Chasmaporthetes Kanti, New Species From

AMERICAN MUSEUM

Novitates PUBLISHED BY THE AMERICAN MUSEUM OF NATURAL HISTORY CENTRAL PARK WEST AT 79TH STREET NEW YORK, N.Y. 10024 U.S.A. NUMBER 2632 SEPTEMBER 30, 1977

HENRY GALIANO AND DAVID FRAILEY Chasmaporthetes kanti, New Species from China, With Remarks on Phylogenetic Relationships of Genera within the Hyaenidae (Mammalia, Carnivora) , 'i- AMERICAN MUSEUM Novitates PUBLISHED BY THE AMERICAN MUSEUM OF NATURAL HISTORY CENTRAL PARK WEST AT 79TH STREET, NEW YORK, N.Y. 10024 Number 2632, pp. 1-16, figs. 1-3, tables 1-3 September 30, 1977

Chasmaporthetes kani, New Species from China, With Remarks on Phylogenetic Relationships of Genera within the Hyaenidae (Mammalia, Carnivora)

HENRY GALIANO1 AND DAVID FRAILEY2

ABSTRACT A new species of Chasmaporthetes is described Within a cladistic framework, Chasmaporthetes is from deposits of Early Pleistocene age in Shansi the sister group of Euryboas; Chasmaporthetes, Province, People's Republic of China. The con- Euryboas, and Lycyaena form the sister group of cept of Chasmaporthetes is extended to include Hyaenictitherium, Hyaena, Pachycrocuta, and "Lycyaena" ("Euryboas") lunensis Del Campana, Crocuta. 1914 but not Euryboas bielawskyi Schaub, 1941.

INTRODUCTION The first description of a hyaenid in the New 1931 and 1939. At the conclusion of the Central World was that of Chasmaporthetes ossifragus by Asiatic Expeditions of the American Museum of Hay (1921) from the Anita Fauna (Irvingtonian), Natural History, Liu Hsi Ku and Kan Chuen-pao Coconino County, Arizona. Stirton and Christian ("Buckshot") were employed by Frick to con- (1940) described a second genus and species tinue field collection in Late Tertiary and Qua- Ailureana johnstoni, from North Cita Canyon ternary deposits of China. The efforts of these (late Blancan), 'Randall County, Texas, which men produced an extensive collection of Late they later (1941) transferred to Chasmaporthetes. Cenozoic fossils from numerous localities, three Chasmaporthetes ossifragus has been recorded in of which yielded specimens of an Asiatic repre- two Florida faunas: Santa Fe River iB, Blancan sentative of the genus Chasmaporthetes. This (Webb, 1974) and Inglis 1A, Irvingtonian (Klein, sample of Chasmaporthetes from China and the 1971; Webb, 1974). Undescribed or poorly pre- additions to the North American hypodigm have served material that possibly are hyaenids have served to clarify the phyletic position of Chas- been mentioned from early Blancan localities maporthetes within the Hyaenidae. in Kansas (Hibbard, 1950); Idaho (Bjork, 1970); The age of the Chinese deposits was deter- Arizona and Mexico (Repenning, 1962). mined by the fauna associated with Chasmapor- The material discussed in the present paper thetes kani. Complete faunal lists of the Frick was assembled by the late Childs Frick between China localities and the justification (or modifi-

'Curatorial Assistant, Department of Vertebrate Paleontology, the American Museum of Natural History, New York, 10024. 2Student, Department of Systematics and Zoology and Museum of Natural History, University of Kansas, Law- rence, 66045.

Copyright © The American Museum of Natural History 1977 ISSN 0003-0082 / Price $1.30 2 AMERICAN MUSEUM NOVITATES NO. 2632 cation) of this age assignment are in preparation SYSTEMATICS by Tedford, Manning, and Galiano. ORDER CARNIVORA BOWDICH, 1821 The authorship of this paper is about equally divided. The senior author was responsible for SUBORDER AELUROIDEA FLOWER, 1869 the general outline and the discussions of generic CHASMAPORTHETES HAY, 1921 affinities and ecological relationships. The junior author organized the paper and was responsible Type species. Chasmaporthetes ossifragus for the description and discussion of Chasmapor- Hay, 1921. thetes. Included species. Chasmaporthetes johnstoni All measurements are in millimeters. Paren- (Stirton and Christian, 1940). C. lunensis (Del theses indicate an approximate measurement. Campana, 1914); C. borissiaki (Khomenko, 1932); Abbreviations used are: C. kani, new species; ? C. nitidula (Ewer, 1955). AMNH, the American Museum of Natural History Revised Generic Diagnosis. Pl larger than in F:AM, Frick Collection, the American Museum other hyaenid genera (absent in Euryboas). of Natural History Anterior accessory cusps of P2 and P3 more QSV, Museum de Lyon, France prominent than in other hyaenids; upper and TRO, Timberlane Research Organization, Lake lower tooth rows curved with premolars slightly Wales, Florida imbricate, as opposed to a straight tooth row UF, University of Florida, Gainesville, Florida as in Euryboas. USNM, National Museum of Natural History, Smithsonian Institution. Chasmaporthetes kani, new species ACKNOWLEDGMENTS Holotype. F:AM 99789, palate, with nearly complete dentition missing only the left P' (lost We are deeply indebted to Dr. Richard H. in life). Tedford of the American Museum of Natural Type Locality. Hsia-chuang, approximately History, and Dr. Germaine Petter of the Museum 20 km. south of Shou Yang, Shou-yang District, National d' Histoire Naturelle, Institut de Paldon- Shansi Province, People's Republic of China. tologie, Paris, France, for their interest and sug- Age. Nihowanian (Early Pleistocene). gestions during the preparation of the present Etymology. Named for Kan Chuen-pao, mem- paper. Grateful acknowledgment is extended to ber of the Central Asiatic Expeditions and pro- Dr. Bjorn Kurtdn, Museum of Zoology, Helsinki, fessional fossil collector for the Frick Laboratory. Finland, for comments on our manuscript and Diagnosis. Ml slightly longer than P4; anterior allowing us to read his generic review in manu- accessory cusps of P2 and P3 relatively weak as script. We also thank Drs. Malcolm McKenna and compared with other Chasmaporthetes species. Eugene Gaffney, and Messrs. Earl Manning and Hypodigm. Hsia-chuang: F:AM 99784, palate George Engelmann, the American Museum of with right 11-35 Cl, Pl-M', left I-3, C', PI; F:AM Natural History, and Dr. Bruce MacFadden, Yale 99785, left ramus with Cl, P2-Ml (P4 broken); University, for their comments and criticisms on F:AM 99786, left ramus with Cl, P2-M,; F:AM various parts of this paper. An additional note of 99787, right ramus with 13, Cl, P2-M,; F:AM gratitude is due Dr. S. David Webb, University of 99788, left ramus with C, (broken), P2-Ml. Ma Florida, Florida State Museum, and Professor Tzu Kou: F:AM 99781, right ramus with 11-3, Chow Min Chen, Institute of Vertebrate Paleon- Cl, P2-M,; F:AM 99790, left ramus fragment tology and Paleoanthropology, Peking, People's with MI. Niu Wa Kou: F:AM 99783, partial Republic of China, who reviewed the completed skull with right 11-3, Cl, PI-Ml, left I-3, Ml; manuscript. Numerous revisions of this manu- F:AM 99782, immature left ramus with I,, script were typed by Mss. Iris Jeromnimon and DI2-3, DCI, DP2-4. Ma Tzu Kou and Niu Wa Kou Annlinn Kruger. Ms. Jacqueline Tung and Mr. are in Yu-she District, Shansi Province. Ernst Heying helped with the translations. Fig- A check was made to determine if any two ures 1 and 2 were drawn by Mr. Ray Gooris in specimens of the hypodigm belong to the same his usual fine manner. individual. This did not prove to be the case. 1 977 GALIANO AND FRAILEY: CHASMAPORTHETES KANI 3

