sign in sign up
<<
Home , Aeluroidea, Banded Linsang, Hemigalinae, Linsang, Poiana (genus), Spotted linsang

University of Nebraska - Lincoln DigitalCommons@University of Nebraska - Lincoln

Papers in the Earth and Atmospheric Sciences Earth and Atmospheric Sciences, Department of

2001 Basicranial Anatomy of the Living Prionodon and (Mammalia, , ), with Comments on the Early Evolution of Aeluroid Carnivorans Robert M. Hunt Jr. University of Nebraska-Lincoln, [email protected]

Follow this and additional works at: https://digitalcommons.unl.edu/geosciencefacpub Part of the Earth Sciences Commons

Hunt, Robert M. Jr., "Basicranial Anatomy of the Living Linsangs Prionodon and Poiana (Mammalia, Carnivora, Viverridae), with Comments on the Early Evolution of Aeluroid Carnivorans" (2001). Papers in the Earth and Atmospheric Sciences. 549. https://digitalcommons.unl.edu/geosciencefacpub/549

This Article is brought to you for free and open access by the Earth and Atmospheric Sciences, Department of at DigitalCommons@University of Nebraska - Lincoln. It has been accepted for inclusion in Papers in the Earth and Atmospheric Sciences by an authorized administrator of DigitalCommons@University of Nebraska - Lincoln. AMERICAN MUSEUM NoVltates PUBLISHED BY THE AMER I C AN M U SEUM OF NAT U RAL H ISTORY CENTRAL PA RK WEST AT 79T H STREET, NEW YORK. NY 10024 Number 3330, 24 pp., 10 figures, 2 tables April 26, 2001

Basicranial Anatomy of the Living Linsangs Prionodon and Poiana (Mammalia, Carnivora, Viverridae), with Comments on the Early Evolution of Aeluroid Carnivorans

ROBERT M. HUNT, JR.'

CONTENTS

Abstract ...... , .. , .. , .. , ., .. , .. , . ... 2 Introduction ...... • . . . . • ...... • . . • ...... • ...... 2 Abbreviatio ns ...... 3 Cranial and Dental Compari sons ...... • ...... 3 Basicranial Anato my of Paiaeoprioll odoll ...... • . . . . • ...... • . . . . . • ...... 7 Basicranial Anato my of Priollodoll and Poiana ..... • .... • ...... •. . . . . • ...... 11 Discussion and Conclusions ...... • ...... • . . • . . • . . . . • ...... 17 Acknowledg nlents ...... • ...... • ..•.. • ....•...... 23 References ...... • •. . .. • . •• _ . _ • • . • • _ ... • ...... •.....•.... • ...... 23

, Research Associ3u: . Division of Pal eoll!ology. American Museum of N a tur~l His.tory; Professor. Geological Sciences. and Cur~to r. Ven cbr.ue Paleontology. University of Nebraska. Li ncoln. NE 68588-0549.

Copyright C Ameri can MUSoC um of Nmur.l l History 2001 ISSN 0003-0082 I Pri ce $3.00 2 AMERICAN MUSEUM NOVITATES NO. 1

ABSTRACT The living Asian , Prionodon pardicolor, shares marked anatomical similarities in basicranium and with the extinct aeluroid, Palaeoprionodon lamandini, from the Quercy ®ssures, France. The living African linsang, Poiana richardsoni, is similar yet slightly more derived in basicranial traits relative to Prionodon pardicolor, and also has basicranial and dental features indicating a relationship to the living genets (Genetta). The basicranial and auditory anatomy of a series Palaeoprionodon-Prionodon-Poiana can be in- terpreted as a morphocline showing the progressive alteration of the form of the petrosal and auditory bulla from the plesiomorphic aeluroid state in the Quercy fossils to a derived con- dition typical of the linsangs (Prionodon, Poiana) and living genets (Genetta). The basicranial anatomy of Genetta, including the structure of the petrosal and auditory bulla, is typical of other of the Viverridae. The other lineages of living viverrids are believed to have undergone a similar transformation in their basicranial anatomical pattern from the plesiom- orphic state present in Oligocene aeluroids, exempli®ed by Palaeoprionodon, to the modern patterns typical of the living subfamilies (including the endemic Malagasy viverrid genera).

INTRODUCTION siomorphic basicranial pattern common to these aeluroid genera (Hunt, 1998). As a re- African and Asian linsangs of the family sult, I became interested to learn whether liv- Viverridae are living, nocturnal aeluroid car- ing linsangs have retained throughout the nivorans, occupying forested environs in the mid- and later Cenozoic the archaic basicra- Old World tropics (®g. 1). The African lin- nial pattern of these Quercy aeluroids, and in sang (Poiana Gray, 1864) is represented by particular, whether they retain the pattern one or at most two species in west . found in Palaeoprionodon, which has often Poina richardsoni and P. leightoni are dis- been allied with the Asian linsangs of the tributed from Liberia to northern Zaire, and Prionodon (Teilhard de Chardin, 1915, the island of Fernando Po (Bioko), according to Rosevear (1974). They are considered here Gregory and Hellman, 1939). as a single species, P. richardsoni, following The living African and Asian linsangs are Pocock (1908), who recognized leightoni as rare : Prionodon pardicolor and P. a subspecies of richardsoni. The Asian lin- linsang are listed as endangered (by the Con- sangs (Prionodon Hors®eld, 1822) are con- vention on International Trade in Endangered sidered to be more diverse: one species (the Species, and by the U.S. Department of the , P. pardicolor) is found from Interior). Although the eastern population of to Indochina, and a second species Poiana richardsoni is not reported as endan- (the , P. linsang) reported gered, the western population is listed by the from and the Malay Peninsula into IUCN (International Union for Conservation , Java, and (Nowak, 1991). of Nature) as of indeterminate status (No- Linsangs nest above ground and eat insects, wak, 1991); whether these eastern and west- small vertebrates, and some plant material ern groups are in contact through the tropical (Rosevear, 1974). forest belt is uncertainÐRosevear (1974) The Old World linsangs are of particular considered them among the rarest of African interest because of their evident anatomical . Therefore, the opportunity to dis- similarity in cranial and dental features to sect the auditory region and describe the ba- one of the oldest fossil representatives of the sicranial anatomy of both Poiana richard- aeluroid Carnivora, Palaeoprionodon laman- soni and Prionodon pardicolor was fortu- dini, from the Oligocene ®ssure deposits of itous and timely. Are the linsangs ``living Quercy, France (Teilhard de Chardin, 1915; fossils''? This study considers the possibility Gregory and Hellman, 1939). A recent study that linsangs are relicts of the Oligocene ae- of Palaeoprionodon and other closely related luroid fauna, preserving a basicranial mor- primitive aeluroids (, Stenoga- phology from a time when the le, ) from the Oligocene and early were in an initial phase of their great Eur- of western Europe identi®ed a ple- asian radiation. 2001 HUNT: AELUROID EVOLUTION 3

Fig. 1. Geographic distribution of the living Asian linsangs (Prionodon pardicolor, P. linsang) and African linsang (Poiana richardsoni), and the Eurasian localities that have produced fossils of the Oligocene aeluroid Palaeoprionodon. 1, Palaeoprionodon lamandini, Quercy ®ssures, France; 2, Hsanda Gol, Mongolia (?Palaeoprionodon); 3, Poiana richardsoni leightoni (western area), P. r. richardsoni (eastern area); 4, Prionodon pardicolor (northern area), Prionodon linsang (southern area).

ABBREVIATIONS h hypoglossal (condyloid) foramen ic internal carotid artery Anatomical L middle lacerate foramen m mastoid A alisphenoid P petrosal a ``apron'' of petrosal plf posterior lacerate foramen ac alisphenoid canal (posterior opening) pp paroccipital process of the exoccipital BO basioccipital R rostral entotympanic BS basisphenoid rp rugose surface of petrosal promonto- c common opening for the hypoglossal and rium for attachment of the caudal en- posterior lacerate foramina in Poiana totympanic in Nandinia d depression in basisphenoid for internal SQ squamosal carotid loop T ectotympanic E caudal entotympanic tf ¯ange of the ventral process thinned by F facet on petrosal promontorium for ec- appression of the caudal entotympanic totympanic tt tensor tympani fo foramen ovale V ventral process of the petrosal promon- gf postglenoid foramen torium 4 AMERICAN MUSEUM NOVITATES NO. 1 x line of caudal entotympanic attachment (formed by the right and left orbitosphe- to ectotympanic noids) is thin and narrow at this locus and Z contact of ectotympanic ¯ange with pe- probably represents the primitive aeluroid trosal promontorium condition for the emergence of cranial nerves

V1,V2, and the optic nerve from the brain- Institutional case. The optic foramen lies in close prox- AMNH American Museum of Natural History, imity to the orbital ®ssure in these aeluroids, New York (Department of Mammalo- in contrast to many arctoid carnivorans in gy) which the optic foramen is placed farther for- FMNH Field Museum of Natural History, Chi- ward along the orbital wall. cago (Department of Mammals) An alisphenoid canal is present, its poste- MNHN MuseÂum National d'Histoire Naturelle, rior opening placed a few millimeters ante- Paris rior to the foramen ovale. In Prionodon and Poiana the maxillary branch of the trigemi- CRANIAL AND DENTAL nal nerve (V ) exits the braincase through the COMPARISONS 2 small foramen rotundum in the cranial wall Cranial measurements of Prionodon, and emerges within the alisphenoid canal.

