New Early Pliocene Cercopithecidae (Mammalia: Primates) from Aramis, Middle Awash Valley, Ethiopia

Home , Pliopapio

PUBLISHED BY THE AMERICAN MUSEUM OF NATURAL HISTORY CENTRAL PARK WEST AT 79TH STREET, NEW YORK, NY 10024 Number 3350, 36 pp., 12 ®gures, 2 tables October 31, 2001

STEPHEN R. FROST1

ABSTRACT The Middle Awash Research Project has collected a large sample of fossil cercopithecids from the Aramis, Kuseralee, and Sagantole drainages in the Middle Awash paleoanthropolog- ical study area of Ethiopia. These sites have been securely dated to 4.4 Ma. The craniodental material from this assemblage supports the diagnoses of two distinct new genera and species, which are described here. Pliopapio alemui is a mid-sized papionin represented by a complete cranium, several partial jaw fragments, and many isolated teeth. Kuseracolobus aramisi is a medium-sized colobine represented by several maxillae, mandibles, and other cranial fragments, as well as by isolated teeth. Stratigraphically associated postcranial remains will be discussed in a separate report. Pliopapio alemui is distinctive from other known African papionins in the combination of its cranial, mandibular, and dental morphology. It lacks the diagnostic facial features of Par- apapio, as well as the ¯attened muzzle dorsum, facial fossae, and maxillary ridges of Papio. Moreover, it does not possess any of the derived dental and cranial specializations of Thero- pithecus. Kuseracolobus aramisi is larger than all modern African colobines, but smaller than all known Cercopithecoides, Paracolobus, and Rhinocolobus. It is distinctive from Cercopi- thecoides and the colobine from Leadu in its symphyseal, corporal, and gonial morphology, and from Libypithecus, Paracolobus, and Rhinocolobus in its facial morphology. This early Pliocene sample ®lls a temporal gap between the terminal Miocene and later Pliocene sites and documents the existence of two new cercopithecid taxa, increasing known diversity in the family.

1 New York Consortium in Evolutionary Primatology, Ph.D. Program in Anthropology, City University of New York, and Division of Paleontology, American Museum of Natural History.

INTRODUCTION this paper are based on the limited number of relatively complete specimens. While the The Cercopithecidae are relatively com- isolated teeth are only diagnosable to sub- mon and diverse in most Pliocene and Pleis- family, they are assigned to the same two tocene fossil assemblages from sub-Saharan species as the more complete material based Africa. During this period, the subfamily on size and a lack of evidence to the con- Cercopithecinae is represented by eight gen- trary. era or subgenera while the Colobinae is represented by at least ®ve (Szalay and Delson, INSTITUTIONAL ABBREVIATIONS 1979; Delson, 1984, 1988, 1994). Few of these taxa, however, are represented in the KNM National Museums of Kenya, Nairobi, Early Pliocene. The locality of Aramis in the Kenya Awash River valley of Ethiopia has yielded NME National Museum of Ethiopia, Addis Ababa, Ethiopia the hominid Ardipithecus ramidus (White et NMT National Museum of Tanzania, Dar es al., 1994, 1995). The large cercopithecid Salaam, Tanzania sample from Aramis and sites of similar age MB Berlin Museum, Berlin, Germany in the Middle Awash adds considerably to our knowledge of cercopithecid evolution SYSTEMATIC PALEONTOLOGY during this period. The craniodental anatomy and systematics of these cercopithecids are CLASS MAMMALIA LINNAEUS, 1758 described here. ORDER PRIMATES LINNAEUS, 1758 Between 1992 and 1999, the Middle INFRAORDER CATARRHINI E. GEOFFROY, 1812 Awash Research Project collected and cata- FAMILY CERCOPITHECIDAE GRAY, 1821 logued 925 cercopithecid specimens from localities within the Aramis, Kuseralee, and SUBFAMILY CERCOPITHECINAE GRAY, 1821 Sagantole drainages. This sample comes TRIBE PAPIONINI BURNETT, 1828 from the Aramis Member of the Sagantole Pliopapio, new genus Formation between the GaÁala Tuff Complex TYPE SPECIES: Pliopapio alemui, new spe- (GATC) and the Daam Aatu Basaltic Tuff cies. (DABT). These tephra have been radiomet- GENERIC DIAGNOSIS: A genus of African rically dated to 4.39 Ϯ 0.03 and 4.39 Ϯ 0.05 papionin, distinguished from Parapapio, Lo- Ma respectively (Renne et al., 1999). The phocebus, and Cercocebus by the presence size, precise stratigraphy, and dating of this of a clear anteorbital drop, although this is sample make it one of the best documented not as distinct as in most Papio and Man- in the African Pliocene. drillus. In this aspect, its pro®le is most sim- From a limited initial sample, Wolde- ilar to that of Macaca, but Pliopapio has a Gabriel et al. (1994) listed three cercopithe- relatively longer muzzle. It is different from cid species at Aramis based on preliminary Papio (Papio), Gorgopithecus, Lophocebus, identi®cations by Delson: cf. Paracolobus Cercocebus, and Mandrillus in that the muz- sp., cf. Colobinae sp. ``A'' (sensu Eck, 1977) zle lacks postcanine and suborbital fossae. and cf. Parapapio sp. On the basis of the The absence of maxillary ridges distinguish- current, vastly expanded craniodental sam- es it from Papio, Theropithecus (Omopithe- ple, two new genera and species are diag- cus) and Mandrillus. The muzzle dorsum is nosed here, one colobine and one cercopithe- saddle-shaped and rounded in paracoronal cine. In the postcranial sample, two and pos- section. In these aspects it is similar to Ther- sibly three morphs can be identi®ed. None of opithecus oswaldi (sensu Leakey, 1993, as is the postcranial specimens are directly asso- used throughout this paper), but is unlike the ciated with any of the craniodental material, ¯attened dorsum and squared paracoronal and their af®nities and functional interpreta- section found in Papio (including Dinopithe- tions will be presented elsewhere. cus) and Mandrillus. Relative to neurocranial This is one of the largest Early Pliocene breadth, the rostrum is narrow in comparison samples of fossil cercopithecids; however, with those of all known African papionins, the diagnoses for the two taxa described in although it is in the lowest end of variation 2001 FROST: CERCOPITHECIDAE 3 for Papio, Mandrillus, and Macaca. Crush- with complete but damaged dentition, found ing in the anterior portion may contribute to by Awoke Amzaye in December 1995. this impression, but cannot account for it en- HYPODIGM: See appendix 1. tirely. HORIZON: All specimens are derived from Unlike those in Parapapio, Cercocebus, the Aramis Member of the Sagantole For- and Lophocebus, the cranial vault is separat- mation, between the GATC and DABT, and ed from the brow ridges by a distinct are therefore dated to 4.4 Ma (Renne et al., ophryonic groove. The temporal lines in the 1999). holotype male remain widely separated and DESCRIPTION: The most complete specimen do not form a sagittal crest, as opposed to is the holotype male skull ARA-VP-6/9332 the situation in Theropithecus, Gorgopithe- (see ®g. 1). The mandible is attached to the cus, Papio (Dinopithecus), and Paradoli- cranium by a thin layer of matrix that pre- chopithecus. vents the two from being safely separated. While the mandibular symphysis is shal- Therefore, the palate and much of the cranial lower and more sloping in pro®le than that base are not available for observation. Of the of most papionins, it is shorter and more main cranial regions, only the left zygomatic rounded than that of Parapapio ado (from is completely missing. The right orbit is Laetoli and Kanapoi) or the papionin from damaged and lacks much of the right half, Lothagam. In these latter taxa the symphysis except for one triangular portion around the is even longer and more sloping, with the zygoma. The right zygomatic arch is mostly incisive alveolar process projecting more an- present, but crushed, missing only the region teriorly, producing a more procumbent inci- around the temporojugal suture. The right sor row. Corpus fossae are absent, distin- maxilla and mandible are weathered, reveal- guishing it from most Papio (Papio)(P. h. ing the roots of the teeth. The other cranial kindae females have very slight fossae, and specimens are all considerably more frag- some P. izodi lack them), Gorgopithecus, mentary. ARA-VP-6/437 (see ®g. 2) is a par- Theropithecus (Omopithecus), Mandrillus, tial right male maxillary fragment with the and Lophocebus. dorsal surface up to the piriform aperture, the The molars are higher crowned, more roots of the canine and fourth premolar, and straight-sided, and less ¯aring on average complete central incisor and third premolar. than those of Papio, Macaca, Mandrillus, ARA-VP-1/1723 (®g. 2) is a partial right fe- and Lophocebus, and far less so than those male maxilla preserving the canine through of Cercocebus. As in most papionins, but un- third molar. ARA-VP-1/1007 (®g. 2) is a like Mandrillus and Cercocebus, the premo- slightly crushed left female premaxillomax- lars are not particularly large relative to the illary fragment with the rostral surface premolars. The mandibular incisors are nearly served nearly to the lateral border of the pir- vertically implanted, whereas those of Par- iform aperture, a damaged lateral incisor, and apapio ado from Laetoli and Kanapoi are complete canine through ®rst molar. more procumbent. Pliopapio alemui is smaller in cranial size ETYMOLOGY: The name refers to the Plio- than all known Theropithecus and Papio, cene age of the taxon and its papionin status. other than P. hamadryas kindae, P. angus- ticeps, and P. izodi, to which it is similar in Pliopapio alemui, new species size. It is sightly larger than all but the largest Macaca, such as M. thibetana and M. ne- SPECIFIC DIAGNOSIS: As for genus. mestrina. This cranial size estimate is con- ETYMOLOGY: Named in honor of Alemu ®rmed by centroid sizes computed from a Ademassu, Paleoanthropology Laboratory sample of 45 cranial landmarks (see ®g. 3). curator at the National Museum of Ethiopia In dental size, it is marginally smaller than responsible for assisting with preparation, Parapapio ado from Laetoli and Kanapoi, molding, and study of all Middle Awash specimens. 2 Sex identi®cations throughout the text are based HOLOTYPE: ARA-VP-6/933. A nearly com- upon the highly dimorphic size and morphology of the plete male cranium and attached mandible upper and lower canines and the lower third premolar. 4 AMERICAN MUSEUM NOVITATES NO. 3350