Material Used for Comparison. Chasmapor- thumbprint-sized depression is present in the thetes ossifragus. AMNH 99098 (cast of holo- zygomatic arch just ventral to the orbit at the type, USNM V-10023), edentulous ramus, from place the dorsal portion of the buccinator muscle the Val Verde Mine near Anita, Coconino Coun- takes its origin. This depression is present in ty, Arizona; AMNH 95208, left ramus with P2-M I Euryboas but not in available specimens of Cro- (cast of UF 18088), from Inglis 1A, Citrus cuta crocuta. County, Florida; AMNH 100078 (cast of TRO The anterior margin of the braincase rises at SF-1) left detached p2, Santa Fe River, Flor- approximately a 75 degree angle from the ida (courtesy of John Waldrop); F:AM 23390, glenoid fossa leaving a large, deep temporal fossa fragmentary left ramus with P2-4, from Ben- behind the orbit. This condition also exists in all son area, Cochise County, Arizona. Chasma- advanced hyaenid genera with some slight varia- porthetes johnstoni: Panhandle Plains Museum tion. no. 2343, holotype left ramus with 13, Cl, P2-M1, The alignment of the foramen rotundum, an- text and figures in Stirton and Christian (1940), terior lacerate foramen, and the optic foramen from Randall County, Texas. Chasmaporthetes parallels the anterior margin of the braincase, lunensis: partial skull with upper dentition, text placing these foramina in a vertically compact and figures in Del Campana, 1914, from Olivola group in which neither the optic foramen (ante- (Val di Magra), Italy. Euryboas bielawskyi: riorly) nor the foramen rotundum (posteriorly) AMNH 101261 (cast of unnumbered holotype extends anterior to the margin of the anterior housed in Falculte des Sciences, Clermont- lacerate foramen. Ferrand, France), mandible with complete den- Ventral to the orbit, the area of origin for the tition from Rocca Neyra, Montagne de Perrier, anterior part of the medial pterygoid muscle France; AMNH 99148 (cast of Mus. Basel, V.A. slopes smoothly to the posterior margin of the 1822), maxillary fragment with P3-4 from Val palate with no evident sculpture for the attach- d'Arno, Italy; Euryboas sp.: AMNH 26995 (cast ment of this muscle. of QSV 53), anterior portion of a skull with The posterior and anterior widths of the pal- nearly complete dentition missing right 11-2, and ates vary, due perhaps to sexual dimorphism. left 11-3, C1, from St. Vallier, France. The width of the palate between the canines may be less than the length of P4 (in F:AM 99784) or DESCRIPTION may be greater than the length of P4 (F:AM Skull. Neither the skull roof nor the portion 99789, see measurements in table 1). In a similar of the skull posterior to the glenoid fossa is pre- fashion, the posterior width of F:AM 99789 is served in the Chinese sample. The remaining part greater than the length of the upper tooth row of the skull is comparable in size to the corre- (C'-P4), whereas in F:AM 99784 it is less. F:AM sponding elements in Crocuta crocuta. 99783, which is more like F:AM 99784 in ro- The postorbital processes are small and blunt bustness of canines and premolars, is inter- as in Palhyaena and Lycyaena. In Hyaena and mediate in both cases. The palatal proportions Crocuta these processes are more prominent. The described above, in Euryboas (QSV 53) are like area behind the postorbital process is not so con- those of F:AM 99789 (see fig. 1). stricted as in Crocuta, but slightly swollen, as in Dentition. The upper tooth row is slightly Lycyaena. The forehead is domed and joins the curved between p2 and P4, not to the degree in muzzle immediately posterior to P3. The muzzle Crocuta, but more than that of Euryboas. C' and is high and flattened dorsally, giving a much Pl lie parallel to the long axis of the skull pro- sturdier appearance to the snout than that of ducing a change in the curvature of the tooth Crocuta crocuta (see fig. 2). row at this point. p2, and to a lesser extent P3, The anterior opening of the infraorbital fora- are slightly rotated medially in F:AM 99783 and men lies above the parastyle of P4. This condition 99789; but in F:AM 99784, P3 shows the greater is seen in many hyaenids but altered in Hyaena, rotation. This rotation of premolars, with imbri- Pachycrocuta, and Crocuta in which the opening cation, increases the illusion of curvature in the of the infraorbital foramen lies above P3. A tooth row. 00 Ca

m C1 (D r- llo W)C 00 0 1- cl~~~~~c 04 tt 0 00 (, 00 _; (N I 0 * N - -4 o000O II 110I x x x x x x- x 4 X x Ch 00 oq cl; N rN 0 10 0 00 cl (ON -- -4 c N (N

