurements of the Type Specimen of Dxodon shoshonensis. Mm. Greatest antero-posterior diameter of symphysis -* 164 " transverse " " " at Py 144 Vertical diameter of ramus at Pj 100 Daeodon calkinsi (Sinclair). Type: Skull, several vertebrte, and portions of fore and hind limbs. Horizon : Upper part of the Prormrycocluvras beds. Locality : Bridge Creek, Wheeler County, Oregon. Locality of Type : Paleontological Collection of University of California, No. 953. From the description and figures of Dxodon [Elotherium) calkinsi (89, p. 134, PI. XV) it is quite evident that the specimen pertains to the genus Dxodon de- scribed l)y Cope in 1878(7, p. 15). Althougli the type of the latter represents an animal of considerably larger size, it agrees perfectly, so far as comparison can be made, with the species described by Sinclair. The character by which D. calkinsi " The alveolar border is brokeu, which naturally reduces the antero-posterior diameter to some extent. PETKRSON : A UKVISION OF THE ENTKL0D0NTI1)/E 05 is specifically to Ije separated from JJ. sJio.'ihonensis is principally the smaller size, so far as can be determined by comparison of the two. Sinclair states that the type of D. calkinsi represents a very old individual and that many of the teeth are shed and Flfi. 19. Outline of Illustration of Skull of D.Todon calkinsi (.Sinclair). ( After Sinclair. ) the alveoli closed. " P^ . . . resembles the smaller premolars of the upper series. Anterior and posterior cingula are well developed on the lower molars. In Mg the hypoconulid is not differentiated from the posterior cingulum, which projects slightly, forming a very small heel. . . . "The mandible is peculiar in the absence of the knol)-like Ijosses on the chin The protuberances beneath P^ are small and deeply cupped. The dependent angle slopes gradually backward without the abrupt downward curvature characterizing E. ingens. ... "The chief point of specific value attaching to the cranium is in the .shape and direction of the jugal processes. These processes are plate-like with a thickened median rib. The free edges, especially the anterior, are thin and sharp. The processes are short, not extending below the lower mandibular l)order. The orbits are posterior in po.sition, their anterior borders lying above the posterior edge of M^'." The unciform and tibia were kindly sent to the writer for purposes of study by Professor .John ( '. Merriam of the University of California. It is at once notice- able that the lunar facet of the unciform is more elevated than in Dinohyus and 66 MEMOIRS OF THE CARNEGIE MUSEUM that the small facet for the magnum on the proximo-radial angle is more nearly radially in position as in the Oligocene genus Arcliscotherium. The unciform and magnum apparently articulated dorsally as in the latter genus. There are other- wise no marked differences between these two bones in Dasodon a.nd DinoJnjus. Measurements of Unciform. Mm. Greatest antero-posterior diameter 43 " transverse " 40 " vertical " 35 The tibia in general resembles that of the Oligocene type rather more strongly than the later Miocene form. The fibula was not coossified with the tibia, although a rough border on the fibular side of the tibia indicates that the two l^ones were in close proximity to each other. On the anterior termination of the inturtrochlear ridge of the distal end of the tibia is a facet, which indicates that upon extreme for- ward Hexure of the tibia this facet articulated with a corresi)onding facet in the large pit on the anterior face of the astragalus, as in DinoJiijns. The general features of the tibia are otherwise quite similar to those in the Entelodonts generally. The proximal end of the tibia is wanting, as is also the external portion of the distal end, in the specimen here described. Measurements of Tibi.a. Mm. Greatest lengtli of the fragment 276 " transverse diameter of shaft, medially 40 " antero-posterior diameter of shaft, medially 32 " " " " " distal end 42 Genus DINOHYUS Peterson. (Plates LV-LXI.) In 1905 the writer sent from the field the original description of this genus to Dr. W. J. Holland, Director of the Carnegie Museum, who suggested the name Dinochwms ; the description was published in Science (78, pp. 211-212). Later JMnochwnis was found to be preoccupied, and a note of correction by the writer appeared the same year in Science (79, p. 719). The skull and lower jaws were published in 1907 (81, pp. 49-51, Pis. XVI-XVII) and the detailed description of the skeleton is found on pages 77 et seq. of this memoir. Dinohyus hollandi Peterson. Tyj'ie : A nearly complete skeleton. Horizon : Lower Miocene (Lower Harrison beds). Localitij : Agate Spring Fossil Quarries (Quarry No. 1), Sioux County, Nebraska. PETERSON : A REVISION OF THE ENTELODONTIDiE 67 Locality of Type : The Carnegie Museum (Catalogue of Verteljrate Fossils, No. 1594). Pnncipal Generic Characters of J)inohyuf< : Median incisor reduced and some- times wanting. Transverse diameter of P- nearly equal to the antero-posterior, Pj with large deuterocone ; the crown subquadrate in outline and the tooth of rela- tively small size; a tendency to increase tlie antero-posterior diameter of M-, and the meta- and hypocones of relatively large size ; lower molars with subequal height of posterior and anterior tuljercles ; the tubercles .separated by narrow cross- valleys ; the trigon lost; all the premolars spaced ; dependent process of the jugal of proportionally small size and the posterior termination of the zygomatic process developed into a strong buttress at the anterior border of the glenoid cavity ; small bony eminences on the chin and a strong knob-like process on the inferior border of the ramus opposite 1\ ; a relatively gradual backward slope of the angle ; relatively short alveolar border of the premaxillary and rather short chin ; verti- cal and transverse canals of the dorsal vertebra; as in Sus; fibula coossified with the tibia ; trapezium absent and mt. V sometimes absent. Genus AMMODON Mar.sh. Although Professor Cope (5, p. 704), Leidy (50, p. 388), and Marsh (57, p. 3), referred to the New Jersey specimen, thei-e was no adequate description of the type until 1893, when the latter author gave a short description together with good illus- trations (63, pp. 409-410). Known Generic Characters of Ammodon : P^ of relatively large size ; large hypo- conulid on M^ ; anterior and posterior tubercles of lower molars subequal in height ; the loss of the trigon; cross-valleys narrow; type representing an animal of large size. Ammodon leidy anum Marsh. Type : l\ ; Neotype : Mg of left •* side. Horizon : Middle Miocene ? Locality : Near Farmingdale, Monmouth County, New Jersey. Locality of Type: Collection of Yale Museum, No. 12040. From Marsh's description (63, pp. 409-410) and also from my observations upon the specimen it seems quite prol)al)le that P^ and Mj which Marsh descriljed belong to the same individual. Working on this hypothesis it would seem that Pj is of greater proportionate diameter than in DiiiohynshoUandi. The poste- rior heel of P^ in Ammodon is also larger than in Dinohyus, which is due to a heavy ''Professor Marsh mistook the left M^ for that of the rifjht siile, a mistake which might lead to coDfasion. 68 MEMOIRS OF THE CARNEGIE MUSEUM cingulum surrounding it in the former genus, but which is absent in Dinohyus. Furthermore, in the type of Ammodon the crown is somewhat liigher and on the posterior face there are two mammilated ridges of considerable prominence Fig. 20. Type of Ammodon 'leidy- anum Marsh. Internal and Crown View of Pj. 5 nat. size. (Collection Yale Museum, No. 12040.) Fig. 21. External and Crown View of Mj of i!ieotype ot Ammodon kidpanum. S nat. size. (Collection Yale Museum, No. 12040.) which take their origin, one at the postero-internal angle and the other at the antero-external angle of the heel, and continue to very near the apex of the crown (see fig. 20). In Dinohyus hollandi there is only a slightly uneven folding of the otherwise smooth enamel on the posterior face of the protoconid, and there are no decided ridges. M-j of Ammodon leidyanum is very similar to that in Dinohyus hollandi. In the for- mer the posterior heel (hypoconulid) is of rela- tively greater development (see fig. 21) than in the latter, and the tooth has a continuous cingulum on the antero-external and antero-internal faces, while the corresponding tooth in Dinohyus has the cingula represented on the postero-external, and but veiy slightly on the external and internal angles of the anterior face. The teeth of Ammodon leidyanum are Fig. 22. External and Dorsal View of larger thaii those in Divohyus and represent an even Fragment of Astragalus of Ammodon? J ^^Ql-e modified type than the latter. nat. size. (Carn. Mus. Cat. Vert. Foss., ttp, .t.i , <•.< i • ,i „ ,,,„ , Untortunately the type oi Ammodon is rather No. 1548.) ■' ■' ^ inadequate and con.sequently the basis of compari- son is very limited, but the characters enumerated above would seem to indicate that the geological horizon in which the New Jersey specimen was found, represents PETERSON : A REVISION OF THE ENTELODONTID^E 69 deposits as late, or perhaps later, tlian the lower Harrison Ijeds of Nebraska, in which the type of Uhiohi/iis was discovered. In this connection it is of interest to note that a portion of an astragalus of an E'idelodont of very large size was found in 1901 on the surface near the top of the upper Harrison beds in the upper Niobrara valley. This fragment (No. 1548, ( 'arn. Mus. Cat. Vert. Foss.) although considerably worn by the elements, plainly indicates characters that are different from those in the corresponding bone of Dlnohijiis. Not only is the bone larger in size, but its sustentacular facet is proportionally of greater vertical convexity and the distal portion of the bone is produced anteriorly to a much greater extent than in DinoJijius. This peculiar feature of the astragalus under description would seem to indicate that the tarsus of this genus (AmmoJonf) had a somewhat different angle. Species the Generic Loc.vpion of Which is Douhtkul. Elotherium imperator Leidy. T/fpe : A canine, an incisor, and a [)ortion of the crown of a premolar tooth. Horizon : Middle John Day beds? Locality : Bridge Creek, and John Day River, Oregon. Locality of Type: National Museum, Washington, D. C. ?■'" The first specimens of Entelodonts, which Professor Leidy reported from the John Day Basin in 1873 (53, pp. 217, 320), are altogether indeterminable generically and specifically and consequently inadequate as types. " Elotlicrinm imperator" should therefore be regarded only as an historical record of the first report of the existence of Entelodonts in the John Day formation. Elotherium superbum Leidy. Tyjje : An incisor tooth. Horizon : Miocene ? Locality : Douglass Flat, Calaveras County, ('alifornia. Locality of Type : Unknown to the writer. The type of " Elother'mm superlmvi'' is altogether inadequate, and its geo- logical horizon is apparently uncertain. The record (49, p. 175 ; 50, p. 388) of this species should be regarded as only possessing value from the standpoint of history and geographical distribution. ™ No catalog number at this writing lias been given to the specimens. 70 memoirs of the carnegie museum The History of the Discovery of the Agate Sprinc; Fossil Quarries, Where the Skeleton of Dinohyus Hollandi was Found. While engaged in field work for the Carnegie Museum on the upper Niobrara River, locally known as Running Water, Sioux County, Nebraska, in 1904, the writer had often been invited by Mr. James H. Cook to visit his ranch, the Agate Spring Stock Farm, located on that stream, some twenty-five or thirty miles east of the Nebraska- Wyoming state line. One day in the latter part of July, I decided to l:)reak camp and go down the river in search of new localities for fossils and also to study the geological features of the neighborhood more fully. As Mr. Cook's ranch was on our way down the stream, it was decided to pay him a visit, and accord- ingly we stopped at his ranch. After a camp-ground had been pointed out to me. Fig. 23. Ranch-House of Mr. James H. Cook, Agate Spring Stock Farm. (From a photograph by Mr. Albert Thomson.) on top of a high butte immediately to the south of the farm buildings, and arrange- ments for wood, water, etc., had been made, the preliminary work of prospecting the neighborhood was at once under way. A day or two later Mr. Harold Cook, the eldest son of Mr. James H. Cook, accompanied the writer to a small elevation some four miles to the east of the farm buildings and immediately beyond the east- PETERSON : A REVISION OF THE ENTELODONTID.E 71 ern limits of the land belonging to the ranch.'-' The talus of this low hill was dis- covered to be filled with fragments of liones, and was afterwards designated as quarry A (See fig. 24, A). On our return to the ranch 1 reported to ^Ir. James H. Cook Fig. 24. View of the Buttes Showins the Location of tlie Oiftereut Quarries. Xos. 1-3, Quarries of the Carnegie Museum. AM, American Museum of Natural History. NU, University of Nebraska. A, Quarry A. (From a photograph by the writer.) that the place which his son had shown me was of much interest and importance to me and that I wished to start the work of excavation on the prospect immediately. This was entirely satisfactory to l\Ir. Cook and his family. In fact there was evident satisfaction on the part of Mr. Cook that I had found something which I regarded as of interest and importance near his farm, and I was accorded every civility, which I could possibly desire. As I wished to be near my work, Ur. (Jook invited me to camp in his "lower field." Accordingly our first camp was pitched on tbe south bank of the stream close to the hill and the operation of excavating in quarry A was begun during the last few days of July. We had worked three or four days in this quarry when I decided to visit the two buttes (since named Carnegie Hill and University Hill by Prof. E. H. Barljour) which lie about tliree hundred yards to the south of the place wbere we were working. One may easily imagine the thrill- ing excitement of a fossilduinter when he finds the talus of the hillsides positively covered with complete bones and fragments of fossil remains. It was with comparatively little effort that I was able to articulate portions of 2' Mr. James H. Cook informed me that fall that tlie hills containing the fossils were on his ground (80, p- 487) ; this statement was found afterward to be founded upon an error, and Mr. Harold Cook has since (1908) filed on this government land. 72 MEMOIRS OF THE CARNEGIE MI^SEUM the feet of Diccrathermm. coolci and Moropus using the disassociated bones picked up in great abundance in the takis. Here then was a veritable wonderland ! I sjaent a considerable portion of the afternoon before I returned to where my assistants, Mr. T. F. Olcott and A. A. Dodd, were at work. Tlie next day I again returned to the hills with wrapping paper, twine, and a sack in order to pack and bring down the bones which I had picked up the previous day ; and also to more fully ascer- tain, if possible, the extent of these deposits. A few days later when I reported my additional find to Mr. Cook, I learned that I was perhaps notlthe original discoverer of the bones in these two hills, as he told me that he had seen bones there as long FlQ. 25. View of Carnegie Hill and University Hill, Showing Quarries No. 1 and No. 2 of the Carnegie Museum on Carnegie Hill, and Quarry of the University of Nebraska to the left. (From a photograph by the writer.) ago as 1890, but always thought tlieni to be of recent origin. In 1908 I was further convinced that tlie bones in these hills had been seen even earlier by Mr. Octave Harris,28 a prominent ranchman of the neighborhood. To Mr. Cook is, however, due much credit, as these deposits of bones would perhaps not have been brought to the attention of paleontologists at this time, had he not invited me to his ranch. In looking over the whole situation it was plain that the deposits required more time and assistance than I then had. Together with this fact I may mention that I had not been well all that spring; in fact I was obliged to leave the field on account of sickness about the middle of August. I wish to here acknowledge, with much appreciation, the kind and sympathetic treatment I received from Mr. Cook and his family during my week of sickness at his ranch. Before leaving for the east ^■^Mr. Harris extended much aid to our party during our last (1908) season's work in the quarry, and the interest which he took in our work is heartily appreciated and most gratefully acknowledged. PETERSON : A REVISION OF THE ENTELODONTID.E 73 I assured Mr. Cook that I would be back the following spring to systematical In- take up the work in the hills, which decision he assured me Avas in accord with his wishes. Early in the spring of 1905 the writer accompanied by Mr. T. F. Olcott resumed work in quarry No. 1 (see fig. 25). Mr. Cook impelled by curiosity had already started in the previous fall to excavate in quarry No. 1 (see Plan of Quarry, PI. LIV, Cook Excavation), but after earnest entreaty by letter to await the coming of more expert help desisted from his work. Many piles of fragments were found on the edge of the opening which had been dug out by Mr. Cook. These fragments were carefully gathered and packed before the commencement of systematic excava- tions. The work being fairly started, Mr. Olcott was left in charge of the quarry, while I returned to Pittsburgh in order to attend to other duties in the museum. Thus was started one of the most important fossil quarries ever found in North America. When I again returned to the field later in the season, INIr. Olcott was still engaged in the same hill, while Professor E. H. Barbour, of the Nebraska State Uni- versity, had been invited by Mr. (Jook to open a quarry in the adjoining hill, after- wards named by Professor Barbour "The University Hill." Much material had been uncovered in quarry No. 1 among which the most important were bones of Moropus, many individuals of Diceratherium and an unusually well preserved, and pretty nearly articulated skeleton of a gigantic Entelodont (see PI. LIV, Plan of Quarry, Blocks Nos. 61, 86 in Sect. 4). The discovery of the latter was a surprise, nothing having hitherto been found representing this family in the lower Miocene of Nebraska. One hind limb and most of the parts of the skeleton anterior to the pelvis were recovered. The pelvis, the lumbar vertebra}, and perhaps also one hind limb^* were unfortunately dug out by Mr. Cook and those assisting him and only portions of four vertebrae and the pubic symphysis of the pelvis were recovered from the fragments left on the edge of the quarry. The skull was almost in contact with the atlas. The cervical and dorsal verte- brfe were found in a successive series back of the skull, the dorsals being articulated by their zygapophyses (see Plate LIV, Sect. 4, Blocks 61, 72, 86). The greater number of the ribs and the sternum were found in close proximity in their relative positions, while one hind limb was found a short distance from the main portion of the skeleton. The fore feet were scattered. There is, then, no doubt that the f(jl- lowing description of the skeleton is that of one individual. In addition to this skeleton thei-e were found, in Sections 16-21 in quarry No. "The head of a femnr, which may or may not pertain to this specimen, was found on the damp. 74 MEMOIRS OF THE CARNEGIE MUSEUM 1, opened in 1908 by the writer, many other bones, fragments, and scattei'ed teeth of Dinohi/us. Some of these fragments were much worn by the elements before they were finally deposited, indicating that they were probably transported by a Fig. 26. At Work at the North End of Quarry No. 1. (From a photograph by Professor Barbour. ) stream of water. When this additional material is used in this paper its catalog number is always mentioned in connection with the description. GEOLOGif'AL Notes. The lower Miocene of western Nebraska and eastern Wyoming has been sub- divided into four horizons. Successively overlying the Oligocene formation these horizons are: (1) the Gering beds; (2) the Monroe fJreek beds; (3) the lower Harrison beds ; (4) the upper Harrison beds. The latter horizon is capped at Spoon Butte and other higli elevations on the divide between the Niobrara and Platte rivers, by a hard stratum of pinkish-gray sandstone (81, p. 23, fig. 1 ; 72f, p. 73). This sandstone is from twenty-five to fifty feet (or about seven to fifteen meters) thick. No fossils have been found in this stratum on Spoon Butte by which its relative age can be determined. It is perhaps of considerably later origin than the Harrison beds, and may even be as late as the Pliocene. For convenience in this connection these beds may be called The Spoon BxMe hcds.