Fig. 1. Holotype of Pliopapio alemui ARA-VP-6/933. All views are in Frankfurt orientation. Lateral view of the left side. Opposite page: dorsal view (above); lateral view of the right side (below). and similar to specimens from Ekora, the Lo- tinct from Papio and Mandrillus. The max- mekwi Member of the Nachukui Formation, illary fossae are also extremely shallow. the Tulu Bor Member of the Koobi Fora For- Once again, this is similar to the above gen- mation, and Unit 2 of the Chiwondo beds. era and to Papio (Dinopithecus) (see Delson Dental dimensions for Pl. alemui are listed and Dean, 1993). From what is preserved in in table 1. ARA-VP-1/1007 and ARA-VP-1/1723 the Rostrum. The infraorbital foramina are females seem to lack these structures as well. only preserved on the left side of ARA-VP- The muzzle dorsum of ARA-VP-6/933 is 6/933. The four foramina are arranged in a largely smooth and saddle-shaped, as it is in superolaterally concave arc, as in Theropi- Theropithecus oswaldi. It is concave in the thecus (Eck and Jablonski, 1987). Relative to sagittal plane and forms a convex arc in par- the orbit they lie roughly in mid-mediolateral acoronal cross section at the level of the sec- position and are placed more closely to its ond molar. This cross section is sharper, how- inferior rim than in Theropithecus. ever, with the nasals forming a more acute In ARA-VP-6/933 and ARA-VP-6/437 angle than they do in T. oswaldi. This is clos- there is little to no development of maxillary er to the cross section of Macaca mulatta, ridges, similar to Parapapio (Freedman, M. nemestrina, or M. thibetana, but with a 1957), Theropithecus (Theropithecus) (sensu relatively longer muzzle. When viewed lat- Delson, 1993), and most Macaca, but dis- erally the muzzle pro®le is concave from gla- 2001 FROST: CERCOPITHECIDAE 5

Fig. 1. Continued. 6 AMERICAN MUSEUM NOVITATES NO. 3350

Fig. 2. Maxillae of Pliopapio alemui. Counterclockwise from top left: ARA-VP-1/1007 lateral view, ARA-VP-1/1723 (reversed), ARA-VP-6/437 lateral view (reversed), ARA-VP-6/437 frontal view, ARA- VP-1/1007 occlusal view, ARA-VP-1/1723 occlusal view. bella to rhinion, displaying an antorbital sal aperture at a margin of approximately 4 drop, and is also concave from rhinion to na- mm, as it does in most larger papionins. Un- sospinale, and ®nally convex from nasospi- like in T. gelada, it does not enter the piri- nale to prosthion. While the entire muzzle is form aperture. The nasal process of the pre- quite long and not unlike that of Papio, the maxilla projects further posteriorly than it length of the segment from glabella to rhi- does in Papio, approaching to within 1.5 cm nion comprises less of the total muzzle of the orbits. The premaxillomaxillary suture length than it does in Papio (see ®g. 4). Rhi- is therefore largely an anterolaterally nion is also considerably more prominent smoothly curving arc. The premaxillae pro- than in Papio or Theropithecus. ject relatively far anteriorly beyond the ca- The sutures of the muzzle are well pre- nine, and there is a modest diastema sepa- served on the left side of ARA-VP-6/933 and rating the canine from the incisors (6.5 mm on ARA-VP-6/437. The premaxillomaxillary on the left side of ARA-VP-6/933). suture follows the superior portion of the na- When viewed superiorly, the muzzle is 2001 FROST: CERCOPITHECIDAE 7

Fig. 3. The x-axis represents different papionin genera or subgenera as follows: Ccb, Cercocebus; Din, Papio (Dinopithecus); Lop, Lophocebus; Mac, Macaca; Man, Mandrillus; Omo, Theropithecus (Omopithecus); Pap, Papio (Papio); Pdo, Paradolichopithecus; Ppa, Parapapio; Pli, Pliopapio; Smo, Theropithecus oswaldi; The, T. gelada. Sample size for each (sub)genus is given in parentheses below each label. The y-axis plots values of centroid size computed from 45 cranial landmarks. Centroid sizes were computed using Morpheus (Slice, 1998). The central bar represents the median, or 50th percentile. The bottom and top of each box represent the value at the 25th and 75th percentiles, respectively, and the whiskers extend to the farthest observation that is less than 1.5 times the length of the box. Any individuals outside of the whisker range are marked separately. much narrower than the neurocranium (see narrower than that of Papio. The nasals of ®g. 5). In comparison to the length of the ARA-VP-6/933 are distorted, but probably neurocranium, the muzzle is longer than in would have formed a straight superior mar- most Macaca or Parapapio and shorter than gin. The premaxillae then bow gently later- in Papio (Papio) and Mandrillus (see ®g. 6). ally to the aperture's widest point just above The muzzle of the female ARA-VP-1/1007 the roots of the incisors, then curve convexly is considerably shorter than that of the male to meet at nasospinale in a relatively acute ARA-VP-6/933. When they are lined up at inferior angle. There is no evidence of ante- the ®rst molar, the mesial edge of the lateral rior nasal tubercles. incisor of ARA-VP-1/1007 is even with the The maxillary dental arcade is preserved middle of the canine of ARA-VP-6/933. in ARA-VP-6/933, and it is preserved from The piriform aperture is preserved in C to M3 in ARA-VP-1/1723 and from I2 to ARA-VP-6/933, partially preserved in ARA- M1 in ARA-VP-1/1007. The dental arcade is VP-6/437, and only a small portion is pre- somewhat distorted in ARA-VP-6/933, but served in the female ARA-VP-1/1007. The appears to have been largely U-shaped, with outline of the piriform aperture is typically the canines marking the bases of an anterior papionin, being roughly ovoid but forming a arc composed of the incisors. The alveolar V at its inferior pole. In breadth it is slightly margins appear to be gently bowed laterally, 8 AMERICAN MUSEUM NOVITATES NO. 3350

TABLE 1 Dental Measurements for Pliopapio alemui and Kuseracolobus aramisi 2001 FROST: CERCOPITHECIDAE 9

Fig. 4. Box and whisker plot of the ratio: length nasion±rhinion/length nasion±prosthion. Labeling is the same as in ®g. 3. with their widest point at the mesial loph of thecus (other than T. gelada and T. oswaldi M2 and narrowest at P3, bulging laterally leakeyi). ARA-VP-6/933 is the only speci- again at the canine, although less so in the men to preserve the zygomatic arches. The females ARA-VP-1/1007 and ARA-VP-1/ anterior surface of its zygoma curves grad- 1723. The molar series forms a short arc, ually and smoothly superoposteriorly, with with the M2 oriented slightly obliquely. The only a very slight depression in the region of premolars are set in a straight line from M1 the infraorbital foramina and maxillozygo- to C. matic suture. This depression is unrelated to When viewed laterally, the maxillary den- any maxillary fossae and is the only feature tition in ARA-VP-6/933 and ARA-VP-1/ to interrupt the otherwise smoothly curving 1723 is basically straight to very slightly surface. The inferior margin of the anterior concave, as in most cercopithecines. The pal- portion of the zygomatic arch is a smooth ate is preserved in ARA-VP-6/933, but it is semicircular curve interrupted by a small py- covered in matrix (which cannot be removed ramidal process where the maxillozygomatic without causing damage to the specimen). suture intersects. The superiormost point of ARA-VP-1/1723 preserves a small part of the inferior margin lies below the lateral edge the palatal process, which is about 0.5 cm in of the orbit, at which point the zygoma depth anteriorly and deepens slightly poste- curves inferiorly again. The temporal surfac- riorly. es do not appear strongly excavated as in Zygomatic Arch. The maxillary root of the Theropithecus, but there is some damage and zygomatic arch arises from above the distal distortion here. loph of M2 in the male ARA-VP-6/933 and In superior view, the zygomatic arches are above the mesial loph of M2 in the female no more laterally ¯ared than in most Macaca ARA-VP-1/1723. This is farther anterior or Papio, but are more smoothly curved. than in Papio, Gorgopithecus, and Theropi- This is particularly notable anteriorly where 10 AMERICAN MUSEUM NOVITATES NO. 3350