00 0--Cl 00 1-Ccli I X)XX 0o t ^ N8 00 t t N 0o Cl u0 F- Cl cl C

cr-ON .0 ^ o- ^ 00 -_ t 4Cl Cl-~ N cr- 00 No N (- x 00 00 CIN x xo Cl - Cl 10 u)~~~~~~~~~~~~~~~~~~~~~0 cli el; Cl a)a) X a) 00 4:U i oo o o l~~~~~~~~~~~x x x x -o r- a) en O -

00~~ o 00- 'ta)o 4-4 00 N * r- rs C (ON o o oN0 N _N Cq, -4 -_ 0C' Clr Cl4 ~, . . . x x C; c' 00 ,- 000 o Cl o rN 1'0 400 o

,It C;

CSO 00 ._ N 0_00 xxxxxo _0 _ _ 1 X 00 X CX 1- CIA N r-a C_ oq C: 11 oq 00 a) 00 . . l a)- 0 cn a..l ~07e 3 a) a) 0~~~n n 00 0e

ZC. Cl C's ;,, 1-40~ 0 atgE) t- 0 c "to_;3~o 0 o 00 .to e- a o0 0 4- 0 0 0 o'a ~~ ~ U: oo ~~~ ~ ~ ~ ~ a 0'a)'a 4- +Z 4 = X ) $C '-4 a) I.5.4 oa t aX X X X X a)~~~ o ~ ^ , a) 0~~~~~~~~ a.) C) saSa) x a)~~~~~~~~~~~~~~~i ~ ~ ~ ~ ~ Lt _ 00tooo -4 0n E Cl _ 04