^ It also appears, judging from the fauna of the upper Harrison beds, that this '" From a recent publication by Dr. Matthew (Science, N. S., Vol. XXIX, No. 735, p. 196, 1909) it would seem that these beds may be regarded as representing the Ogallala formation. PETERSON : A REVISION OF THE ENTEI.ODONTID.K 75 horizon of the Miocene exposed along the upper Niobrai-a, should be regarded as the base of the middle Miocene, or a horizon filling, in part at least, the hiatus between the lower and the middle Miocene. Places of non-conformity between the upper and lower Harrison beds are frequently found, the first of which observed by theAvriterin 1901 is situated immediately east of the Niobrara-Wyoming State line. m Spoon BuJIe Beds ' ^f . Upper Harrison D^ds 11 200|t. Lowtr Harrison B«J» 200|t Monrot Creek Beds 300fl Gering BeJs aooct Oliaocene Beds Fia. 27. Diagram matio Section of the Miocene Beds in Western Nebraska and Eastern Wyoming. This is in reality the type locality of the upper Harrison beds, which were called to J. B. Hatcher's attention ])y the writer in 1901. Hatcher, then ( 'urator of Paleontology of the ('arnegie Museum, decided these beds to be the " Nebraska beds " of Hcott. Viewing these beds from the standpoint of certain portions of the fauna (81, p. 56) taken together with lithological characters, the sediment being usually of a darker color, it would seem that a natural division between the lower and middle Miocene may be established at the contact of these two horizons. As has been stated elsewhere (80, p. 487 ; 81, p. 41) the Agate Spring Fossil Quarries are located in the lower Harrison beds. The origin of the deposit in which the fossil bones of these quarries are found is most likely a stream deposit. In sup- port of this view may be cited the fact tbat the parting plane underneath the layer carrying the Ijones is a few shades darker in color than that layer itself, and the indslone ab&oluTely I of |"05SiU Fia. 28. Section of Face of Quarry of University of Nebraska at tlie End of tlie Season of 1908, showing channels in the stratum supporting tlie fossiliferous layers. The Nos. 1-20 indicate sections 5 ft. in length. bones are often closely packed in irregular channels and pockets, which vary in thickness from two or three to eighteen or twenty inches (8 to 50 cm.). The sur- faces of the bones are sometimes highly polished and worn, and there are numerous 76 MEMOIRS OF THE CARNEGIE MUSEUM fragments of bones which are worn down to the shape of a pebble. These conditions could have been brought about in no other way than by the action of water in motion. There was here most likely a stream of considerable magnitude, instead of a lake, as was suggested by the writer (80, p. 49). The fossil-bearing bed has all the appearance of a stream the bottom of which was covered by quick-sands. In the quarry located on " University Hill " (see fig. 28) there are distinct evidences of narrow and rather shallow channels, which were separated by sandbars of greater or lesser dimensions. The bone-layer of this quarry is thus, irregularly separated, although confined to one general level throughout portions of the base of the hill. It seems reasonable to suppose that in the immediate vicinity of the Agate Spring Fossil Quarries the ancient fauna of the region found a favorite crossing of this stream. The remains of the unfortunate animals which attempted to ford the river under unfavorable circumstances furnish the records which the collector is fortunate in securing, and which give a glimpse into the struggle for existence in those early times. In this connection it is interesting to note certain topographical features in the vicinity of the Agate Spring Fossil Quarries. In this region the general direction of the Niobrara River is east and west. Very often the side canons of this stream ex- tend in a direction contrary to that which one would naturally expect. The draws and canons on the north side extend diagonally southeastward and in a perfectly natural direction towards the main stream, while very often the canons on the south side extend diagonally northwestward or up-stream, which is generally regarded as a feature of stream piracy. This anomalous feature was quite perplexing to me for some time, and it was not until I had studied the character of the sediment in con- nection with the topography that any tangible explanation was reached. It is a well-known fact, which has been observed by Darton, Hatcher, Peterson, and others, that in these Miocene beds are masses of pipey concretions of greater or less extent, which always have their long axes directed in a nearly northwest-and-southeast direction. The canons to the north of the Niobrara would have no difficulty in cutting their way through to the river in a natural course, while those on the south side of the stream would sometimes have to accommodate themselves to the up- stream dii-ection which is the course of the least resistance to the erosive elements. In this way there are developed sharp ridges and hogbacks, nearly always in the north- west and southeasterly direction on either side of the river. Chimney-rocks and pillars of numerous shapes are seen along the course of the Niobrara River for some forty miles in this general locality. The entire Miocene section, and especially the lower Harrison beds, are appa- PETERSON : A REVISION OF THE ENTELODONTID.B 77 rently of considerably later age than the John Day formation in which Entelodonts have been discovered, judging from the fauna which has been published elsewhere (81 pp. 34-56). That some genera of the E^itelodoiitidie continued to the close of the lower Miocene is now well established ; indeed it would perhaps not be surprising to find them represented in the middle and possibly in the upper INIiocene. Dinohyics, which was discovered in the upper part of the lower Miocene beds (lower Harrison beds), was in all probability followed in the Miocene in New Jersey, by Arnmodon leidyanum., which is without much doubt a later form. Detailed Description of Dinohyus Hollandi Peterson. The Superior Dentition. (Plates LVII and LVIII.) The median pair of incisors are lost in the type, but the alveolus of the right side is complete, and indicates a proportionally small tooth.''"'' I^ is present and is much worn on the apex and the internal face ; the tooth has a crowded position, and its antero-posterior diameter is greater than the transverse. P- is the largest of the incisors and is about twice the size of I- ; it has received much wear on the postero-lateral angle from contact with the inferior canine, and the apex is also much worn. The antero-posterior diameter of this tooth is, as in its prede- cessor, the greatest. All the incisors are more crowded than in the Oligocene forms generally. It is also notice- able that the transverse groove immediately above the enamel band, which Professor Scott refers to (87, p. 274), is very well developed in the type, and could not have been caused by the wear of the lower teeth. As in other forms of the family, the canine tooth is very large, recurved, and has an enormous fang. In the present specimen the apex of the tooth has received con- siderable wear as has also the antero-internal face. There is a light cingulum on the posterior face, but externally the enamel is smooth. The first, second, and third superior premolars in Dinohijas are in general quite similar in shape to the corresponding teeth in Entelodon magnum of Europe. They ™'' a premaxillary bone (No. 2145, Carn. Mus. Cat. Vert. Foss.) in the collection from the Agate Spring Fossil Qaarries, (Quar. No. 1, Sect. 19), which is here referred to Dinohym, has only two alveoli, the median incisor having dropped out and the alveolae closed. Fia. 29. Obliqae Side Views of Premaxillary Bone of D. hollandi Peterson. (Carn. Mas. Cat. Vert. Fobs., No. 2145.) 78 MEMOIRS OF THE CARNEGIE MUSEUM are, however, relativelj^ small, P- especially so, and the whole series is more spaced, while in the latter genus the second premolar forms a^losed continuous series with the posterior teeth, and P- only is widely separated hy a diastema. In Dinoln/us F- and P- are widely separated and P- and P- are separated only by a very short diastema. Back of P- the dentition forms a continuous closed series. pi is of relatively large size with a simple crown, the apex of which is placed well forward the cingulum is quite heavy on the antero-internal angle and the Fig. 30. Upper Right Canine of Type of Diiiohytts hoHandi Peterson. ^ nat. size. 1594. ) (Cam. Mus. Cat. Vert. Foss. , No. heel slopes rapidly behind. P^ is of about the same size and shape as V-, except that the antero-internal cingulum is slightly more develojDed and extends further toward the external face of the tooth. The enamel of this tooth is less crinkled than that of pi. P- has a relatively smaller antero-posterior and a greater transverse diam- eter than in the Oligocene forms. This transverse enlargement of the tooth PETERSON : A REVISION OF THE ENTEI.OPONTID.T; 79 causes a characteristic trihedral cross-section, which is very much less apparent in the forms from the Oligocene. Tlie tooth is slightly larger than P^. The entire crown is practically taken up by the large protocone. The cingulum has developed into a narrow shelf on the antero-internal angle and continues in a mucli less con- spicuous manner on the internal lace, terminating in the heavy cingulum posteriorly. I'i has no cingulum externally, which is also true of all the other premolars, and its transverse is slightly greater than its antero-posterior diameter. The tooth has a proto- and a deuterocone. The latter is relatively more developed than in Eiitelo- don. There is a heavy cingulum posteriorly and on the antero-external angle a prominent basal lieel is developed, l>ut there are no external or internal cingula. The molars of D'moliym are more specialized than those of ArcJurofherinm, as is shown by the fact that the tubercles of the crowns are more depressed and consoli- dated, and the posterior portion of M^ is relatively more developed than in the Oligocene genus. The crown of M^ in the type specimen is much worn, so that it shows two large and irregular tracts with tlie bottom of the cross-valley extending beyond the triturating surface, especially on the inner portion of the tooth. There are heavy cingula anteriorly and posteriorly and a less developed cingulum exter- nally, while internally the tooth is smooth. M^ is the largest of the series ; the principal difference Ijetween it and the corresponding tooth in the Oligocene forms (especially those from the lower Oligocene) is revealed in the interruption of the internal exit of the cross-valley by a heavy rounded ridge (see PI. LVIII, fig. 1) which was undoubtedly developed from the cingulum. The anterior border is entirely surrounded by a heavy cingulum, as is also the postero-external angle, while on the external face immediately back of the exit of tlie cross-valley thei-e is very slight indication of a cingulum. The posterior portion of M^ is composed of (1) metacone, (2) hypocone, and (3) metaconule ; the first is of much larger size than the second, and the third is very inconspicuous, in fact it is entirely absent on the corre- sponding molar of the right side of the type specimen. The cross-valley of M^ is interrupted in the same manner as is the case in the preceding tooth, and in addition this tooth has a mammillary cingulum which to some extent also closes up the external exit of the cross-valley. This cross-valley is not closed up in the speci- mens of Archifothermm known to me from the lower Oligocene. There is in the present type a heavy anterior cingulum on M^, but no posterior cingulum ; the back part of the tooth having played a more important part in the make-up of the ci'own in tliis genus than is seen in older types. The antero-external tubercle of the last upper molar in Mntelodon magnum has reached a greater development than in Archxolhci-inin and in this respect is more 80 MEMOIRS OF THE CARNEGIE MUSEUM nearly like what is seen in Dinohi/us. The posterior portion of this tooth in the latter genus is not unlike that of the European form, and shows a more advanced development than in Archieotheriuvi. Altogether in Archxotherium the crowns of the molars are somewhat lower, the apices of the different points of the grinding surface are less united, and the teeth as a whole are less modified than in Enfelodon magmim and Dinohyus hoUandi. Measurements. Mm. Length of alveolar border of iiiaxillary and premaxillary 465 Distance from median incisor to Mi 330 Length of molar series 132 Antero-posterior diameter of I^ 28 Transverse " " I^ 24 " " " I» 41 Antero-posterior " " I- 35 " " " canine at base 50 Transverse " " " " " 53 Antero-posterior " " Pi 39 Transverse " " pi 22 Antero-posterior " " P^ 38 Transverse " " PS 22 Antero-posterior " " F- 42 Transverse " " P^ 33 Antero-posterior " " PA 37 Transverse '' '• Pi.... 40 Antero-posterior " " Mi 42 Transverse " " M' 44 Antero-posterior " " M^ , 45 Transverse " " M* 48 Antero-posterior " " M^- 45 Transverse " " M^ 47 The Inferior Dentition. (Plate LVIII.) The incisors are much worn, but their outlines present more obtusely pointed crowns than in the older forms and the median pair is proportionately smaller, so that the gradual enlargement from ly to I3 is in about the same proportion as in the superior series. The teeth are crowded and occupy a more nearly transverse position in the front of the jaws than is seen in the Oligocene forms generally. The crown of ly is much worn ; and its antero-posterior exceeds its transverse diameter, so that it is oval in cross-section. lo is about twice the size of the median incisor and is also oval in cross-section, its antero-posterior being greater than its transverse PKTERSON : A REVISION OF THE KNTELODONTID.K 81 diameter. There is a heavy cinguluin oii the posterior face which extends half way around the base of the lingual side. I a is nearly twice as large as the tooth just described, but otherwise the two are quite similar. On the external faces of I2 and I3 are deep grooves across the faces, just below the enamel band, which were not caused by the wear of the upper teeth. Inasmuch as the canines of both sides have a similar and much greater worn area externally, which could not have been reached by any of the upper teeth, it further supports the contention of Professor W. B. Scott and C. F. Brackett, that these animals used the lower tusks in digging in the ground for roots (87, footnote, p. 275). They may also have l)een used for stripping off the foliage of shrubbery. There is a slightly developed cingulum on the posterior face of the canine, otherwise the enamel is quite smooth. The crown is higher and more pointed than that of the upper canine, but in general appearance and size it differs little from that tooth. ^' Py is separated from the canine by a short diastema ; the tooth is implanted in the jaw by two strong fangs which coalesce for some distance below the crown. The latter is simple aixl is rather low when compared with that of Archxotheriam. On the internal face the enamel is crinkled, but externally the tooth is quite smooth. Posteriorly there is a heavy cingulum and anteriorly there is a prominent oblique ridge, which originates at the antero-internal angle and continues outward and upward to the apex, constantly diminishing in prominence in its upward course. The apex of the crown in the type is considerably worn, but it is quite easy to ascer- tain its contour which was very much more obtuse than in Archieotherium. There is a considerable diastema between Py and l\. The latter is also implanted with two roots which are, however, more divergent and do not coalesce below the crown as in 1\. The crown of P^ is of about the same size as the crown of the preceding tooth, but its apex is less worn and there is a mammillated ridge on the anterior and posterior faces of the protoconid. There are prominent anterior and posterior cingula-and the enamel on the internal face is thrown into light folds, while the buccal face of the tooth is smooth. P3 is, as usual, the largest in the series ; it is not crowded in the alveolar border, but is separated from r^ by a very short dia- stema. The apex of the crown is next highest to that of the canine and in general form it is not unlike that of P2. P4 has a heavy heel which agrees in general with Py in Entelodon magnum. Altogether Py of the present genus presents a different outline from that of the corresponding tooth in the earlier genera, which fact is wholly due to the increased transverse diameter of the tooth posteriorly. '" The left lower canine is disproportionately short and stubby and was evidently broken oS at the apex dnriug the life of the animal. OZ MEMOIRS OF THE CARNEGIE MUSEUM The inferior premolars of Entclodon are relatively large, more regular in size, and Pg- and P^ are entirely surrounded bycingula ; while in DinoJiyus the premolar teeth as a whole are smaller, P3 being conspicuously the larger of the series and P3 and P^ have the cingula less strongly developed. The molar series in Dhiohyus is very crowded and occupies less than one half of the antero-posterior diameter of the alveolar border. As in other allied genera, the molars are of relatively small size in comparison with the jaw and they increase Fig. 31. Crown View of Inferior Dentition, Left Side. Type specimen of Dinokyus Iwlhindi Peterson, i nat. size. (Cam. Mus. Cat. Vert. Foss., No. 1594.) in size jDOsteriorly. My is consideraljly worn in the type, so that the quadrituber- cular pattern is converted into two heavy transverse tracts having irregular outlines and with only a remnant of the cross-valley on the inner half of the tooth. On the antero-internal angle the cingulum is developed into a small basal tubercle. Oppo- site the cross-valley there are short and smoothly rounded cingula externally and internally. On the postero-external angle is a cingulum, which is continuous with a similar short cingulum on Mg. The posterior face of the tooth is closely crowded against the anterior face of Mj. In the type the quadritubercular pattern of Mj is plainer than is the case in the tooth just described, but trituration has already caused the appearance of the solid cross-crests ; the anterior being slightly more elevated than the posterior. An- teriorly there is a deep, narrow fissure, which is caused by projecting horns from the proto- and metaconids. On the external face the cingulum extends from the postero-external l)ase of the protoconid and continues around the hypocon"id, termi- nating in the posterior basal heel.'" There is no cingulum on the external face of the protoconid, nor on the internal face of the inferior molars. The proto- and metaconids of M^ are very little worn, so that the separation is still present, but it is reduced to a very narrow and shallow fissure on account of the '^In ray preliminary notes on (Dinochmrus) Dhiohyus hoUandi, Science, Vol. XXII, p. 212, 1905, 1 made the state- ment that "There is no cingulum on the inferior molars," which is incorrect. PETERSON : A REVISION OF THE ENTELODONTIDiE 83 transverse development of the tubercles. The fissure which corresponds to that de- scribed on the anterior face of M^^ is present on M^ l)ut is entirely closed in by the lateral horns of the proto- and metaconids. There are slight cingula on the antero- external angles and the external face of the hypoconid is surrounded by a heavy mamniiHated cingulum. Posteriorly there is a heavy basal ledge and a minute hypoconulid which is located at the base (posteriorly), between the hypo- and ento- conids. One of the more significant characters of the lower dentition is seen in the relatively greater development of the hypo- and entoconids, which causes the nar- rowing of the cross-valley between the latter and the anterior tubercles. There were found in quarry No. ], the third lower molars of the right and the left rami (Vert. Foss., 1835 ; 2114), which differ from Mg in the type of IHnohi/tis, by having the hypoconulid well developed. The cross-valley Ijetween the anterior and posterior tubercles is also interrupted externally l)y a heavy cingulum. The postero-external face of these teeth are otherwise quite smooth, while in Divohi/iis h)Uandi, there is a continuous cingulum from the cross-valley to the posterior face of the hypoconid. These teeth may possibly represent another genus, but are in- adequate as types. The molars of Entehdon differ from those of Dmolujus in having the postero- internal cusp (entoconid), especially on Mo and M3 much less developed. The ledge produced by the cingulum on the posterior margin of the crown, which is so prominent in Dinohifus, is only very slightly represented in Entelodon. The cingula are also better represented on the molars of Entelodon than in those of the American genus. The most important feature of the lower dentition is seen in the greater development of the posterior portion of the crown of M^ in Dinohi/us than in Entelodon which in the former is distinctly more specialized than in the latter genus. This same feature of M ^^ is also observed on comparing I)ivohi/t(s with ArcJueotherinm. Measurements. Hm. Length of inferior dentition 465 " from median incisor to M, 325 " of molar serie.s 137 Antero-posterior diameter of ly 16 Transverse " " ly 12 Antero-posterior " " Ij 25 Transverse " " Ij 19 Antero-posterior " " I5 at the hase of crown 34 Transverse " " Ij " " " 27 Antero-posterior " ofcanine " " " 48 Transverse " " " " " 45 84 MEMOIRS OP THE CARNEGIE MUSEUM Antero-posterior diameter of Py at the base of crown 38 Transverse " " Py " " " 21 Antero-posterior " " P^ " " " 40 Transverse " " P^f " " " 21 Antero-posterior " " P7 " " " 54 Transverse " " P^ " " " 28 Antero-posterior " " Pj " " " 45 Transverse " " Pj " " " • 28 Antero-posterior " " My " " " 42 Transverse " "My " " " 34 Antero-posterior " " M^ " " " 47 Transverse " " M^ " " " 40 Antero-posterior " " M^ " " " 50 Transverse " " Mj " " " 39 The Milk Dentition. (Plate LIX.) In section 12, block 66 (see Plan of Quarry, PI. LIV), were found the remains (No. 2137, Carn. Mus. C'at. Vert. Foss.) of the skull, and close by in another block, a portion of the lower jaw of a young individual, with the dentition partly in place. An incisor is represented only by the crown ; the tooth being too young for the formation of a fang. This incisor is pressed by the matrix close to the canine and has a simple conical crown slightly crinkled at the base and it is very thinly covered by enamel. The only noticeable difference between this tooth and more adult forms of the canine is its strongly serrated anterior and posterior ridges and the crinkled enamel. By excavating the alveolar border at the antero-internal angle of the deciduous canine the crown of the permanent canine was found. .The point of the crown of the permanent P- appears in the circular alveolus immediately back of the canine. It is thus seen that the temporary P^ had a single fang and was crowded out quite early. Professor Scott has also (87, p. 276) pointed out a similar occurrence observed by him in a young specimen ot Archxotlwriwn which would tend to firmly establish the fact that there are four deciduous premolars in the upper jaw of at least two genera of this family. Deciduous P- of the specimen under description is represented only by the alveolus, while the deciduous P^ is in position in the right maxillary and Pi again represented by alveoli. In the left jaw temporary P* has been placed in position. Deciduous P- of this specimen has two large cusps and two smaller tubercles. The anterior cusp (protocone) is a bluntly pointed pyramid which occupies the anterior half of the crown, while the postero-internal cusp (tetartocone) is smaller, and is confined to about one quarter PETERSON : A REVISION OF THE ENTELODONTID.'E 85 Fio. 32. Sup- posed Deciduous Upper First Pre- molar. Nat. size. (Cam. Mus. Cat. Vert. Foss., No. 2P27.) of the tooth at its postero-internal portion. The two small cusps are nearly on a direct transverse line and immediately external to the tetartocone. Anteriorly and posteriorly there are very heavy cingula ; externally and internally the cingulum is only moderately developed. The deciduous P^ Avas found quite close to its position and there is comparatively little, or no doubt, that this tooth belongs with this specimen. The tooth is molariform, as in the Oligocene genus (87, p. 287). It is strikingly similar to M^ of the permanent series, but has the antero-external portion of the cingulum developed into a more decided cusp.