Fig. 5. Box and whisker plot of the ratio: width of muzzle at M1/M2 contact/biporionic width. Labeling is the same as in ®g. 3. they are more posteriorly angled than in Pa- slightly laterally ``drooping'' appearance. pio. The zygomata of the latter genus jut out There are no supraorbital notches. more sharply, perhaps due to greater maxil- The interorbital breadth is narrow, and gla- lary fossa development. The scar for the or- bella is not prominent. There is some damage igin of the masseter muscle is visible in in this area, but nasion was probably the an- ARA-VP-6/933 and terminates anteriorly teriormost point on the frontal. The orbits very close to the maxillozygomatic suture. themselves are largely mediolaterally oval in The posterior termination is not preserved, outline, being relatively short and broad. The but must have been anterior to the zygoma- lacrimomaxillary suture seems to lie just at ticotomporal suture as there is no scar on the the orbital rim, and the lacrimal fossa was zygomatic process of the temporal. likely contained entirely in the lacrimal bone. Orbital Region. The orbital region is only Calvaria. The calvaria is only preserved preserved in ARA-VP-6/933. Internally both in ARA-VP-6/933. It is relatively globular in orbits are occupied by matrix. The supraor- overall shape, with its greatest width at the bital torus is relatively prominent, but thin external auditory meatus. It is generally lack- superoinferiorly. It is mildly V-shaped in su- ing in superstructures and is considerably perior view and separated from the neurocra- broader than the muzzle. When viewed in nium by a broad ophryonic groove. Unlike Frankfurt horizontal, the frontal bone rises in Papio, T. oswaldi, and larger Macaca, above the supraorbital torus and achieves its there are no bulges above the torus at the maximum height about 1 cm anterior to midpoints of the orbits. In frontal view, the bregma. The cranial vault remains at this superior orbital rim and torus rise only very height until about 2.5±3 cm posterior to breg- slightly lateral to the sagittal plane, then ma. The temporal lines are faint and widely curve inferiorly, giving the torus a mildly su- separated, curving posteriorly less than 1 cm periorly bowed surface and the orbits a medial to the lateral orbital margins. Poste- 2001 FROST: CERCOPITHECIDAE 11

Fig. 6. Box and whisker plot of the ratio: length inion±nasion/length nasion±prosthion. Labeling is the same as in Fig. 3. rior to this, they remain subparallel, approx- to the sagittal plane and appear nearly round imating only slightly posteriorly. In conjunc- in cross section. tion with the weak temporalis development, Facial Hafting. The only specimen where postorbital constriction is slight and the tem- the relationship between the face and neu- poral fossae are narrow. The nuchal crests rocranium can be assessed is ARA-VP-6/ are slight to nonexistent at inion, but become 933. The glenoid fossa lies closely in line rather large laterally, having their greatest with the alveolar plane. The glenoid fossa is width behind the external auditory meatus. only slightly more elevated than in Papio, Viewed posteriorly, the vault is taller than but less so than in Theropithecus. Its position that of Theropithecus, which is broad and is not unlike that in Parapapio cf. jonesi low, but is similar to that of Papio or Ma- from Hadar (see Szalay and Delson, 1979: caca. 345). The face is less klinorhynch than that Basicranium. The basicranium of ARA- of Papio (Papio), but also less airorhynch VP-6/933 is largely covered in matrix, and than that of Theropithecus gelada. the foramen magnum is obscured by an ar- Mandible. ARA-VP-6/933 preserves most ticulated atlas. The occipital plane is proba- of the mandible, with considerable damage bly inclined at about 45Њ in Frankfurt orien- to the right side. ARA-VP-1/73 (see ®g. 7) tation. The mastoid processes do not appear is a male mandible with most of the corpora to be prominent. The postglenoid processes and symphysis intact. The inferior margin is may be closely approximated to the glenoid missing posterior to the symphysis. The com- fossae, but this is dif®cult to tell, and it is plete dentition is present other than the left impossible to see whether they are separated canine through right i1. ARA-VP-1/133 (see by a sulcus, as in T. oswaldi darti. The ex- ®g. 7) is a considerably distorted and crushed ternal auditory meatus are basically normal female mandible lacking the left ramus, but 12 AMERICAN MUSEUM NOVITATES NO. 3350

Fig. 7. Mandibles of Pliopapio alemui. Top row left to right: ARA-VP-1/73, ARA-VP-1/133, ARA- VP-1/1006 (two pieces). 2nd row: ARA-VP-1/563, ARA-VP-1/740, ARA-VP-1/548, ARA-VP-1/740, 2001 FROST: CERCOPITHECIDAE 13

Fig. 8. Pro®le views of symphyses of Pliopapio alemui, Parapapio ado from Laetoli, and P. ado from Kanapoi. Top: MB 1938±1 (cast, reversed), ARA-VP-1/563. Middle: NMT-LAET 74/322b (cast), ARA-VP-1/73. Bottom: KNM-KP-286 (reversed). preserving most of the right. The inferior venile mandible with most of the corpus and margins are largely intact, and the left and the right dp3-m1 and the left dc-m1. ARA- right p4-m3 are present. ARA-VP-1/1006 VP-1/548 (see ®g. 7) is a right juvenile cor- (see ®g. 7) preserves separate and partially pus with dp4 and m1 in place, and the tips crushed female right and left corpora with all of the crowns of i1-c just beginning to of the dentition other than the right central emerge from their crypts. incisor through the left canine. ARA-VP-1/ The symphysis slopes at an angle similar 563 (see ®g. 7) is a female symphysis with to that of Macaca fascicularis. This is more some of the left corpus and the dentition sloping than in many papionins when viewed from the right i1 through the left m2, and the in pro®le, but less so than the symphysis of right p3. ARA-VP-1/740 (see ®g. 7) is a ju- Parapapio ado from Laetoli (Leakey and

← ARA-VP-1/548 (reversed). 3rd row: ARA-VP-1/73, ARA-VP-1/563. 4th row: ARA-VP-1/133 left corpus (1349 is an old number), ARA-VP-1/1006 left corpus. Bottom row: ARA-VP-1/133 right corpus, ARA-VP-1/1006 right corpus. 14 AMERICAN MUSEUM NOVITATES NO. 3350