4 'e

00 o

ON c) Lo ON

00 CN to cd 0- m

5 6 AMERICAN MUSEUM NOVITATES NO. 2632

FIG. 2. Chasmaporthetes kani, new species F:AM 99783, lateral view, X 1/2. The incisors are proportioned as those ofPal- one or two labial cusps separated by a basin hyaena or Lycyaena. I1 and 12 are subequal in which occludes with the single-cuspid talonid size; 12 is the larger. I1-3 have faint anterior cingu- of M1. la. It and 12 have large posterior cingula; that of The lower jaws are reminiscent of the type of 13 is barely noticeable. E. bielawskyi, in shape and size. The ramus has a C' is of normal hyaenid shape with two small deep, upright symphyseal region that extends be- ridges, one anteromedial and one posterior, run- low the natural curve of the ramus producing a ning the lengtlh of the crown. doubly curved ventral margin, i.e., the margin is All the cheek teeth have faint labial and lin- convex beneath the canine as well as beneath the gual cingula, as in all members of the hyaenidae. carnassial. A single large mental foramen is Pl, absent in Euryboas, is proportionally placed beneath the anterior root of P2. The mas- larger than in any other hyaenid genus. The im- seteric fossa is moderately deep, the indistinct brication of p2 and P3, as described above, may be anterior border of the masseteric fossa is beneath the result of accommodating this unusually large the talonid of M1. The angular process, the area tooth. Pl is about half the length of p2, essen- of insertion for the medial pterygoid muscle, is tially unicuspid, has a convex labial surface, and rounded and does not extend posteriorly to the a concave lingual surface. Pl of F:AM 99784 has condyle. This corresponds to the poor demarca- very distinct anterior and posterior accessory tion on the skull of the origin of the anterior cusps. portion of the medial pterygoid muscle as dis- p2 and P3 have more prominent anterior acces- cussed previously. The mandibular foramen is sory cusps than those in Euryboas, although that placed about midway on a line between the of p2 may not be clearly evident. The postero- talonid of M1 and the angle of the ramus. Like lingual margins of p2 and P3 are broadly ex- the mental foramen, the mandibular foramen is panded, creating a lingual shelf on these teeth. large. P4 has a protocone that is round in occlusal The lower incisors are proportioned as those outline, which does not reach to the anterior of Lycyaena and Euryboas. I1 and 12 are small margin of the tooth. In contrast to Euryboas, the and subequal in size, 12 is larger. 13 has a small protocone does extend to the anterior margin of posteroexternal accessory cusp. The lower canine the tooth. The tip of the paracone is at the ap- has weak anterior and posterior ridges running proximate center of the tooth. The parastyle is the length of the crown. larger than in Euryboas, comprising slightly over P1 is absent, as in all Chasmaporthetes and one-quarter of the length of P4. Euryboas species examined except for the curi- ml, as in Euryboas, has one lingual cusp and ous exception of C. borissiaki, where this tooth is 1 977 GALIANO AND FRAILEY: CHASMAPORTHETES KANI 7 present only on the right side of the mandible. American species, known only from rami, al- The presence of PI is variable in Lycyaena. though later authors (Repenning, 1967; Savage The lower cheek teeth become progressively and Curtis, 1970; Kurten, 1971) have suggested larger from P2 to Ml. Each has a faint cingulum synonymy with the European genus Euryboas around the base of the crown. On P3 and P4 this Schaub, 1941. cingulum may be enlarged lingually to the poste- Schaub (1941) described a mandible from rior accessory cusp, as in other species of Perrier-Rocca Neyra, France, as the type specimen Chasmaporthetes and Euryboas. As in other of Euryboas bielawskyi and referred a partial Chasmaporthetes species, the premolars are tri- maxilla with P34 from Val d'Arno, Italy, and cuspid, the cusps are placed on the long axis of several limb elements from these and other local- each tooth. The size of the anterior accessory ities to this species. Schaub believed that Eury- cusp on P2-4 is also variable. It may be barely boas bielawskyi and Chasmaporthetes johnstoni, discernible on P2 and small but evident on P3, or based on lower dentitions, were closely related small on P2 and clearly evident on P3. In either but did not discuss C. ossifragus. Schaub also case, the anterior and posterior accessory cusps compared the partial maxilla of E. bielawskyi are of equivalent size on P4. with the type specimen, a skull, of "Lycyaena" The variation in prominence of the anterior lunensis Del Campana, 1914, from Olivola (Val di accessory cusps on P2 and P3 iS possibly related Magra) Italy. Schaub (1941) noted the large para- to sexual dimorphism as those rami in which the style and posteromedial position of the proto- anterior accessory cusps of P2 and P3 are most cone on P4, the long, thin premolars, and the prominent also have the more robust premolars larger anterior accessory cusps on the upper pre- and canines. Generally, P2 to P4 are oval in out- molars of "Lycyaena" lunensis. Other than the line although in several individuals, again prob- shape of the premolars, which we use as a species ably due to sexual dimorphism, the lower pre- character, the above features are also seen in the molar widens posteriorly. Unfortunately, as no upper dentitions of Chasmaporthetes kani and other population samples of Chasmaporthetes are are used in our revised generic diagnosis of known, comparison of these features is impos- Chasmaporthetes. sible. Viret (1954), with no large sample to aid his The length of the largest M1 (F:AM 99788) is interpretation of variation, recognized the ge- greater than that of the type ofE. bielawskyi and neric identity of Euryboas. He believed that the that of C johnstoni. The lower carnassial is al- differences between "Lycyaena" lunensis and most evenly proportioned in occlusal outline, ex- Euryboas bielawskyi, as discussed by Schaub cept for a slight medial constriction between the (1941), and present in new material he described, protoconid and paraconid. The protoconid is could be nothing more than variation among higher than the paraconid. The metaconid is ab- individuals of different sizes and ages. Viret sent, and the talonid is unicuspid. (1954) therefore synonymized E. bielawskyi with "Lycyaena" lunensis and referred a muzzle (QSV 53) from St. Vallier, a partial maxilla (Uni- DISCUSSION versity of Clermont no. 3285) from Sen6ze, and The genus Chasmaporthetes, type species C. several isolated teeth possibly of the same indi- ossifragus, was proposed by Hay, 1921, for a vidual (University of Clermont, unnumbered) fragmentary, edentulous (roots only present) from Perrier-Etouaires, to Euryboas lunensis. ramus from the Val Verde Mine, near Anita, The uniformity of characters in the sample of Coconino County, Arizona. Stirton and Christian Chasmaporthetes from China and their similarity (1940) described a second ramus, with a com- with "Lycyaena" lunensis Del Campana, in con- plete dentition, from North Cita Canyon, Ran- trast to Euryboas bielawskyi Schaub, lead us to dall County, Texas, as the type specimen of believe that Schaub (1941) was correct in sepa- Ailuraena johnstoni which they later (1941) rating "Lycyaena" lunensis and Euryboas bielaw- transferred to Chasmaporthetes. The genus has skyi. Furthermore, the similarity between the until now been restricted to these two North lower dentitions of Chasmaporthetes kani and 8 AMERICAN MUSEUM NOVITATES NO. 2632 those of C. johnstoni and C ossifragus, and be- The size relationship of P4 to Ml quickly sepa- tween the upper dentitions and skulls of C kani rates these three species of Chasmaporthetes and and "Lycyaena" lunensis indicate that these four is probably a derived condition in both C species represent a discrete taxonomic unit which johnstoni and C ossifragus. M1 is only slightly is separable from the original concept of Eury- larger than P4 in C kani (the presumed primitive boas as proposed by Schaub (1941). condition also seen in Lycyaena and Euryboas) Of the specimens referred to "Euryboas" but much larger (nearly 25%, Klein, 1971) in C lunensis by Viret (1954), we would place one, ossifragus and slightly smaller than P4 in C. the partial maxilla from Seneze, in Chasma- johnstoni (Stirton and Christian, 1940). porthetes. This partial maxilla has an alveolus The features of the type specimen of Chasma- for P', a large anterior accessory cusp on P3, a porthetes lunensis, a skull and upper dentition, large parastyle, and a posteromedially directed which are held in common with C. kani and in protocone of P4, and a curved tooth row as is contrast to the maxillary fragment referred to seen in C kani and C lunensis. The size of P1, as Euryboas bielawskyi by Schaub (1941) and the indicated by the alveolus, and the length of P3, muzzle and teeth referred to "E." lunensis, sensu relative to P4, are less than is seen in either C lato, by Viret (1954), are the presence of a large kani or C lunensis and this specimen may not be P1, the larger parastyle and the position of the referable to any described species of Chasma- protocone on P4, the slightly curved tooth row porthetes. with the imbricate premolars, and the high, dor- The features of the lower dentition of sally flattened muzzle. The shape of the muzzle Chasmaporthetes kani, which are also seen in C is one of the more unusual features of C lunensis ossifragus and C johnstoni, in contrast to the in that the premaxillae rise above the level of the type mandible of Euryboas bielawskyi, are the maxillae creating a saddle, or flat area, between larger anterior accessory cusps on P2-P4, the the snout and the domed forehead. This may be more curved tooth row, and the slight imbrica- due to crushing but is duplicated in the most tion of the premolars. complete skull of C. kani, F:AM 99783, and in The lower premolars of Chasmaporthetes kani complete contrast to the muzzle of Euryboas, as and C johnstoni differ from those referred to C. seen in QSV 53 from St. Vallier, in which the ossifragus in having lower, less pointed, proto- muzzle slopes smoothly upward from the pre- conids. The anterior accessory cusps on PI in maxillae to the highly domed forehead. This may both C kani and C johnstoni are less well devel- be a generic character of Chasmaporthetes but as oped than in P1 of C. ossifragus. These are prob- it is questionable and can be examined in only ably derived (apomorphic) characters of C. ossi- three specimens, two of Chasmaporthetes and fragus. In C kani and C ossifragus, the lower one of Euryboas, we have not included it in our premolars widen posteriorly; in P3 and P4 this revised generic diagnosis of Chasmaporthetes. widening is most evident in the appearance of a The upper dentition of Chasmaporthetes kani large posterior cingulum. In the lower premolars and C. lunensis differ in the size and shape of P2 of C johnstoni, the widest part of each tooth is and P3. In C. lunensis these teeth do not widen near the central cusp. This more "cat-like" ap- posteriorly and in C kani, i.e., they lack large pearance is possibly a derived feature of C internal cingula, and they are relatively longer johnstoni. (the length of P2 is about two-thirds that of P4 and P3 is about five-sevenths that of P4, whereas 'Other specimens discussed by Viret (1954), a com- in C kani these proportions are one-half to plete muzzle (QSV 53) from St. Vallier and several iso- three-fifths for p2/p4 and one-half for P3/P4). The lated teeth including P3 and P4, from Perrier, are external cingula on p24 of Clunensis may be referable to Euryboas. The muzzle lacks P', has straight heavier than on C kani but this is often tooth rows, the anterior accessory cusps on P2 and P3 individ- are barely noticeable, and (as also seen in a P4 from ually variable among carnivores. Judging from Perrier) the parastyle of P4 iS smaller than in Chasma- the alveolus, P' of C lunensis was about the same porthetes and the protocone is directed anteromedially size as that of C kani. and extends as far forward as the anterior margin of As Chasmaporthetes ossifragus and C john- the tooth. stoni are known from rami and lower dentitions 1 977 GALIANO AND FRAILEY: CHASMAPORTHETES KANI 9 only and C. lunensis only from a skull, it is pos- cusps of Lycyaena. p2 and P3 of "Lycyaena" niti- sible that C. lunensis is synonymous with one of dula, in fact, show a greater development of the the North American species. The large P3, relative anterior accessory cusps than is seen in known to P4, of C. lunensis would lead one to expect a species of Euryboas or Chasmaporthetes. As correspondingly large P4, relative to Ml, as is seen Chasmaporthetes differs from Euryboas in having in C. johnstoni. The lower premolars of C. john- a greater development of the anterior accessory stoni are also thin as are the upper premolars of cusps, that genus may be the more correct refer- C. lunensis. Only the discovery of additional ence. The large size of the parastyle and the material of either C. lunensis or C. johnstoni can placement of the protocone on P4 support this verify this possible synonymy. referral although the recovery of additional The type specimen of "Hyaena" borissiaki material, which would illustrate the condition of Khomenko, 1932, is an excellent partial skeleton Pl and the degree of curvature of the tooth rows, from the Moldavian Republic, U.S.S.R. It agrees would make this referral more certain. with our generic diagnosis of Chasmaporthetes in having a large P' (although smaller than in C. REMARKS ON GENERIC RELATIONSHIPS kani or C. lunensis and possible plesiomorphic), Simpson (1945) placed Chasmaporthetes in large accessory cusps on the premolars and in the the Felidae, incertae sedis; a conclusion with position of the protocone on P4. However, the which we disagree but which serves to introduce tooth row is as in an lower straight Euryboas, the taxonomic problems surrounding Chasma- apparent symplesiomorphy, and the snout is porthetes. Various genera have been postulated as evenly sloping (also as in Euryboas) an apparent near relatives of Chasmaporthetes: Ictitherium convergence. Although we have not examined (Hay, 1921), Euryboas (Kurten, 197 1), and Per- the material, this species appears to be referable crocuta (Hendey, 1975), although many authors to Chasmaporthetes sensu stricto, in which it would be differentiated from the other species (Savage and Curtis, 1970; Repenning, 1967) have tended to recognize the similarities between Eury- the small size of the near of by M', equal lengths boas and Chasmaporthetes. Hendey (1975) cited P4 and M1, and the variable presence of P1 gaps in geographic distribution between these two (present on the right ramus of the type but not genera and believed that the similarities between the left). Euryboas Schaub, 1941, and Chasmaporthetes The occurrence of hyaenas in the New World Hay, 1921, might be due only to convergence as a of from has been regarded result immigration and not to close relationship. The discovery of over the Asia Bering Isthmus (Hendey, 1975) Chasmaporthetes in China contradicts Hendey's sometime during the Late Tertiary, although as (1975) hypothetical migration into North far as hyaenids are concerned direct evidence for America of a species of Percrocuta Kretzoi, 1938, this hypothesis was lacking. The Frick material instead ofEuryboas. Chasmaporthetes was appar- a gap in our of this fills crucial understanding ently generically distinct before its entry into event, as it establishes a Holarctic distribution for Chasmaporthetes during the Early Pleistocene. North America and shares a number of derived characters with Euryboas rather than with An interesting note on the distribution of Percrocuta.' hyaenas, and a possible African record of Chasmaporthetes, is "Lycyaena" nitidula Ewer, 'The taxonomic position ofPercrocuta is uncertain al- 1955 ("L." silberbergi nitidula in Ewer, 1967), though it is probably less closely related to advanced from the Swartkrans and Sterkfontein caves, hyaenas than has been proposed (Hendey, 1975). Per- South Africa. This species is certainly more cor- crocuta appears to be an early parallel of the more ad- referable to either De Beau- vanced hyaenids. Percrocuta retains a number of primi- rectly Euryboas (as tive characters, such as a large metacarpal I (Hendey, mont, 1967, proposed) or Chasmaporthetes than 1974; metacarpal I is vestigial in all advanced hyaenids); to Lycyaena in that it has a deep mandibular yet is highly derived, even exceeding other hyaenids, in symphysis, a large, single mental foramen on the certain features such as reduction of the protocone of ramus, and large anterior accessory cusps on P4; and has uniquely derived characters, such as a pre- P2-4 in contrast to the more doglike chin, two maxillary-frontal contact in the skull, which are not mental foramina, and much smaller accessory seen in other hyaenids. 10 AMERICAN MUSEUM NOVITATES NO. 2632