Poiana, and two skulls of Quercy Palaeo- Both V2 and the blood vessels traveling in prionodon illustrate their similarity in size the canal exit the skull directly below the or- and proportion (table 1). All four skulls have bital ®ssure at the anterior opening of the ca- a basilar length of ϳ6±7 cm and display a nal, which sometimes is confused with the similar form (®gs. 2, 3): a slender, tapering true foramen rotundum.2 This arrangement is rostrum; a gradually ascending forehead; and also true of Genetta. Palaeoprionodon, al- relatively large, wide orbits in which the an- though somewhat damaged in the orbital re- terior part is ¯oored by the maxilla (®g. 3). gion, appears to have been very similar to The braincase is conspicuously expanded rel- the living linsangs in the con®guration of the ative to the narrow rostrum, and the surface orbital wall and the placement of these fo- of the braincase indistinctly registers the to- ramina (Hunt, 1998: ®g. 4A). pography of the neocortex. The cerebrum has The palate of the linsangs and Palaeo- not grown backward to cover the cerebellum prionodon is triangular in ventral view, with in Prionodon and the Quercy skulls, and this a narrow anterior part that expands to its is re¯ected in a distinct separation between maximum width between the upper carnas- the anterior and posterior parts of the brain- sials (®g. 2B). The upper dentition of these case (a constriction in the skull occurs be- genera is delicate: the are uniform in tween the two regions). In Poiana there is a somewhat greater posterior overgrowth of 2 My identi®cation of these foramina follows the in- the cerebellum by the cerebrum than in terpretation put forward earlier by Pocock (1916a), based upon his detailed skull dissections. In many car- Prionodon. A triangular occiput is common nivoran species, including the living linsangs, V2 passes to all four skulls; in the midline the vermis through the cranial wall via a small foramen that opens of the cerebellum produces a low elevation into the alisphenoid canal (this canal always lies external dorsal to the foramen magnum. A cranial en- to the braincase and has anterior and posterior aper- tures). Pocock regarded the hidden internal opening in docast of Prionodon pardicolor has been de- the cranial wall for V as the true foramen rotundum, scribed and ®gured by Radinsky (1975), 2 and the more visible external opening that transmits V2, comparing it with that of Poiana and other once it has joined with the vessels of the alisphenoid living viverrids. canal, as the anterior aperture of the alisphenoid canal. In the living linsangs and in the plesio- Thus, in some carnivorans with an alisphenoid canal (alar canal of Miller et al., 1964), the foramen rotundum morphic Quercy (Palaeopriondon, Steno- cannot be seen when looking at the orbit, but remains plesictis) and St.-GeÂrand (Stenogale) aelu- entirely hidden by the bony exterior wall of the alisphe- roids, the foramen rotundum, orbital ®ssure, noid canal. However, in some (e.g., otariid and optic foramen are deeply recessed in the ) the alisphenoid canal does not incorporate the true foramen rotundum for V2 but remains a separate sphenoid in a common elliptical depression bony tube, and in this case the anterior opening of the located in the posteroventral orbital wall. In alisphenoid canal and the foramen rotundum actually ap- these carnivorans, the interorbital partition pear as separate openings visible in the posterior orbit. 2001 HUNT: AELUROID CARNIVORE EVOLUTION 5

Fig. 2. Skulls of the Quercy Palaeoprionodon (MNHN Qu 9370), the Asian linsang Prionodon pardicolor (AMNH 163595), and the African linsang Poiana richardsoni (AMNH 51438), from left to right. (A) dorsal view; (B) ventral view. Scale bar in this and all subsequent ®gures is 1 cm. shape and size and are arranged in a trans- tocone and diminutive parastyle. The M1 is verse row (I3 is only slightly larger than the not present in the Palaeoprionodon skulls interior incisors). P1-3 are thin bladelike but, based on placement of the alveoli, prob- teeth; Prionodon still retains a plesiomorphic ably was similar in form to M1 in the lin- double-rooted P1, whereas P1 in Poiana and sangs where this tooth is a small narrow tri- Genetta is single-rooted. P4 is a typical angle with a prominent parastyle. However, shearing with a small cuspate pro- M1 in Prionodon retains a distinct paracone 6 AMERICAN MUSEUM NOVITATES NO. 1

Fig. 3. Skulls of the Asian linsang Prionodon pardicolor (AMNH 163595, above) and African linsang Poiana richardsoni (AMNH 51438, below) in lateral view. Figures 3±10 are stereophotographs. and metacone and is not as narrow as in don: both have a tall trigonid with the para- Poiana, in which the paracone-metacone are conid-protoconid shearing blade set apart subsumed in a thin crest and are no longer from a well-developed, conical, posterolin- distinct cusps. I have not observed M2 in any gually directed metaconid, accompanied by a of the living linsangs, but a vestigial M2 oc- reduced, low, slightly basined talonid. But an curs in some individuals of Palaeopriono- even more striking similarity exists in the don, and M2 has been reported in some spec- form of m2 in these two genera. The m2 of imens of Poiana richardsoni (Ewer, 1973; the Quercy genus is distinguished by a small Rosevear, 1974). In the linsangs, the palatal low trigonid, in which the metaconid is set choanae are closed (¯oored by bone) for 4± well apart (posterolingually displaced) from 5 mm posterior to the M1s, whereas in Pa- the protoconid-paraconid and separated from laeoprionodon this closure is not present. them by a prominent valley (Hunt, 1998: ®g. In the mandible, the (p2-4) of 6h). This same derived form of m2 occurs in the living linsangs are thin bladelike teeth, Prionodon pardicolor, but in the single in- each with anterior and posterior cingulum dividual of Poiana richardsoni that I was cusps and a posterior accessory cusp; in the able to examine, the m2 trigonid lacks the Quercy genus the form of p3-4 is like that of strongly displaced metaconid, and has the the linsangs, but the p2 is usually reduced in three trigonid cusps closely grouped as the size and thus often lacks these additional points of an equilateral triangle. Neither Pa- cusps. Remarkably, Prionodon shows the laeoprionodon nor the species of living lin- same reduction in size of p2 (relative to p3) sangs retain an m3. seen in Palaeoprionodon, but the reduced p2 Among the living viverrids, the taxa most does not occur in Poiana. The p2 of Poiana closely resembling the linsangs in dentition, is tall, not low, and in this trait, and in the body form, and pelage are the genets (Ge- form of the entire row, it is clearly netta genetta and related species, Kingdon, like that of the genets. The p1 is double-root- 1977). Genets appear to be African linsangs ed in Prionodon but is reduced to a single- scaled to larger size, differing only in such rooted tooth in Poiana and Genetta. minor features as retention of M2 and in sub- There is a very close correspondence of tle variation in pelage patterns. m1 form in Prionodon and Palaeopriono- Thus, initial comparison of cranial and 2001 HUNT: AELUROID CARNIVORE EVOLUTION 7

TABLE 1 Comparative Cranial Measurements (in mm) of Prionodon pardicolor, Palaeoprionodon lamandini, and Poiana richardsoni (Carnivora, Viverridae).