** Anteriorly there are three cusps on the crown : one internal, one external, and one directly an- terior. The two former are of equal size, while the latter is quite small and occupies the anterior exit of the longitudinal valley. Pos- teriorly there are two subequal cusps which are separated from the anterior cusps by a deep cross-valley. The tooth is nearly surrounded by a cingulum. The chief feature of the maxillary bone is its comparatively short antero-posterior diameter. The premolar series is set closer to- gether in the alveolar border than is the case in the adult ; the preor- bital foramen, however, seems to occupy the same relative position as in the latter. The palatine plate is quite thick, but its transverse diameter is small. The milk dentition of the lower jaw is represented only by the last molar (No. 2137A, Carn. Mus. Cat. Vert. Foss.). The crown of this tooth is composed as usual of three transverse pairs of cusps and a small talon developed by the cingulum on the posterior face. The tooth is very long and narrow with broad transverse valleys separating each pair of tubercles, while the longitudinal valley is much less distinct. The crown is supported by two strong fangs which are bifurcated near their lower extremities. The posterior portion of a lower jaw (No. 2137A, Carn. Mus. Cat. Vert. Foss.) was found near the tooth described above and undoubtedly belongs to it. In this mandibular fragment the posterior alveolus for dPf is still preserved, and into this the tooth referred to fits quite perfectly and is regarded by the writer as belonging to the same individual (see PI. LIX, fig. 3). When the jaw is exca- vated from the inner side back of the last milk-molar there is seen the complete crown of My, and only the anterior portion of M... The crown of My is directed upward in the usual way, while that of M.j is directly inward, and would have to rotate outward and upward in order to erupt through the alveolar border in the usual manner. As is seen from the illustration (I'l. LIX, fig. 3) the two anterior » Professor Scott evidently mistook the anterior part of the tooth he described for Ihe posterior portion ( /. c, p. 276) 86 MEMOIRS OF THE CARNEGIE MUSEUM cusps of Mt7 are well united and apparently not developed independently, as in young titanotheres.^^ The posterior portion of My is not represented at all, as the tooth is too young. The jaw fragment was somewhat worn by the elements before it was finally imbedded in the bone-bearing layer of the quarry. The coronoid process is, how- ever, complete, and is low and broad antero-posteriorly. The condyle is on a hori- zontal line with the alveolar Ijorder and the angle is seen to project quite strongly downward below the ventral border of the jaw. The dental foramen is very large and is situated immediately in front of, and below, the condyle. The temporal fossa is high and very shallow. An isolated upper tooth, which was found in the Agate Spring Fossil Quarries (Quarry No. 1) by Mr. Harold J. Cook and presented to the writer for publication, appears to be a deciduous upper premolar (dP^-) of the right side (fig. 33). This tooth is surprisingly similar to an upper tooth of a cai'- nivore (canid) and was so regarded by the writer for some time. Dr. W. D. Matthew kindly made a systematic comparison with the material in the large collection of the American Museum and suggested that it was per- haps a deciduous tooth of an Entelodont. The addi- tional material secured in the Agate Spring Quarries Fig. 33. Crown View of Deciduous (i^j-ji^g ^1,^ ggason of 1908 enables US to correctly deter- P' hetiSiAe, Dhwhyusfiollandi. Nat- t • i 8i^e. (Harold cook'8 Collection.) ^ine this tooth and it is here illustrated in order to assist the student. The chief peculiarity of this tooth is that instead of having three cusps on the posterior portion of the crown, as the tooth described above, it has only two, and in this respect agrees better with the tooth of Archseotherium described by Scott (87, p. 276). The portion of the posterior root which remains unbroken is abruptly pointed and short, having characters one would expect to find in a milk-tooth. Measurements. Mm. Great diameter of the crown 43 Diameter across the paired tubercles 27 The Skull. (Plates LV-LVIII.) The orbits in the cranium of Dinohyus are so placed that the eye looked directly forward in an unusual manner (see Plates LVI, LX, and LXI). The eye is located "Hatcher, Annals Carnegie Mnsenm, Vol. 1, pp. 260-261, 1901. PETERSON : A REVISION OF THE ENTKLODONTIDK 87 well back in the skull, the face is very long, and the brain-case short. Tliere is a well developed sagittal crest, and an overhanging occiput.''"'' The descending broad and thin plate below the orbit is quite small when compared with some Oligocene species, and the zygomatic process of the jugal terminates in a heavy Ijuttress, which forms the anterior border of the glenoid cavity. Tiie comparatively gentle down- ward sweep of the angle of the lower jaw and the relatively small process on the ventral border of the chin are other features of importance. The base of the skull has received considerable injury by crushing, and the sutures are entirely closed, but the character of the different bones are generally simi- lar to, and agree quite closely with, the descriptions of those of the American Entelo- donts given by Professors Leidy (44, pp. 57-67), Scott (87, pp. 278-287), and others. The basioccipital is short, slightly keeled ventrally, and hasai'ough area at its junc- tion with the basisphenoid. The condylar foramen is of considerable size, and, as in A. mortoni, is placed some distance in front of the condyle. The exoccipital has a great transverse diameter at the base of the occiput and narrows rapidly superiorly. The condyles are well separated by broad notches superiorly and inferiorly, and their transverse diameters are a little less than three-fifths greater than the vertical, which agrees better with the American species than with Entdodon magnum oi Europe. The foramen magnum is transversely oblong in a greater degree than is the case in the Princeton specimen, but this may in part be due to crushing. The paroccipital process is quite heavy, trihedral in cross-section, and terminates in a truncated and rugose free end ; its position is similar to that in Archicotherium mortoni figured V»y Professor Scott (87, PI. XVIII, fig. 2), but apparently relatively shorter. The stylomastoid foi-amen and the pit for the tympano-hyal occupy apparently the same relative positions as in the Oligocene genus (see PI. LVII), /. e., anterior and external to the condylar foramen. The superior wing-like processes of the supra-occipital are crushed laterally, but I judge that in their normal condition they assume characters similar to the American forms generally. The basisphenoid, as in Archicothertum ingeiis, is narrow and not keeled ventrally. There is a canal in the median line, just at the point where the basisphenoid is concealed by the union of the palatines and pterygoids ; this is also present in A. mortoni and has been figured by Scott (1. c, PI. XVIII, fig. 1 ). Whether or not there was a tympanic bulla in Dliio- hyus cannot be determined from the material at hand."* The pterygoids are much com- pressed laterally, but they extend well down and terminate in short and backwardly directed hamular processes. The foramen ovale is of considerable size and occupies '* This overhanging of the occiput in the type is due, to some extent, to crushing. '"In a well preserved skull of Dinoliym hollandi collected in the Agate Spring Quarries ( University Hill Quarry) by Professor Barbour in 1908, there are no tympanic buUx. 88 MEMOIRS OF THE CARNEGIE MUSEUM practically the same position as in the Oligocene genus, i. e., just lateral to the ridge of the alisphenoid which leads to the pterygoids. The skull is distorted in this re- gion, so that the peculiar conditions of the pterygoids, which Professor Scott describes (87, p. 283), cannot be ascertained from this specimen. Tlie parietals extend well down on the side of the cranium and terminate superiorly in the high and sharp sagittal crest, which is rounded from before backward in a manner similar to what is observed in some other American forms. The sutural contact between the frontal, supraoccipital, and the squamosal is entirely obliterated. The frontals are very broad and much inflated over the orbits, terminating laterally in a heavy process, which meets its fellow of the jugal ; the two are well coossified and form a heavy vertical bar, which completely encloses the orbit posteriorly. The orbit is of large size and faces forward and slightly outward. Immediately in front of the sagittal crest is a broad canal, which extends obliquely upward and backward, apparently piercing the cranial wall (see PL LVI). The exit of this canal leads into a depres- sion on the frontals, which is of considerable depth and triangular in outline. This depression extends outward laterally by means of two large grooves, one on either side, running parallel with the temporal ridges, and by means of a broad anterior groove, which is continuous with the depressed median area of the frontals. On account of the enormous inflation of the frontals over the orbit and the great lateral extent of the postorbital process the eye was placed lower down and had a more direct forward look than in the older types of this family (see Pis. LV, LVI). The zygomatic process of the squamosal has an enormous development, which seems to be relatively quite as great, if not greater, than is the case in the earlier forms. The lambdoidal crest descends steeply to a point immediately above the base of the paroccipital process ; the border of the arch then rises again to form a broad, thin process similar to that in Sus, the peccary, and other ungulates. An- teriorly the zygomatic process forms the sutural contact with the jugal at the base of the orbital process. The temporal fossa thus has a tremendous transverse diameter, which is one of the chief characteristics of the cranium in this family. The glen- oid cavity is well defined and its anterior border is provided with a strong buttress (the zygomatic process of the jugal referred to above), which seems to be a mark of evengreater specialization in this genus, than in the John Day forms, and has already been referred to in previous papers (81, pp. 49-51). The jugal, as has been stated by Marsh, Scott, and others, constitutes one of the most extraordinary features of the skull in this family. Anteriorly the jugal extends but little on the side of the face, uniting by suture with the maxillary and the lachrymal. In its backward extension the upper margin forms the lower half PETKRSON : A RKVISION OF THE ENTELODONTID.i; 89 of the border of the orbit, and the lower margin descends rapidly to form the dependent, broad, and laterally compressed process, which is in Dlnohiju^ consider- ably smaller than is the case in many earlier species. This anomalous process is directed slightly outward as well as downward, and no doubt varies in shape in different species, if not in different individuals ; at all events this appears to be true of the Oligocene forms.^' Superiorly the jugal sends out a postorljital process, which is coossified with that of the frontal, as stated above and the posterior process is received by the zygomatic process of the squamosal (see PI. LV). The lachrymal covers a considerable portion of the side of the face and articu- lates anteriorly with the maxillary, superiorly with the frontal, not with the nasal, and inferiorly with the jugal. The bone supplies a portion of the anterior border of the orbit. The lachrymal tubercle is quiet prominent. The long and slender nasals articulate superiorly with the spear-shaped proc- esses of the frontals, and laterally with the maxillaries and the premaxillaries. At their junction with the frontals the median area of the nasals is very convex transversel}'^ and on the sides of the muzzle the nasals have a slight inward bend to accommodate them to the concave sweep of the muzzle in front of the orbits. More anteriorly the nasals are regularly convex from side to side and gradually decrease in width ; their anterior ends are slightly damaged in the type, but it is seen that their free ends are slightly separated in the median line, and also that they projected slightly beyond the superior border of the premaxillaries. The premaxillaries are more truncated anteriorly than in the known Oligocene forms. This causes the reduction of the median pair of incisors stated above, and the anterior border ascends slightly more rapidly than in the earlier forms. Super- iorly the premaxillaries have a long contact with the nasals and posteriorly they are received by the oblique border of the maxillaries. The premaxillaries are of large size, but are relatively smaller than those in Hippopotanius and 8us. The incisive foramina are large and the palatine processes of the premaxillaries are only thin bony bridges, which are suturally connected in the median line, and extend back opposite the posterior portion of the canine. The maxillaries have a great antero-posterior diameter and the sides of the long and comparatively narrow muzzle are largely made up of these bones, although the lachrymal and the jugal cover a considerable space. The alveolar border forms an almost straight antero-posterior line until the canine is reached ; at this point the maxillary has a rapid outward curve to accommodate the root for the large canine. The maxillo-preraaxillary suture is somewhat short in comparison with that in older ■" Theae prooesses are preseut ia all of the Aiuericau species, of wliicli complete material, represeDting this por- tion of the skull, has been found. 90 MEMOIRS OF THE CARNEGIE MUSEUM forms. The infraorbital foramen is above the posterior part of P-^. The palatine plates are comparatively narrow transversely and have a considerable concavity in the same direction. In the region of the palatines the skull is much crushed and the sutures are entirely obliterated, so that the outlines cannot be traced. The posterior nares are long and narrow, and extend anteriorly to opposite the anterior portion of M-. From Kowalevsky's illustrations (38, Plates XVI ; XVII, fig. 5) it is evident that the base of the skull and the occiput in Entelodon is quite different from what is observed in the American forms. Judging from Kowalevsky's figures, the occip- ital condyle in the European form has a much greater vertical diameter in propor- tion to its transverse, the region of the paroccipital process extends further below the condyle, and the process itself is different in shape, being more compressed antero-posteriorly, expanded transversely, projecting well downward, and termi- nating in a rounded point. The base of the zygomatic process of the squamosal is in contact with the paroccipital process not unlike what is seen in Sus (See fig. 2 on page 45). The anterior face of the paroccipital process (see Kowalevsky, PI. XVI), where we should expect to find the solidly fused region between the base of the paroc- cipital process and the postglenoid similar to that in the American Entelodonts, shows a remarkable similarity to what is seen in Sus, and it would not be surprising to find the external auditory meatus situated higher up on the cranium in Entelodon than is the case in the American species. In examining the posterior view of the skull of Entelodon {I. c, PI. XVII, fig. 5) it is at once observed that the summit of the supraoccipital is greatly expanded and again much contracted midway between the top and the condyles, and on either side of the median line is a long and narrow excavation, which is not present in the American forms. Another similarity to Sus, and also to some extent to Hippopotaimcs, is seen on the superior border of the fora- men magnum, where the continuation of the supraoccipital sends down two blunt projections, which impart an irregular upper border to the foramen. This latter feature does not appear in the American forms. The mandible is much elongated to conform with the extremely long muzzle. The horizontal ramus is, as in the Oligocene forms, nearly straight, and is character- ized by considerable depth and thickness. The inferior border has one large knob-like tubercle in the middle region beneath P^ and M^, and only a trace of a tubercle opposite the posterior face of the symphysis. The angle is also extended below the horizontal line of the inferior border, but not to the extent seen in some earlier forms. These knob-like processes which are given off from the ventral border of the mandible are proportionally small, especially the anterior pair, when PETERSON : A REVISION OF THE ENTELODOXTID/E 91 compared with some of the large species {Pelonax ramosinn) from the upper Ohgocene. In the latter form these knobs on the lower jaws are particularl}^ long and heavy, indicating with other peculiarities a divei-ging line at that time. In Dinohijus the alveolar border has a rapid outward turn at the canine to accommodate the enormous root of that tooth. The chin is square and the symphysis shows no mark of division between the two rami. There are two mental foramina : the anterior, which is the larger, is situated below P^ ; and the posterior below P^_r. The ascending rami are low and rapidly flare outward in order to meet the widely separated glenoid cavities of the squamosals, which is a peculiarity most nearly approaching what is observed in Hippopotamus. The condyles are greatly convex antero-posteriorly and their transverse diameter is considerable ; they are only very slightly elevated alxjve the alveolar border. The coronoid process is low, but broad, and its free end terminates in an enlarged truncated rugosity. The temporal fossa is deep and extends across nearly the entire antero-posterior face of the ascending ramus, but does not reach very low ; its inferior border is developed into a heavy ridge, which unites with the base of the condyle externally. The dental foramen is of rather small size and is a little posterior to the middle antero-posterior diameter of the sigmoid notch and on a level with the alveolar border. The h3'oidean arch of Dinohi/ns holhnidi is represented b}- a number of frag- ments and three very nearly complete bones, the stylohyoid and both ceratohyals. Fig. 34. (1) Inner View of Right Ceratohyal; (2) Inner View of Styloliyal ; (3) External View of Stylohyal. Belonging to type of D. holhmdi Peterson. Nat. Size. (Cam. Mas. Cat. Vert. Fobs., No. 1594.) The different bones appear to be of somewhat smaller proportions than in Archxo- therium ingcns described by Professor Scott. The shaft of the stylohyoid is a sinuous rod with prominent and spiral ridges (of which one extends the entire length of the 92 MEMOIRS OK THE CARNECxIE MUSEUM shaft) and expanded upper and lower ends. The upper half of the shaft is tri- hedral in cross-section, while the lower half is subcylindrical. In Ar^;ha'otheri^^m, according to Scott (87, p. 286), the upper portion of the shaft is compressed later- ally and has a more oval section lower down. The cerato-hyal is shorter than the st3dohj'oid in about the same proportion as in Archscotherium. The proximal end is expanded and cup-shaped, and immediately below the enlarged end the shaft becomes quite thin transversely, l)ut broad antero-posteriorly. The lower half of the shaft is still more oval in section and terminates in an enlarged end for car- tilaginous attachments. The bone as a whole is very like that in Ardntothermm. Measurements. Mm. Greatest length of skull 883 Distance from postorbital process of jugal to and including premaxillaries... 