The taxonomic position of Chasmaporthetes The group which includes Chasmaporthetes as as well as other generic relationships within the well as the modern hyaenids (branching point 3) Hyaenidae can perhaps be most easily discussed is advanced with respect to its sister-group, con- and visualized in a cladistic framework (sensu taining Palhyaena, in having: Hennig, 1966; and Brundin, 1966). In a cladistic (A) deeper jaws analysis, reduced to its most salient features, (B) reduced P1 and M2 groups are considered to be naturally related (C) the palate broadened anteriorly (monophyletic) if they can be shown to share (D) Pl crowded between Cl and p2 characters which are presumed to have newly (E) an enlarged anterior mental foramen or ab- arisen (are shared derived = synapomorphic) in sent posterior mental foramen their most recent, common, although hypothetical, ancestor from primitive character states. Chasmaporthetes, Lycyaena, and Euryboas, as These monophyletic groups, supported by one or recognized by De Beaumont (1967) in another more synapomorphies, constitute testable hy- phraseology, comprise a monophyletic group potheses of a kind not generally provided by a whose synapomorphic characters include (branch- phylogenetic "tree." Falsification can be pro- ing point 4): vided by contradiction of homology or polarity (A) small or absent M2 and PI of the characters utilized. As a principle of in- (B) small or absent metaconid on M1 quiry ". . . agreement in characters must be inter- (C) reduction of the labial portion of the poste- preted as synapomorphy as long as there are no rior cingulum on P4 grounds for suspecting origin to be symplesio- (D) trenchant, catlike, premolars morphy (shared primitive) or convergence" (E) dental series in straight or only slightly (Hennig, 1965). The outcome is a display of curved rows relationships in which maximum parsimony is Chasmaporthetes is most closely related to achieved, in that the greatest numbers of synapo- Euryboas (together forming a natural group at morphic characters requiring the fewest parallel branching point 5) in sharing: and convergent characters have determined the groupings. We chose to use a cladistic representa- (A) M2 absent tion as it is clear, concise, and follows from a (B) protocone of P4 reduced (C) I3 slightly enlarged stated set of rules. (D) P1 absent The hyaenids, exclusive of the problematical (E) Ml enlarged Percrocuta group, can be characterized as having (F) metaconid of M1 absent (see figure 3, point 1): (G) talonid of M1 unicuspid (A) a large conical parastyle on P4 (H) M2 reduced (B) the protocone of P4 placed directly lingual to (I) a single, large mental foramen beneath P2 the parastyle (J) reduced angular process on the ramus (C) a broad posterior cingulum on P4 differs from in (D) a short basicranium and a relatively elongate Chasmaporthetes Euryboas face having (point 6): (E) a reduction of metacarpal I and metatarsal I (A) a high, long muzzle (plesiomorphic) as op- (F) long, slender, doglike limbs posed to a shorter muzzle in Euryboas (apo- Furthermore, the advanced genera of hyaenas, morphic) (B) Pi enlarged (apomorphic) versus Pl lost in including the living species, form a phyletic Euryboas (also apomorphic) group in that (branching point 2) they share: (C) larger anterior accessory cusps on P24 (apo- (A) a high, triangular occipital crest morphic) (B) long canines (D) larger parastyle on P4 (C) wide and robust premolars (E) the protocone of P4 iS lingually directed and (D) the absence of an entepicondylar foramen in does not extend to the anterior limit of the the humerus tooth (apomorphic) as opposed to an antero- (E) frontal sinuses expanded posteriorly lingually directed protocone in Euryboas, 1977 GALIANO AND FRAILEY: CHASMAPOR THETES KANI I1I