TAXON Prionodon Paleoprionodon lamandini Poiana pardicolor richardsoni Measurement AMNH 163595 Qu 9348 Qu 9370 AMNH 51438 Basilar length 60.2 67.4 71.7 61.2 Rostral widtha 13.8 14.8 14.9 12.0 Braincase, greatest width 24.1 26.9 27.8 24.7 Palatal width between P4 paracones 12.7 14.7 16.5b 11.8 Width between postorbital processes 13.8 13.1 15.3 16.8 Width between mastoid processes 21.8 24.6 24.9 21.4 Width between condyloid foramina 8.7 9.6 10.1 9.1c Ectotympanic, length 6.8 8.8 8.9 6.6 Ectotympanic, width at center 3.1 4.6 4.9 3.9 Length, anterior auditory regiond 5.9 8.1 8.2 5.6 Length, posterior auditory regione 7.0 6.6 6.9 9.2 Condylobasal length 64.3 70.0 74.1 65.2 Condylobasal lengthf 68.0 (Prionodon pardicolor,Nϭ 7) 71.3 (Prionodon linsang,Nϭ 9) Condylobasal lengthg 67.6 (Poiana richardsoni,Nϭ 7) a Measured between infraorbital foramina. b Estimated measurement. c Condyloid foramina are placed within the posterior lacerate foramina. d Measured from apex of ventral promontorial process to anterior limit of ectotympanic. e Measured from the apex of the ventral promontorial process to the posterior limit of the caudal entotympanic chamber of the bulla. f From Pocock (1933). g From Rosevear (1974). dental anatomy of the linsangs with the 9370 preserves one ectotympanicÐthis bone Quercy fossils suggests the possibility of re- is absent from the opposite side, allowing a lationship, one in which Prionodon is most full view of the petrosal and middle ear in similar to Palaeoprionodon; Poiana is more this species. Detailed descriptions of the ba- derived in its morphology, showing many sicranial structure of these two skulls are pre- correspondences with GenettaÐthe basicran- sented in Hunt (1998). Here I illustrate Qu ium provides particularly relevant informa- 9348 (®g. 4), and compare it with the basi- tion on the nature of this relationship. crania of the living linsangs. The hallmark of the auditory region of Pa- BASICRANIAL ANATOMY OF laeoprionodon is a large, centrally placed, PALAEOPRIONODON blocky petrosal, whose promontorium ex- I have been able to study the basicranial tends below the plane of the basioccipital to anatomy of two skulls of Palaeoprionodon form a prominent ventral process (®g. 4, V). lamandini from Quercy, France, in the col- This process buttresses the lateral margin of lections of the MuseÂum National d'Histoire the basioccipital. The apex of the promon- Naturelle, Paris. Both preserve the basicran- torium with its ventral process divides the ium in very good condition: MNHN Qu auditory region into anterior and posterior 9348 retains both ectotympanic bones (one parts. The anterior part is covered by the in the life position, demonstrated by the crescent-shaped ectotympanic bone: the tip placement of the ectotympanic crura in sock- of the anterior crus rests in a small depres- ets in the squamosal bone), and MNHN Qu sion in the squamosal; internal to this is a 8 AMERICAN MUSEUM NOVITATES NO. 1 broader concavity for the anterior face of the aeluroids (Hunt, 1987: 32): a small, unin¯at- ectotympanic. The posterior crus is attached ed hyaline cartilage that encloses a posterior to the post-tympanic process of the squa- auditory space of very small volume. When mosal that forms the anterior face of the mas- Nandinia's caudal entotympanic is removed toid process. from the auditory region, its perimeter of at- The petrosal promontorium is so large in tachment to the ectotympanic, petrosal, and these early aeluroids that the posterior limb rostral entotympanic is marked by character- of the inwardly tilted ectotympanic crescent istic ridges and rugose surfaces. Because the makes contact with the petrosal (®g. 4C, Z). areas for the attachment of the caudal ento- This contact is marked by a prominent facet tympanic to the ectotympanic and petrosal in on the surface of the promontorium, imme- Palaeoprionodon are nearly identical to the diately lateral to the ventral process (®g. 4B, same areas in Nandinia (compare ®gs. 4 and F). This distinctive juxtaposition of ectotym- 5A), it is virtually certain that a connective panic against the robust petrosal is an im- tissue element of similar form and dimen- portant anatomical characteristic of the ge- sions, either ®brous or cartilaginous, formed nus. the caudal entotympanic in Palaeopriono- However, it is the posterior part of the au- don. This presumably cartilaginous entotym- ditory region behind the promontorium that panic would have covered the posterior au- displays a striking plesiomorphic morpholo- ditory region, forming its ¯oor, and then, as gy unique to Palaeoprionodon among fossil in young Nandinia, extended forward as a aeluroids. Many living aeluroids (particularly strip of tissue along the medial rim of the felids and viverrids) have a greatly length- ectotympanic (®g. 5B, C). In both Nandinia ened auditory region behind the promonto- and Palaeoprionodon there is a gap between rium, and this elongated space is occupied by the inner margin of the ectotympanic and the a markedly in¯ated bony entotympanic cap- rostral entotympanic. This strip of connective sule, the caudal entotympanic of Van der tissue would ®ll this gap, intervening be- Klaauw (1922). In contrast, the earliest- tween the inner edge of the ectotympanic and known aeluroid crania from Quercy (Palaeo- the ventral edge of the rostral entotympanic prionodon, Stenoplesictis) and St.-GeÂrand (compare ®gs. 4B, C and 5B, C); the rostral (Stenogale) have short posterior auditory re- entotympanic is preserved as a small osseous gions of very small volume, which would wedge, separated from the ectotympanic, in have been covered by a small, unin¯ated some Quercy Palaeoprionodon (®g. 4, R). caudal entotympanic (Hunt, 1998). It is true This condition can be observed in both ju- that in the fossil Quercy crania no ossi®ed veniles and young adults of Nandinia (®g. 5, caudal entotympanic element has been pre- R). served that would establish its actual size and Thus, the auditory bulla of Palaeoprion- shape. However, although the posterior au- odon was made up of three ontogenetic ele- ditory region remains open and uncovered in ments: (a) a small, wedge-shaped, osseous these Quercy aeluroids, the rim of the ecto- rostral entotympanic in the anterointernal tympanic and the surfaces of the petrosal and corner of the auditory region; (b) a slightly surrounding basicranial bones demonstrate in¯ated (widened) bony ectotympanic cres- conclusively that a small plesiomorphic cau- cent applied to the surface of the petrosal; dal entotympanic was present (®g. 4B, x and and (c) a ®brous or cartilaginous caudal en- black triangles). It was either too loosely at- totympanic element of small size, covering a tached to remain with the skull during fos- posterior auditory area of very small volume. silization or was formed of an unossi®ed Palaeoprionodon exhibits a plesiomorphic connective tissue, such as cartilage, which aeluroid auditory cachet, identi®ed by the decayed and was not preserved. form of the petrosal and its ventral process, Among the living aeluroids, only the Af- the ectotympanic resting against the robust rican palm , Nandinia binotata, retains promontorium, and particularly by the mod- a cartilaginous caudal entotympanic in the est volume of the posterior chamber of the adult (®g. 5). This is the most plesiomorphic bulla enclosed by an unexpanded and unos- caudal entotympanic known among living si®ed caudal entotympanic. 2001 HUNT: AELUROID CARNIVORE EVOLUTION 9

Fig. 4. The basicranium of Palaeoprionodon (MNHN Qu 9348) from Quercy: (A) ventral, (B) posteroventral, and (C) posterolateral views. In (B) small black triangles indicate line of attachment of the caudal entotympanic to the margin of the petrosal. For abbreviations in this and subsequent ®gures, see pages 3±4. 10 AMERICAN MUSEUM NOVITATES NO. 1

Fig. 5. Basicranium and auditory region of the living African aeluroid Nandinia binotata. (A), juvenile female from Akenge, Zaire (AMNH 51450), ventral view; (B) medial view of AMNH 51450; (C) neonate, from Medje, Zaire (AMNH 51472). The ectotympanic is separated from the osseous rostral entotympanic by an intervening strip of connective tissue representing the anterior continuation of the caudal entotympanic. 2001 HUNT: AELUROID CARNIVORE EVOLUTION 11