Delson, 1987) and Kanapoi (Patterson, Dentition. The incisors are fairly large rel- 1968), and considerably less so than in the ative to the molars, which is typical for most papionin from Lothagam (Leakey et al., in papionins. The I1 is broad, ¯aring, and spat- press). The incisive alveolar plane is oriented ulate in anterior view. The I2 is more asym- nearly vertically, whereas it projects more metrical and not as broad, with a small lateral anteriorly in the above-mentioned taxa. The tubercle. The lower incisors have straight incisor row is thus nearly vertical in Plio- mesial and distal borders in anterior view, so papio alemui, whereas the incisors of the that they are less ¯aring than the uppers. The others are more procumbent, with the central lateral border of the i2 is more laterally incisor projecting well beyond the lateral. curved than that of the i1. As is typical of The projecting alveolar process of P. ado cercopithecines, they lack lingual enamel and produces a symphysis that is quite different the labial surface stands out, yielding a in pro®le from that of Pl. alemui (see ®g. 8). somewhat chisel-like occlusal surface. The The symphysis is pierced by a median mental labial surface is often ``squared'' in occlusal foramen. There appear to have been very view. The canines are typical of cercopithe- faint, triangular mental ridges. The superior coids, being highly sexually dimorphic and transverse tori in ARA-VP-1/73 and ARA- having a mesial groove on the uppers that VP-1/133 extend posteriorly to the middle of extends onto the surface of the root. P4 in superior view. Both superior and in- The upper premolars are typical bicuspid ferior transverse tori are well developed. papionin teeth. The p3 is a highly sexually The Middle Awash mandibles show only dimorphic tooth. The paraconid is less well slight or no development of corpus fossae. developed than in the colobine, and the male Although there is some damage to the infe- mesiobuccal ¯ange is signi®cantly longer rior margin in ARA-VP-6/933, it appears that than that of the females. The p4 develops a the deepest point was relatively anterior, per- small mesiobuccal ¯ange in some males haps under p4, and that the inferior margin (e.g., ARA-VP-6/933), has more of a talonid curved gently convex down. The inferior than the p3, and has a fairly high lingual margin is thus anteriorly divergent. The notch. oblique line emerges near the level of the The molars in general are high-crowned mesial lophid of m3 or the distal lophid of for a papionin, with relatively little ¯are, es- m2. The extramolar sulcus is smooth and pecially in comparison to the Kanapoi sam- weakly developed. The gonial region is ple. Cusp relief above the lower lingual/up- unexpanded. If present at all, the mylohyoid per buccal notch is high for a papionin, but line is poorly developed. lower than in colobines. Accessory cuspules Viewed superiorly, the tooth rows are are often present in the lingual notch. In the nearly parallel along their lingual surfaces, upper molars, the lingual cusps are elevated from M3 to P3, with the canine slightly me- relative to the central basin and seem to be dial and the incisors curving sharply medi- connected by continuous, well-developed ally. In lateral view there is a normal curve postproto- and prehypocristae. The mesial of Spee (i.e., the tooth row is concave up- loph is wider than the distal. The M2 is often ward). the largest of the upper molars. The lower The ramus is well preserved only in ARA- molars have normal low relief and higher lin- VP-6/933. It is back-tilted, although less so gual notch. The buccal cusps tend to be fairly than in Papio but more so than in Macaca columnar, with the mesial and distal foveae or Theropithecus (Theropithecus). The cor- being pinched, although not to the extent of onoid process is even with or slightly higher those of Theropithecus. The ¯oor of the me- than the condyle, from which it is separated dian buccal notch seems to slope downward by a shallow semicircular mandibular notch. distally. On the m1±2, the distal cingula de- There is a deep triangular fossa below the velop a very small hypoconulid 6±10% of coronoid process on an otherwise relatively thc time depending on scoring. In the m3, smooth lateral surface. The masseteric tuber- the hypoconulid is generally tightly pressed osity is faint, and the whole area of its at- against the hypoconid, so that the distal buc- tachment is not heavily scarred. cal notch is very constricted compared to the 2001 FROST: CERCOPITHECIDAE 15 mesial. Additionally, the distal buccal notch ramen. This distinguishes the genus from rarely preserves any ``shelf'' at the base. both Rhinocolobus (at least as known from The dI2 has a crown that is basically spat- the Shungura and Koobi Fora Formations) ulate, low in height, broad, and angles me- and Cercopithecoides, which possess one. In sially. The root is broad and labiolingually lateral view, the symphysis is deep with a ¯attened. The dC is a mesiodistally elongate vertical pro®le. The corpus is quite deep and tooth, with a crown that is approximately tri- robust overall and it deepens posteriorly. In angular in labial view. The dc crown has a its depth and robusticity, the corpus is very prominent central cusp that is labiolingually different from the Leadu colobine, Procolo- compressed and a crest that extends mesially bus, Cercopithecoides, Semnopithecus, Py- from its apex. Distally there is a small ac- gathrix, and Presbytis. It is more like Colo- cessory cuspule. In general, the deciduous bus, Paracolobus, and Rhinocolobus, but is premolars are similar to adult molars, but not as deep as the corpora of the last two narrower, with more lateral ¯are, and genera. It is further differentiated from Par- loph(id)s that are more weakly developed acolobus and Rhinocolobus by the presence than the adult teeth. In addition, the upper of larger prominentia laterales, similar to dPs have relatively larger mesial and distal Nasalis (including N. (Simias)). The gonial foveae. The mesial fovea is particularly large region is expanded, separating it from the and elongate on the dP3. The dp3 protolo- Leadu colobine, Cercopithecoides, and Pro- phid is much narrower than the hypolophid. colobus, but far less so than in Paracolobus There is also a well-developed preprotocris- mutiwa. tid, and what may be a paraconid, yielding a In the dentition, the I1 crown does not mesial fovea that is triangular in shape. The ¯are in anterior view, so that the apex is not dp4 is more similar to an adult m1, but nar- signi®cantly wider than the base. This is dis- rower with a relatively longer mesial fovea. tinct from the ¯aring I1 in Procolobus. The P3 protocone is not reduced, as in Cercopi- SUBFAMILY COLOBINAE JERDON, thecoides, Colobus, and Rhinocolobus. The 1867 distal lophid of the m3 is typically narrower than the mesial, as is typical for the Pres- Kuseracolobus, new genus bytina and likely primitive for the Colobinae TYPE SPECIES: Kuseracolobus aramisi, new (Szalay and Delson, 1979). Three individuals species. out of 30 in the K. aramisi sample show hy- GENERIC DIAGNOSIS: A genus of colobine polophids that are wider than their protolo- monkey with a broad interorbital area, as is phids, and most are near to even. Most of the typical for colobines, but distinguishing it m3 hypolophids are wider relative to their from Libypithecus, Nasalis, and Rhinocolo- protolophids than those of Mesopithecus. bus (especially considering size). The projec- ETYMOLOGY: Named after the Middle tion of the lower face anterior to the zygo- Awash locality of Kuseralee Dora (``Place of matic arches in K. aramisi is generally sim- the Kusera tree'' in the Afar language). ilar, in proportion to overall cranial size, to that of Cercopithecoides, Mesopithecus, Tra- Kuseracolobus aramisi, new species chypithecus, and the undescribed colobine from Leadu (NME AL2±34, which is asso- SPECIFIC DIAGNOSIS: As for genus. ciated with a partial skeleton; see Delson, ETYMOLOGY: Named after the type site of 1994: 40). In comparison to Colobus, the Aramis. lower face of K. aramisi is less projecting, HOLOTYPE: ARA-VP-1/87. A nearly com- and it is distinctly less so than in the long- plete male mandibular corpus with the entire faced genera Paracolobus, Rhinocolobus, postcanine dentition except the right m1, as- Dolichopithecus, and Nasalis. In pro®le, the sociated with left and right C, left c, and a maxillary alveolar margin completely lacks left maxillary fragment with M1±3 found by the airorhynchous shape of Semnopithecus. A. Asfaw on December 18, 1992 (see ®g. 9). As is typical for most colobines, the man- HYPODIGM: See appendix 2. dibular symphysis lacks a median mental fo- HORIZON: All specimens are derived from 16 AMERICAN MUSEUM NOVITATES NO. 3350

Fig. 9. Holotype of Kuseracolobus aramisi ARA-VP-1/87. Occlusal view. Opposite page: lateral views, left side above, right (reversed) below. the Aramis Member of the Sagantole For- of the root of the zygoma, part of the lateral mation, between the GATC and DABT, and aspect of the face and piriform aperture, and are therefore dated to 4.4 Ma (Renne et al., most of the palatal process. A small fragment 1999). of the right maxilla with M1±2 and some of DESCRIPTION: The best cranial specimens, the palatal process is also preserved, along ARA-VP-6/1686 and KUS-VP-2/70 (see ®g. with the glabellar portion of the frontal and 10), are both females. ARA-VP-6/1686 pre- an isolated left I1. The holotype, ARA-VP- serves both maxillae and premaxillae below 1/87, preserves a left maxillary fragment the middle of the zygomatic process of the with M1±3 and the roots of P4, a small part maxilla, but little of the palate. The entire of the palatine process, and the very base of dentition is preserved other than the right I2 the zygomatic process (see ®g. 10). ARA- and M2. KUS-VP-2/70 preserves a left pre- VP-1/6 is a male left maxilla preserving P3- maxillomaxillary fragment with C-M1, most M3 and the root of C (see ®g. 10). It is highly 2001 FROST: CERCOPITHECIDAE 17