N( (a,

FIG. 3. A cladogram expressing the relationships of hyaenid genera.

which reaches as far forward as the anterior (A) enlarged I3 margin of P4 (plesiomorphic) (B) reduced Pl (F) tooth rows curved with the premolars slightly (C) massive canines imbricate, i.e., the posterior margin of a pre- (D) enlarged or swollen anterior cusp on P3 molar sits labially to the anterior end of the following premolar (apomorphic). Hyaena Brisson, 1762, Pachycrocuta Kretzoi, 1938, and Crocuta Kaup, 1828 (branching point The apomorphic features of the modern 8), in relation to Hyaenictitherium, share: (A) hyaenids and Hvaenictitherium' the sister group the absence of P1 and M2, (B) reduced frontal (branching point 7) to'Lycyaena, Euryboas, and sinuses, and, possibly, (C) shortened hind limbs. Chasmaporthetes are: Lack of associated skeletal material for all genera 'In the use of Hyaenictitherium Kretzoi 1938, type prevents certain recognition of the last character. species H. ("Ictitherium") hyenoides Zdansky 1924, we One of the more interesting observations in differ in opinion from Ficcarelli and Torre, 1970, who this expression of relationships is that the brown would retain this species in Ictitherium. hyaena, "Hyaena" brunnea, is more closely re- 12 AMERICAN MUSEUM NOVITATES NO. 2632

TABLE 2 Some Morphological Comparisons of Hyaenid Genera Palhyaena Lycyaena Euryboas Chasmaporthetes

Palate Long and narrow Long and slightly broad Short and broad Long and broad

I 3 Small Small Slightly enlarged Slightly enlarged, large in North American forms Pi Small Small Absent Very large p2 Long and slender Long and slender Long and slender Long and slender

P3 Long and slender Long and slender, faint Slightly reduced in Long and slender, anterior cusp length, faint anterior prominent anterior cusp cusp Ml Small Very small Smaller than Lycyaena About the same size as in Euryboas

M2 Small Very small when present Absent Absent

Mandible Long and slender Long and slender, mental Long and deep, a single Same as in Euryboas foramen beneath P2 large mental foramen large beneath P2

Pi Small Very small, sometimes Absent Absent (except in C. absent borissiaki) P2 Long and slender Long and slender Short and slightly robust Long and slender with small anterior cusp P3 Long and slender Long and slender Long and slender, slightly Long and slender, more robust than in anterior cusp present Lycyaena P4 Strong lingual Lingual cingulum re- Same as in Lycyaena Lingual cingulum re- cingulum, duced, anterior and duced, anterior and faint anterior posterior cusps posterior cusps cusp prominent prominent

Ml Metaconid re- Metaconid very small, Metaconid absent, tal- Metaconid absent, tal- duced, tal- talonid very small onid unicuspid and onid unicuspid and onid small and bicuspid trenchant trenchant and bicuspid M2 Small Absent Absent Absent lated to Pachycrocuta and Crocuta than to (B) the anterior cusp of P3 formed into a ridge Hyaena hyaena. Although superficially very simi- (C) talonid of M1 small and metaconid of Ml lar in appearance, and sharing a number of plesio- reduced or absent morphic features such as having a metaconid and Furthermore, "Hyaena" brunnea and Pachy- talonid on the and brown bicuspid Ml, striped crocuta (branching point 10) each have a larger hyaenas apparently do not form a natural group. P2, deeper jaws, and less reduced M1 than does The apomorphic features that "Hyaena" brunnea Crocuta. with and Crocuta shares Pachycrocuta (branching A comparison of some morphological charac- point 9) are: ters ofChasmaporthetes and other hyaenid genera (A) large conical P3's is given in table 2. 1 977 GALIANO AND FRAILEY: CHASMAPORTHETES KANI 13

TABLE 2 - (Continued)