BASICRANIAL ANATOMY OF nal; a postglenoid foramen is absent, indi- PRIONODON AND POIANA cating loss of the vein exiting the cranium at that point to supply the external jugular ve- A skull of Prionodon pardicolor was nous drainage (which is present in Palaeo- made available for dissection by the Depart- prionodon); the form and placement of the ment of Mammalogy, American Museum of glenoid fossa for the mandible and its rela- Natural History. It is a female collected in tion to the auditory bulla and foramen ovale 1938 from Mt. Victoria, Chin Hills, Burma, is the same; the mastoid region is not devel- at an elevation of 2800 m (AMNH 163595). oped as a prominent process or shelf; the ec- The form and size of the skull and the ex- totympanic (anterior) chamber of the bulla pression of many anatomical details are ex- lies directly in front of the caudal entotym- tremely similar to those of the Quercy Pa- panic (posterior) chamber; and the ectotym- laeoprionodon, and this is re¯ected in the ge- panic bone is a slightly expanded element neric name initially given to the fossil taxon resting on the petrosal promontorium. by Filhol (1880: 1579). Previous studies Pocock (1916b) and Bugge (1978) de- have emphasized the anatomical similarities scribed the path of the internal carotid artery between Prionodon and Palaeoprionodon in the auditory region of a number of viver- (Teilhard de Chardin, 1915; Gregory and rids but did not include the linsangs. The in- Hellman, 1939), but did not include a com- ternal carotid artery in both genera takes es- parison of their basicranial anatomy. Al- sentially the same course: it enters the audi- though relevant distinctions can be found be- tory region about midway along the length tween the two genera, the basicranial ana- of the bullaÐthe point of entrance can be tomical pattern of the living Prionodon par- observed at the posterior end of the ventral dicolor, including construction of the process of the promontorium (®g. 6). The ar- auditory bulla, shows an evident relationship tery runs between the ¯anged ventral process to the pattern in Palaeoprionodon. and the caudal entotympanic (remaining ex- For comparison with the Asian linsang ternal to that element, hence extrabullar), Prionodon, a skull of the African linsang, next travels along the lateral surface of the Poiana richardsoni, was also made available ventral process, and then follows the ventral for studyÐits auditory bulla already partially edge of the rostral entotympanic, turning me- dissected. It is a male collected in 1910 from dially to enter the cranial cavity at the middle Medje, Zaire (AMNH 51438), at an un- lacerate foramen. It is partially or entirely en- known elevation in the northeastern part of closed in a bony tube along the ventral bor- the west African rain forest (Allen, 1924). der of the rostral entotympanic. (Pocock Figure 6 compares the basicrania of the [1916b] mistakenly thought that the rostral Asian linsang (Prionodon pardicolor, entotympanic of viverrids was part of the ec- AMNH 163595) and the African linsang totympanic bone.) The anteriorly directed ar- (Poiana richardsoni, AMNH 51438). In gen- tery makes an abrupt 180Њ change in course eral the two linsangs share a similar anatom- at its anterior terminus in the auditory re- ical pattern. Some differences are evident: (a) gionÐhere the artery forms a loop nested in the condyloid (hypoglossal) foramen remains a depression in the basisphenoid, turning separate from the posterior lacerate foramen backward to enter the middle lacerate fora- in Prionodon, but is incorporated in that fo- men. Thus, during its passage through the ramen in Poiana; and (b) the ossi®ed caudal auditory region, the internal carotid artery entotympanic element of the auditory bulla never takes an intrabullar course: the artery is more in¯ated in Poiana than in Prionodon, does not penetrate the caudal entotympanic and in the former has grown forward over element to enter the posterior chamber of the the ectotympanic and backward toward the bulla, and because it remains con®ned to the paroccipital process to a greater degree than edge of the rostral entotympanic, it does not seen in Prionodon. In both of these charac- enter the anterior bulla chamber. ters, Prionodon is the more plesiomorphic Prionodon has a robust, blocky petrosal taxon. centrally situated in the auditory region (®g. In both genera there is an alisphenoid ca- 7). Its promontorium is elevated to form a 12 AMERICAN MUSEUM NOVITATES NO. 1

Fig. 6. Basicrania of Prionodon (AMNH 163595, A) and Poiana (AMNH 51438, B) in ventral view. The bony ¯oor of the posterior chamber of the auditory bulla has been removed on one side in each individual to show the ectotympanic resting on the petrosal promontorium, and the size of the posterior chamber formed by the caudal entotympanic. Note the more expanded or in¯ated posterior chamber in Poiana relative to Prionodon.

well-developed ventral process (V) that but- ventral process of Palaeoprionodon, reshap- tresses the edge of the basioccipital (®g. 7B). ing it into a low, laterally compressed ¯ange The margins of the basioccipital are bent tightly applied to the basioccipital. This downward and applied to the medial face of ¯ange becomes even more prominently de- the petrosal. Although the posterior face of veloped and bladelike in other genera of liv- the promontorium is more smoothly rounded ing viverrids (e.g., Genetta, Civettictis), but in the Asian linsang, the form of the Prion- remains in an incipient state in Prionodon.In odon petrosal is much like that of Palaeo- Poiana the promontorium and the ¯ange are prionodon. However, in Prionodon, the ven- similar to these features in Prionodon, but tral process of the promontorium has a some- the condition of the ¯ange in Poiana is what more derived appearance than seen in slightly more derived and begins to approach the Quercy genus: it is apparent that the liv- the more developed petrosal ¯ange observed ing Asian linsang has modi®ed the blocky in such living viverrids as Genetta. 2001 HUNT: AELUROID CARNIVORE EVOLUTION 13

Fig. 7. Dissection of the auditory region of Prionodon pardicolor (AMNH 163595) in oblique lateral view. The posterior chamber of the bulla has been opened (A), revealing the in¯ected dorsal margin of the caudal entotympanic applied to the petrosal, and the ectotympanic resting on the promontorium anterior to the round window. Removal of the ectotympanic and rostral entotympanic (B) allows an unrestricted view of the petrosal promontorium with its robust ventral process forming an incipient ¯ange buttressing the edge of the basioccipital. Note that the caudal entotympanic covers the ventral process of the promontorium when in place.

The degree of thinning of the ventral pro- auditory bulla of viverrids evolved, the pos- cess of the petrosal in living viverrids (re- terior chamber increased in volume by ex- sulting in the eventual development of the pansion of the caudal entotympanic element. bladelike ¯ange) is an index of the amount During this process of relative growth, ap- of caudal entotympanic expansion. As the plication of the enlarging caudal entotym- 14 AMERICAN MUSEUM NOVITATES NO. 1 panic to the petrosal promontorium altered Quercy Palaeoprionodon share a similarly the form of the ventral process in both Prion- shaped ectotympanic bone, inclined inward odon and Poiana. toward the midline to the same degree, and In the Quercy aeluroid, Palaeoprionodon, resting on the promontorium. The posterior the unossi®ed caudal entotympanic was sim- limb of the ectotympanic is directly applied ilar in form to the cartilaginous entotympanic to the surface of the promontorium immedi- of Nandinia (®g. 5; also Hunt, 1987: ®gs. 6, ately anterior to the round window. Contact 7B, 15±16), based upon a comparison of the is made over a linear distance of ϳ3±4 mm well-preserved auditory anatomy of MNHN in Prionodon (®g. 6, Z), but in the Quercy Qu 9348 and 9370 with that of Nandinia. carnivore the contact is restricted to the lat- The caudal entotympanic was a small, unex- eral part of the promontorium at the location panded element that covered the posterior of the facet (®g. 4, Z). In Prionodon this con- auditory region behind the ectotympanic. Al- tact extends farther medially, creating a more though it extended anteriorly along the me- complete partition between the anterior and dial rim of the ectotympanic to form part of posterior chambers of the bulla. Both Poiana the medial wall of the bulla, it did not alter and Prionodon share this more internally ex- the form of the ventral process of the pro- tended application of the ectotympanic rim montorium. But in Prionodon, the anteriorly to the promontorium (®g. 6, Z), which seems migrating and enlarging caudal entotympan- to be a derived trait, whereas the condition ic, accompanied by ossi®cation of the ele- in the Quercy genus is considered to repre- ment, compressed the ventral process to cre- sent the initial (rudimentary) application of ate the incipient ¯ange. Continued enlarge- the ectotympanic to the petrosal. Also, the ment and anterior growth of the caudal en- ectotympanic in Prionodon is relatively and totympanic chamber eventually produced the absolutely smaller than in Palaeoprionodon more derived ¯ange observed in Poiana and (table 1), so that in the Quercy genus the in other living viverrids. The conversion of large ectotympanic and massive petrosal the ventral process of the promontorium in combine to produce an anterior bulla cham- Poiana into an incipient ¯ange proceeds ber of somewhat greater volume relative to from the posterior part of the process for- the Asian linsang. ward, and the creation of the ¯ange is caused Just as in the Quercy genus, a small, by the margin of the caudal entotympanic el- wedge-shaped bony rostral entotympanic ®ts ement pressing only into the posterior part of tightly into the anterointernal corner of the the process, while not advancing craniad. auditory region in Prionodon. Along its ven- This demonstrates that an advancing caudal tral margin runs the internal carotid artery, entotympanic element can produce such a enclosed in a bony tube, with the exception ¯ange as it enlarges by a process of relative of a short section of the lateral wall of the growth. tube that is not ossi®ed and is closed by con- It is interesting that in a juvenile Poiana nective tissue. The medial rim of the ecto- richardsoni (AMNH 51440) the ectotympan- tympanic is in¯ected, and contacts and fuses ic is even larger in proportion to the caudal with the edge of the rostral entotympanic. entotympanic than in the adult, and the rel- The Asian and African linsangs share the ative size of these two bulla elements in this same form and spatial relationship of the ec- juvenile effectively duplicates the size of the totympanic and rostral entotympanic. In both ectotympanic and (hypothetical) caudal en- genera the inner margin of the ectotympanic totympanic elements in Palaeoprionodon. contacts the ventral edge of the rostral en- AMNH 51440 also shows that the caudal en- totympanic. The two bulla elements fuse totympanic in a juvenile is con®ned to the along this line without intervention of the posterior auditory region behind the ventral caudal entotympanic, and when the ectotym- process of the promontorium and has not ini- panic is carefully detached from the auditory tiated any signi®cant forward growth in or- region, the rostral entotympanic accompanies der to cover the ectotympanic. This is also it, the two elements delicately attached along true of Prionodon pardicolor. the inner rim of the ectotympanic bone (®g. The Asian linsang Prionodon and the 8). In both genera, the internal carotid artery 2001 HUNT: AELUROID CARNIVORE EVOLUTION 15