Fig. 9. Continued. 18 AMERICAN MUSEUM NOVITATES NO. 3350

Fig. 10. Maxillae of Kuseracolobus aramisi. Top row: KUS-VP-2/70, ARA-VP-6/1686. Middle row: ARA-VP-1/87, ARA-VP-1/6, ARA-VP-1/1686. Bottom row: KUS-VP-2/70, ARA-VP-1/1686 (reversed). 2001 FROST: CERCOPITHECIDAE 19 damaged, however, revealing the roots of the The maxillary dental arcade is best pre- teeth. served in ARA-VP-6/1686, and although dis- Kuseracolobus aramisi is larger in cran- torted bilaterally, it is reasonably intact on iodental size than Colobus, Libypithecus, and the left. KUS-VP-2/70 and ARA-VP-1/87 Mesopithecus, but smaller than Cercopithe- also preserve partial tooth rows. The dental coides (other than a new species from Loth- arcade forms a gentle arc from M3 to C, be- agam to be described by Leakey et al., in ing widest around the M1/M2 contact, with press), Paracolobus, Rhinocolobus, and Dol- no tooth deviating from this line. There is a ichopithecus. It is similar in size to Nasalis, sharp angle at the canines, and the incisors the colobine from Leadu, the smaller colo- form a relatively straight, ¯at arc between the bine at Hadar, and larger subspecies of Sem- canines. The palate is partially preserved in nopithecus. Colobines of similar dental size KUS-VP-2/70 and appears to have been fair- also occur in the Omo Shungura Formation ly shallow and ¯at. It is slightly deeper in the in Members B, C, D, and G, from the Tulu male ARA-VP-1/87, which preserves a small Bor Member at Koobi Fora, and from the part of the palatal process. Upper Laetolil beds. It is also similar in den- Mandible. The mandible is best preserved tal size to Pl. alemui. The dental dimensions in the male specimen ARA-VP-1/87, which of K. aramisi are given in table 1. retains much of the corpus and most of the Frontal. A small part of the glabellar area rami, although the margins and gonion are is preserved in KUS-VP-2/70 and ARA-VP- damaged and the condyles are lacking. Ex- 1/13. Both are similar in overall morphology cept for the right m1, the entire postcanine to the Leadu colobine. The ®rst specimen is dentition is preserved. ARA-VP-1/5 (see ®g. a female and has an interorbital breadth of 11) is the symphysis of a male with the left 12.1 mm; the second is of unknown sex and c-m1 and right p3-m1. ARA-VP-1/290 (see has an interorbital breadth of 10.8 mm, com- ®g. 11) is probably a subadult male symphy- pared with 13.8 mm in the male Leadu spec- sis with the left i1±2 and p4-m1 and the imen. Both Aramis specimens preserve na- crowns of the canines and p3s erupting. sals that are slightly pointed anterosuperiorly. ARA-VP-6/796 (see ®g. 11) is the symphysis Maxilla. The most complete specimens are of a female with the left m1 through right ARA-VP-6/1686 and KUS-VP-2/70, but the p4. ARA-VP-1/1774 (see ®g. 11) preserves male ARA-VP-1/87 also preserves a left the right corpus down the inferior margin un- maxillary fragment. The root of the zygoma der the m1±3. ARA-VP-1/564 (see ®g. 11) appears to be positioned above M1, or the preserves part of the corpus below the mo- M1/M2 contact in both male and female lars, but none of the margin. Other specimens specimens. This placement of the zygoma is preserve more fragmentary portions. slightly more anterior than that found in C. The symphysis is quite steeply sloping, williamsi from Koobi Fora and in most C. rather deep overall, and has a vertical pro®le. williamsi from South Africa. The piriform Both transverse tori are robust. The superior aperture is more vertically inclined and the transverse torus extends posteriorly to the rostrum shorter than in the male cercopithe- distal portion of p3, and is fairly steeply slop- cines ARA-VP-6/437 and ARA-VP-6/933. ing. Anteriorly, the symphysis lacks a me- The piriform aperture is quite narrow and dian mental foramen. tall, and the plane of its outline forms an an- A shallow fossa is present on the lateral gle slightly more than 60Њ with the alveolar surface of the corpus. This is largely due to margin in the subadult KUS-VP-2/70 and ap- the presence of lateral bulging near the in- proximately 45Њ in ARA-VP-6/1686. The inferior margin, which is the widest part of the ferior portion of the piriform aperture is corpus. This can be most clearly seen in the sharply V-shaped. In superior view, the pre- female ARA-VP-6/796, but also in ARA- maxillae form a squared-off rostrum. The VP-1/290. This morphology is unlike that of premaxillae would have been relatively short the Leadu colobine, AL231±1a (a specimen overall and generally similar in outline to from Hadar most likely to be the same taxon those of the Leadu colobine and KNM-ER as the Leadu colobine), and Cercopithecoi- 4420. des (particularly KNM-ER 4420), where the 20 AMERICAN MUSEUM NOVITATES NO. 3350

Fig. 11. Mandibles of Kuseracolobus aramisi. Top row: ARA-VP-1/7, ARA-VP-1/70, ARA-VP-1/ 564, ARA-VP-1/1774, 2nd row: ARA-VP-6/796, ARA-VP-1/5, ARA-VP-1/290. 3rd row: ARA-VP-6/ 796, ARA-VP-1/5, ARA-VP-1/290. Bottom row: ARA-VP-6/796, ARA-VP-1/1774 (reversed). 2001 FROST: CERCOPITHECIDAE 21

Fig. 12. Lateral views of mandibles from the Leadu colobine AL2±34 (above) and Kuseracolobus aramisi ARA-VP-1/87 (below). corpus is the widest at mid-height. The men- lobus. In ARA-VP-1/87 the right corpus tal foramen seems to be single in all of the deepens posteriorly from p3 to m3, although mandibles recovered, and its position varies there is damage below the m1 through mesial from the below the m1/p4 contact in ARA- m3. ARA-VP-1/1774 deepens from m1 to VP-1/1774 to the p3/p4 contact in the juve- m3 and preserves a bulge below the m2. The nile ARA-VP-1/290. The corpus is fairly gonial area is partially preserved in ARA- deep overall, especially compared with the VP-1/87 and is expanded, although not to the Leadu colobine (see ®g. 12), AL231±1a, or extent seen in Paracolobus mutiwa. This is Cercopithecoides from both East and South quite distinct from the comparatively unex- Africa. It is shallower but thicker than man- panded gonial area seen in the Leadu colo- dibles of Paracolobus mutiwa and Rhinoco- bine and Cercopithecoides. 22 AMERICAN MUSEUM NOVITATES NO. 3350

Viewed superiorly, there is a wide extra- nected by sharp cross-lophs. On the upper molar sulcus, and the oblique line blends into molars, the paraloph is broader than the hy- the corpus at mesial m2 or distal m1, com- poloph, but less so than in cercopithecines. parable to that in the Leadu colobine and The buccal notch has a ``crease'' reaching AL231±1a. There is no strongly marked toward the cervix from the buccal notch. A ridge at the anterior limit of the masseter distal ®fth cuspule is variably present on the scar. The mandibular dental arcade, although M3. The upper molars are all roughly similar slightly distorted, is best preserved in ARA- in overall size and generally arranged in a VP-1/87. It forms a parabolic arch, except straight line. The lower molars have very that the area across the incisors is ¯attened. deep lingual notches with high cusps. The Dentition. The dentition overall is typical distal cingulum of m1±2 forms a distal cus- for colobines. The upper incisors are smaller pule 6±8% of the time, depending on scor- and far less ¯aring than in papionins. The ing. On the m3, the hypoconulid is well de- mesial and distal margins of the I1 crown are veloped, and there is typically (62±92%, de- roughly parallel and slant mesially, and the pending on scoring) a tuberculum sextum as widest part of the crown is approximately at well. This contributes to the presence of a mid-height. Lingually, there is a cingulum well-developed distal lingual notch between around the base. The I2 is caniniform in the hypoconulid and the entoconid. There is crown shape and also has a lingual cingulum. also a well-developed distal buccal cleft. The The lower incisors possess lingual enamel metalophid is usually wider than the proto- and are small, peglike teeth compared with lophid on m1±2 but generally not on m3 (al- those of papionins. The i1 lingual surface is though it is occasionally). These lophid pro- shoveled, and the crown is slightly ¯aring in portions for K. aramisi are typical of Asian anterior view. The i2 has a crown that is nar- colobines (which may be the primitive state rower overall, is more of a parallelogram in for the subfamily; see Szalay and Delson, outline, and possesses a distal cuspule, or 1979) but different from extant African co- ``lateral prong'' (Delson, 1975). The canines lobines. are typical of cercopithecids, being compar- Of the deciduous dentition, the di2 is pos- atively large teeth and highly sexually di- sibly known in ARA-VP-1/2092. It looks morphic. like a miniature of the permanent i2, with a The upper premolars are typical bicuspid relatively narrow crown and a distal prong. teeth. The protocone of the P3 is usually pre- This is distinct from normal papionin mor- sent and often large, but is sometimes re- phology, where the lower dIs are rather duced. The P3 is generally more triangular broad. Of the deciduous premolars, upper in occlusal outline, and the P4 often has a bit and lower dP4s are known. The dP4 is much more of a talon. The p3 is sexually dimorphic like the M1, but far more ¯aring, with more as for most cercopithecids, particularly in the approximated cusps. The mesial and distal development of the mesiobuccal ¯ange. It is foveae are relatively longer than in the mo- also typical of colobines in that the paraconid lars. The dp4 is similar to the m1, but is nar- is generally more pronounced than in cer- rower relative to its length, with a metalo- copithecines. The male mesiobuccal ¯ange is phid that is wider in comparison to the pro- shorter than those of male cercopithecines tolophid than in m1. The lophids are not (although longer than female cercopitheci- quite as well developed as those of the mo- nes), is more inferiorly directed, and the tal- lars, but are better developed than those of onid extends more lingually. The p4 is a the deciduous premolars of cercopithecines. more molariform tooth, but may develop a mesiobuccal ¯ange in males. There is also a BODY WEIGHT AND SEXUAL DIMORPHISM greater amount of cusp relief (i.e., the differ- ence between the height of the cusps and the Based on the available dental material that lowest points of the crown between them) is assignable to sex, body weight has been than is the case in cercopithecines. estimated here for the Aramis cercopithecids The molar crowns are only slightly ¯aring utilizing the methods and prediction equa- with tall, widely spaced cusps that are con- tions of Delson et al. (2000). In their study, 2001 FROST: CERCOPITHECIDAE 23