528 " " " " " " " " " occipital condyle 305 Length of alveolar border of maxillary and premaxillary 465 Distance from median incisor to Ml 330 Greatest transverse diameter of condyles 165 Vertical diameter of condyle 56 Greatest transverse diameter of skull, approximately 580 Transverse diameter of muzzle in front of orbits, approximately 175 " atP4 132 Greatest length of mandible 725 Depth of mandible, inferior border of angle to and including condyle 200 " " " at M3 120 " " " " P^ 100 Vertebral Column. Atlas. — The atlas, on the whole, is very similar to that of the Oligocene genus. The more important differences are here given. The posterior exit of the vertebrar- terial canal is shifted lower down on the posterior face of the transverse process than in Archxotherium. In the latter genus the position of this foramen is higher up, or, nearly on the superior face at the base of the transverse process.^ In Dino- Jtijiis the facets for the axis are relatively higher and narrower, the cavities for the occipital condyles shallower, the neural canal narrower, and the neural spine some- what more prominent than in ArcJmotherium. The transverse process does not pro- ject backward nearly as much as in the Princeton specimen, and consequently gives to the atlas of the latter a relatively greater total antero-posterior diameter. Axis. — The axis of Dinohyus differs from that of Arch n'othermm in having the ** In a specimen (No. 1665) of the Carnegie Mnsenm collection, from the Titannthermm beds, Sand Creek, Sionx Co., Nebraska, this foramen is placed lower down than in the Princeton specimen. PETERSON : A REVISION OK THE ENTKT.ODONTIDil?, 93 articular facets for the atlas more wedge-shaped ; the facets are also of relatively less transverse and greater vertical ^* diameter, and the transverse process is heavier. This enlargement of the transverse process is especially noticeable at the superior Fig. 35. Anterior and Superior Views of Atlas of D. hollandi Peterson. ] nat. size. base over the posterior exit of the vertebrarterial canal where there is a blunt and rugose tubercle. The well defined ridge, which extends backward from the odon- toid process along the floor of the neural canal in Archivofhcrmvi,, is in Dinohyus Fig. 36. Posterior and Inferior Views of Atlas of D. Iiolland I Peteraon. \ nat. size. converted into a broad rounded surface, which slopes backward and is interrupted by a rough ridge, which extends across the entire floor of the neural canal some "The vertical diameter seems to liave developed superiorly, as the ventral projection is of about the same pro- portion as in Arc/in'olheiium. 94 MEMOIRS OF THE CARNEGIE MUSEUM distance back of the odontoid process. Back of this ridge is a ti'ansverse excava- tion, 8 or 10 mm. in antero-posterior diameter, which is again succeeded by an elevated area along the posterior portion of the floor of the neural canal. As in Fig. 37. Lateral, Posterior, and Anterior Viewa of Axis of D. hoHandi. J nat. size. ArchseotJieriurii, the pedicles of the neural arches are not pierced by foramina for the passage of the second pair of spinal nerves. These canals are more lateral and are bridged over superiorly by heavy arms of the transverse processes which unite with the sides of the pedicles. TIdrd Cervical. — The third cervical vertebra differs from that in Archieotherium (87, p. 290) by having a proportionallj' much less developed neural spine, a sharper inferior keel, and the bony bridge, which bounds the vertebrarterial canal externally, Fig. 38. Lateral View of Third Cervical of D. hollandi Peterson. \ nat. size. Fin. 39. Posterior and Lateral Views of Fourth Cervi- cal of D. hollandi Peterson. J nat. size. narrower. As Professor Scott has stated (/. c, p. 290) in connection with his study of Archaeotherium, there are no foramina for the spinal nerve through the anterior portion of the pedicles as in Su.i. Fimrth Cervical. — The neural spine of this vertebra is damaged, but I infer that, as in the preceding vertebra, it is relatively less developed than in the Oli- gocene genus. The inferior lamella has a slightly different shape, being more deci- dedly separated from the diapophysial process and is less developed than in Arclixo- therium (87, p. 290). In Dinohyus there are deep excavations on the inferior side I'ETEIfSON : A REVISION OF THE ENTELODONTID.E 95 of the base of the diapophyses, which are scarcely indicated in the Princeton speci- men. The diapophyses are fully as well developed as in Archeotherium, and the inferior keel is sharper than in the latter genus. Fifth Cervical — As in the preceding vertebra, the neural spine is lost. The centrum is relatively much more opisthoccclous, and the inferior lamella is slightly less developed and less everted than in Archicotherium. The inferior keel is sharp Fig 40. Posterior and Lateral Views of Fifth Cer- vical of D. holtandi Peterson. J nat. size. Fig. 41. Inferior View of Fifth Cervical of /). hoUandi Peterson. \ nat. size. and continues to the postero-inferior border of the centrum, where a blunt tubercle is located on either side of the median keel. In Arcluvotherium the keel is heavier and it terminates posteriorly in a heavy rugose mass, which indicates a relatively Fig. 42. Posterior (1), Lateral (2), and Anterior (3) Views of the Sixth Cervical of /). /ioHa»(Zi Peterson. } nat. size. heavier attachment for muscles than in the Miocene genus. In the latter genus the diapophyses present a more trihedral section and the ends are less expanded than in ArchxotheHum. 96 MEMOIRS OF THE CARNEOIE MUSEUM Sixtk Cervical. — The base of the neural spine of the sixth cervical vertebra is present, and tiie size indicates a proportionally less developed process than is the case in the corresponding bone in Archasothe- rium. As in the preceding vertebra, the cen- trum is relatively more opisthocoelous than in the sixth cervical of the Princeton specimen. The inferior keel is also sharper and terminates in front in a swollen area. The anterior face of the centrum extends down over this area, which adds to the convexity of the centrum in a peculiar manner, quite unusual in most of the Artiodactyla, and is not unlike what is seen in the recent horse. The diapophyses are of aljout the same proportionate size as in Archaso- therium, but the inferior lamellaj are relatively more developed, which is remarkable when one bears in mind that the neural spine is relatively less developed, while the neural arch has about the same proportions in both genera. Seventh Cervical. — As in the preceding ver- tebrae the neural spine of the seventh cervical vertebra of JHnohyus is relatively smaller than in Archseotheriuin. The postzygapophysis is placed higher up than in Arclnvotherium, but is less concave and looks outward more decidedly than in the Oligocene genus. The pedicle is more rounded and the postero-inferior portion of the centrum is proportionally somewhat heavier, and has decided tubercles on either side of the median line, which are absent in Archxothe- rinm. The transverse jjrocess has the same pi'oj^ortions as in the Princeton specimen, and, as in that specimen, is not perfor- ated by the vertebrarterial canal. First Dorsal Vertebra. — In the type of Dinuhyns Jiollaiidi there are fourteen dorsal vertebne which were found in the quarry articulated with one another by their zygapophyses. p^^ ^^ j^,^^;^^ yj^^ Tlie first dorsal vertebra is characterized by a long and very of Seventh cervical of n. heavy neural spine greatly exceeding in length that of the ''""'""'' p^'^^""- i "»' seventh cervical. The neural spine has about the same rela- tive size as in the Princeton specimen, but the postzygapophysis is more distinctly separated from the transverse process by a constricted area, and the centrum is Fig. 43. Posterior and Lateral Views of the Seventh Ceivical of D. hoUandi Peterson. \ nat. size. PETERSON : A REVISION OF THE ENTELODONTID.'E 97 more opisthocoelous. Tliis vertebra is, except in these minor details, quite similar to the corresponding bone in Archwotherinm. Second Dorsal Vertebra. — The neural spine of the second dorsal vertebra is higher than that of the first, but not nearly so robust. In relative proportions the spine agrees quite well with that of Archieotherium. The vertebra as a whole Flu. 45. Lateral and Posterior Views of First Dorsal of I), hollandi Peterson. ', nat. size. Fio. 46. Lateral View of Second Dorsal of D. Iiul- litndi Peterson. J nat. size. Fio. 47. Lateral View of Third Dorsal of I), hotlandi Peterson. 'i nat. size. differs in some important particulars from that of A rcli;votheriurii, viz.: the vertical diameter of the centrum is relatively greater, owing to the unusual development of the inferior keel ; the anterior half of the keel is developed into a heavy and rugose tubercle, which projects well below the anterior face of the centrum ; the prezyga- pophyses nearly meet in the median line, forming a narrow in-egular emargination, while in Archieotherimn they are more widely separated. The posterior face of the 98 MEMOIRS OF THE CARNEGIE MUSEUM centrum is decidedly more concave than in the Oligocene genus. The general out- line of the bone, except in the matter of its deeper centrum, is quite similar to that of the Princeton specimen. TJiird Dorsal Vertebra. — The tip of the spine of the third dorsal is broken off, but it was evidently as high as that of its predecessor, and only slightly lighter. %l V ^ Fig. 48. Lateral View of Fourtli Dorsal of D. hollandi Peterson. } nat. size. Fig. 49. Lateral View of Fifth Dorsal of D. hollandi Peterson. } nat. size. One of the most important differences between the vertebral column of Dinohyus and that of ArchieotJieriumi is seen in the arrangement of the canals through the sides of the pedicles and the vertical canals through the superior portion of the neural arch, similar to that in Sas and Bison americanus. In Professor Scott's admirable memoir (87, p. 290) it is stated that the pedicles of the neural arch are not perforated by vertical canals as in Sas. While this is to a certain extent true, it is seen upon very clo.se scrutiny of the specimen, that some of the dorsals have the vertical as well as the transver.se canals ^Ji'csent, although they are relatively smaller, are more irregular in their position, and could not have possessed the functional PETERSON : A REVISION OF THE ENTELODONTIDvE 99 importance which they have in tlie Miocene genus. Back of the second dorsal in DinoJi-yus hollandi the arrangement of these canals is almost identical with that in Sus, the vertical canal being even relativel}- larger. The costo-vertebral canals are very narrow but deep excavations, which lead directly into these large horizontal and vertical canals, forming an excellent protec- tion for the spinal nerves. The keel of the centrum is fully as well developed as in the preceding vertebra. In Archscothermm the corresponding vertebra has a heavy and rugose area on the anterior half of the keel, but the ventral projection is not nearly so great. Fourth Dorsal Vertebra. — In Dinohyus this vertebra is more nearly like that of the Archfeotherium skeleton from Princeton than any of the preceding dorsals. The vertical diameter of the centra back of the second dorsal in Archieothcrium rapidly increases, so that the centrum of the fourth has about the same proportionate diam- eter as in Dinohyus. The arrangement of the foramina through the walls of the neural canal, which are sometimes absent in Archxotherium ingens, is the chief char- acter showing a difference between the vertebrte of the two animals. Fifth Dorsal Vertebra. — The top of the neural spine of the fifth dorsal is com- plete, but unfortunately the contact is lost at the lower half of the spine. The length of the spine is estimated by a gradual slope from the second to the seventh dorsal vertebrte, which have complete neural spines. The general construction of this bone differs so little from that in Archtcofheriam that a description seems hardly necessary.^" Sixth and Seventh Dorsal Vertebrse. — With the exception of the broader and undoubtedly longer neural spine in the sixth dorsal vertebra, the seventh differs very little from it, and the two are so similar to the fifth dorsal that a separate description seems superfluous in this connection. With the exception of the per- forations of the transverse processes, the deep median furi-ow near the base and rugose surfaces on the anterior faces of the neural spines, and the somewhat more sharply keeled centra, there are no differences from the corresponding vertebriB in Archwothermm described by Professor Scott. Eighth Dorsal Vertebra. — This vertebra has no ventral keel on the centrum, it being quite smoothly rounded below, and consequently having a smaller vertical depth than the vertebrte anterior to it. The median furrow or groove on the anterior face of the neural spine extends well up and is still quite deep on the fragment we possess, proving that the spine itself must have attained a consider- *" In cleauing off the matrix for a closer examination of the transverse process of the fifth dorsal vertebra in Archivulhenum ingens, which Professor Scott described, it is found that there is a small foramen on the upper face of the neural arch which is apparently connected with the horizontal canal as in Sus. 100 MEMOIRS OF THE CARNEGIE MUSEUM able length. The tubercular facet for the rib is of large size and concave in the antero-posterior direction. The transverse process is heavy, as is also the upper portion of the neural arch. The pedicle and transverse processes are perforated by the canals referred to above. Nirdlt Dorml Vertebra. — The ninth dorsal is represented only by a portion of the neural spine, half of the neural arch, and the upper third of the centrum of FlCJ. 50. Lateral View of Sixth Dorsai of D. hollandi Peterson. \ nat. size. Fig. 51. Lateral View of Seventh Dorsal of D. hollandi Peterson. J nat. size. Fig. 52. Lateral View of Eighth Dorsal of D. hollandi Peterson, i nat. size. the left side. The capitular and tubercular facets are only slightly smaller than in the preceding vertebrtc. The transverse process is of aljout the same size, but the perforations are somewhat smaller than in the eighth dorsal vertebra. Tenth, Eleventh, and Twelfth Dorsal Vertehrx. — The tenth, eleventh, and twelfth dorsal vertebrae are represented only by fragments, but enough remains to supply essential characters. The centra are depressed and apparently have no ventral keels, and the neural spines still continue to have the median groove which extends from the base upward for some distance on the anterior face of the spines. On the tenth dorsal the rib-facets possess considerable size, Ijut they rapidly diminish in the following vertebrre. The neui-al spines also show a rapid reduction, but the perfo- PETERSON : A REVISION OF THE ENTEI.ODONTID.T; 101 rations of the pedicle and transverse processes continue to be of large size up to the thirteenth dorsal. There is only a slight suggestion of the metapophysis on the twelfth dorsal vertebra. Thirteenth Dorsal Vertebra. — In the region of the thirteenth dorsal there is an important change in the dorsal series. The neural spine was still, no douV>t, of the dorsal type, but short. The anterior zygapophysis is also of the dorsal type, while the postzygapophysis fits into a nearly /'' \ typical lumbar articulation of the succeeding vertebra. The \ \ transverse process is quite heavy and has an almost flat sur- •.. \, face for the tuberculum of the rib. The metapophysial ■', area is divided into three small tubercles and the superior aspect of the transverse process is on the whole very rugose. The vertical perforation is very small and is placed near the anterior margin of the neural arch. The horizontal canal for the transmission of the spinal nerves, which perforates ^^'°- ^s- Lateral view of Thirteenth Dorsal of D. hoi the wall of the neural arch between the posterior capitular ^^^^. ^^^^^^ , ^^^ ^.^^ and the tubercular facets is of large size. This vertebra appears to correspond most closely with the vertebra described by Professor Scott as the eleventh thoracic vertebra in Archnvtherium ingens (87, p. 293). Fourteenth Dorsal Vertebra.— The fourteenth dorsal of Dinohi/us agrees most closely with the one which Professor Scott regards as the twelfth of Archiridherium ingens. In Z);7io/(v/«s this vertebra differs in many important characters from either the twelfth or thirteenth dorsals in the Princeton specimen in which the thirteenth is the last dorsal according to Scott. From both the twelfth and the thirteenth in the latter specimen the fourteenth thoracic vertebra in Dinohnus differs in having a very well developed transverse process, a large perforation through the sides of the neural arch back of the capitular facet for the last rib, and the absence of anapophyses. The superior portion of the neural spine is broken off', but its broad antero-posterior aspect indicates that in shape it (rlosely resembled the lumbar vertebrse. Tlie pos- terior zygapophysis is identical in form with that of the twelfth (thirteenth?)" m Archieothermm. As Scott has shown in his memoir, page 293, the cros.s-section of the zygapophyses of the lumbar vertebrte present an S-like outline which causes the development of large episphenial processes. The neural canal is transversely broad, but low. The centrum is well rounded and has no ventral keel or prominent rugosi- ties for muscular attachments. The facet for the last ril/is of large size, oval in "The writer is of the opinion that Archieolherium from the Oligoceue may have had fourteen thomcic (dorsal) vertebrie thus agreeing in its vertebral formula with Dlnoliym. 102 MEMOIRS OF THE CARNEGIE MUSEUM outline, and placed in an oblique position. The metapophyses are prominent on this vertebra. Lumbar Vertebrae. — There are only four lumbar vertebrae represented in the type specimen of Dinohyus, but from the fragments at hand it is very plainly shown that the last lumbar has been lost.^^ There are in the splendidly preserved speci- FlQ. 54 . Anterior, Inferior, and Lateral Views of Fourteenth Dorsal of D. hollandi Peterson. J nat. size. Fig. 55. Lateral and Interior Views of First Lnmbar of D. hollandi Peterson, i nat. size. men of ArchcVotherium in the Princeton collection, six lumbar vertebree, which un- doubtedly also must have been the number in the genus Dinohyus. This arrange- ment is thought to be correct in view of the evidence at hand from the Miocene and Oligocene genera and also from the fact that in Sus scrofa and the Pecora there are sometimes twenty thoraco-lumbar vertebme : fourteen dorsals and six lumbars. Fig. 56. Lateral View of Second Lumbar of D. hollandi Peterson, i nat. size. Fig- 57. Lateral and Inferior Views of Third Lumbar of !>. hollandi Peterson. } nat. size. The centra of the four lumbar vertebrae preserved in the type are all present and in two instances portions of the arches and of the zygapophyses are also present. The neural spines and the transverse processes were unfortunately all lost in the proc- " In the spring of 1905, the fragments of the lumbar vertebrtc, the sacrum and a portion of the pubic symphysis of the pelvis was found by the writer on the edge of the quarry. These parts were dug out during the excavation in quarry No. 1, in the fall of 1904, by Mr. Cook and his assistants. PETERSON : A REVISION OF THE ENTELODONTID.E 103 Fio. 58 Inferior View of Fifth Lum- bar of D. hollandi Peterson. J nat. size- ess of collecting.^-' The neural spines are, however, restored from another specimen (No. 2193B) collected in the Agate Spring Quarries (Quarry No. 1) in 1908 and their shape (see Plates LX, LXI) is thought to be nearly correct. The centra in the anterior portion of the series are somewhat different from those in Archseotherium, having a transversely broader aspect ventrally and a very rugose surface on the sides. Posteriorly the cen- tra become more depressed and the rugosities on the sides are divided into an anterior and a posterior area by a decidedly broad and shallow groove, extending obliquely from the posterior intervertebral notch to the ventral portion of the centrum, where it fades away, leaving a sharp ridge on the anterior margin of the exit of the groove. The neural arch of the first lumbar vertebra is pierced by a canal immediately posterior to the base of the transverse pi'ocess. On the second lumbar is a similar canal on the right, while on the left side there is no canal present. In another specimen (No. 2126^ Carn. Mus. Cat. Vert. Foss.) of Dinohjus hol- landi is a last lumbar vertebra which approximates the jjroper size of the type speci- men. This vertebra presents some interesting differences from the corresponding bone in the Princeton specimen. Of these differences the inoi'e important ones are : the proportionately lighter postzygapophyses and the much heavier transverse processes in Dinohjus; the enlarged transverse process of the last sacral vertebra in Dinohyus, which is due in a great measure to the heavy and rugose posterior border, somewhat after the analogy of some of the Perissodadijla {Equus). In the last lumbar ^^-^"^'"W ^^— ^ Fig. 59. Posterior, Lateral and Anterior Views of Sixth Lumbar of D. hvllaiuli Peterson. } nat. size. (Carn. Mus. Cat. Vert. Foss., No. 2126.) vertebra of Archimtherium iiKjena the transverse process has in its posterior border near its base a small and rugose tubercle, which apparently had a tendency to meet "A series of four posterior lumbar vertebra; (No. 2139B) was found in quarry No. 1 darinj; the last (1908) season which furnishes some additional information. There were one and most probably two additional lumbar vertebne in front of this series, judging from the absence of vertebra; with no perforations of the neural arches as in the first and second lumbais of the type and also from the characters of the fourteenth dorsil, which is present in this series. 