Miocene Cercopithecoidea from the Tugen Hills, Kenya

lable at ScienceDirect

Journal of Human Evolution 59 (2010) 465e483

Contents lists avai

Journal of Human Evolution

journal homepage: www.elsevier .com/locate/ jhevol

Miocene Cercopithecoidea from the Tugen Hills, Kenya

Christopher C. Gilbert a,b,*, Emily D. Goble a, Andrew Hill a

aDepartment of Anthropology, Yale University, P.O. Box 208277, New Haven, CT 06520-8277, USAb Yale Institute for Biospheric Studies, Environmental Science Center, 21 Sachem Street, P.O. Box 208105, New Haven, CT 06520-8105, USA

a r t i c l e i n f o

Article history:Received 5 November 2009Accepted 1 May 2010

Keywords:ColobineCercopithecineVictoriapithecusDatingPapioninCraniaPostcrania

* Corresponding author.E-mail addresses: [email protected] (C.

edu (E.D. Goble), [email protected] (A. Hill).

0047-2484/$ e see front matter � 2010 Elsevier Ltd.doi:10.1016/j.jhevol.2010.05.005

a b s t r a c t

Miocene to Pleistocene fossiliferous sediments in the Tugen Hills span the time period from at least15.5 Ma to 0.25 Ma, including time periods unknown or little known elsewhere in Africa. Consequently,the Tugen Hills deposits hold the potential to inform us about crucial phylogenetic events in Africanfaunal evolution and about long-term environmental change. Among the specimens collected from thisregion are a number of discoveries already important to the understanding of primate evolution. Here,we describe additional cercopithecoid material from the Miocene deposits in the Tugen Hills sequence,including those from securely dated sites in the Muruyur Beds (16e13.4 Ma), the Mpesida Beds(7e6.2 Ma) and the Lukeino Formation (w6.2e5.7 Ma). We also evaluate previously described materialfrom the Ngorora Formation (13e8.8 Ma). Identified taxa include Victoriapithecidae gen. et sp. indet., cf.Parapapio lothagamensis, and at least two colobines. Specimens attributed to cf. Pp. lothagamensis wouldextend the species’ geographic range beyond its type locality. In addition, we describe specimens sharingderived characters with modern African colobines (Tribe: Colobina), a finding that is congruent withprevious molecular estimates of colobine divergence dates. These colobine specimens represent some ofthe earliest known members of the modern African colobine radiation and, in contrast to previoushypotheses, suggest that early African colobines were mainly arboreal and that semi-terrestrial LateMiocene and Plio-Pleistocene colobine taxa were secondarily derived in their locomotor adaptations.

� 2010 Elsevier Ltd. All rights reserved.

Introduction

The modern cercopithecoid subfamilies, the Cercopithecinaeand the Colobinae, are estimated to have diverged in Africa some-time during the Middle Miocene (e.g., Raaum et al., 2005; Ting,2008). The major tribes within each subfamily are estimated tohave diverged soon after, between 12.3 and 8.2 Ma in the Middle toLate Miocene (Raaum et al., 2005; Ting, 2008). Unfortunately, thereare very few known terrestrial fossil deposits in Africa during thesecritical time periods for cercopithecoid evolution (Hill, 1999). Assuch, the origins of the major living cercopithecoid groups remainpoorly documented.

The Tugen Hills region in the northern Kenyan Rift Valley, westof Lake Baringo, is one of the few fossiliferous areas that span thetime period in question, and it has provided a wealth of fossils overthe past four decades or so, from deposits spanning the Miocenethrough the Pleistocene (Hill, 2002). A number of important fossil

C. Gilbert), emily.goble@yale.

All rights reserved.

primates have been recovered from this area, including the earliestspecimen attributed to the genus Homo (Hill et al., 1992; Sherwoodet al., 2002a), and what was for a number of years the oldest knownhominin (Hill, 1985; Ward and Hill, 1987). In addition to fossilhominins, the Tugen Hills have also produced other notableprimate specimens, in particular a partial skeleton of the apeEquatorius (Ward et al., 1999; Kelley et al., 2002; Sherwood et al.,2002b), possibly the latest representative of Proconsul (Hill et al.,1985; Hill and Ward, 1988; Deino et al., 1990), some of the latest-occurring specimens of the primitive cercopithecoid family Vic-toriapithecidae (Hill et al., 2002), the earliest colobine monkey,Microcolobus (Benefit and Pickford, 1986), one of the earliestmembers of the Theropithecus radiation and type of Theropithecusbaringensis (Leakey,1969; Eck and Jablonski,1984,1987; Delson andDean, 1993; Gilbert, 2008), and the type and nearly completeskeleton of the Pliocene colobine Paracolobus chemeroni (Leakey,1969; Birchette, 1982).

While some of these important cercopithecoid specimens havebeen named and described (see references above), and somecomments on their geological context published (Gundling andHill, 2000; Behrensmeyer et al., 2002; Hill et al., 2002; Kingstonet al., 2002), a number still remain to be documented. This paper

Figure 1. Geological section of the Tugen Hills. Cercopithecoid taxa found in theMuruyur Beds, Ngorora Formation, Mpesida Beds, and Lukeino Formation are docu-mented to the right of the section. Taxa listed within the same formation are notnecessarily from the same level or contemporaneous (see also Table 1).