TABLE 2 Distribution of Characters in Various Cercopithecid Generaa

subfamily-speci®c equations for each sex 12 kg (range 10±15 kg) and 8.5 kg (7±10 were derived based on a large cercopithecid kg), respectively. These estimates yield a ra- sample with associated or literature-compiled tio of male-to-female body mass of 1.4:1. body weights. These equations were used This value is similar to that seen in such ex- here on the Aramis dental material that is tant species as Lophocebus albigena, Maca- assignable to sex and subfamily. The results ca sylvanus, M. nemestrina leonina, and Cer- are based on the means of several estimates copithecus aethiops pygerythrus. M. sylvanus based on equations for dental measurements is similar in body weight to Pl. alemui, found to be most accurate at predicting body whereas the others are smaller. weights of known samples (for a thorough Body weight estimates have been made for discussion, see Delson et al., 2000). Males Kuseracolobus aramisi in the same manner. and females of Pliopapio alemui are esti- Males and females of this population are es- mated to have had a mean body weight of timated to have had mean body weights of 24 AMERICAN MUSEUM NOVITATES NO. 3350 approximately 18 kg (range 14±22 kg) and Parapapio aff. ado and an unnamed colobine 12 kg (10±14 kg), respectively. This yields a in their faunal list based on more extensive ratio of 1.5:1 for male-to-female body recent collections. The Kanapoi mandible ®ts weight. This level of dimorphism is similar well into the variation seen at Laetoli. KNM- to that of most subspecies of Semnopithecus KP 286 possesses the distinctive symphyseal entellus, Pygathrix (Rhinopithecus) roxel- morphology of the Laetoli material, and the lana, and Procolobus badius oustaleti. Of molars show an even greater level of ¯are these taxa, only S. entellus approaches K. ar- than does the Laetoli material. There are also amisi in weight (Delson et al., 2000). two specimens of a small colobine of unknown af®nity (Leakey et al., 1995). The site DISCUSSION of Ekora is probably similar in age to Kan- apoi and has produced a single broken cer- The Middle Awash sample described here copithecine lower molar, KNM-KP 287, sim- ®lls a temporal gap between the latest Mio- ilar to the papionin from Kanapoi in size and cene and oldest known Pliocene localities. morphology (Leakey and Leakey, 1976). Sites of broadly similar age to Aramis with From farther north in the Turkana basin fossil cercopithecids are relatively few in there is a small papionin mandible, KNM- sub-Saharan Africa, and none of them con- WT 16752, from the Nachukui Formation tains specimens that can be unambiguously west of Lake Turkana. It has been assigned allocated to either of the Aramis taxa. The to cf. Parapapio aff. ado by Harris et al. Early Pliocene (3.7±3.5 Ma) Upper Laetolil (1988). This mandible is from just above the beds in Northern Tanzania have produced a Tulu Bor tuff, and thus dates to ca. 3.4 Ma. sizable sample in which four species of cer- It seems to lack the symphyseal morphology copithecids are represented. Included in this of P. ado. Fragmentary remains of a larger collection is a small papionin larger in dental papionin, allocated to Parapapio whitei size than Pl. alemui. This sample includes based on size and a lack of mandibular cor- the type specimen of Cercocebus ado (Hop- pus fossae, have also been recovered from wood, 1936). Leakey and Delson (1987) re- this level, but they are substantially larger assigned it to the genus Parapapio (which is than Pl. alemui. At Koobi Fora there is a best diagnosed on facial morphology, absent small mandibular fragment with the second at Laetoli), as P. ado, largely based on an and third molars, KNM-ER 3122, from the absence of mandibular corpus fossae. There Tulu Bor Member near the Toroto Tuff dated are two specimens of a larger papionin also to about 3.3 Ma. It is from a papionin of present. A colobine tentatively assigned to an similar size to those from Aramis and Kan- unnamed species of Paracolobus is the most apoi (personal obs.). Also from the Tulu Bor abundant colobine at Laetoli. A smaller co- Member, and thus between 3.4 and 2.68 Ma, lobine, similar in size to K. aramisi, is pre- is KNM-ER 3041, a colobine mandibular sent but represented by only a few isolated fragment with m1±3 that is similar in dental teeth. size to K. aramisi (personal obs.). The site of Kanapoi is closer in age to Ar- The Hadar Formation is downstream from amis at approximately 4.1 Ma and has pro- Aramis in the Awash Valley. In the Middle duced a papionin mandible, KNM-KP 286, Pliocene portion of the formation (spanning collected in 1966 and described as Parapa- 3.4±2.9 Ma) four species of Cercopithecidae pio jonesi on the basis of molar dimensions are known. Theropithecus oswaldi darti is by Patterson (1968). Leakey and Leakey the most common species (Eck, 1993). A (1976) reassigned this specimen to the same species of Parapapio is also present, termed species as the small papionin from Laetoli, P. cf. jonesi by Delson (Szalay and Delson, Cercocebus ado, based on the mandibular 1979; Delson, 1984). This taxon is near in morphology. This specimen was tentatively size to Pliopapio alemui, but it possesses the moved to the genus Parapapio along with distinctive cranial features of Parapapio. the Laetoli material by Leakey and Delson Two colobines are also present. One is a (1987) and was thus designated cf. Parapa- large colobine, tentatively assigned to Rhin- pio aff. ado. Leakey et al. (1995) listed cf. ocolobus turkanaensis (Delson, 1994) and 2001 FROST: CERCOPITHECIDAE 25 clearly different from K. aramisi. A smaller Pliopapio alemui, Parapapio ado, Parapa- colobine, similar in size to K. aramisi, is also pio jonesi (all of which are broadly similar present. Based on a nearly complete mandi- in size and molar morphology), or any other ble (AL231±1a) and a partial humerus well represented taxon. The mandible from (AL222±14), this taxon is likely to be the the Nachukui Formation lacks the diagnostic same species as the Leadu colobine (AL2± symphyseal morphology of P. ado but is too 34 and the associated partial skeleton). This fragmentary to allocate to Pliopapio based taxon is quite different from K. aramisi in its on anything other than size. mandibular morphology, as discussed above. Colobines similar in size to K. aramisi Additionally, the Leadu colobine was prob- also occur at Hadar in the Sidi Hakoma ably fairly terrestrial in its locomotor behav- Member, Leadu, the Omo Shungura Forma- ior (see Delson, 1994), much more so than tion in Members B, C, D, and G, the Tulu is tentatively suggested by any of the post- Bor Member at Koobi Fora, and the Upper crania found at Aramis. Laetolil beds (Delson, 1994; Leakey, 1987; There is an isolated lower third molar from Leakey and Delson, 1987). The colobine Unit 2 of the Chiwondo Formation, which from Leadu and Hadar is clearly different has been biostratigraphically dated to be- from K. aramisi. All of the others are essen- tween 3.5 and 4.5 Ma (Kaufulu et al., 1981). tially represented by isolated dental remains This tooth is from a papionin similar in size and thus lack the morphological features that to Pl. alemui and the Kanapoi papionin distinguish among K. aramisi, the Leadu co- (Frost and Kullmer, in prep.). In South Af- lobine, and other colobine taxa. rica, there is only the earliest Pliocene or One possible explanation for the absence Late Miocene Quarry E at Langebaanweg, of the Aramis taxa at other sites is that these from which only two isolated cercopithecine taxa were endemic to northern Ethiopia. Giv- premolars have been recovered (Grine and en the lack of sites of similar age to Aramis, Hendey, 1981). let alone sites that sample a similar habitat, North of the Sahara, the Mio-Pliocene a more likely hypothesis for their absence in sites of Sahabi and Wadi Natrun have each the fossil record outside Aramis is sampling. produced a single species of colobine and WoldeGabriel et al. (1994) discussed the en- cercopithecine (Szalay and Delson, 1979; vironment at Aramis and found that faunally Meikle, 1987). The cercopithecines at both it was most consistent with a closed habitat. sites are represented by isolated teeth placed They noted the high abundance of primates, in the genus Macaca largely based on ge- particularly colobines. Other East African ography (Delson, 1975). The colobine, Li- Pliocene sites have high proportions of cer- bypithecus markgra®, is represented by the copithecids, such as the Shungura Formation relatively complete holotype cranium at the (BobeÂ, 1997). Aramis, however, is different latter site and a few teeth and postcrania at from all other East African Pliocene sights the former (Delson, 1994). Libypithecus has in the very high abundance of colobines rel- also been listed from the Middle Awash, but ative to cercopithecines (56% colobines at from lower in section (Kalb et al., 1982). Aramis vs., e.g., 2±11% for different mem- This diagnosis was based on a mandible from bers of the Shungura Formation; see Eck, the Kuseralee Member (older than Aramis 1976). The cercopithecid sample at Kanapoi, and currently under analysis by Y. Haile-Se- which is most similar in age to Aramis, is lassie), which probably represents a taxon dominated by papionins, with very few co- other than Libypithecus. lobines present (Leakey et al., 1995). The Neither of the taxa described in this paper only African site that approaches Aramis in can be positively recognized outside of the this respect is Laetoli, where colobines rep- Middle Awash. P. ado is recognized at Lae- resent 36% of the identi®able cercopithecids. toli and Kanapoi, but this species is quite dis- At Laetoli, however, cercopithecids make up tinct from Pl. alemui. The fossils from the only a small proportion of the total fauna Chiwondo beds in Malawi and from Koobi (Leakey and Delson, 1987). Thus, in its com- Fora are clearly small papionins, but they are bination of abundant cercopithecids, and too fragmentary to be diagnosed as either most of those cercopithecids being colobines, 26 AMERICAN MUSEUM NOVITATES NO. 3350