104 MEMOIRS OF THE CARNEGIE MUSEUM a corresponding surface on the anterior face of the pleurapophysis of the first sacral vertebra and is somewhat simihir to what is found in Hipjwpotamus. The trans- verse process of tlie last lumbar vertebra in Arcluvothcrium is otherwise quite light, as is tlie case in all the preceding lumbar vertebrse. In the specimen under descrip- tion the heavy posterior border of the transverse process of the vertebra does not come in contact with the anterior face of the pleurapophysis of the first sacral, but had a cartilaginous attachment with the ilium. The neural spine has a vertical position and its anterior and posterior borders gi'adually taj^er from the base of the spine to the summit. The latter is rounded, slightly enlarged, and rugose. The neural canal is broad, but low, and its floor is occupied in the median line by a heavy ridge, which extends antero-posteriorly the entire length of the superior face of the centrum. The centrum is depressed, but possesses a considerable transverse diameter ; its posterior face is larger than its anterior. The prezygapophyses have prominent surfaces for the episphenial processes of the vertebra in advance of it, and these processes of the postzygapophyses are also quite prominent. Sacrum. — The sacrum is composed of three well coossified vertebrae.^' The centrum of the first sacral vertebra is quite large, but the succeeding two rapidly Inferior and Superior Views of Sacrum of D. hollandi Peterson. Vert. Foss., No. 2126.) nat. size. (Cam. Mus. Cat. decrease in size, so that the posterior face of the centrum of the last vertebra has the transverse diameter less than one half of the anterior face of the first sacral. The pleurapophyses, which are very heavy on the first sacral, decrease in the posterior sacrals as rapidly as the centra. Unfortunately the neural spines are lost, . "In examining the sacrum of the Princeton specimen, Arcli.rolherium ingens, No. 10885, it is seen that on its ven- tral face there are tliree distinct centra, which are very firmly coossified. Professor Scott (87, p. 294) mistook the two posterior centra of this sacrum for one only. PETERSON : A REVISION OF THE ENTELODONTID.E 105 but their bases indicate that they were coossified so as to form one soHd spine. The neural canal is of consideral)le size a nd its outline anteriorly is subtrian- gular ; it diminishes rapidly backward agreeing in its diameter with all the other measurements of the sacrum. The chief differences revealed by the sacrum of Dmohyus when compared with that of Arducothcrhim are the relatively smaller antero-posterior and transverse diameters of the centrum, the actual measurements of length and breadth in the former being only very little greater than those of the latter genus, notwithstanding the smaller size of Arclixotheriuni. The vertical diam- eter of the anterior face of the centrum more nearly corresponds to that of the Oli- gocene genus. The prezygapophyses have, as in Archteotheriuni, the additional articular surfaces for the episphenial processes of the last lumbar vertebra. From material acquired in 11J08 I am able to give complete illustrations (see figs. 60-62) and also to more fully describe parts of the sacrum, which are partially or wholly lost in the t^'pe specimen. The reduced diameter, especially the antero-posterior, of the centra in the sacrum observed in the type specimen of Dinohi/us Jiollandi is repeated in other individuals. In No. 2126 (Carn. Mus. Cat. Vert. Foss.) there is a complete sacrum (see figs. GO-62) and its measurements agree quite closely with those of the type. Fio. 01. Anterior and Posterior Views of Sacrum of D. hollandi Peterson. Foss., No. 2126.) I nat. size. (Carn. Mus. Cat. Vert. Anteriorly the neural canal, which is triangular in outline, is of large size and rapidly decreases posteriorly. All the sacral foramina are of large size. The two anterior spinous processes are very heavy and quite solidly coossified, e.Kcept at their extremities,"" which are slightly separated and also greatly enlarged. The spine of the '^lu a smaller though older specimen (No. 2139) the summit of the spines is solidly coossitied and much ex- panded laterally. 106 MEMOIRS OF THE CARNEGIE MUSEUM last sacral vertebra is entirely free from those in front of it and all of its diameters, though of considerable size, are much reduced, when compared with the heavy spines of the first two vertebrae. The postzygapophyses are quite well developed and bear distinct articular facets for the prezygapophyses of the first caudal vertebrae. The posterior extension of the pleurapophyses exhibits a free area back of the iliac con- FlG. 62. Lateral View of Sacrum of D. hoi- landi Peterson. | nat. size. (Cam. Mns. Cat. Vert. Fobs., No. 2126.) Fig. 63. Anterior, Superior and Inferior Viewsof Caudal from about the Middle of the Tail of D. hol- landi Peterson. § nat. size. (Carn. Mus. Cat. Vert. Foss., No. 2135.) Fig. 64. Superior and Lateral Viewsof Posterior Caudal of D.hollandi Peterson. S nat. size. (Cam. Mus. Cat. Vert. Foss., No. 1823.) tact which is fully as large if not larger than is the case in Arclixotherium ingens (see fig. 62). Caudal Vertehrsc. — The caudal vertebrae of the type of Dinohyus are not present. From several isolated bones found scattered throughout the bone-bearing layer of the quarry and also from a vertebra (No. 2135) found in the same horizon north of the Niobrara River, it is plain that the tail was of practically the same proportionate length as that of the Oligocene genus. Cuts of this bone and of one of the posterior bones of the tail found in quarry No. 1 are here given (figs. 63 and 64). Ribs. The ribs on both sides (especially the ones in the anterior portion of the thoracic cavity) are well represented in the type of Dinohyus hoUandi. The first rib is some- what sub-cylindrical proximally and broadens almost immediately below the head, while in Arclnvotherium ingens it continues rod-like to a greater distance distally and flares out more suddenly at the ventral end. In the anterior portion of the series PETERSON : A REVISION OF THE ENTELODONTID.T; 107 the ribs are broad, with compressed sessile heads, and large tubercles, which are separated from the lieads by deep cavities in the same manner as in Archseolhcriiim. The fragments of the posterior ribs indicate that they were perhaps proportionately Fig. 65. (1) First Rib ; (2) Second Rib ; (3) Fifth Rib. From type of V. hotlandi Peterson, i nat. size. broader than those in ArcJueotheriiom. In their relative lengths the ribs are not un- like those of the Oligocene genus, those behind being relatively perhaps somewhat longer. Sternum. There are six sternebrte present in the type. The structure of the sternum is unique and can hardly be compared with that of the Suidfe. The presternum has an unusual vertical diameter, it is compi-essed transversely and the side view pre- sents an irregular and curious outline. The superior border of the presternum is slightly concave and terminates in front in a heavy and rugose tubercle ; anteriorly the bone is strongly emarginate, producing the superior tubercle which has just been mentioned and an additional, larger, inferior tubercle (see No. 1 in fig. HG). Ventrally the bone again has an emarginate border which terminates posteriorly in a rapidly expanded, truncated, and very rugose process. The contact for the succeeding segment continues, from the truncated process mentioned, obliquely forward and upward and terminates in a slightly enlarged area which forms the contact for the second pair of ribs. The attachments for the first pair of ribs are situated on the lower half of the bone, nearer the anterior tban tbe posterior border. In Archwothcrium ingens the pre- sternum has the same general outline as in l)lnnhi/us, but the borders, especially the 108 MEMOIRS OP THE CARNEGIE MUSEUM anterior, are less emarginate, which gives to the manubrium of the former genus a less striking appearance. The first segment of the mesosternum has a curious structure. It is compressed laterally and its vertical is much greater than its an tero- posterior diameter. The side view shows a Q-shaped outline with the straight border behind. The segment is expanded on its borders, those behind being the heaviest. The antero-posterior Fig. 66. (1) Lateral View of Sternum. (2) Inferior View of Sternum. Type of D. hoUandi Peterson. } nat. size. diameter of the bone is slightlj^ greater ventrally than dorsally. The bone differs greatly from that in Archfcothcrium ingens which has a relatively greater diameter vertically and a more angular outline laterally. The following two segments in the sternum of DinoJiyus are coossified and their vertical diameter rapidly decreases, while their transverse diameter increases. The transverse diameter of the dorsal border is twice that of the ventral, and the posterior face of the fourth segment very suddenly expands laterally. The succeeding two segments of the mesosternum are broader than deep and the last segment is the broadest and shallowest. The dorsal face is considerably broader than the ventral and the last two segments are slightly concave both dorsally and ventrally. The posterior border of the last segment is injured by erosion, but the suture separating the following segment (xiphisternum) is plainly indicated (see fig. 66, No. 2). The sternum as a whole is not unlike that of Arclixotlierium, but it possesses a proportionally smaller antero-posterior diameter, which seems to indicate that the thorax of Dinohyus was relatively shorter. Sternal Ribs. — There were four ossicles found with the sternum which are evidently the sternal ribs. They are much enlarged at one end and taper rapidly to the other (the attachment for the sternum) and have rugose faces for a heavy cartilaginous covering. PETKRSON : A REVISION OF THE ENTELODONTID^ 109 Measurements. Mm. Atlas, greatest antero-posterior diameter 152 " " transverse " 337 " " vertical diameter including neural spine 127 transverse diameter of articulation for condyle 166 " " " " " " axis 144 Axis, greatest antero-posterior diameter 140 " " transverse " 145 " " vertical diameter including neural spine 160 " length of odontoid process .' 44 " transverse diameter of centrum posteriorly 78 " vertical " " " " 78 Third cervical, greatest antero-posterior diameter 110 " " transverse diameter 162 " " vertical " 148 " " transverse " of centrum posteriorly 80 " " vertical " " " " including ven- tral keel 90 Fourth cervical, greatest antero-posterior diameter 105 " " " transverse " 144 " " " vertical " approximately 145 " " transverse diameter of centrum posteriorly 80 " " vertical " " " " including ven- tral keel • 98 Fifth cervical, greatest antero-posterior diameter 105 " " " transverse " 145 " " " vertical " approximately 160 " " transverse diameter of centrum, posteriorly 80 " " vertical " " " '' 97 Sixth cervical, greatest antero-posterior diameter 110 " " " transverse " approximately 140 " " " vertical " " 2G0 '' " transverse diameter of centrum posteriorly 76 " " vertical " " " " 88 Seventh cervical, greatest antero-posterior diameter 118 " " " transverse " 125 " " " vertical " including neural spine.. 318 " " antero-posterior diameter of centrum 86 " " transverse diameter of centrum, posteriorly 91 " " vertical " " " '' 82 First dorsal, greatest vertical diameter 480 " " antero-posterior " of centrum 70 " " greatesttran8ver.se " " " posteriorly 93 " " vertical diameter of centrum, postei'iorly 69 Secontl dorsal, greatest vertical diameter 512 110 MEMOIRS OF THE CARNEGIE MUSEUM Mm. Second dorsal, antero-posterior diameter of centrum, posteriorly 65 " " greatest transverse " " " " 89 " " vertical diameter of centrum, posteriorly 65 Seventh dorsal, greatest vertical diameter 365 " " antero-posterior " of centrum 67 " " greatest transverse " " " posteriorly 62 " " vertical diameter of centrum, posteriorly. 67 First lumbar, antero-posterior diameter at the zygapophyses 90 " " " " " of centrum 65 " " transverse diameter at anterior zygapophyses 98 " " " " of centrum, posteriorly 68 " " vertical " " " " 54 Sacrum, greatest antero-posterior diameter, approximately 150 " antero-posterior diameter of centrum 105 " transverse diameter of centrum, anteriorly 66 " vertical " " " " 42 " transverse " " sacrum, posteriorly 31 " vertical " " " " 22 " greatest transverse diameter of sacrum 165 Ribs. Greatest length of first rib 430 •' " " second rib 467 " " sixth rib 620 " " " eleventh rib 710 " " " last rib, approximately 54© Sternum. Greatest antero-posterior diameter 505 " " " " of presternum 172 " vertical " " " 148 " transverse " " " anteriorly 35 " " " " " posteriorly 13 " antero-posterior " " first segment in mesostei-num 60 " vertical " " " " " " 112 " transverse " " " " " " 40 Antero-posterior diameter of sixth sternebrse, approximately 90 Transverse " " " " 82 Vertical " " " " approximately 24 Greatest length of largest sternal rib 100 " transverse diameter at largest end of sternal rib 44 The Fore Limb. The comparative lengths of the different elements of the Hmbs in Dinohyus, when compared with Archseotherium, in the judgment of the writer reveal characters PETEK80N : A REVISION OF THE ENTELODONTID.E 111 of sufficient importance to warrant, when the many other differences are also taken into account, the generic separation of the two forms. From the very complete description of Archieotherium ingens by Professor Scott it is quite plain that the Oligo- cene genus was already capable of high speed, which appears to have been a require- ment of these animals throughout the successive geological ages in which they existed. We naturally look for a reduction in the length of the humerus and femur accompanied by the retention or increase of the length of the lower portions of the limbs among the later survivors of the family. The limbs of Dinohyus which fortu- nately were found with the skeleton, show a greater advance than is found in Archeo- thcrium ingens*^ in those characters which are necessary for speed and endurance. Fig. 67. External View of Right Scapula of Type of D. hullandi Peterson. } nat size. "The philosophy of the increase in length of the lower part of the limb in recent hoofed mammals has been dis- ouBsed by Dr. Matthew in the Memoirs of the American Museum of Natural History, Vol. I, Part 7, p. 432 (1901). 112 MEMOIRS OF THE CARNEGIE MUSEUM Scajiula. — The scapula of Dinohi/us is relatively higher than that of the Prince- ton specimen from the Oligocene. Its neck also has a relatively greater transverse diameter. In other respects very little difference in the general make-up of this bone of the fore limb is shown in the two genera. The pre- and postscapular fossje are divided b}' the spine in very nearly the same way in Dinohyus and Archseothe- rium, while according to Scott (87, p. 298), the John Day form, Boochceras humer- ostis Cope, has a much broader blade with the pre- and postscapular fossae of nearly equal width. Measurements. Mm. Scapula, greatest length 560 " " width 335 " transverse diameter of neck 93 " antero-posterior diameter of glenoid cavity 86 " transverse " " " " 67 Humerus. — The greater tuberosity of the humerus of DinoJnjus was unfortu- nately lost in the process of collecting, so that its elevation above the head is only conjectural.^' The bone is otherwise complete and its length is relatively less than that of Bodclicems humerosus from the John Day formation, and considerably less than that of Archiwthermm. Indeed the humerus of J)rauhi/us is even proportion- ately somewhat shorter than is the case in Bos and Equus caballus, while that of Archxotherium is relatively longer than in these two recent genera. The proximal end of the humerus in Dinohijus is, as in ArcJuvotherium, of great antero-posterior diameter ; its transverse diameter, though less than the antero-posterior, is also great, perhaps relatively greater than that in Archxotherium. The head is large and takes up a considerable portion of the proximal end. The greater tuberosity is very massive and evidently terminated above in a heavy lilunt process, as in other represen- tatives of the family. The lesser or internal tuberosity is unfortunately also dam- aged, but its anterior border indicates a strong margin which bounded the bicipital groove. The latter is broad and displays a large, smoothly convex tubercle, over which the tendon for the biceps muscle spread. The deltoid ridge is prominent and extends well down on the shaft. The distal end is much exjianded, especially trans- versely, as the supinator ridge and the surfaces for the attachments of the ligaments are well developed. The internal epicondyle which is on the point of disappearance in Archxotherium is still smaller in Dinohyus. The intercondylar ridge, which is fully as prominent as in Archxotherium, is shifted even moi-e outwardly than in that genus, and the trochlea has a more modernized appearance. The ridge which bounds the supratrochlear fossa externally is more prominent than in Archxotherium; the differ- " From material recently { 1908) acquired in the Agate Spring Fossil Quarries, it is very evident that the dotted lines in Fig. 68 are too high. PETERSON : A KRVISION OF THE ENTKLODONTID.E 113 ence in the relative depth and size of the supratrochlear, and the anconeal fossas in the two genera is small. Fio. 68. External and Posterior Views of Humerus of Fu). G9. Katlial and Anterior Views of the lllno- Type of D. kollandi Peterson. ] uat. size. Tbe dotted lines radius of Type of D. hollaiidi Peterson. ', nat. size. in the cut are too high. Measurements. Mm. Humerus, greatest length, approximately 480 " head of humerus to aud including distal end 444 '' middle of head " " " " " 425 " antero-posterior diameter, pro.ximal end, approximately 175 " transverse " " " " 120 " " " distal " " 120 " antero-posterior " " " " 105 114 MEMOIRS OF THE CARNEGIE MUSEUM Radius and Ulna. — As has been intimated above, the ulna-radius in Dinohyus is proportionately longer than in Archseotherium. ingens ; it is also more curved or bowed forward ; the olecranon process is more truncate ; and the sulcus for the exten- sor tendon is less clearly defined. The humeral articulation of the radius differs onl}' in minor details from what is seen in the Princeton specimen. The most noticeable of these minor differences are : the relatively narrower articular surface external to the intertrochlear groove, and a rugose pit, which interrupts the humeral articulation near the posterior margin. The transverse diameter of the shaft of the ulna-radius is shortest immediately below the head ; from this point it rapidly in- creases distally, and the distal extremity Jias only a slightly greater diameter than the shaft of the compound bone. All traces of the suture between the radius and ulna are practically lost, except at the proximal and distal ends. The deeply channeled external face of the shaft shown in Archxotheriam ingens is not so notice- able in Dinohyas. The ridge which separates the pyramidal and lunar facets is more developed, and the ridge separating the latter facet from that for the scaphoid is less developed than is the case in the Princeton specimen. In this way the peculiar condition found in Archseotherium, ingens in which the ulna occupies the entire proximal face of the pyramidal and also effects a lateral contact with the lunar is even more emphasized in Dinolii/us hollandi. The oblique direction of these ridges and the articulating surfaces to receive the carpus are practically nearly the same as in the Oligocene genus. Between the radius and ulna on the posterior face is a deep pit to accommodate the posterior hook-like process on the proximal end of the pyramidal when the manus is flexed backward. This pit is less developed in Arclurotherittrn. The pisiform facet in DiiwJii/iis as in ArchxiAlieriurii is continuous with that of the pyramidal. Measurements. Hm. Ulna-radius, greatest length 605 " head of radius to distal cad 475 ' ' length of olecranon process 157 " transverse diameter of head of radius 105 " " " " shaft immediately below head of radius 76 " greatest transverse diameter of distal end 117 " " antero-posterior diameter of distal end 83 The Manus. Scaphoid. — The scaphoid is not present in the manus of the type specimen of Dinohgus. Judging, however, from the proximal articulation of the magnum and the corresponding articulation of the radius, which are quite broad, it is evident PETERSON : A REVISION OP THE ENTKLf)DONTrDJ? 