Hyaenictitherium Hyaena hyaena "Hyaena" brunnea Pachycrocuta Crocuta

Long and slightly Long and slightly Short and broad Very short and Very short and broad broad broad broad Slightly reduced Large and robust Large and robust Large and very Large and robust robust Slightly reduced Small and rounded Small and rounded Small and rounded Small and rounded Small and slender Small and slender, Small and robust Small and robust Very small and strong anterior robust cusp Robust with enlarged Robust with very en- Large and conical Large and conical Very large and anterior cusp larged anterior conical cusp Slightly reduced Small Smaller than in Very small Very small when when compared H. hyaena present with Palhyaena Very small Absent Absent Absent Absent Variable, but never Long and deeper than Short and deep, a Same as in "H." Short and shallow, very deep, mental in Hyaenictitheri- single large brunnea a single large foramen beneath um, a single large mental fora- mental foramen P2 large foramen beneath men beneath beneath P2 P2 P2 Very small, some- Absent Absent Absent Absent times absent Small and slender Small, strong anterior Short and robust Short and robust Very short and cusp robust Short and slightly Robust with promi- Large and conical Large and conical Very large and robust nent anterior conical cusp Small lingual cingu- Lingual cingulum cusp Large and very Large and very Small and slender lum cusp some- present, weak pos- robust robust times present, terior cusp and anterior posterior strong anterior cusps prominent cusp Same as in Palhyaena, Small, talonid reduced Large, metaconid Large, metaconid Very large, meta- but larger and bicuspid and talonid rarely present, conid absent, small, talonid talonid some- talonid uni- bicuspid times unicuspid cuspid Very small Absent Absent Absent Absent

ECOLOGICAL ROLES typically nonhyaenid features of Chasma- The ecological equivalency of the borophagine porthetes can also be seen in the borophagine dogs in the New World with the Old World Aelurodon taxoides group of McGrew (1944). hyaenids is a well-known relationship (see for ex- The premolars of both genera form nearly ample Romer, 1966, and Kurten, 1971). The straight rows, have strong anterior and posterior comparison of roles is, however, usually confined cusps, and are long and thin. P1 of each is excep- to the more characteristic members of each tionally large within each subfamily. The meta- group, Crocuta and Borophagus, and overlooks conid of Ml is reduced or absent; the talonid is an almost parallel evolutionary history of the unicuspid or nearly so in each genus. Both genera two groups. For example, the characteristic and have well-developed jaws with heavy chins. 14 AMERICAN MUSEUM NOVITATES NO. 2632

Whereas Chasmaporthetes and Aelurodon tax- related to Euryboas. Chasmaporthetes, Euryboas, oides lengthen and strengthen the premolars and and Lycyaena form a natural group in sister- premolar area of their jaws, Crocuta and Boro- group relationship to Hyaenictitherium, Hyaena, phagus have reduced the anterior premolars and Pachycrocuta, and Crocuta. developed a large, conical piercing tooth. Among the Pliocene and Pleistocene hyaenids Chasmaporthetes and Aelurodon taxoides may there were apparently two major adaptive trends. have been occupying the ecological role now One, the Chasmaporthetes group, had more elon- filled by wolves and hunting-dogs (Lycaon). gate premolars, straighter tooth rows, a strong Other ecological equivalencies between genera chin, and no development of bone-splitting pre- of borophagine dogs and Tomarctus and hy- molars. The second, in which all modern hyaenas aenids are given in table 3. and Hyaenictitherium are included, was a bone- crushing habitus involving the acquisition of SUMMARY AND CONCLUSIONS heavy premolars and a short, wide skull concomi- Chasmaporthetes kani, new species, from de- tant with a curved tooth row, including a large posits of Nihowanian age (Early Pleistocene) in conical piercing tooth, P4, in Pachycrocuta and Shansi Province, People's Republic of China, is Crocuta. Theoretically, these conditions could the first member of this genus found in Asia and have developed independently a number of times the first species for which both upper and lower within hyaenid genera. However, in identifying dentitions are known. "Lycyaena" lunensis (Del these adaptations and hypothesizing that they are Campana, 1914), known from the holotype skull the result of major adaptive shifts in only two only, from the Villafranchian Val d'Arno of lines of hyaenas, we are assuming that the pres- Italy, is transferred to Chasmaporthetes as its af- ence of these characters in a hyaenid genus re- finities with the North American late Blancan sults from direct inheritance and demonstrates its and Irvingtonian species, C. ossifragus Hay, 1921, inclusion within a monophyletic group. These and C. johnstoni (Stirton and Christian, 1940), character suites are derived characters at branch- can now be recognized. Euryboas bielawskyi ing points 4 and 7 in our cladogram (fig. 3) and Schaub, 1941, is the Villafranchian type species become primitive characters at lower taxonomic of Euryboas. levels. Other derived characters, not specifically Chasmaporthetes is apparently most closely related to either of these trends, such as the loss

TABLE 3 Morphological and Ecological Equivalencies between Genera of the Hyaenidae and Borophaginae

Hyaenidae Borophaginae Morphological Features Role

Crocuta Borophagus- Short-faced, large conical "Spotted hyaena" Osteoborus piercing tooth, reduced bone-crushing carni- anterior premolars, hori- vore zontal ramus shallow Pachycrocuta Aelurodon haydeni Large, short-faced, massive "Large spotted hyaena" brevirostris dentition, horizontal large bone-crushing ramus deep carnivore Lycyaena-Euryboas- Aelurodon taxoides Long-faced, P1 large, pre- "Large hunting dog" Chasmaporthetes group-Strobodon molars trenchant, hori- strong-jawed (grasp- zontal ramus deep ing) carnivore Palhyaena Tomarctus Long muzzle, slender rami, "Jackal-coyote" unmodified premolars, un- generalized canoid reduced metaconid and form talonid on MI 1 977 GALIANO AND FRAILEY: CHASMAPOR THETES KANI 15