Aramis is unique, which may suggest a hab- rican ancestral morphotype would then be itat different from other known localities. more Macaca-like, but with a longer rostrum, as in Pl. alemui. Thus, the anteorbital PHYLOGENETIC RELATIONSHIPS drop present in Papio, Theropithecus, Man- drillus, Pliopapio, and possibly Gorgopithe- The phylogenetic positions of the taxa de- cus would be primitive for the subtribe. scribed here have not yet been completely Since Parapapio has been identi®ed based analyzed, as this task requires an extensive on facial material at the late Miocene site of analysis of the family. However, it is possible Lothagam (Leakey et al., in press), this al- to offer hypotheses based on the evidence ternative would require that the Pliopapio presented above. The Asian and North Afri- lineage extend back at least as far. can genus Macaca is recognized by most The Colobinae are generally divided into modern studies as the sister taxon to a mono- two subtribes, the Colobina in Africa and the phyletic sub-Saharan African papionin clade Presbytina in Asia (Szalay and Delson, (Szalay and Delson, 1979; Harris and Diso- 1979). The only cranial features typically tell, 1998; Tosi, et al., 2000). Among the ex- used to diagnose the two subtribes are that tant genera, there appear to be two main the P3 has a reduced protocone and that the clades, one consisting of Mandrillus and m3 has a distal lophid that is wider than the Cercocebus and the other of Papio, Thero- mesial lophid in the Colobina, whereas the pithecus, and Lophocebus (Harris and Diso- Presbytina possess a larger P3 protocone and tell, 1998; Fleagle and McGraw, 1999). The narrower m3 distal lophid. For both of these suprageneric relationship of the fossil forms, features, the derived state is that of the Af- however, remains controversial (Delson, rican forms. How the many extinct African 2000). Within the sub-Saharan African pa- forms ®t into this framework is dif®cult to pionins there are four main groups, which discern. As stated above, these features are may represent clades. The ®rst group, Par- both intermediate in Kuseracolobus aramisi, apapio, is often regarded as the most primi- with the P3 protocone being variably reduced tive genus and as a potential ancestral form and the m3 distal lophid being approximately from which the other genera radiated (Szalay equal in width to the mesial. This may in- and Delson, 1979). No fossils known to date dicate that K. aramisi is partially evolved to- can be unambiguously identi®ed as belong- ward the derived state for these features and ing to the second group, the Mandrillus/Cer- therefore may indicate some af®nity with the cocebus clade. The third group consists of African clade. This is a hypothesis that clear- Papio, Theropithecus, and possibly Gorgo- ly needs more thorough analysis. pithecus. Lophocebus forms the fourth Within the larger fossil colobines from Af- group, as its position relative to Papio, Ther- rica there appear to be two groups. One of opithecus, and Gorgopithecus is ambiguous. these is a long-faced group composed of Par- In relation to these groups, two possible acolobus and Rhinocolobus (but not neces- phylogenetic positions seem likely for Plio- sarily Libypithecus). In addition to large size papio alemui. One option is that anteorbital and long faces, this group is further diag- drop may be a synapomorphy uniting Plio- nosed by having mandibular corpora that are papio with the Papio/Theropithecus clade to deep and narrow in breadth, that deepen pos- the exclusion of Parapapio. Within this teriorly, that are ¯ared at gonion (except less group, however, Pliopapio is clearly distinct in P. chemeroni, but greatly in P. mutiwa), from the former genera, as each possesses and that have tall rami. This group is further derived features it lacks. A second, although distinguished from the second group by be- less likely option, is that Pl. alemui may be ing generally more arboreal. The second a primitive member of its tribe, either an an- group is a short-faced and terrestrial group cestor to or sister taxon of the remaining gen- composed of East African Cercopithecoides era. This position requires that the rostral and possibly the Leadu colobine. This sec- pro®le and facial hafting of Parapapio be au- ond group, along with South African Cer- tapomorphies, derived relative to the primi- copithecoides, shares a mandibular morphol- tive African papionin morphotype. The Af- ogy that includes a shallow but very broad 2001 FROST: CERCOPITHECIDAE 27 corpus, a shallow and vertical symphysis, to T. D. White and the Middle Awash Re- and a short but deep ramus with a relatively search Project for their hard work in collec- unexpanded gonion. Unfortunately, K. ar- tion and preparation of the material (partic- amisi does not ®t well into either group. The ularly for the weeks of work spent in prep- diagnostic features and geographic distribu- aration of ARA-VP-6/933) that made this tions of all Afar and Turkana Basin taxa are study possible. I thank T. D. White, D. discussed in more detail in Frost (2001). DeGusta, and Y. Haile-Selassie for their help in many ways during my study of this ma- SUMMARY terial, including some of the identi®cations. T. D. White, D. DeGusta, Y. Haile-Selassie, A sample of 925 fossil cercopithecids has K. McNulty, E. Delson, and M. G. Leakey been collected from the GATC±DABT inter- reviewed previous drafts of this paper, and val of the Sagantole Formation of Ethiopia. due to their comments it is much improved. This is the largest collection from the Early I am grateful to the Kenya National Muse- Pliocene of Africa described to date, and ums for access to specimens in their care, contains two new species, each in a new ge- and special thanks to M. G. Leakey for all nus. Pliopapio alemui is distinct from other her help at the KNM and for kind access to known papionin genera in its cranial mor- the Lothagam and Kanapoi material and to a phology. This species is similar in dental size pre-publication draft of her manuscript for to Parapapio ado from Laetoli and Kanapoi, the Lothagam cercopithecids. I thank E. Del- but is distinct from these samples in its man- son for the use of photographs by Chester dibular morphology. Kuseracolobus aramisi Tarka of ARA-VP-1/87, ARA-VP-1/133, is similar in overall size to a medium-sized ARA-VP-1/564, and ARA-VP-1/1006. Spe- colobine from Leadu, near Hadar. K. aramisi cial thanks to Henry Gilbert for photographs is different in its mandibular morphology of ARA-VP-1/5, ARA-VP-1/133, ARA-VP- from this latter species. Other colobine teeth 6/1686, and ARA-VP 6/679. I also thank of similar size occur across East Africa, but Chester Tarka and Lorraine Meeker for ad- their af®nity is impossible to determine given vice and help with the ®gures; any remaining the existence of at least two similarly sized defects are entirely my doing. E. Delson has species. None of the available morphology my particular gratitude for his knowledge- aligns this species with any other known co- able guidance, comments, advice, and sup- lobines. Neither of these species can be de- port, and for unlimited access to his collec- ®nitively identi®ed outside of the Middle tion of fossil cercopithecid casts. I thank the Awash. This is probably due to a lack of sites Division of Paleontology of the American of similar age and environment to Aramis. Museum of Natural History and its Chair- The phylogenetic placements of both species man, Mark Norell, for access to space and are currently unresolved, but Pl. alemui may facilities and for accepting this research for be more closely related to Papio and Ther- publication in Novitates; Meng Jin and Rob opithecus than to Parapapio. K. aramisi does Voss of the AMNH Publications Committee not clearly ®t within the Colobina or Pres- and two reviewers are thanked for their help- bytina. It is also not clearly allied with either ful comments on earlier versions of the man- the long-faced fossil genera or with the short- uscript. er-faced (semi-)terrestrial forms. This research was supported by the Wen- ner-Gren Foundation (Gr.6436), the L.S.B. ACKNOWLEDGMENTS Leakey Foundation, a NYCEP fellowship I thank the Ethiopian Centre for Research (NSF BIR 9602234), and a predoctoral fel- and Conservation of Cultural Heritage, the lowship from the Smithsonian Institution, Ethiopian Ministry of Information and Cul- Evolution of Terrestrial Ecosystems Consor- ture, and the National Museum of Ethiopia tium. I gratefully acknowledge these sources for permission to study the material in their here. Additionally, I acknowledge the L.S.B. care. My thanks also go to Ato Alemu Ade- Leakey Foundation for Research into Human massu for all of his kind help in the Paleo- Origins and the PSC-CUNY Faculty Re- anthropology lab at the NME. I am grateful search Award Program (grant 6±68479) for 28 AMERICAN MUSEUM NOVITATES NO. 3350 support of Eric Delson's research on Middle scaling in extinct and extant taxa. An- Awash fossil cercopithecids in 1997, which thropol. Pap. Am. Mus. Nat. Hist. 83: permitted me to accompany him on a short 160 pp. visit to the NME to begin study of these fos- Eck, G. G. sils. 1976. Cercopithecoidea from Omo group de- posits. In Y. Coppens, F. C. Howell, G. L. Isaac, and R.E.F. Leakey (eds.), Ear- REFERENCES liest man and environments in the Lake BobeÂ, R. Rudolf Basin: 332±344. Chicago: Univ. 1997. Hominid environments in the Pliocene: Chicago Press. an analysis of fossil mammals from the 1977. Diversity and frequency distribution of Omo Valley, Ethiopia. Ph.D. diss., Omo group Cercopithecoidea. J. Hum. Univ. Washington, Seattle. Evol. 6: 55±63. Delson, E. 1993. Theropithecus darti from the Hadar 1975. Evolutionary history of the Cercopithe- Formation, Ethiopia. In N. G. Jablonski cidae. In F.S. Szalay (ed.), Approaches (ed.), Theropithecus: the rise and fall of to primate paleobiology. Contrib. Pri- a primate genus: 15±76. Cambridge: matol. 5: 167±217. Basel: S. Karger. Cambridge Univ. Press. 1984. Cercopithecid biochronology of the Af- Eck, G. G., and N. G. Jablonski rican Plio-Pleistocene: correlation 1987. The skull of Theropithecus brumpti among eastern and southern hominid- compared with those of other species of bearing localities. Cour. Forschung- the genus Theropithecus. In Les faunas sinst. Senckenb. 69: 199±218. Plio-Pleistocenes de la Basse Vallee de 1988. Chronology of South African australo- l'Omo (Ethiopie). Tome 3, Cercopithe- pith site units. In F. E. Grine (ed.), Evo- cidae de la Formation de Shungura: lutionary history of the ``robust'' aus- 11±122. Cahiers de Paleontologie, Tra- tralopithecines: 317±324. New York: vaux de Paleontologie Est-Africaine. Aldine. Paris: Editions du CNRS. 1993. Theropithecus fossils from Africa and Fleagle, J. G., and W. S. McGraw India and the taxonomy of the genus. 1999. Skeletal and dental morphology sup- In N. G. Jablonski (ed.), Theropithecus: ports diphyletic origin of baboons and the rise and fall of a primate genus: mandrills. Proc. Nat. Acad. Sci. 96(3): 157±189. Cambridge: Cambridge Univ. 1157±1161. Press. Freedman, L. 1994. Evolutionary history of the colobine 1957. Fossil Cercopithecoidea of South Afri- monkeys in paleoenvironmental per- ca. Ann. Transvaal Mus. 23: 121±262. spective. In A. G. Davies and J. F. Frost, S. R. Oates (eds.), Colobine monkeys: their 2001. Fossil Cercopithecidae of the Afar de- ecology, behavior and evolution: 11± pression, Ethiopia: Species systematics 43. Cambridge: Cambridge Univ. and comparison to the Turkana Basin. Press. Ph.D. diss., City Univ., New York. 2000. Cercopithecinae. In E. Delson, I. Tat- Frost, S. R., and O. Kullmer tersall, J. A. Van Couvering, and A. S. In prep. Fossil Cercopithecidae from the Chi- Brooks (eds.), Encyclopedia of human wondo Beds, Malawi. evolution and prehistory, 2nd ed.: 166± Grine, F. E., and Q. B. Hendey 171. New York: Garland. 1981. Earliest primate remains from South Delson, E., and D. Dean Africa. S. Afr. J. Sci. 77: 374±376. 1993. Are Papio baringensis R. Leakey, Harris, E. E., and T. R. Disotell 1969, and P. quadratirostris Iwamoto, 1998. Nuclear gene trees and the phylogenet- 1982, species of Papio or Theropithe- ic relationships of the mangabeys (Pri- cus? In N. G. Jablonski (ed.), Thero- mates: Papionini). Mol. Biol. Evol. 15: pithecus: the rise and fall of a primate 892±900. genus: 125±156. Cambridge: Cam- Harris, J. M., F. H. Brown, and M. G. Leakey bridge Univ. Press. 1988. Stratigraphy and paleontology of Plio- Delson, E., C. J. Terranova, W. L. Jungers, E. J. cene and Pleistocene localities west of Sargis, N. G. Jablonski, and P. C. Dechow Lake Turkana, Kenya. Nat. Hist. Mus. 2000. Body mass in Cercopithecidae (Pri- Los Angeles Cty. Contrib. Sci. 399: 1± mates, Mammalia): estimation and 128. 2001 FROST: CERCOPITHECIDAE 29