Fig. 8. Rostral entotympanic fused to the inner margin of the ectotympanic in Prionodon (AMNH 163595), medial view, ventral at top. Dashed line indicates path of the internal carotid artery within the ventral edge of the rostral entotympanic. travels along the ventral edge of the rostral color from Tonkin and (FMNH element and does not enter the middle ear 39175, 39176, 75814, all males) show an in- cavity. The application of the relatively termediate degree of in¯ation. Skulls of unexpanded ectotympanic to the small rostral Prionodon linsang from Borneo (FMNH entotympanic element in Prionodon and 88606, male; 8371, female) at the southeast- Poiana is an aeluroid character state also pre- ern limit of the genus in de- sent in most other living viverrids, hence velop the most in¯ated caudal entotympanic does not distinguish these genera. chambers. Inspection of Prionodon linsang The American Museum skulls of the Asian in other North American collections indicates and African linsangs can be immediately that this species has a consistently more in- separated by the different degree of expan- ¯ated caudal entotympanic bulla than P. par- sion of the caudal entotympanic. Among all dicolor.InP. linsang the caudal entotym- living viverrids, Prionodon pardicolor shows panic has grown forward, nearly covering the the closest correspondence to Palaeopriono- medial border of the ectotympanic, whereas don in the small volume of its posterior bulla in P. pardicolor the caudal entotympanic re- chamber (®g. 7). Measurements of the length mains largely posterior to the ectotympanic, of the posterior auditory region (table 1) in- migrating forward along its posterointernal dicate that P. pardicolor is proportionately margin in slight but varying degree in dif- similar to the Quercy genus, and differs in ferent individuals. this regard from Poiana (®g. 9). There is less An estimate of the relative amount of cau- expansion or in¯ation of the caudal entotym- dal entotympanic in¯ation in Prionodon and panic in Prionodon relative to Poiana, and Poiana is further indicated by the relation- during the ontogeny of these linsangs a pro- ship of the paroccipital process to the pos- gressive increase in in¯ation of this element terior wall of the caudal entotympanic. In occurs (e.g., a juvenile of Poiana richard- plesiomorphic carnivorans (e.g., Nandinia, soni, AMNH 51440, a male with milk teeth Daphoenus, Cynodictis, Amphicynodon, collected in 1914 at Niapu, Zaire, shows less Mustelictis, nimravine ), the paroccipital in¯ation of the caudal entotympanic than in process is not applied to the bulla, but rather the adults). exists as a rodlike process diverging postero- In Prionodon the relative amount of cau- ventrally from the exoccipital. In most living dal entotympanic in¯ation appears to in- viverrids, including Poiana, posterior growth crease from northwest to southeast over the and expansion of the caudal entotympanic geographic range of the genus. Skulls of creates a ®rm contact between entotympanic Prionodon pardicolor from (FMNH and the paroccipital process of the exoccip- 35463, female; 35464, male) at the north- ital, and as a result the process has become western geographic limit of the species dis- broadened and dorsoventrally expanded into play the least in¯ation. Skulls of P. pardi- a thin sheet of bone that covers most of the 16 AMERICAN MUSEUM NOVITATES NO. 1