115 that the scaphoid in the Miocene genus had a relatively greater transverse diam- eter than in ArcJueotherium from the Oligocene. In 1908 the Carnegie Museum field party was fortunate in finding (see plan of the quarry, PI. LIV, Sect. 19, No. 47) a fore foot which was almost completely I ^ ^ ■f Fia. 70. Posterior ( 1 ) and Anterior (2) Views of the Right Fore Foot of Type of /). hollandi Peterson. } nat. size. ps, facets for palmar sesamoid. articulated. This specimen (No. 2136, Carn. Mus. Cat. Vert. Foss.) adds materially to our knowledge, inasmuch as the fore foot of the type specimen was found scat- tered, though in close proximity to the skeleton. The scaphoid as a whole has a small transverse diameter, while antero-posteri- orly it has the greatest diameter of all the carpal bones, which is due chiefly to the large palmar process. The more important characters in which this bone differs from that of ArcJiHotheriurn, described by Professor Scott (87, pp. 302-303), are the 116 MEMOIRS OF THE CARNEGIE MITSEUM relatively greater vertical diameter, the enlarged projection on the dorso-ulnar angle, which causes a slightly greater transverse diameter in this region, and also the absence of the third facet for the lunar on the ulnar face. The articular surface for the radius is divided into two parts, an antero-external and a postero-internal, as in Archscotherhim. The postero-internal facet is saddle-shaped, concave antero-j^osteri- orly, and constitutes the principal articulating surface of the proximal face. The antero-external facet is small, convex, continuous with the former facet, and descends steeply to conform with the anterior portion of the articular surface on the lunar On the ulnar face of the scaphoid there are two articulating facets for the lunar, one dorsal and one palmar, while in Archfcoiherium there are three. The dorsal facet is located on the prominent point which overhangs the ulnar surface of the radius and articulates posteriorly with the proximal radial face of the lunar by a broader surface than is the case in Archvotlierium. The lunar facet, which is dorsal and distal in Archieothcrium., is, as stated above, absent in Dinohyus. This facet though small is quite distinct in the former genus, while in Dinohyus, the further development of the prominent ridge on the proximal face of the magnum has so effectually separated the scaphoid and lunar in this region that the two bones perhaps very rarely, if ever, came into actual contact. The palmar facet for the lunar is of large size and occupies the entire lunar face of the palmar process of the scaph- oid. Between the different facets the lunar face of the scaphoid is much excavated and extensively arched in the antero-posterior direction. The latter character is chiefly due to the excessive outward turn of the large j^almar process. Distally the scaphoid has three facets : one posterior for the trapezoid, one anterior for the magnum, while on the ulnar angle there is a long narrow facet which also articu- lates with the high and antero-posteriorly directed ridge on the proximal face of the magnum. The articular surface for the trapezoid is concave antero-posteriorly and is continuous with the surfaces for the magnum. The two latter surfaces are irreg- ularly convex and concave. Radially the bone is quite rugose and gently convex, while the anterior face is much more convex from side to side. From Kowalevsky's illustration (38, PI. XXVI, fig. 23) it would seem tlmt the scaphoid of EnteJodon has a smaller antero-posterior diameter, when compared with its vertical dimension, than is seen in the American forms. In Anthracotherium (see Kowalevsky, Paleontographica, Vol. XXII, p. 299 ; PI. XI, fig. 38) the scaphoid has a remarkably close similarity to the scaphoid in the Entelodontidiv, as has been pointed out by Kowalevsky and Scott (87, p. 303). The scaphoid in Sus and Hip- po2')otamus on the other hand is more unlike that in the Entcloduntidir, being broader, and of smaller antero-posterior diameter. PETF.RSON : A REVISION OF THE ENTELODONTID.E 117 Lunar. — The lunar of Dlnohyus is distinctly more modified than is the case in Entelodon and Archvolhermm. In Dirwhjius the vertical and antero-posterior diameters in relation to the transverse, are greater than in tlie Princeton specimen, a character naturally to be looked for in a later form. As in Archieotherium, tlie facet for the radius is divided into two parts, an anterior and a posterior; the former has a less abruptly rounded transverse lidge than in Archivolherium, and the latter has a relatively greater antero-posterior diameter and is deflected on the uhiar side to a greater degree. The dorsal face has a slightly dif- ferent outline from that figured by Kowalevsky (38, PI. XXVI, figs. 21-32), and also different from that which ap- pears in the Princeton specimen. The principal ditterence in outline from that of the older forms is caused by the convex facet for the magnum on tlie radial face. In ArchmAhemim, the facet is altogether the reverse, being concave instead of convex (see fig. 71). In Dinohyus it is seen that both the magnum and the unciform have yielded to tlie development and modification of the beak of the lunar so tliat the lateral contact of the two is entirely lost dorsally, and the lower end of the lunar is almost in touch with the proximal end of the third metacarpal (see fig. 70, 2). The beak which is more bluntly pointed than in ArclLWotfiermm is also apparently shifted farther toward the radial side than in the latter genus. The palmar face has a more oblique outline above ; other- wise it is ({uite similar to that of the Oligocene genus. The ulnar side has two facets for the cuneiform, which are almost identical with those of Arch gcuilierium. The dorso-proximal facet for the scaphoid is more clearly defined than in Arcfnv- otherium and the changed condition of the dorsal facet for the oiiierium crassum Marsii. < uat. , , T , . . ,1 ]• 1 r r j.\ \ size. (Carn. Mns. Cat. Vert. magnum, as stated above, gives to the radial tace oi tlie bone ^ ., . ^ ° ' ' ^ Foss., No. 1 ()(),").) a more nearly vertical position than in the latter genus. The posterior portion of the facet for the magnum is recurved downward and developed into a hook on the posterior radial angle, wliich is relatively larger than in the Princeton specimen. Meahurements. Mm. Lunar, greatest antero-posterior diameter CO " " transverse " anteriorly 45 " " " " posteriorly 50 " " vertical " anteriorly 60. " " " " posteriorly 45 Fia. 71. Anterior View of Portion of Fore Foot of Archa'- 118 MEMOIRS OF THE CARNEGIE MUSEUM Pyramidal — In Dmohijus the posterior ascending tubercle of the facet for the uhia is more prominent than in Archseothermm, which results in a greater concavity antero-posteriorly in the former. The facet for the ulna extends down on the bone in an unusual manner and hence it is proportionally greater than in Archieotherium, Hip- 2)opotamus, or Shs. The pisiform facet is more obhque than in the Oligocene form. Below this facet is a large and deeply excavated sulcus which extends parallel to the sharp ridge on which is located the above mentioned facet. The radial face has two facets for the lunar ; the supero-anterior one is plane, subtriangular in outline, and inclined obliquely upward and outward ; the postero-inferior facet is convex antero-posteriorly and concave supero-inferiorly and separated from the ulnar facet by a sharp ridge, while in Archeotheriuvi the two facets are separated by a sulcus of considerable diameter. The distal articular facet is similar to that of the Princeton specimen in every respect. Measurements. Mm. Pyramidal, greatest antero-posterior diameter 61 " transverse " posteriorly 43 " vertical " anteriorly 43 " " posteriorly 64 Pisiform. — There is no pisiform present with the type of Dinohyus. In another specimen (No. 2136, Carn. Mus. Cat. Vert. Foss.) the pisiform was found nearly in posi- tion and is here described in connection with the type specimen. There are only slight differences in the shape of the pisiform oi Dinohyus and that of the older American types. The facet for the ulna extends slightly further back and is surrounded by a much heavier border so that the proximal face of the shaft has a deep excavation while in Archfeotherium it is more evenly rounded. The facet for the pyramidal is larger than that for the ulna. From the attachment of the carpus the bone is strongly recurved backward and radially as in Archwotheriim. and the free end is enlarged into a thick knob. The bone, though comparatively small, is relatively of slightly larger size than that of the Ohgocene genus. The pisiform of Entelodon, as Scott (87, p. 304) has pointed out, is very different in its outline, though not altogether unlike that of the American forms. Scott suggests that the irregular shape of the pisiform which Kowalevsky (38, PI. XXVI, fig. 27) figures "might be due to disease." It would, however, seem from the illus- tration that the facets also differ in having an irregular convexity while in the American forms they are obliquely concave. The pisiform of Anthracotherium is considerably larger and somewhat different in form, having a more decided shaft, a less club-shaped free end, and apparently differently shaped articular facets for the PETERSON: A REVISION OF THE ENTELODONTIDTE 119 carpus and ulna (Palaeontographica, Vol. XXII, PI. XI, fig. 58). In Stis the pisiform is shorter, thinner, and deeper, while in Hippopotamus it has a general similarity to the pisiform of the EntelodontidiE. Measurements. Mm. Pisiform, greatest length 70 " vertical diameter of proximal end 32 " transverse " " " " 31 " " " " distal " 32 " vertical " " " " 30 Trapezium. — The trapezium of Dinohyas is no doubt entirely wanting, judg- ing from its rudimentary condition in the Oligocene genus," and its entire absence in the John Day form. Professor Cope states (10, pp. 62-63) that there is no trace of a facet for a trapezium on the trapezoid of Bodchosnis hunierosus, which was veri- fied by me on examination of the type specimen in the American Museum.'" Trapezoid. — Though absent in the type of Dlnohyus this bone was found in posi- tion on the magnum of a fore foot, found in section 19 (see plan of quarry, PI. LIV, No. 47) and is here described. The transverse diameter of the trapezoid (No. 2]36,Carn. Mus. Cat. Vert. Foss.) is only two millimeters greater than in Archieotheriarn, while the antero-posterior diameter is proportionately somewhat greater. The bone is consequently much com- pressed laterally and has a subquadrate outline, when viewed from its ulnar face. Radially the proximal face is more rounded before backward, by the articulating sur- face for the scaphoid, which gives this (radial) face a more irregular quadrate appear- ance. The transverse diameter of the bone is greatest proximally, and it gradually tapers distally, terminating in a border, which is rather narrow and rounded inferiorly. The facet for the scaphoid is obliquely convex antero-posteriorly and extends well back on the posterior face of the bone. The ulnar face bears three facets for the mag- num, two of which are distal ami one proximal. The proximal facet is well defined near the dorsal face and continues backward as a smooth rounded ridge on the proxi- mo-ulnar angle. The two distal facets are a palmar and a dorsal, and may be regarded as practically one facet, as they are divided only by a very shallow, almost imper- ceptible, groove, while in Archxotherhmi there is a moderately large median swelling on the ulnar face which puts the two facets at a considerable angle with each other. "* There is, in the Princeton specimen described by Professor Scott, a trapezium with a well formed facet for the trapezoid, and the bone, though laterally compressed, is of considerable size antero-posteriorly and extends well down over Mc. II, but does not articulate with the latter. (See fig. 10 on page 55.) "In my note book I find the following statement regarding the trapezium of Bo'uchceriis : " The trapezoid has a small pit radially which may have lodged a trapezium, but there is no distinct facet." 120 MEMOIRS OF THE CARNEGIE MUSEUM Between the proximal and distal facets the surface of the bone is very slightly excavated while the corresponding face on the magnum is very deeply excavated, forming a large sinus. Radially there is but one small facet : that for the rudimentary Mc. II (see fig. 72). This facet is situated near the dorsal face and has a radial position, while that in Archxothe- tA rium described by Professor Scott (87, p- 305) is directly on the distal face. There is no facet S'i'i: fi'^ '\'i'-^ tero-posteriorly, and terminates in elevated and xj^^u;^ : ' jli fibular portion of the astragalar facet is rapidly 1 ^| ' t_^i/}^ deflexed externally and ternnnates abruptly at v Jy^-&!.^' ^' "M the fibular angle, while the tibial portion is gently '• -^ ; \ ^ \iIM/ concave transversely and greatly concave antero- ^■' y%^-^f^^^^^l&^yW posteriorly. The fibular face has three facets for f ent.c "' Type of d. hoiiandi Peterson. J nat. . . , . size, /-facet. £c — ecto-ouneiform. .He - toward coossification between the navicular and ., mesooaneiform. cuboid, as tiie articulating facets are reduced and the surfaces very rugose throughout. The dorsal face is plane except the rugose and sharp fibular angle which abuts against the cuboid. On the anterior portion of the tibial face the bone is convex vertically, and more so posteriorly. This convex por- tion is suddenly succeeded by a prominent and rugose area for muscular attachments. 136 MEMOIRS OF THE CARNEGIE MUSEUM The plantar hook is quite prominent and extends well below the distal facets, while in Archicotherium it is much reduced, and is, as Professor Scott states, " hardly more than a roughened ridge." The anterior portion of the distal face is taken up by the facet for the ectocuneiform. The latter facet is divided into two areas by fibular and tibial sulci leaving a narrow isthmus midway between the two. The small facet for the mesocuneiform, on the tibial angle in Arcliseotheriuin, is entirely absent in Dinohyus. Another interesting difference in Dinohyus is seen in the very much less prominent tubercle on which the facet for the entocuneiform is located ; the facet being almost on the same level with that for the ectocuneiform, and sad- dle-shaped, with the posterior margin deflected on the base of the plantar hook. The navicular of Entelodon magnum (38, PI. XXVII, figs. 34 and 37) is rela- tively shorter antero-posteriorly, and has a less developed palmar hook than in Dinohyus. In Sfiis the navicular is quite different, being broad and articulating dis- tally with the three cuneiforms, as is also the case in Hippopotanms. In the latter genus, however, the navicular has no palmar hook, and the articulating surface between the navicular and cuboid is situated near the palmar face. Measurements of Navicular. Mm. Greatest anteror-posterior diameter 73 " transverse " 41 " vertical " auteriorly 33 " " '' posteriorly 55 Entoctmeiform. — Unfortunately the entocuneiform is lost in the type. Judging, however, from the space on the navicular, and cuneiform when in position, this bone was of considerable height, rather compressed trans- versely and antero-posteriorly, and in genei'al appar.- entlysomewhat differentfrom tliat in Archicotherium. An entocuneiform of Dinohyus (No. 2139a, Carn. Mus. Cat. Vert. Foss.) which was found in quarry No. 1, section 20 (See plan of quarry, PL LIV), is of an animal of approximately the same size as the type specimen. The bone differs some- what from the corresponding bone in Archxothe- rium, being dorsally less produced and more de- FiG.80. (1) External View of Right En- velopcd in the palmar direction ; it is not pointed tocuneiform of A. ingens Leidy (Princeton distally, and its navicular articulatiou is less ob- Musenm, No. 10885). (2) Same View of , . tx • j> ii ^ i i i xi j. xi lique. It is further to be observed that there are Corresponding Bone in D. hoUandi Peterson '■ (Cam. Mus. Cat. Vert. Foss., No. 2139a). two facets. One near the proximal end for the PETKRSON : A REVISION OF THE ENTELODONTIDyE 137 mesocuneiform (fig. 80, nic), and another small one near the distal end for Mt. II. These are not present in Archxothermm ingens (the Princeton specimen). The distal fibular angle of the bone under description is entirely taken up by a large facet, which is slightly convex in all directions, and articulates with a corresponding facet on the palmar process of Mt. III. The bone as a whole has a great vertical diameter, while antero-posteriorly and transversely it is rather compressed, especially in the latter diameter. Scott (78, p. 316) has already stated that the entocuneiform has not yet been found in connection with Anthracothermm or with Entelodon. In Sxls the ento- cuneiform is larger distally than proximally, the palmar tubercle is very prominent and the bone as a whole is altogether different from that in Archseoihermni or Dinohyus, having a large and very prominent articular facet for Mt. II. In Hippo- potamus the entocuneiform is also heavier, and broader than in the latter genera. Mesocuneiform. — The mesocuneiform is entirely ankylosed with the ecto- cuneiform, but its outline, especially on the palmar and tibial angle, is clearly defined (Fig. 79, Mc). There is, however, no articular facet on the proximal end for the navicular as in Archxotherhm and there is no facet for Mt. II, but on the proximal fibular angle is a facet for the entocuneiform. Ectomneiform. — The ectocuneiform is of considerable size and (I-shaped in general outline viewed distally or proximally. The entire proximal face is occupied by the facet for the navicular, and the distal face is taken up by the facet for Mt. III. On the palmar tibial angle of the distal face is a very minute articular surface for Mt. II. The fibular face is very rugose back of the facet for the cuboid, which is located near the dorsal face and occupies the entire vertical diameter of the bone ; there is no facet for the cuboid on the proximo-palmar angle, as in Archseotherium. In this respect the ectocuneiform of Dinohyus is more like that in some peccaries, while in ^Fippopolamus the ectocuneiform articulates with the cuboid only by the large palmar-fibular facet. The antero-posterior diameter of the facet on the distal face for Mt. Ill is proportionately greater than in Archseotherium, and, as in that genus, the facet is in its dorsal portion obliquely convex from side to side, and near the plantar region obliquely concave antero-posteriorly. The fibular border of the facet is deeply emarginated by a sulcus. Measurements of Ectocitneiform. Mm. Greatest antero-posterior diameter including mesocuneiform 52 " transverse diameter ^-' " vertical " ^^ The general outline of the ectocuneiform of ^7ite/o(Zon magnum (38, PI. XXVII, 138 MEMOIRS OF THE CARNEGIE MUSEUM figs. 34, 35, 37) is quite similar to that of Dinohyus, but the distal articulation is interrupted on the fibular angle and it is plainly seen from the figure that Mt. Ill, did not come in contact with the cuboid as in Dinohyus. Gthhoid. — The cuboid is relatively large and reveals some interesting and important differences from that of the earlier types. The articular facets for the calcaneum and astragalus occupy the proximal end in nearly equal proportions, the former facet being slightly wider,^" which is directly the opposite of what is the case in Archieothermm ingens, where the astragalar facet is the larger of the two. The cuboid of the type of Bobchcerus humerosns more nearly approaches that of Dinohyus so far as the two specimens can be compared. The astragalar facet of the present genus is plane transversely, deeply concave antero-posteriorly, and the articular surface is confined to the anterior and posterior parts, with an interrupted median area twelve millimeters long. A deep pit is located on the fibular side of this interrupted sinus, which separates the calcaneal and astraga- lar facets at this point. The calcaneal facet is less interrupted than that in Archimtherium, as described by Scott, and continues obliquely from in front upward and backward, terminating near the summit on the postero-tibial angle of the ascending palmar hook. The tibial border of the dorsal face is heavy and very rugose ; the fibular angle is also 2:)rominent and abruptly convex antero-posteriorly. The tibial face is rugose and in addition the dorsal facet for the navicular has developed a projecting ledge " so that together with the palmar facet, which also occupies a similar though much larger ledge, the cuboid and navicular bones are very strongly interlocked. Below the navicular facet near the dorsal angle is a small facet with square outlines, which articulates with the ectocuneiform. The latter bone was undoubtedly connected posteriorly by cartilage only, as the corre- sponding surfaces on the two bones are very rugose with no distinct facets, such as are found in Archxotherinm ingens. Immediately below the facet for the ectocunei- form there is a third facet on the tibial angle.'