of M2, M2, and P1, are then assumed to be charac- Ewer, R. F. ters occurring in parallel at branching points 5 1955. The fossil carnivores of the Transvaal and 8 rather than indicative of relationship be- caves. The lycyaenas of Sterkfontein tween the Crocuta group (exclusive ofHyaenicti- and Swartkrans, together with some therium) and the Chasmaporthetes group (exclu- general considerations of the Transvaal fossil hyaenids. Proc. Zool. Soc. Lon- sive of Lycyaena). don, vol. 124, part 4, pp. 839-857. The Plio-Pleistocene evolution of Old World 1967. The fossil hyaenids of Africa-a reap- hyaenids paralleled the Mio-Pliocene history of praisal. In Bishop, W. W., and J. D. borophagine canids in the New World. From an Clark, Background to evolution in Af- early, undifferentiated stock of small, unspecial- rica. Chicago, Univ. Chicago Press, pp. ized, doglike carnivores, two major trends devel- 109-123. oped in both groups. One, which included Ficcarelli, G., and D. Torre Chasmaporthetes in the Hyaenidae and Aeluro- 1970. Remarks on the taxonomy of hyaenids. den taxoides in the Borophaginae, expanded the Palaeont. Italica, vol. 66 (new ser., vol. doglike habitus, and probably filled the role of 36), pp. 13-33. 19 pls. Gregory, W. K. the wolf or hunting-dog of today. The second, in 1914. Locomotive adaptations in fishes illus- which are included living hyaenas and Boropha- trating "Habitus" and "Heritage." Ann. gus itself, developed heavy premolars and brachy- New York Acad. Sci. (for 1913), p. cephalic skulls associated with the bone-crushing 267. habitus of modern hyaenas. 1951. Evolution Emerging. New York, Mac- Millan Co., two vols., 1013 pp. Hay, 0. P. LITERATURE CITED 1921. Descriptions of species of Pleistocene Bjork, P. R. Vertebrata, types or specimens of most 1970. The Carnivora of the Hagerman Local of which are preserved in the United Fauna (Late Pliocene) of southwestern States National Museum. Proc. Idaho. Trans. Amer. Phil. Soc., n.s., vol. U.S.N.M., vol. 59, pp. 599-642. 60, part 7, pp. 3-54. Hendey, Q. B. Bock, W. J. 1974. The Late Cenozoic Carnivora of the 1965. The role of adaptive mechanisms in the south-western Cape Province. Ann. S. origins of higher levels of organization. Africa Mus., vol. 63, pp. 1-369. Syst. Zool., vol. 14, no. 4, pp. 272-287. 1975. Relationships of North American hy- Brundin, L. aenas. S. Africa Jour. Sci., vol. 17, p. 1966. Transantarctic relationships and their 187. significance, as evidenced by chirono- Hennig, W. mid midges with a monograph of the 1965. Phylogenetic systematics. Ann. Rev. subfamilies Podonominae and Aphro- Entomol., vol. 10, pp. 97-116. teniinae and the austral Heptagyiae. K. 1966. Phylogenetic systematics. Urbana, Univ. Sven. Vetenskakad. Handl., ser. 4, no. Illinois Press, 263 pp. ll,pp. 1-472. Hibbard, C. W. De Beaumont, G. 1950. Mammals of the Rexroad Fauna from 1967. Observations sur les Herpestinae (Viver- the Upper Pliocene of southwestern ridae, Carnivora) de l'Oligocene su- Kansas. Trans. Kansas Acad. Sci., vol. perieur avec quelques remarques sur des 44, pp. 265-313. Hyaenidae du Neogene. Arch. Sci., Khomenko, P. Geneve, vol. 20, fasc. 1, pp. 79-108. 3 1932. Hyaena borissiaki n. sp. aus der Rusillon- pls. Fauna Bessarabiens. Trav. Inst. Paleo- Del Campana, D. zool. Acad. Sci. URSS, vol. 1, pp. 81- 1914. La Lycyaena lunensis n. sp. dell'ossario 136. pliocenico di Olivola (Val di Magra). Klein, J. G. Palaeont. Italica, vol. 20, pp. 87-104. 1 1971. The ferungulates of the Inglis 1A Local pI. Fauna, Early Pleistocene of Florida. 16 AMERICAN MUSEUM NOVITATES NO. 2632

M.S. thesis. Univ. Florida, Gainesville. radiometric dating. Geol. Soc. Amer. Kretzoi, M. Special Paper, no. 124, pp. 207-231. 1938. Die Raubtiere von Gombaszbg nebst Schaub, S. einer Ubersicht der gesamtfauna. Ann. 1941. Ein neues Hyaenidengenus von der Mus. Nat. Hungary, vol. 31, pp. 88-157. Montagne der Perrier. Eclog. Geol. Hel- Kurten, B. vetiae, vol. 34, pp. 279-286. 1971. The Age of Mammals. New York, Co- Simpson, G. G. lumbia Univ. Press, 250 pp. 1945. The principles of classification and a McGrew, P. O. classification of mammals. Bull. Amer. 1944. The Aeluroden saevus group. Field Mus. Nat. Hist., vol. 85, xvi + 350 pp. Mus. Nat. Hist., Geol. Ser., vol. 8, no. Stirton, R. A., and W. G. Christian 13, pp. 79-84. 1940. A member of the Hyaenidae from the Miller, A. H. Upper Pliocene of Texas. Journ. Mam- 1949. Some ecological and morphological mal., vol. 21, no. 4, pp. 445-448. considerations in the evolution of 1941. A iluraena Stirton and Christian referred higher taxonomic categories. In, Mayr to Chasmaporthetes Hay. Ibid., vol. 28, E., and E. Schiuz (eds.), Ornithologie als no. 4, p. 198. biologische Wissenschaft, p. 84. Viret, M. J. Repenning, C. A. 1954. Le loess a Bancs Durcis de Saint-Vallier 1962. The giant ground squirrel, Paene- (Dr6me), et sa faune de mammiferes marmota. Jour. Paleont., vol. 36, pp. Villafranchiens. Nouvelles Archives du 540-556. Museum d'Histoire Naturelle de Lyon, 1967. Palearctic-Nearctic mammalian dis- fasc. 4, pp. 1-200. 33 pls. persal in the Late Cenozoic. In, Hop- Webb, S. D. kins, D. M. (ed.), The Bering land 1974. Pleistocene Mammals of Florida. bridge, Stanford, Stanford Univ. Press, Gainesville, Univ. Florida Press, 270 pp. pp. 288-3 11. Zdansky, 0. Romer, A. S. 1924. Jungtertiare Carnivoren Chinas. Palae- 1966. Vertebrate Paleontology. Chicago, Univ. ont. Sinica, series C, vol. 2, fasc. 1, pp. Chicago Press, 468 pp. 1-149, 33 pls. Savage, D. E., and G. H. Curtis 1970. The Villafranchian Stage-Age and its

r..,t