Fig. 9. Dissection of the auditory region of Poiana richardsoni (AMNH 51438) in oblique lateral view (compare with ®g. 7). The posterior chamber of the bulla has been opened (A), showing the in¯ected dorsal margin of the caudal entotympanic applied to the petrosal, and the ectotympanic resting on the promontorium anterior to the round window. Removal of the ectotympanic and rostral entotym- panic (B) reveals the petrosal promontorium with ventral process produced as a ¯ange buttressing the edge of the basioccipital. The ¯ange is more developed in Poiana than in Prionodon, and the caudal entotympanic element is more in¯ated. posterior bulla wall. A discrete rodlike par- ly a rudimentary contact between the bulla occipital process no longer exists in these vi- and the base of the paroccipital process. Con- verrids. But in Prionodon, only a minimal sequently, the base of the process is only amount of posterior expansion of the poste- weakly indented by the expanding bulla, and rior bulla chamber (formed by caudal ento- the tip of the process remains recognizable, tympanic) has taken place, and there is mere- diverging slightly ventrad from the bulla (Po- 2001 HUNT: AELUROID CARNIVORE EVOLUTION 17 cock, 1933). In primitive aeluroids, such as link. Its auditory region not only can be de- Palaeoprionodon and the living palm civet, rived from that of the Oligocene aeluroid, Nandinia, the unossi®ed caudal entotympan- Palaeoprionodon, but also can be interpreted ic attaches near the base of the paroccipital as a transitional state between the Quercy ge- process, and the process remains a rodlike nus and the living African linsang, Poiana. structure, descending from the skull, and is Both Prionodon and Poiana exhibit an early not thinned and applied to the bulla wall. stage in the development of the modern vi- According to this interpretation, Priono- verrid auditory region found in living Ge- don is arrested at an evolutionary stage of netta, which is similar to the auditory region bulla development only somewhat more ad- of the other living viverrids. vanced than Palaeoprionodon, and evidently Of particular importance is the form of the derived from that stage. The principal mor- petrosal promontorium in Prionodon pardi- phological changes necessary to modify the color and Poiana richardsoni (®g. 10): the bulla of the Quercy aeluroid to arrive at the ventral process of the promontorium is pro- bulla morphology of Prionodon involved os- duced as an incipient ¯ange on the medial si®cation of the caudal entotympanic that margin of the Prionodon petrosal. Poiana forms the posterior chamber of the bulla, ac- displays a similar but more developed ¯ange. companied by a small amount of in¯ation of Yet both petrosals are very little modi®ed the chamber. The modest in¯ation of the cau- from the plesiomorphic aeluroid state found dal entotympanic in Prionodon resulted in a in the Quercy Palaeoprionodon. Other ana- slight change in shape of the ventral process tomical features of the auditory region of of the promontorium, creating a nascent con- these two living genera are correlated with ®guration of the petrosal ¯ange that became the differing degree of development of the more prominently modi®ed and developed in ¯anged ventral processÐplesiomorphic char- other living viverrids. The auditory anatomy acters in Asian Prionodon are consistently seen in Prionodon can be explained as an more derived in African Poiana: Prionodon intermediate morph between the character pardicolor has (a) a modest in¯ation of the state found in Palaeoprionodon and that of caudal entotympanic chamber of the bulla; Poiana. (b) a discrete hypoglossal foramen separated from the posterior lacerate foramen; and (c) DISCUSSION AND CONCLUSIONS an independent paroccipital processÐPoiana The basicranial anatomy of the Asian lin- has a more developed petrosal ¯ange, a more sang, Prionodon pardicolor, is of particular in¯ated caudal entotympanic, and has signi®cance in a consideration of the early merged the hypoglossal with the posterior evolution of aeluroid carnivorans. Previous- lacerate foramen. The greater amount of ex- ly, a study of the oldest known fossil aelu- pansion of the caudal entotympanic in roids from Quercy and St.-GeÂrand (Palaeo- Poiana has caused the paroccipital process to prionodon, Stenoplesictis, Stenogale, Hap- spread over the rear wall of the bulla, but in logale, Anictis) indicated the existence of a Prionodon, because there is less expansion primitive morphology of the petrosal and au- of the bulla, a vestige of the independent pro- ditory bulla common to several of these early cess remains. aeluroid taxa (Hunt, 1998). Skulls with well- There are other anatomical traits that sug- preserved basicrania are known for Palaeo- gest linsangs conserve a number of plesio- prionodon, Stenoplesictis, and Stenogale. morphic viverrid features. Prionodon is more The auditory regions of these fossil aeluroids conservative in these features relative to compare in anatomical grade with the living Poiana. archaic aeluroid, Nandinia binotata, but no For example, the surface of the neocortex connecting or intermediate morphological of Prionodon and Poiana retains one of the stage existed to link the archaic and the mod- simplest patterns among viverrids (Radinsky, ern basicranial types. The basicranium of 1975): there are only two major sulci, the Prionodon pardicolor, particularly the audi- coronolateral and suprasylvian, and in Prion- tory bulla and petrosal, provides that critical odon the cerebellum shows only a minimal 18 AMERICAN MUSEUM NOVITATES NO. 1 amount of overlap by the cerebrum, whereas don, also support a close relationship be- in Poiana the amount of overlap is greater. tween the Quercy genus and the Asian lin- The perineal scent glands are absent in sang. Prionodon pardicolor (Pocock, 1915a), but G. G. Simpson (1945) had placed Prion- appear to be present in a rudimentary state odon in a tribe, Prionodontini, and Poiana in in Poiana (Pocock, 1933: 970; Rosevear, the Viverrini closely related to Genetta. Such 1974: 222). Pocock quite reasonably argued segregation of Prionodon essentially follows that the absence of perineal glands in such Pocock's (1933) separation of the genus (as viverrids as Prionodon, Fossa, and Crypto- a subfamily Prionodontinae) following his procta, as well as their absence in herpestids, discovery that perineal glands were absent in felids, and hyaenids, indicated that the an- both sexes. Rosevear (1974) thought that the cestral aeluroid lacked such glands, which similarities between the Asian and African were later independently acquired in various linsangs were convergent and without evi- viverrid lines. Pocock's argument receives dence of af®nity. Here I would argue that the support from the observation that perineal basicranial similarities, coupled with evident glands differ in position and in anatomical correspondence in size and form of these an- detail in the various living viverrids that have imals, their dental and cranial traits, and pel- these glands (the details of glandular struc- age and external anatomy of the feet empha- ture and position are presented in a series of sized by Pocock, suggest a phylogenetic af- papers by Pocock, 1915a, 1915b, 1915c, ®nity among living Asian and African lin- 1915d, 1915e; Pocock, 1915d, clearly pre- sangs and genets, an af®nity shared at a more sents his rationale for the evolution of peri- basic level with the archaic aeluroids of the neal glands.) Considering the array of ple- genus Palaeoprionodon. In this sense, then, siomorphic basicranial features in Priono- because one of these living aeluroids, Prion- don, the absence of perineal glands is in- odon pardicolor, can be closely identi®ed structive. In Prionodon the placement of the with an Oligocene genus from Quercy, and in proximity to the scrotum, and the was much like it in size, body form, and in vulva to the anus, stands in contrast to their details of cranial and dental anatomy, this separation in viverrids with perineal glands, species of Asian linsang is reasonably con- and appears to be plesiomorphic for aelu- sidered a ``living fossil'', a modern proxy for roids, since this state also occurs in felids, an Oligocene grade of aeluroid evolution that herpestids, and Nandinia (but not in hyaen- has shown little change in ϳ25±30 million ids). years. Earlier studies of the linsangs by Mivart The combined anatomical evidence sug- (1882) and Pocock (1915a) established that gests the existence of a clade that includes the color and pattern of the pelage and the the Oligocene Palaeoprionodon as a ple- external anatomy of the feet and rhinarium siomorphic stem taxon, and the more derived also supported a relationship between the living genera, Prionodon, Poiana, and Ge- Asian and African species. My own obser- netta, as member taxa of a hypothetical mor- vations of the pelage of the two genera are phocline. Gregory and Hellman (1939) in agreement (brown with black markings, placed both Prionodon and Palaeoprionodon mostly spotted, with some striping on the in their subfamily Prionodontinae, and this shoulders, and rings around the tail). The subfamily I believe can reasonably incorpo- ability to retract the claws is present in both rate not only the Quercy Palaeoprionodon, genera, and the texture of the fur and their but also the species of living linsangs and general appearance and behavior particularly genets (table 2). Gregory and Hellman reminds one of many of the small living fe- (1939: 335) thought that the skull of Palaeo- lids. prionodon could represent the ancestor ``of Finally, the pronounced similarity in de- all the diverse subfamilies of the Viverri- tails of the dentition of Palaeoprionodon and dae''. Study of the Quercy fossils suggests a Prionodon (form of m1, m2, reduction of an- more limited role for Palaeoprionodon near terior premolars), and the retention of ple- the ancestry of only the linsangs and genets. siomorphic double-rooted P1/p1 in Priono- The viverrine viverrids (, Viverricula, 2001 HUNT: AELUROID CARNIVORE EVOLUTION 19

Fig. 10. Final stage of the dissection of the auditory bulla of Prionodon (AMNH 163595, A) and Poiana (AMNH 51438, B) in ventral view. Note the robust ventral process of the promontorium in Prionodon, only incipiently modi®ed as a ¯ange appressed against the basioccipital, hence similar to the form of the ventral process in Palaeoprionodon (compare with ®g. 4). In Poiana the ¯ange has been further modi®ed as a thin blade and has been extended fore and aft to a greater degree than in Prionodon. The caudal entotympanic is more in¯ated in Poiana, broadly contacting the paroccipital process, whereas in Prionodon the caudal entotympanic is not as expanded, and the process still retains a vestige of its primitive rodlike form. 20 AMERICAN MUSEUM NOVITATES NO. 1

Civettictis) appear to be closely related to the the diversity that existed during the early his- Prionodontinae because of their marked sim- tory of the family, con®ned as it is to the Old ilarity in dental and basicranial anatomy. But World, where much of the viverrid record of the other principal subfamilies of viverrids, fragmentary jaws and teeth comes from ex- the and , proba- ceptional burial settings such as the ®ssure bly stem from other early aeluroids of Eo- deposits at Quercy, La Grive, and Winter- cene or Oligocene age. shof-West, and the freshwater limestones of Pocock (1916b) astutely remarked that the St.-GeÂrand and Oeningen. The few European Viverridae, as originally conceived by Flow- viverrid fossils, coupled with a rather meager er (1869) and Mivart (1882), was ``a hetero- Asian record, make it dif®cult to determine geneous group including all the aeluroids if the living genera are of considerable anti- which are not obviously cats or ''. quity, or if the species diversity evident today Part of this heterogeneity derived from inclu- in southeast Asia and Africa only recently sion of the in Viverridae; today, came into existence. herpestids have been removed from Viverri- Because many early viverrids were prob- dae by general consensus (Gregory and Hell- ably arboreal and thrived in subtropical to man, 1939; Hunt, 1987, 1991; Wyss and tropical forests, where skeletal remains often Flynn, 1993; McKenna and Bell, 1997). Re- are infrequently preserved in the fossil re- cent studies of the aeluroid Carnivora have cord, their past history could continue to re- recognized the morphologic uniformity of main poorly known. However, as some vi- herpestids, as well as the living felids and verrids occupied more open environments in hyaenidsÐa highly uniform basicranial mor- semiarid to arid regions during the climatic phology characterizes each family, and on- cooling of the Neogene, particularly in tec- going molecular/biochemical studies also tonically active regions in south Asia and the support the integrity of these groups (Wurster African rift, where sedimentation in restrict- and Benirschke, 1968; Wurster-Hill and ed basins was taking place, the probability of Gray, 1975; Wayne et al., 1989). But viver- preservation of these species increased and a rids are consistently seen as more diverse, an modest record has developed in these set- observation that would appear to be sup- tings. But the scarcity of viverrid fossils in ported by the variety of subfamilies and gen- these sedimentary environments suggests era created for these by Pocock that they were never very abundant, with a and other students, many genera including species diversity probably similar to the pre- only a single species. sent. Are the viverrids a monophyletic group? This perspective envisions the preserva- What evidence exists to suggest that viver- tion of Palaeoprionodon and other aeluroid rids are not simply a polyphyletic aggregate fossils at Quercy as the chance result of the of primitive aeluroids? I would argue that development of karst terrain in southern three lines of evidence support monophyly: France in the and early Oligocene. (a) the morphology of the basicranium, par- Karst ®ssures serve as particularly effective ticularly the ontogenetic development and traps for small carnivores. Indeed, the karst adult form of the auditory bulla and petrosal ®ssures at Quercy sample a rather diverse ae- (Hunt, 1974, 1987); (b) the general tendency luroid array of taxa [the early felids Proail- to develop perineal glands in the family (Po- urus and Stenogale, a -like aeluroid cock); and (c) molecular-biochemical evi- Haplogale, the enigmatic aeluroids Steno- dence indicating relationship among genets, plesictis and Anictis (Teilhard de Chardin, , and paradoxures (Wayne et al., 1989). 1915; Hunt, 1998)]. The Quercy ``window'' But the tracing of living viverrid genera and provides a brief, geographically limited species backward in time beyond the later glimpse into an aeluroid radiation that we Cenozoic has met with little recent success. know was not con®ned to western Europe The relatively poor fossil record of viver- but was probably unfolding throughout the rids is in part responsible for this situation Eurasian landmass, based upon the fossil ae- (Petter, 1974; Hunt, 1996a, for a summary of luroid record beginning to appear at eastern the viverrid fossil record). We know little of Asian localities such as Hsanda-Gol, Alag 2001 HUNT: AELUROID CARNIVORE EVOLUTION 21