^^ This facet articulates in a promi- nent and quite important manner with a process of Mt. Ill which extends above the head of Mt. IV similar to, though less prominent than, that in the manus (figs. 70, 78). These articulations are not present in the peccary, Sus, Hijj'pojwtaimis, or Entelodon magnum, but on Mt. Ill in Archasotherium there is a very slight indication ^* On page 448 (38) Kowalevsky says that the cuboid of Entelodon magnum has a narrower facet for the astragalus than for the calcaneum, which is also well shown in his illustrations (PI. XXVIII). *' In Boiiclioerus humerosus this projecting ledge is more developed than in the Princeton specimen from the Oligo- cene, but is not so prominent as in Dinohyus. ''In the type of Boodicerus humerosus this facet is not present on the cuboid, but in a smaller individual from the green sandstone (middle beds) of the John Day formation (Yale Museum No. 12765 ; Collector, L. S. Davis, 1875) this facet is fairly well developed on the cuboid. PETERSON : A REVISION OF THE ENTELODONTID^ 139 of a facet on the fibular angle of the head, which is quite prophetic of the conditions found in Dinohyus. The anterior portion of the distal face is taken up by the facet for Mt. IV, that for Mt. V being entirely absent in the type. The palmar face is greatly excavated above the hook ; the latter is very prominent and covers the entire lower portion of the palmar face. Distally there is on the tibial angle a facet of considerable size for the palmar process of Mt. IV. The vertical diameter of the cuboid- is relatively less than in the Princeton specimen, which in this respect as well as in the deeper tendinal sulcus on the fib- ular side of the Carnegie Museum specimen, seems to agree more nearly with Ente- lodon magnum (38, pp. 448-149, PI. XXXVII, figs. 34, 35, and 36). In the latter species the facet on the palmar hook for the palmar process of Mt. IV is absent, and the facet for Mt. V on the fibular-distal angle is present, just the reverse of what appears in Dinohi/us, in which there is no facet for Mt. V, and there is a facet present on the palmar hook for Mt. TV. Measukemento of Cuboid. ^fm. Greatest antero-posterior diameter 70 " transverse " -50 " vertical " anteriorly .5.5 " " " posteriorly 71 . Metatarml II. — The second metatarsal is reduced to a very small oblong nodule about half the size of Mt. II in the Princeton specimen. The nodule is held in place chiefly by the distal end of the entocuneiform, for which it has an articulating facet of considerable size on the palmar face. On the proximal tibial angle is a small facet for the ectocuneiform. Measurements of Second Metatarsal. .Mm. Greatest length "22 " transverse diameter 13 ' ' antero-posterior diameter 7 Metatarsal III. — The third metatarsal is larger and slenderer than ^Ic. Ill, but the articular facets for I\It. IV on the third metatarsal are sui'prisingly similar to the corresponding facets for Mc. IV. Thus the fibular angle of the head rises to a prominent tubercle which overlaps the head of Mt. IV and articulates with the cuboid in much the same manner in which Mc. Ill articulates with the unciform. This character of the pes seems to be a modification ])erhaps brought about since Oligocene times, as the genus Archxotlicrmm has only a very faint suggestion of a facet for the cuboid, and has the fibular angle of the head of Mt. Ill only very 140 MEMOIRS OF THE CARNEGIE MUSEUM slightly higher than the tibial angle.^^ The proximal facet for the ectocuneiforra is slightly concave transversely in the dorsal region, and convex antero-posteriorly in the palmar region to conform with the facets of the ectocuneiform already described. On the postero-fibular angle of the head is a very minute facet which articulates with Mt. II. Below this facet is a rugose shallow channel, which receives the second metatarsal. The palmar process is strongly developed, and, as in Archasotherium, bears three facets ; the one on the tibial side is for the entocuneiform, the one on the fibular side is for Mt. IV, and the one on the palmar end is for a large sesamoid. The facet for Mt. IV, which is near the dorsal face is larger and more pronounced in Dinohyus than in the Princeton specimen. The shaft is quite flat, long, and slender on the palmar and fibular faces, round on the tibial and dorsal faces, and is similar to that of Archasotherium, except for its greater rugosity on the fibular face, which is a marked character, showing a tendency toward the coossification of the two functional digits in Dinohyus. The large tubercle which is developed on the fibular border of the dorsal face just above the trochlea in Archseothermm, is in Dino- hyus developed into a swollen area which occupies the entire dorsal face of the shaft, causing a deep depression distally. This depression is the superior boundary of the distal trochlea. The latter is, as in Archieotherium, rather low and narrow. The carina, which is confined principally to the plantar side, is plainly indicated on the dorsal face by a faint ridge, which continues to the extreme proximal border of the articulation. On either side of the carina is a shallow, depression, which causes a slight lateral convexity in the tibial and fibular portions of the trochlea. Measurements of Third Metatarsal. Mm. Greatest length 240 ' ' ' antero-posterior diameter of head 55 " transverse " " 40 Antero-posterior diameter o' sliaft medially 30 Transverse " " " 30 " " " trochlea 40 Antero-posterior " " " 45 Metatarsal 1 V. — The fourth metatarsal is of very nearly the .same length and thickness as the third. The head does not rise as high as that of the third, but the palmar process is, as in Archasotherium, longer than that of the third metatarsal. Dorsally the articular facet for the cuboid is slightly concave in all directions and rises into an antero-posterior convexity in the palmar region. There is a large facet for the cuboid on the end of the palmar process ; the tibial face of the process is ■'"Some species of the John Day formation more nearly approach Dinohyus in the matter of the articular facets of the cuboid and Mt. III. PETEKSON : A REVISION OF THE ENTELODONTID.E 141 occupied by the facet for Mt. Ill ; and immediately behind and at almost right angles is a third facet, which corresponds to that on Mt. Ill and which supports a palmar sesamoid. There is no evidence of a facet for Mt. V on the fibular angle of the head. Dorsally the shaft is quite flat on its dorsal, tibial, and fibulai- faces, while the palmar face is unevenly rounded. Towards the distal end the shaft is more cylindrical, the palmar face being l^y far tlie flattest. Distally the bone is similar, including the trochlea, to that of the third metatarsal. Measurements of Fourth Met.\tarsal. Mm. Greatest length 237 " antero-posterior diameter of head 59 " transverse " " 41 " " " shaft medially 28 Antero-posterior " " " " 32 " " " trochlea v. 33 Transverse " " " 37 Metatarsal V. — Judging from the total absence of facets lor the fifth metatarsal on the cuboid and Mt. IV it would seem that this element is entirely wanting in Dinohyus. There is, however, a minute scale-hke eminence on the fibular angle on Mt. IV, which may, or may not, represent the fifth metatarsal. A water-worn fragment of a fourth metatarsal which I refer to Dmohi/'iis (No. 1927, Carn. Mus. Cat. Vert. Foss.) was found in the Agate Spring Fossil Quarry, No. 1, where the type of DmoA^H.s- was found. On the postero-fibular angle of the facet for the cuboid of this metatarsal the articulation is flexed downward in such a manner as to suggest the presence of a small fifth metatarsal in this individual. Palmar Sesamoid. — The palmar sesamoid is a large bone, triangular in outline, compressed antero-posteriorly, and much expanded, especially in the tibial and distal directions. The palmar surface is rugose and on the dorsal face near the fibular angle are two facets for the third and fourth metatarsals described above. There are facets on the palmar processes of the metatarsals in ArchtvotJierium which indicate the presence of similar sesamoids. Measurements of Palmar Sesamoid. Mm. Greatest vertical diameter 38 " transverse " 32 ' ' antero-posterior diameter 12 Phalanges. — The proximal and median phalanges are proportionately longer and narrower than tho.se of the manus ; otherwise they are similarly de{)rcssed antero-posteriorly and exj^anded transversely. The articulation for the metatarsal 142 MEMOIRS OF THE CARNEGIE MUSEUM on the proximal phalanx is divided into two parts by a faint groove, which extends across the face and terminates in the deep plantar notch. On either side of the latter are facets on the extreme plantar angle of the proximal end which abuts against the sesamoids. The distal trochlea is, as in the manus, simply convex antero-posteriorly and concave transversely. The median phalanx is quite similar to that in the manus. The ungual phalanx is short, high, and somewhat com- pressed. They suggest strongly those of the recent camels though proportionally larger. Measurements of Phalanges. Proximal phalanx. Mm. Greatest length 78 " transverse diameter of proximal end 42 " antero-posterior " " " " 3.5 " transverse " " distal " 37 " antero-posterior " " " " 23 Median phalanx. Mm. Greatest length 57 " transverse diameter of proximal end 36 " antero-posterior " " " " 35 " transverse " " distal " 34 " antero-posterior " '' " " 30 Ungual phalanx. Mm. Greatest length 39 " vertical diameter, approximately 30 ' ' transverse ' ■ 27 Sesamoids. — The sesamoids, though of considerable size, are much smaller than those of the manus. They are heaviest anteriorly and taper to an obtuse hook pos- teriorly. The dorso-proximal angle carries an articular surface which abuts against a corresponding surface on the proximo-plantar angle of the first phalanx described above. Restoration of Dinohyus Hollandi. (Plates LX, LXI.) The most conspicuous features of the skeleton of Dinohi/us hollandi are the disproportionately large skull, the long muzzle, the heavy neck, the high withers, and the elongated limbs. Many other characters are impossible to fully appreci- ate without seeing the skeleton mounted in full relief The thorax has nearly the same proportions as in Bos tcmrus, but the lumbar region is shorter. In 8us the lumbar region is relatively longer than in Dinohyus, and the latter genus is, in this PETERSON : A REVISION OF THE ENTELODONTID.'E 143 respect, more analogous to the Hippopotamus. The sternum is of enormous size and there are osseous sternal ribs present. The fore limbs are powerful and the skeleton in this region is not unlike that of Bison americanus. In the region of the fifth, sixth, and seventh cervicala the neck was of enormous vertical diameter. There are only four lumbar vertebrre present derived from the t\'pe and a fragment of the pubis rej^resents all that was found of the pelvis. These parts were found in a pile of fragments left on the edge of the quarry by Mr. f'ook who started to excavate in the fall of 1904. The pelvic material, which was found in quarry No. 1 in 1908, is used in this restoration and its anatomical features are thought to be practically correct. The right femur and certain bones of the manus were also found in quarry No. 1 and are here used for the purpose of accuracy. The right cuboid, the entocuneiform and Mt. IV are derived from other individuals. The caudal region is restored throughout. The skeleton Avas mounted by Mr. Serafino Agostini of the paleontological staff, to whom much credit is due for the skill he has displayed in his work. In 1894 Professor Marsh published (64, pp. 407-408, PI. IX) a restoratiou which is based on rather incomplete material and as a consequence a number of errors occur, some of which were pointed out by Professor Scott in his memoir on "ElotJieriuni" ( 87, pp. 320-321 ). In Dinohyus hollandi there are unquestionably four- teen dorsal vertebra), which were found in a continuous series. The character of the anterior sacral vertebra in comparison with the lumbars present in the type of JJino- hyus hollandi seems to indicate that there must be two vertebrae missing. Two verte- bra) have therefore been inserted (1. 3 and 1. 6) which appears justifiable. Further- more, it is quite evident from Professor Scott's publication (87), as well as from my own observations on the same material, that there are six lumbar vertebra) in Arcliico- therium ingens. It would then seem that Marsh (1. c.) was correct as regards the number of presacral vertebra) of his restoration, but made a mistake in the division of the dorso-lumbar series, while Scott's restoration (fig. 9) shows only thirteen dor- sal vertebra). In examining the Princeton specimen it occurred to me that the spines in the posterior dorsal region are too abruptly shortened and that there was probably an additional dorsal vertebra. In comparing the articulated skeleton of Dinohyus with that of Archn'oiherium ingens (see fig. 9) it is at once seen that the former has a relatively larger head, shorter and heavier body, higher withers, and a shoi'ter pelvis. Furthermore, the sagittal crest is higher posteriorly, which indicates a greater vertical diameter of the neck immediately back of the skull. Thus, it is seen that the Miocene genus had a relatively heavier neck than its Oligocene predecessor. 144 MEMOIRS OF THE CARNEGIE MUSEUM Measurements of the Restored Skeleton. Cm. Length from premaxillary of skull to the posterior end of ischium 288 Length of cervical region 51 " " dorsal " 91 '' " lumbar " 38 " " sacral " approximately 12 " " sternum 52 Height at 2d dorsal vertebra 177 " at superior border of ilium 140 " " end of ischium 113 A model in full relief of Dinohyus hoUamJi, a photograph of which is given on Plate LXII, has been made by Mr. Theodore A. Mills under the direction of the writer. An examination shows that this figure is at variance with the painting prepared by Mr. Charles R. Knight under the direction of Professor Henry Fair- field Osborn (72a, p. 713). Arch'xotherium is represented in that figure as having long stifle bristles on a very broad neck, with the ears as in the Suhrx, and as having heavy pendant wattles hanging from the bony tubercles of the inferior margins of the mandibles. Although the bristles and the broad neck shown in Knight's illus- tration may be correct, there is reason to believe that the ears were placed lower down and were more drooping, judging from the position of the external auditory meatus, which in the true pigs is directed upward and slightly outward, while in those American genera of the Entelodontida^, in which this part is known, the auditory meatus is directed nearly horizontally outward and has a prominent over- hanging border on the upper side formed by the squamosal. In the restoration by Mr. Mills the dependent processes on the inferior border of the mandibular rami of Dinohyus are represented as supporting muscles, it being believed by the writer that these processes, at least in the genus Dinohyus, existed for the attachment of muscles in order to give required strength to this portion of the head. The model, of which the figure is given, is of course to a certain extent conjectural, but is believed by the writer to very fairly represent the animal as it was in life, and he takes occasion to compliment Mr. IMills upon the skill which he has shown in carrying out the suggestions made to him. Conclusion. While the osteology of at least two genera of this interesting family is now quite completely before us, I make no attempt to present a succession of species, as i-t seems to me that we have not yet the necessary material in order to satisfactorily study the more detailed questions of the phylogeny of this group. Aymard and Pomel with the extremely limited material representing this family at their com- PETERSON : A KEVISION OF THE ENTELODONTID^ 145 maiid, were constrained to place it among the living swine, where it has since, with only one or two exceptions (Siibursi of de Blainville and Arctodon of Leidy), gener- ally been placed by different writers. That the family is a " collateral branch " of the Suidfe, which " branched off in early Eocene time " as Marsh says (64, p. 406), is from all the evidence at hand unquestionable. " Schlosser has referred the genus \_Elotherium'] to the bunodont division of the family Anthracoiheriidx, which family he derives from an Eocene stock common to the Anthracothcriidse, the Anaplotheriidit\ the Hippopotamidse, and the Suidas" (87, p. 322). Scott says {I. c.) : "The genus lElotheriuni] is so far specialized that it implies a long ancestry, not a member of which is, as yet, certainly known, although thei'e are certain Eocene genera, wdiich throw some light upon the problem." Stehlin in his discussion of the dental struc- ture of the Entelodoniidx (90, pp. 121-123) concludes in a footnote (p. 123) by say- ing: "jedenfalls liegt die Vorgeschichte der Elotherien noch sehr im Dunkeln." It is apparent that, although we are well acquainted with the general anatomical fea- tures of this family, we can for the present at best only speculate in a general way as to the early history of the group. From the work of Schlosser it is plain that he regards the early Tertiary of the Old World as the home of the ancestral line of the Entelodontidx, which may possibly prove to be true, unless " we may suppose." as Matthew has stated in connection with his study of the genus Ancodon,"' " tliat from a diffusion center in Northern Asia early stages in the evolution of [this] phylum [also] reached Europe," and that the American form reached North America at the beginning of the Oligocene. As for the known forms in the early American Tertiary there is not one which can properly be regarded as truly ancestral to the family. Achttnodon of the Bridger, and A. uintense of the Uinta are already too far advanced in the modification of their dentition, having Ijut three premolars, while the limbs of A. umtense would seem to have retained a more primitive condition, having " four [? functional] digits in the pes" (72, p. 105). The suggestive resemblances of the known characters of Achxnodon found in the Bridger formation, and A. uintense of the Uinta, should, I think, be regarded only as pointing to a remote relationship ; indeed it would seem that no great violence is done in referring these general resemblances to parallelism, which was no doubt a greater factor in moulding characters than is sometimes real- ized. The pi'imitive bunodont teeth, the characteristic glenoid cavity of the skull, the expanded zygomatic arches, and the elongated muzzle of these earlier Tertiary forms are not to be dismis.sed so easil}', as they are certainly most suggestive of the Entelodoutidte. But, as has been suggested by others (87, p. 322) there is no more «'Bull. Am. Muesnm of Natural History, Vol. XXVI, p. 4, 1909. 146 MEMOIRS OF THE CARNEGIE MUSEUM reason why we should not expect a parallelism here as well as, for instance, in the odontoid process of the axis, the selenodont molars, the reduction of the lateral digits, and many other features, which have been independently acquired by sev^eral dis- tinct lines. In regard to the ancestry of the Entdodontidx Dr. Herluf Winge (92a, p. 1 34) seems to have reached a somewhat similar conclusion to that of Schlosser. From a study of the dental, cranial, and other skeletal characters of the Enielodontidx, Winge (/. c, p. 141) seems also to think that they perhaps were less in the habit of rooting in the earth than their ancestors and that, instead of the usual plant-food of the Ungulates, which they perhaps entirely excluded, they accustomed themselves to mixed, or flesh-food. In fact Winge {I. c, p. 141) states that the incisors, canines, and premolars served as grasping-tools and weapons similar to those of carnivores ; and (/. c, p. 142) that it is also not altogether unlikely that they lived in a similar manner to that of dogs, preying on live animals, or, when they discovered carcasses, fruits, etc., that they fed on them. That these animals actually seized live animals as their food, as Dr. Winge seems inclined to believe, may well be doubted, while that they fed on carrion and fruits, as well as on plants, is not altogether unlikely. T^ydekker has very suggestively said (53a, p. 7(S) that the " food of the higher selenodont pigs consisted in great part of leaves and grass (which require finer tri- turation . . .) while their bunodont allies feed, as we know, more generally on roots and tubers, and occasionally on animal matter. Hence it is probable that the muzzles of most of the selenodonts were less elongated than in the true pigs, which require to turn up the soil to obtain nutriment." It has been, I think, quite con- clusively shown, by Scott (87, p. 278), and also in the description of the dentition of J)in(jh!)us in the present paper, that the wear of the lower canines and incisors could not have been caused, in this animal, except by the habit of digging up roots. The dental structure of Tetraconodon, from the Siwalik hills of India, precludes its introduction into the family Erdelodontidx and it has quite correctly been placed in a distinct family by Lydekker (53a, p. 78). This genus is apparently quite far removed from the Entelodontidx as is Achvnodun of the American Eocene. The phyletic and geologic position of the family Entelodontidas may then be expressed in a general way as follows : PETERSON : A REVISION OF THE ENTELODONTID^E 147 Middle Miocene Lower Miocene Oligocene Eocene Amraodon ! Dinohyus Pelonax \ \ \ \ .