Kalb, J. E., E. B. Oswald, A. Mebrate, S. Tebedge, Lothagam: Dawn of humanity in east- and C. J. Jolly ern Africa. New York: Columbia Uni- 1982. Stratigraphy of the Awash Group, Mid- versity Press. dle Awash Valley, Afar, Ethiopia. Meikle, E. Newsl. Stratigr. 11(3): 95±127. 1987. Fossil Cercopithecidae from the Sahabi Kaufulu, Z. M., E. S. Vrba, and T. D. White Formation. In N. T. Boaz, A. El-Arnou- 1981. Age of the Chiwondo Beds, northern ti, A. W. Gaziry, J. de Heinzelin, and Malawi. Ann. Transvaal Mus. 33: 1±8. D. D. Boaz (eds.), Neogene paleontol- Leakey, M. G. ogy and geology of Sahabi: 119±127. 1987. Colobinae (Mammalia, Primates) from New York: Alan R. Liss. the Omo Valley, Ethiopia. In Les fau- Patterson, B. nas Plio-Pleistocenes de la Basse Vallee 1968. The extinct baboon, Parapapio jonesi, de l'Omo (Ethiopie). Tome 3, Cerco- in the Early Pleistocene of northwest pithecidae de la Formation de Shun- Kenya. Breviora 282: 1±4. gura: 148±169. Cahiers de Paleontolo- Renne, P. R., G. WoldeGabriel, W. K. Hart, G. gie, Travaux de Paleontologie Est-Af- Heiken, and T. D. White ricaine. Paris: Editions du CNRS. 1999. Chronostratigraphy of the Mio-Plio- 1993. Evolution of Theropithecus in the Tur- cene Sagantole Formation, Middle kana Basin. In N. G. Jablonski (ed.), Awash Valley, Afar Rift, Ethiopia. Theropithecus: the rise and fall of a pri- Geol. Soc. Am. Bull. 111(6): 869±885. mate genus: 85±123. Cambridge: Cam- Szalay, F. S., and E. Delson bridge Univ. Press. 1979. Evolutionary history of the primates. Leakey, M. G., and E. Delson New York: Academic Press. 1987. Fossil Cercopithecidae from the Lae- Tosi, A. J., J. C. Morales, and D. J. Melnick tolil Beds, Tanzania. In M. D. Leakey 2000. Comparison of Y chromosome and and J. M. Harris (eds.), The Pliocene mtDNA phyolgenies leads to unique in- site of Laetoli, northern Tanzania: 91± ferences of macaque evolutionary his- 107. New York: Oxford Univ. Press. tory. Mol. Phylogenet. Evol. 17(2): Leakey, M. G., and R.E.F. Leakey 133±144. 1976. Further Cercopithecinae (Mammalia, White, T. D., G. Suwa, and B. Asfaw Primates) from the Plio/Pleistocene of 1994. Australopithecus ramidus, a new spe- East Africa. Fossil Verteb. Africa 4: cies of early hominid from Aramis, 121±146. Ethiopia. Nature 371: 306±312. Leakey, M. G., C. S. Feibel, I. McDougal, and A. 1995. Corrigendum: Australopithecus rami- Walker dus a new species of early hominid 1995. New four-million-year-old hominid from Aramis, Ethiopia. Nature 375: 88. species from Kanapoi and Allia Bay, WoldeGabriel, G., T. D. White, G. Suwa, P. Ren- Kenya. Nature 376: 565±571. ne, J. de Heinzelin, W. K. Hart, and G. Heiken Leakey, M. G., M. Teaford, and C. V. Ward 1994. Ecological placement of early Pliocene In press. Cercopithecidae from Lothagam. In M. hominids at Aramis, Ethiopia. Nature G. Leakey and J. M. Harris (eds)., 371: 330±333. 30 AMERICAN MUSEUM NOVITATES NO. 3350 Pl. alemui APPENDIX 1 List of All Specimens Assigned to For each specimen, the catalog number, sex (if diagnosable), side, and element are listed. The catalog number includes the collection locality ahead of therespectively. slash; Associated thus, elements ARA-VP-1/73 share a is catalog from number. locality Aramis 1. The abbreviations KUS and SAG stand for Kuseralee Dora and Sagantole, 2001 FROST: CERCOPITHECIDAE 31 ) Continued APPENDIX 1Ð( 32 AMERICAN MUSEUM NOVITATES NO. 3350 ) Continued APPENDIX 1Ð( 2001 FROST: CERCOPITHECIDAE 33 K. aramisi APPENDIX 2 Fields are as described for Appendix 1. List of All Specimens Assigned to 34 AMERICAN MUSEUM NOVITATES NO. 3350 ) Continued APPENDIX 2Ð( 2001 FROST: CERCOPITHECIDAE 35 ) Continued APPENDIX 2Ð( 36 AMERICAN MUSEUM NOVITATES NO. 3350 ) Continued APPENDIX 2Ð(