Tsab, and Ergil Obo (Dashzeveg, 1996). Pa- TABLE 2 laeoprionodon is not necessarily the ances- Classi®cation of the Aeluroid Families tral viverrid, as Gregory and Hellman Viverridae and Nandiniidae claimed, but is only one of a stem group of early aeluroids that includes Haplogale, Order Carnivora Bowdich, 1821 Division Aeluroidea Flower, 1869 Stenogale, Stenoplesictis, and Proailurus in Family Viverridae Gray, 1821 Europe, and Asiavorator, Shandgolictis, Proailurus, and additional poorly known Subfamily Prionodontinae Pocock, 1933 species in eastern Asia (Hunt, 1998). The Palaeoprionodon Filhol, 1880 plesiomorphic aeluroid basicranial pattern Prionodon Hors®eld, 1821 common to several of these genera suggests Poiana Gray, 1864 Genetta G. Cuvier, 1816 that the same pattern probably was shared by other early aeluroids distributed across Eur- Subfamily Gray, 1821 asia at this time, many as yet unrepresented Viverra Linnaeus, 1758 by cranial remains. Viverricula Hodgson, 1838 The abrupt appearance of Quercy aelu- Osbornictis Allen, 1919 roids at the beginning of the Oligocene, fol- Civettictis Pocock, 1915 lowing the Grande Coupure event in Europe, Subfamily Euplerinae Chenu, 1852 yet their complete absence in the Eocene fau- Fossa Gray, 1864 nas that precede that event, has suggested to Doyere, 1835 several investigators that aeluroids immigrat- Subfamily Cryptoproctinae Gray, 1864 ed to Europe from elsewhere. It is possible Cryptoprocta Bennett, 1833 that increasing continental aridity in central Subfamily Hemigalinae Gray, 1864 Asia during the Oligocene could have ac- companied the development and radiation of Hemigalus Jourdan, 1837 Diplogale Thomas, 1912 aeluroids in that region, leading eventually to Chrotogale Thomas, 1912 their spread westward across the Eurasian Cynogale Gray, 1837 landmass. The importance of the fossils that we can now view through the Quercy-Mon- Subfamily Paradoxurinae Gray, 1864 golia ``window'' into early aeluroid evolu- F. Cuvier, 1821 tion is found in the similar of these Paguma Gray, 1831 Arctictis Temminck, 1824 geographically dispersed fossil species, and Arctogalidia Merriam, 1897 in the uniformity of the known basicranial Macrogalidia Schwartz, 1910 patterns among various genera, the sum of the evidence indicating that these aeluroid Family Nandiniidae Pocock, 1929 fossils document an initial phase of the Eur- Nandinia Gray, 1843 asian aeluroid diversi®cation. Of particular signi®cance, however, is the fact that these Oligocene aeluroids do not in- The Viverrinae most often include the genera clude clear and direct morphological ante- Viverra, Viverricula, Civettictis, Osbornictis, cendents to modern viverrid species (with the Genetta, Poiana, and occasionally the Mal- exception of Palaeoprionodon). The Oligo- agasy Fossa. All are restricted to or largely cene aeluroids display an archaic basicranial con®ned to Africa except for Viverra and pattern and hypercarnivorous dentitions that Viverricula, which are found in southeast are much different from the derived auditory Asia (a large species of Viverra, however, oc- anatomy and dentitions of many living vi- curred in Africa in the Plio-Pleistocene, see verrids. Petter, 1963; Hunt, 1996b). Viverrines are Within the Viverridae, several subgroups dentally conservative, maintaining a shearing have been consistently recognized since the carnassial pair and functional molars (M1-2, 19th century (Mivart, 1882; Gregory and m1-2). Hellman, 1939; Simpson, 1945; McKenna With the exception of the viverrids isolat- and Bell, 1997): the subfamilies Viverrinae, ed on Madagascar (Fossa, Eupleres, Cryp- Paradoxurinae, and Hemigalinae (table 2). toprocta; for a review see Petter, 1974), the 22 AMERICAN MUSEUM NOVITATES NO. 1 remaining living genera are found in south- joined or geographically adjacent popula- east Asia and the islands of and tions that occur in spatial continuity over are placed in the subfamilies Paradoxurinae large geographic regions. The wide-ranging (Paradoxurus, Paguma, Arctictis, Arctogali- terrestrial character of these viverrids (al- dia, Macrogalidia) and Hemigalinae (Hemi- though most climb well and utilize trees) galus, Diplogale, Chrotogale, Cynogale). suggests a genetic continuity over their rang- Certain lines of evidence, such as the endo- es, with less tendency to fragment into allo- cranial casts of viverrids studied by Radinsky patric species. Viverra-Viverricula comprises (1975) suggest that the paradoxures may be a group of similar, closely related species ex- made up of closely related genera. Both sub- tending from Nepal and eastern to families are composed of species that occupy south China, Indochina, and continuing forested settings; are largely nocturnal and southward to Malayasia and the Indonesian arboreal; have perineal glands (reduced or islands. They differ little except in size. The lost in some males); are usually plantigrade, large bush civet, Civettictis, despite its con- with the entire sole of the hindfoot in contact siderable geographic range from the southern with the substrate; and have evolved modi- border of the Sahara southward to Namibia ®ed and molars in which the typ- and South Africa, exists as a single species ical shearing function is modi®ed to a crush- (C. civetta) in both forest and savanna en- ing mode, and the posterior molars can be vironments. At least nine species of genets quite small and reduced in size. Paradoxures (Genetta) are said to occupy forests, grass- and hemigalines seem to be geographically lands, and savannas over nearly all of Africa restricted relict species, derived from an ear- (Nowak, 1991); this species diversity has lier viverrid radiation in eastern AsiaÐan as- been called into question (Rosevear, 1974), semblage of diverse taxa that survived in because genets are known to be highly var- tropical-subtropical settings as these environ- iable in size, pelage color, and even cranial ments became restricted to lower latitudes characters (Allen, 1924). It is likely that a during mid- and late-Cenozoic global cool- number of the named species may in fact be ing. The climatic oscillations of the late Pli- capable of interbreeding and are not repro- ocene and Pleistocene probably contributed ductively isolated. Much of the variation ob- to the species diversity of the paradoxures served among the genets may be the result and hemigalines. Lowering of global sea lev- of the widespread distribution of the species el during glacial maxima provided land mi- over the African continent, and the conse- gration routes from the Asian mainland quent adaptation of populations to local con- southward to the islands of the Sunda shelf, ditions. Kingdon (1977) notes that diversi®- permitting the movement of south Asian cation in African genets has produced a num- mammals into the islands of Indonesia. ber of distinct ecological species adapted to When warmer interglacial intervals ¯ooded particular environments (some of these spe- the Sunda shelf, creating the islands of Su- cies have produced hybrids in captivity). To matra, Java, Borneo, and the numerous what extent the diversi®cation of African vi- smaller satellite islands, the isolation of vi- verrids is due to the ¯uctuation in tropical verrid populations on these islands must have forested environments during the Quaternary resulted in genetic fragmentation of once- glacial episodes remains an interesting and continuous populations. With suf®cient time unresolved question. and continued isolation, the fragmented pop- The considerable diversity of paradoxures, ulations could serve as a potential stockpile hemigalines, and the viverrines Viverra and for the derivation of new viverrid species. Viverricula, seems somehow linked to the The geographic restriction of paradoxures episodic isolation of land areas in Indonesia and hemigalines to forested tropical environ- and adjacent southeast Asia during the cli- ments virtually assures that this scenario oc- matic oscillations of the late Cenozoic. On curred in Indonesia in the later Cenozoic. the other hand, the African viverrids Civet- The more terrestrial viverrid species of Af- tictis and Genetta, although widely distrib- rica (Genetta, Civettictis) and southeast Asia uted over the continent and exhibiting evi- (Viverra, Viverricula) are made up of con- dent populational variation, are without the 2001 HUNT: AELUROID CARNIVORE EVOLUTION 23 production of marked generic diversity chaux du Quercy. C. R. Acad. Sci., Par- equivalent to that seen in southeast Asian vi- is 90: 1579±1580. verrids. Flower, W. H. Malagasy viverrids include the cranially 1869. On the value of the characters of the and dentally plesiomorphic species Fossa base of the cranium in the classi®cation fossa and the highly derived and specialized of the Order Carnivora. Proc. Zool. Eupleres and Cryptoprocta, that are not in Soc. London 1869: 4±37. Gregory, W. K., and M. Hellman evidence anywhere outside Madagascar, and 1939. On the evolution and major classi®ca- likely represent endemic lineages evolved in tion of the civets (Viverridae) and al- isolation on the island. An improved fossil lied fossil and recent Carnivora: a phy- record of viverrids from the Neogene of Eur- logenetic study of the skull and denti- asia and Africa will be necessary to under- tion. Proc. Am. Philos. Soc. 81(3): stand the origins of these diverse viverrid 309±392. subgroups. Hunt, R. M., Jr. 1974. The auditory bulla in Carnivora: an an- ACKNOWLEDGMENTS atomical basis for reappraisal of carni- vore evolution. J. Morphol. 143: 21±76. I am grateful to Dr. L. Ginsburg, MuseÂum 1987. Evolution of the aeluroid Carnivora: National d'Histoire Naturelle, Paris, for per- signi®cance of auditory structure in the mission to study the crania of Quercy aelu- nimravid Dinictis. Am. Mus. Nov- roids, and to Dr. N. Simmons, American Mu- itates 2886: 74 pp. seum of Natural History, New York, for the 1991. Evolution of the aeluroid Carnivora: vi- loan of the skulls of Prionodon and Poiana, verrid af®nities of the Miocene carni- and for allowing the dissection of the audi- voran . Am. Mus. Novitates tory regions of representative individuals. 3023: 34 pp. 1996a. My thanks to Dr. Bruce Patterson, Field Mu- Biogeography of the Order Carnivora. In J. Gittleman (ed.), Carnivore Behav- seum of Natural History, Chicago, for the ior, Ecology and Evolution 2: 485±541. opportunity to study the linsang material in Ithaca, NY: Cornell Univ. Press. their collections. For ®gure 1, I thank illus- 1996b. Basicranial anatomy of the giant viver- trator Angie , University of Nebraska rid from `E' Quarry, Langebaanweg, State Museum, Lincoln. South Africa. In K. Stewart and K. Seymour (eds.), Palaeoecology and Pa- REFERENCES laeoenvironments of Late Cenozoic Mammals: Tributes to the Career of Allen, J. A. C.S. Churcher: 588±597. Toronto: 1924. Carnivora collected by the American Univ. Toronto Press. Museum Congo Expedition. Bull. Am. 1998. Evolution of the aeluroid Carnivora: di- Mus. Nat. Hist. 47(3): 73±281. versity of the earliest aeluroids from Bugge, J. Eurasia (Quercy, Hsanda-Gol) and the 1978. The cephalic arterial system in carni- origin of felids. Am. Mus. Novitates vores, with special reference to the sys- 3252: 65 pp. tematic classi®cation. Acta Anat. 101: Kingdon, J. 45±61. 1977. East African Mammals: An Atlas of Dashzeveg, D. Evolution in Africa. Vol. III(A), Car- 1996. Some carnivorous mammals from the Paleogene of the Eastern Gobi Desert, nivores. New York, NY: Academic Mongolia, and the application of Oli- Press, 476 pp. gocene carnivores to stratigraphic cor- Klaauw, C. J. van der relation. Am. Mus. Novitates 3179: 14. 1922. Uber die Entwickelung des Entotym- Ewer, R. F. panicums. Tijdschr. Ned. Dierk. Ver- 1973. The Carnivores. Ithaca, NY: Cornell een. 18: 135±174. Univ. Press, 494 pp. McKenna, M. C., and S. K. Bell Filhol, H. 1997. Classi®cation of Mammals Above the 1880. Sur la deÂcouverte de mammifeÁres nou- Species Level. New York: Columbia veaux dans les deÂpoÃts de phosphates de Univ. Press, 631 pp. 24 AMERICAN MUSEUM NOVITATES NO. 1