\ Da-odon Boochoerus / / / / / / / Archseotherium \ \ / \ / Entelodon / / / DiAOEAM Expressing the Phyletic and Geologic Position of Entelodonluise. As Poiael (74, p. 1085), Leidy (50, p. 174), Marsh (64, p. 408), Kowalevsky (38, p. 450), 8cott (87, p. 322), Schlosser (83, p. 80), and others have shown us, this family has in the pigs and Hippopotamics its nearest''^ representatives of the recent forms, while the direct ancestors are yet to be found. In the meanwhile we have seen that during the relatively short geological time in which we are able to trace the family in Europe and America there are certain anatomical differences, indicating lines of divergence. These lines probably point to habits due to the varied environ- ments during the life-history of the group. We have, for example, in the European genus Entelodon, a form with enlarged premolars indicating a diverging step from Arclmother'mm. It has further been pointed out that the Ijase of the skull of IMelodon is very greatly different from that of the American forms ; and also that the trigon of the lower molars is entirely absent, while in Archxotherium it is quite plain in young, but fully adult specimens. Pelonax ravbosuvi of the upper Oligocene has tremendously heavy chin-processes, a very prominent angle, and single-rooted first and second lower premolars. T)a;o(Jon of the John Day formation has no chin-processes at all and has a light angle of the lower jaw ; while in tlie Miocene of Nebraska we have in Dinohi/us a form with very small chin-processes, and a gentle sweep of the downward projection of the angle approaching what is seen in Dnodoit. The median upper incisors of D'mo- hyus are distinctly reduced and are in fact sometimes wanting, having been perhaps shed quite early and the alveoli clo.sed uj). In confining our,selves to these characters of the mandible alone and leaving out Entelodon whose generic position can hardly Ije dcnibted, it would seem that there is a variation of importance, when we consider the fact that all the specimens "' The relationship to tliese recent forms is a very remote one. 148 MEMOIRS OF THE CARNEGIE MUSEUM found with these parts present in the Oh'gocene formation have the chin pi'ocesses present. This cliaracter tlien cannot be regarded as only of sexual importance so far as the Oligocene forms, which are most abundantly found, are concerned. It then remains to be seen whether certain forms like ArcJucotherium cktvum with the relatively small and posteriorly placed chin-processes led up to such forms as ])seo- don or if we will yet find in the lower Oligocene a form minus these protuberances. The discovery of such a form would indicate that the various lines even to quite small details of differences were already well established in older Tertiary time. While the variation in the canine teeth, the dependent processes of the lower jaw, and the jugal of the skull in the different .species of the lower Oligocene, may seem comparatively small, they are nevertheless of some importance when it is shown that certain ■ other parts, as for instance the vertebral column, are quite varied. In Archicothcrium ingcns from the lower Oligocene, the neural arches of the dorsal vertebroe are imperfectly pierced by canals, while in Archwotherlum crassuin from the same horizon, the horizontal canals are apparently present as in Bos taunts. In Dinah ijus from the Miocene we know that the neural arches of the dorsals are not only perforated by horizontal canals, but by vertical canals as well ; a character which is identical with what is found in Sus. While the geographical distribution of this family in Asia is at present only a matter of speculation, it is clear that Europe and especially North America were occupied by it. Thus we have seen that while their remains are comparatively abundant on the flanks of the Rocky Mountains, evidence of their existence is not altogether wanting in California and New Jerse}' (" E. siqyerhiiyi," Ammodon leidij- anum). From the lower Oligocene upward, and before the close of the Miocene, then, they occupied certain areas of the North American continent from the Pacific to the Atlantic coasts. Judging from their anatomy, their undoubtedly omnivorous dental structure, their elongated limbs, and their distribution, it is higlily probable that these animals were capable of comljating adverse conditions when occasion required change. While we know that the representatives of this family continued in America to a later geological age than was recently believed and that there were distinct genera and species during the known existence of the group from the lowei* Oligocene to the middle Miocene, there is yet much to be ascertained and we are far from having solved the whole history of this interesting branch of the suborder Artiodadi/Ia. When a more extensive survey of Asia, especially its northern part, is made, and its Tertiary fauna becomes better known, we probably will be supplied with much valuable information on obscure points in connection with the study of the vertebrate paleontology of the Tertiary formations of the continents of the northern hemisphere. PETERSON : A REVISION OF THE ENTELODONTID/E 149 BIBLIOGRAPHY. 1. Aymard, a. Ann. Soc. d'Agric, Sci., et Commerce du Puy, tome XII, for 1842- 1846, pp. 227-242, 1 plate. Reprint, pp. 1-43, 1 plate, 1848 (^Enielodon). 2. Ami H. M. « On Some Extinct Vertebrata from the Miocene Rocks of the North- west Territories of Canada, recently described by Professor Cope." Science, Vol. XVIII, p. 53, 1891. 3. Bush, Lucy P. Professor Marsh's Monograph of the Dwocerata. Am. Jour. Sci. (3), Vol. XXIX, p. 193, fig. 27, 1885 {Elotherium [clavum'] crasmm). 4 Notes on the Dates of Publication of Certain Genera of Fossil Verte- brates. Am. Jour. Sci. (IV), Vol. XVI, pp. 97-98, 1903 (Entelodon). 5. Cope, E. D. Geology of New Jersey, Appendix C, p. 740, 1868 {Elotherium). Q Report on the Vertebrate Paleontology of Colorado. U. S. Geol. & Geog. Surv. Terr, for 1873, Vol. I, Chapter IV, pp. 504-505, 1874 {Pelonax ramosus). 7. Bull. U. S. Geol. and Geog. Survey Terr., 1, Bull. No. 1, p. 27, 1874 {Elotherium ramosum). 8. Ann. Kept. U. S. Geol. & Geog. Surv. Terr, for 1873, p. 463, 1874. 9. Proc. Amer. Philos. Soc, Vol. XVIII, p. 77, 1879. Also in Pal. Bull. no. 30, p. 15, Dec. 3, 1878 {Daodon shoshonenKtu). 9a. American Naturalist, Vol. XII, p. 833, 1878. 10. Observations on the Faunae of the Miocene Tertiaries of Oregon. Bull. U. S. Geol. & Geog. Surv. Terr., Vol. V, pp. 59-66, 1879 {Boochcvrus hvmeros^is). 11 Extinct Mammalia of Oregon. American Naturalist, Vol. XIII, p. 132, 1879. 12. On Some Characters of the Miocene Fauna of Oregon. Proc. Am.. Philos. Soc, Vol. XVIII, p. 77, 1879 {D(codon). 13 The Systematic Arrangement of the Order PerLssodactyla. Proc. Am. Philos. Soc, Vol. XIX, p. 397, 1881 {Dicodon). 14 On a New Deposit of White River Age in Dakota. Proc. Am. Philos. Soc, Vol. XXI, p. 217, 1883. A letter from Professor Cope to the Secretary of the Am. Philos. Society {Elotherium ramo- sum). 15 On the Structure of the Feet in the Extinct Artiodactyla of North America. Proc. Am. Philos. Soc, Vol. XXII, p. 22, 1884. 16. The Vertebrata of the Tertiary Formations of the West. Book I, U. S. Geol. Surv. of the Terr., p. 713, 1884 {Dcvodon). 150 MEMOIRS OF THE CARNEGIE MUSEUM 17. 17a. 18. 19. 20. 21. 22. 23. Cope, E. D. Ibid. Proc. Am. Assoc. Adv. Sci., 33d Meeting, Phila., p. 483, 1885. Notice of Fossil Remains from the Tertiary of Canada. American Naturalist, Vol. XIX, p. 163, 1885 {Entelodon mortoni). The Vertebrata of the Swift Current Creek Region of the Cypress Hills. Ann. Rept. of the Geol. & Nat. Hist. Survey of Canada, Appendix I, p. 84c, ] 885 {Elotherhmi mortoni). The Perissodactyla. American Naturalist, Vol. XXI, p. 1063, 1887 (DfcocZow). — • The Artiodactyla. American Naturalist, Vol. XXII, pp. 1088-1090, 1888. The Mechanical Causes of the Development of the hard parts of the Mammalia. Jour. Morphology, Vol. Ill, p. 137-290, 1889 (Elo- therium Boocharus). P. 168, Illustrations of the distal ends of radius, humerus, and front and radial views of mauus of Boocliwrus humerosus. Vertebrata of Swift Current River, No. III. American Naturalist, Vol. XXIII, p. 629, 1889 {Elotherium coardatum). On Vertebrata from the Tertiary and Cretaceous Rocks of the North- west Territory. I. The Species from the Oligocene or Lower Mio- cene Beds of the Cypress Bills. Geol. Survey of Canada. Contri- butions to Canadian Palteontology, Vol. Ill, p. 20, PI. XIV, fig. 3, 1891 {^Elother'mm arclatum). Dawkins, W. B. Fossil Mammals in North America. Nature, Vol. II, pp. 119-120; 232-233, 1870. Dana, J. D. Review of J. D. Whitney's " Auriferous Gravels of the Sierra Nevada and California." Am. Jour. Sci. (3), Vol. XVIII, p. 233, 1879. DfipfiRET, Charles. The Evolution of Tertiary Mammals, and the Importance of their Migration. Comptes Rendus des Stances de Acad6mie des Sciences ; t. CXLI, p. 702 ; t. CXLII, p. 618 ; t. CXLIII, p. 1 120. Trans- lated by Johanna Kroeber. In the American Naturalist, Vol. XLII, Nos. 494, pp. 109-114 ; 495, pp. 166-171, and 497, pp. 303-307, February, March, and May, 1908 {Entelodontlda). 26. FiLHOL, H. Ftudes des Mammifc-res Fossiles de Ronzon (Haute-Loire). Annales Sci. Geol., tome XII, art. 3, pp. 190-240, 1882 (Elotherium). 27. Flower. H. Hunterian Lectures. Artiodactyla. Nature, Vol. VII, pp. 428-430, 1873. 28. Flower, W. H., & Lydekker, R. An Introduction to the Study of Mammals Living and Extinct, p. 292,1891. 29. Gaudry, a. Note on the Solid-hoofed Pigs in the Society's Collection. Proc. Zoiil. Soc. London, p. 33, 1877 (Entelodon). 24. 25. 25a. PETERSON : A REVISION OF THE ENTELODONTID.15 151 30. Gaudry, a. 31. Gervais, p. 31a. 32. Greene, F. V. 33. Grinnell, G. B., 34. Hayhen, F. V. 35. HOERNES, p. 36. Hay, O. P. 37. KOWALEVSKY, ^V 38. 43. 44. 39. King, C. 40. Leidy, J. 41. 42. Les encliainements du Monde Animal dans les Temps Geologiques, Mammif^res Tertiaires, Paris, 8°, p. 92, 1878 {Entchdon). Zoologie et Paleontologie Franjaises. Pi. XXXII, fig. 12, p. 194. 1859. Deiixieme edition, 4°, Paris, 1859 (Entelodon). Zoologie et Paleontologie Gen^rales, II, 1876, pp. 74-75, PI. X, figs. 2, 2a. Chemical Investigation of Remains of Fossil Mammalia. Am. Jour. Sci. (2), Vol. XYI, pp. 16-20, 1853. & Dana, E. S. On a New Tertiary Lake Basin. Am. Jour. Sci. (3), Vol. XI, pp. 226-228, 1876. Explorations under the War Department. Proc. Acad. Nat. Sci. Phila., Vol. X, p. 157, 1858. Zur Kenntniss der Milchbezahnung der Gattung Entelodon Aym. Sitzungsber. Akad. Wissensch. Wein, CI, pp. 17-24, 1 plate, 1892. Bibliography and Catalogue of the Fossil Vertebrata of North America. Bull. No. 179, U. S. Geol. Surv., 1902. Pp. 654-656, Elothcrlum, Entelodon, Archceotherium, " Arctodon," Pe1ona:c, Boochrcms ; p. 656, Ammodon ■ p. 633, Da'odon, Daledon. Monographic der Gattung Anthracotherium Cuv. und Versuch einer Natiirlichen Classification der Fossilen Hufthiere. Palfeontogra- phica, Vol. XXII, p. 258, 1873 (Entelodon). Osteologie des Genus Entelodon Aym. Pakeontographica, Vol. XXII, pp. 41 5-450, with Plates XVI, XVII, XXV-XXVII, 1876. Systematic Geology. U. S. Geol. Exploration, 40th parallel. Vol. I, p. 411, 1878. Observations on Two New Genera of Fossil Mammalia, Eucrotophus Jacksoni, and Archaiotherium mortoni. Proc. Acad. Nat. Sci. Phila., Vol. V, p. 90; 92-93, 1850. Description of the Genus Arctodon. Proc. Acad. Nat. Sci. Phila., Vol. V, p. 278, 1851. Report upon some Fossil Mammalia and Chelonia from Nebraska. Sixth Ann. Rcpt. Regents Smithsonian Institution for 1851, j). 64, 1852. David D. Owen's Rept. of a Geological Survey of Wisconsin, Iowa, and Minnesota, and incidentally a portion of Nei)raska Territory. " Description of the Remains of Extinct Mammalia," etc., by Leidy. 1852. Pp. 558-563, PI. X, figs. 1-3; PI. XI, fig. 1 {Archa-o- therium mortoni) ; p. 572 [Arclueotherium robustum). Ancient Fauna of Nebraska. Smithsonian Contributions to Knowl- edge, Vol. VI, pp. 57-66, Pis. VIII, IX, X, figs. 1-7, 1853. [Archceotherium mortoni). 152 45. Leidy, J. 46. 47. 48. 49. 50. 51. 52. 53. 53a. Lydekker, R. 53b. 54. — 55. 56. Lambe, L. M. 57. Marsh, O. C. MEMOIRS OF THE CARNEGIE MUSEUM Ueiuarks ou a Collection of Fossil Mannaalia from Nebraska. Proc. Acad. Nat. Sci. Phila., Vol. VI, p. 392, 1853 {Entelodon). Synopsis of Extinct Mammalia, the Remains of which have been Dis- covered in the Eocene Formations of Nebraska. Proc. Acad. Nat. Sci. Phila., Vol. VII, p. 157, 1854 {Entelodon). Notice of Some Remains of P^xtinct Vertebrated Animals. Proc. Acad. Nat. Sci. Phila.-, Vol. VIII, p. 164, 1856 {Entelodon). Rectification of the References of Certain of the Extinct Mammalian Genera of Nebraska. Proc. Acad. Nat. Sci. Phila., Vol. IX, p. 175, 1857 {Elothenim). Indications of an Elotherium in California. Proc. Acad. Nat. Sci. Phila., Vol. V, p. 177, 1868 {Elotherium superbum). The Extinct Mammalian Fauna of Dakota and Nebraska. Jour. Acad. Nat. Sci., Phila., 1869, Vol. VII, pp. 175-192; 388, PI. XVI {Elotherium mortoni) ; pp. 192-194; 388, PI. XXVII, figs. 8-11 {Elotherium ingens) ; p. 388 {Elotherium Leidyanum) ; p. 388 {Elotlteriiim superbum). Remarks on a Collection of Fossils from Dallas City, Oregon. Proc. Acad. Nat. Sci. Phila., p. 112, 1870 {Elotherium superbum). Report on the Vertebrate Fossils of the Tertiary Formations of the West. U. S. Geol. Siirv. of Wyoming and Portions of Contiguous Territories. 2d (4th) Annual Rept., p. 353, 1871. The Extinct Vertebrate Fauna of the Western Territories. Report of the U. S. Geol. Surv. of Terr. F. V. Hayden, Washington, 1873. Pp. 124-125; 320, PI. A^II, figs. 28, 29 {Elotherium mortoni); pp. 217 ; 320, PI. II, figs. 3, 4 ; PI. VII, fig. 27 {Elotherium imperator) ; p. 320 {Elotherium ingen>i). Molar Teeth and other Remains of Mammalia. Palseontologica Indica, Ser. I, p. 78, 1876 {Tetrdconodontidcv). Siwalik Selenodont Suinse. Palseontologica Indica, Ser. X, Vol. II, p. 146, 1883 {Entelodontidce). Catalogue of the Fossil Mammalia in the British Museum, Part II, p. 249, 1885 {Entelodon, Elotherium, Oltinotherium, Archceotherium.) A Geographical History of Mammals, pp. 161, 162, fig. 35 {Elotherium crassum). The Progress of Vertebrate Palaeontology in Canada. Trans. Royal Soc. of Canada, Vol. X, pj). 29, 51, 1904 {Elotherium coardatum). Notice of Elotherium leidyanum and Rhinoceros matutinus. Proc. Acad. Nat. Sci. Phila., Vol. XXII, p. 3, 1870. (No description.) PETERSON : A REVISION OF THE ENTELODONTID.E 153 58a 59. 60. 61. 62. 63. Proc. Acad. Nat. Sci. Phila., Am. Jour. Sci. (3), Vol. V, p. Am. .Jour. Sci. (.3), Vol. 58. Marsh, O. C. Description of Lophiodon validu.s. p. 10, 1871 (Elolhermm). Notice of New Tertiary Mammals. An 487, 1873 (^Elotherium crassum). Notice of New Tertiary Mammals. III. VII, p. 534, 1874 [Elolhermm bathrodon) Introduction and Succession of Vertebrate Life in North America. Ad- dress before Am. Assoc. Adv. Sci. 1877. Am. Jour. Sci. (3), Vol. XIV, p. 363, 1877. (Also in Popular Science Monthly, March and April, 1878.) Dinocerata. Monographs U. S. Geol. Surv., Vol. X, 1884, fig. 7-5 (Elotherium [clavum'j crassum). The Gigantic Mammals of the Order Dinocerata. 5th Annual Kept. U. S. Geol. Surv., Washington, D. C, p. 1085 {Elotherium [davum'] crassum). Description of Miocene Mammalia. Am. Jour. Sci. (3), Vol. XLVI. 1893, p. 408, PI. VII {Elotherium crassum); p. 409, PI. IX, fig. 1 {Elotherium davum) ; pp. 409-410, PI. IX, figs. 2-3 {Ammodon leidyanum) ; p. 410, PI. IX, fig. 4 {Ammodo7i bathrodon); p. 410 {Ammodon potens). Restoration of Elotherium. Am. Jour. Sci. (3), Vol. XLVIII, pp. 407-408, PI. IX, 1894 {Elotherium crassum) ; also in Geological Magazine (4), Vol. I, pp. 294-295, PI. X, 1894. Vertebrate Fossils of the Denver Basin. Monographs U. S. Geol. Surv., Vol. XXVII, p. 527, PI. XXX {Entelodon crassus). A Provisional Classification of the Freshwater Tertiary of the West. Bull. Am. Mus. Nat. Hist, Vol. XII, 1899, p. 50 {Protelotherium uintense) ; p. 59 {Elotherium, Pelonax, Ammodon) ; p. 63 {Dcvodon); p. 64 {Boochosrus). Fauual Lists of the Tertiary Mammalia of the West. U. S. Geol. Surv., Bulletin 361, Appendix, pp. 91-120, 1909 (pp. 104, 106, 108, Elotheriidce ; p, 113, Dinohus hollandi). The John Day Fossil Beds. Harper's Monthly Magazine, Vol. CTI, p. 587, 1901, Illustration of Skull, p. 586 {Elotherium calkinsi). Tertiary Faunas of the John Day Region. University of California Publications. Bull. Dept. Geol., Vol. V, pp. 171-205, 1907 (p. 186, Elotheriidw). Bull. No. 53, Part II, U. S. National Museum, 1907, p. 24 {Arduco- therium, Arctodon) ; p. 35 {Elotherium). Nicholson & Lydkkker. Manual of Paheoutology. Vol. II, p. 1323, 1889. 64. 65. 66. Matthew, W. D. 66a. 67. Merriam, J. C. 67a. 68. Merrill, G. P. 69. 154 MEMOIRS OF THE CARNEGIE MUSEUM 70. OsBORN, H. F. 71. 72. 72a. 72b. 72c. 72d. 72e. 72f. 75. 76. 73. POMEL, A. 74. 77. Palmer, T. S. 78. Peterson, O. A. Achsenodon, an Eoceue Bunodout. Contribution E. M. Mus. Geol. and Arch. ; Princeton College, Bull. No. 3, p. 23, 1883 {Entelodon). The Rise of the Mammalia in North America. Address Am. Assoc. Adv. Sci., Madison, Wis., 1893. Studies from the Biological Laboratories of Columbia College, Zoology, Vol. I, p. 41. Fossil Mammals of the Uinta Basin. Bull. Am. il/ws. Nat. Hist., Vol. VII, pp. 102-105, figs. 16, 17, 1895 (Frotelotherium, p. 105). Prehistoric Quadrupeds of the Rockies. The Century Magazine, Vol. LII, 1896, p. 715; illustration, p. 713 {Elotherinm). Casts, Models, Photographs, and Restorations of Fossil Vertebrates. Pamphlet issued by the Dept. of Vert. Pal., Am. Mus. Nat. Hist., p. 16, 1898. Correlation between Tertiary Mammal Horizons of Europe and America. Ann. N. Y. Acad. Sci., Vol. XIII, No. 1, pp. 1-72, 1900 {Entelodon [Elothevium], p. 22). Ten Years Progress in the Mammalian Palseontology of North America. Extrait des Comptes Rendus du 6 Congrfis Intern, de Zoologie. Session de Berne, 1904. P. 93, Diagrammatic table. .{Elothei-iidcv.) Tertiary Mammal Horizons of North America. Bull. Am. Museum of Natural History, Vol. XXIII, pp. 237-253, 1907. Cenozoic Mammal Horizons of Western North America. U. S. Geol. Sur. Bull., 361, pp. 1-90, 1909. Sur un Nouveau Genre de Pachydermes Fossiles {Elotherium) Voisin des Hippopotames. Arch. Sci. Phys. et Nat., Geneve, Vol. V, pp. 306-308, 1847 (Elotherium). Note sur un Nouveau Genre de Pachydermes du Basin de la Gironde {Elotherhun magnum). Bulletin de la Societe Geologique de France, Vol. IV, pp. 1083-1085, 1847. Note sur le Genre Hyopotamus Owen et sur les Anthiterium en Gen- eral. Arch. Sci. Phys. et Nat. Gen6ve, Vol. VII, p. 325, 1848. Catalogue Methodique et Descriptif des Vert6br6s Fossiles D^couverts dans le Basin Hydrographique Superieur de la Loire, et surtout dans la Vallee de son Affluent Principal, I'Allier, p. 88, 1853. Index Generum Mammalium, U. S. Dept. Agri. Biological Survey, North American Fauna No. 23, 1904. P. 115 (Archaotherium); p. 117 {Arctodo7i) ; p. 139 {Boochccrus) ; p. 214 (Dceodon) ; p. 257 (Elotherium); p. 260 (Entelodon); p. 473 (Oltinotherium) ; p. 520 (Pelonax). Preliminary Note on a Gigantic Mammal from the Loup Fork Beds of Nebraska. Science, N. S., Vol. XXII, pp. 211-212, 1905 (Dinochcerus). PETERSON : A REVISION OF THE ENTELODONTIDiT; J 55 79. Peterson, O. A 80. 81. 82. Rogers, O. 83. schlosser, m. 84. 85. Scott, W. B. 86. 87. 88. Scott, W. B., & 89. Sinclair, W. J. 90. Stehlin, H. G. 91. Steinmann, G., 92. Trouessart, E. 92a. Winge, H. A Correction of the Generic Name Dinochrerus. L. c, Vol. XXII, p. 710, 1905 (Dinohyns). Tiie Agate Spring Fossil Quarry. Annals Carnegie Museum, Vol. Ill, p. 491, 1906. The Miocene Beds of western Nebraska and Eastern Wyoming and their Vertebrate Faunte. Ann. Car. Mus., Vol. IV, pp. 49-51, Pis. XVI-XVII, 1 907 (Dinohyus). Verzeichniss der bisher Bekannten Fossilen Siiugethiere. Bericht Na- turwiss. Vereins. f. Svvaben und Neuburg (a. V.), XXXII, 1896. P. 204 (Elotheriuin) ; [Boochwrus) ; p. 194 Tkvodon syn. of Titan- otherium. Beitriige zur Kenntniss der Stammesgeschichte der Hufthiere. Morph. Jahrb., Bd. XII, p. 80, 1887 (Entdodon). Beitriige zur Kenntniss der Siiugethierreste aus den Siiddeutschen Bohn- erzen. Geol. und Paleont. Abhandlungen, N. F., Band V, p. 55, 1902 {Entdodon). [Communication to Princeton Science Club Relative to the Hind Foot of the American Entdodon.'] Science, Vol. Ill, p. 266, 1884. On the Osteology of Elothermm. Comte Rendu des seances du trois- i6me Congrfes International de Zoologie, Leyden, 1896, pp. 317- 319. The Osteology of Elothermm. Trans. Am. Philos. Soc, Vol. XIX, pp. 273-324, 1898. PI. XVII (restoration of Elotherium inycns) ; PI. XVIII, figs. 3-12, E. Ingens; figs. 1-2 {Elothermm mortoni). OsBORN, H. F. Preliminary Account of the Fossil Mammals from the White River Formation Contained in the Museum of Comparative Zoology. Bull. Mus. Comp. Zool., Vol. XIII, p. 156, 1887 {En- telodon). New and Imperfectly Known Rodents and Ungulates from the John Day Series. Bull. Dept. Geology, Univ. California, Vol. IV, pp. 132-134, PI. XV, 1905 {Elotherium calkinsi). Uber die Geschichte des Suiden-Gebisses. Erster Theil. Abhandl. Schweizer. Palaontolog. Gesellsch., Vol. XXVI, pp. 121, 122, 123, 1899 (Elotherium). & Dr)DERLEiN, L. Elemente der Paliiontologie. P. 793, 1890 {Entdodon). L. Catalogus Mammalium, 1898, pp. 808-809 {Elotherium); p. 810 {Bodch(crm) ; p. 740 {Lwodon syn. of Titanotherium). Jordfundne og nulevande Hovdyr (Ungulata) fra Lagoa Santa, Minas Geraes, Brazilien. Med Udsigt over Hovdyrenes indbyrdes Slaegt- 156 MEMOIRS OF THE CARNEGIE MUSEUM 94. 95. skab. Aftryk af E Museo Lundii, Kjobenhavu, Danmark. Pp. 70, 134, 142-143, 1906 (p. 70, Entelodontini). 93. Weber, M. Die Siiugetiere. Eiufiibriing in Die Anatomie und Systematik der Kecenteu und Fossilen Mammalia. 1904, p. 653 (Elotheriidse) ; figs. 469, 470 (Elotherium ingens). Whitney, G. D. The Animal Remains, not Human, of the Auriferous Gravel Series. Mem. Mus. Comp. Zool., Vol. VI, p. 244, 1879. WiLCKENS, M. Ubersicht iiber die Forschungen auf dem Gebiete der Palaontologie der Haustiere. V. Die Schweinartigea Tiere (Suiden). Biolog. Centralbl., Vol. V, pp. 210, 216, 217, 1885. Outlines of Vertebrate Palffioutology for Students of Zoology. Uni- versity Press, 1898, p. 343 {Elotherium) ; p. 344 (illustration of Elotherium crassuin). On the Divisions of the White River, or Lower Miocene, of Dakota. Bull. Am. Mus. Nat. Hist., Vol. V, p. 97, 1893. The Extinct Caraelidse of North America, and some Associated Forms. Bull. Am. Mus. Nat. Hist., Vol. X, 1898, p. 120; footnote (^Elotherium). 99. Zittel, K. a. Handbuch der Palaeontologie, 1893, IV Band, p. 304 (Daledon syn.of TUanotherium) ; p. 335 (Elolherium) ; p. 337 {Bobchcerus). 96. Woodward, A. S. 97. WORTMAN, J. L. 98. 158 MEMOIRS OF THE CARNEGIE MUSEUM EXPLANATION OF PLATE LIV. Agate Spring Fossil Quarries. Plan of Quarry No. 1. Bl. = " Block " and associated with numbers indicates relative size of blocks taken out of the quarry. Sections 2 A, to 21, indicate relative position of each block and specimen as they were taken out. Plate LIV. Sec.6 Sec. 5 Sec. 4 Sec.3 Sec. 2 Seel Sec, I A MEMOIRS Carnegie Museum, Vol. IV, Sec2| Sec. 20 S^' 9 S^a Sec 7 Sec 6 Sac 5 Sec,4 Sec 3 SecE Sec I Sec lA Sec.SA AGATE SPRING FOSSIL QUARRIES ai.o H-i OS .093 .VI jov ,MU33uM aiaanHAD ?.momU L-U^- ^t ! 4-^^-4- 4- i-pH- f'n u I OTiryiq^g axAOi 2H YHHAUO