Miller, M. E., G. C. Christensen, and H. E. Evans 1916b. On the course of the internal carotid ar- 1964. The Anatomy of the . Philadelphia: tery and the foramina connected there- W. B. Saunders, 941 pp. with in the skulls of the and Mivart, S t. G. Viverridae. Ann. Mag. Nat. Hist. 17: 1882. On the classi®cation and distribution of 261±269. the Aeluroidea. Proc. Zool. Soc. Lon- 1933. The rarer genera of Oriental Viverridae. don 1882: 135±208. Nowak, R. M. Proc. Zool. Soc. London 1933: 969± 1991. Walker's Mammals of the World (5th 1035. ed.) 2: 1149±1151. Baltimore: Johns Radinsky, L. Hopkins Univ. Press. 1975. Viverrid neuroanatomy: phylogenetic Petter, G. and behavioral implications. J. Mam- 1963. Etude de quelques ViverrideÂs (Mam- mal. 56(1): 130±150. mifeÁres, Carnivores) du PleistoceÁne in- Rosevear, D. R. feÂrieur du Tanganyika. Bull. Soc. GeÂol. 1974. The Carnivores of West Africa. Br. Fr. (ser. 7) 5: 265±274. Mus. (Nat. Hist.) Publ. 723: 220±228. 1974. Rapports phyleÂtiques des ViverrideÂs Simpson, G. G. (Carnivores ®ssipeÁdes). Les formes de 1945. The principles of classi®cation and a Madagascar. Mammalia 38(4): 605± classi®cation of mammals. Bull. Am. 636. Mus. Nat. Hist. 85: 350. Pocock, R. I. 1908. Report upon a small collection of Teilhard de Chardin, P. Mammalia brought from Liberia by Mr. 1915. Les carnassiers des phosphorites du Leonard Leighton. Proc. Zool. Soc. Quercy. Ann. PaleÂontol. 9(3,4): 1±90. London 1907: 1037±1046. Wayne, R. K., R. E. Benveniste, D. N. Janczews- 1915a. On some of the external characters of ki, and S. J. O'Brien the genus Linsang, with notes upon the 1989. Molecular and biochemical evolution genera Poiana and Eupleres. Ann. of the Carnivora. In J. Gittleman (ed.), Mag. Nat. Hist. 16: 341±351. Carnivore behavior, ecology, and evo- 1915b. On the feet and glands and other exter- lution 2: 465±494. Ithaca, NY: Cornell nal characters of the Viverrinae, with Univ. Press. the description of a new genus. Proc. Wurster, D. H., and K. Benirschke Zool. Soc. London 1915: 131±149. 1915c. On the feet and glands and other exter- 1968. Comparative cytogenetic studies in the nal characters of the paradoxurine gen- Order Carnivora. Chromosoma 24: era Paradoxurus, Arctictis, Arctogali- 336±382. dia, and Nandinia. Proc. Zool. Soc. Wurster-Hill, D. H., and C. W. Gray London 1915: 387±412. 1975. The interrelationships of chromosome 1915d. On some external characters of Gali- banding patterns in procyonids, viver- dia, , and related genera. rids, and felids. Cytogenet. Cell Genet. Ann. Mag. Nat. Hist. 16: 351±356. 15: 306±331. 1915e. On some of the external characters of Wyss, A. R., and J. J. Flynn the palm-civet (Hemigalus derbyanus, Gray) and its allies. Ann. Mag. Nat. 1993. A phylogenetic analysis and de®nition Hist. 16: 153±162. of the Carnivora. In F. S. Szalay, M. J. 1916a. The alisphenoid canal in civets and hy- Novacek, M. C. McKenna (eds.), aenas. Proc. Zool. Soc. London 1916: phylogeny (placentals) 32± 442±445. 52. New York: Springer.

Recent issues of the Novitates may be purchased from the Museum. Lists of back issues of the Novitates and Bulletin published during the last ®ve years are available at World Wide Web site http://nimidi.amnh.org. Or address mail orders to: American Museum of Natural History Library, Central Park West at 79th St., New York, NY 10024. TEL: (212) 769-5545. FAX: (212) 769- 5009. E-MAIL: [email protected]

a This paper meets the requirements of ANSI/NISO Z39.48-1992 (Permanence of Paper).