C.C. Gilbert et al. / Journal of Human Evolution 59 (2010) 465e483466

is the first in a series reviewing, identifying, and describing addi-tional cercopithecoids from the Tugen Hills. The fossils described inthis paper were recovered from deposits ranging fromw16 to 6Ma,from the Middle to the Late Miocene. As these fossils are found instrata spanning a critical period in cercopithecoid evolution, theyhave the potential to inform us about the origins and evolution ofthe living cercopithecoids. If molecular estimates are correct, wemight expect to find some of the earliest members of the modernsubfamilies and tribes within the Tugen Hills material.

Tugen Hills stratigraphy and chronometric framework

The Tugen Hills are formed of a large tilted fault block within thenorthern Kenya Rift Valley, to the west of Lake Baringo (Hill, 2002).They extend about 150 km along the rift floor, and in the faultscarps and the foothills to the east are exposed around 3000 m ofrocks, much of which are sedimentary and contain fossils. Thefossiliferous sediments range in age from around 16 Ma to a fewhundred thousand years ago, with only a few significant interrup-tions. These deposits sample a number of time periods otherwiseunavailable in Africa. Systematic collection of the fossils began inthe mid 1960s, with the work of the East African GeologicalResearch Unit (EAGRU) as well as with an ancillary and related teamof graduate students e of which AH was the first e working withBill Bishop from 1968. Fossil collections weremade by these groupsand then subsequently by the Baringo Paleontological ResearchProject (BPRP). This project was founded in 1980 by David Pilbeamand has worked in collaboration with the National Museums ofKenya continually ever since. From 1985 onwards, BPRP has beenbased at Yale with AH as director. BPRP has also built on the earlywork of EAGRU to establish a finely resolved stratigraphic andtemporal framework for the fossil occurrences (Tauxe et al., 1985;Hill et al., 1985, 1986, 1991, 2002; Deino et al., 1990, 2002; Hill,1995, 1999, 2002; Behrensmeyer et al., 2002; Kingston et al., 2002).

The sediments are partitioned into a number of formations ormembers, separated and defined largely by volcanic units. Fromoldest to youngest the principal ones are: Muruyur Beds(w16e13.4 Ma), Ngorora Formation (w13e8.8 Ma), Mpesida Beds(w7e6.2 Ma), Lukeino Formation (w6.2e5.7 Ma), ChemeronFormation (w5.31e1.6 Ma) and Kapthurin Formation(w0.6e0.2 Ma). New primates described here come from theMiocene part of the sequence, from the Muruyur Beds, MpesidaBeds and Lukeino Formation, sites ranging in age from 15.6 to5.7 Ma (Fig. 1). Most of these specimens, even those that preserveparticular or novel anatomical features important for assessingfunction or phylogeny, also largely derive significance from theirgeologic age. Consequently the following descriptions are arrangedchronologically, under headings corresponding to the geologicalformations from which they come. Also, so that this may forma complete review of the Tugen Hills cercopithecoid material, wemake brief reference to previously described specimens within thischronological sequence. Short introductory sections under eachrock unit heading provide more detailed discussion about thestratigraphical and chronological situation of relevant fossil sites.

The Tugen Hills sites were originally recorded by EAGRU invarious and idiosyncratic ways. For example JM 511 is the site ofKibingor, discovered by John Martyn. This method was systema-tized by adding a prefix to the site number indicating the EAGRUmapping area inwhich it was found. JM 511 became 1/511. After theformation of BPRP we created a database cataloging old sites andour new ones; 1/511 became BPRP#1. BPRP numbers are givenwherever possible for the sites mentioned below, but a few of theolder localities have not yet been incorporated in our catalog, andthe EAGRU format numbers are therefore provided, such as 2/228for Aragai, for example.

Materials and methods

A full list of specimens by taxon and anatomical element is givenin Table 1 along with a summary of age. All material was identifiedby CCG and EDG to the lowest taxonomic level possible. All speci-mens are housed in the National Museums of Kenya, Nairobi.Relevant measurements were taken using digital calipers andrecorded to the nearest tenth of a millimeter. They are provided inTables 2e8.

Muruyur Beds

The Muruyur Beds extend in age from 16 to 13.4 Ma, but all ofthe primate specimens described here derive from a single horizon(BPRP#89) (Hill et al., 1991; Behrensmeyer et al., 2002). It isa remarkably dense and laterally extensive bone bed consisting offossils belonging to a great diversity of taxa of a great range of sizes,and Victoriapithecus specimens have previously been recorded ascoming from the site but not described (Behrensmeyer et al., 2002).This fossiliferous horizon is very well constrained by radiometricdates to between 15.83 � 0.04 Ma and 15.59 � 0.03 Ma(Behrensmeyer et al., 2002). Primate fossils also come from otherlevels in the Muruyur Beds, for example, an Equatorius partialskeleton from between 15.58 and 15.36 Ma (BPRP#122), but nocercopithecoids are yet known from that site (Ward et al., 1999;Kelley et al., 2002; Sherwood et al., 2002b; Patel et al., 2009). Fivevictoriapithecid specimens are described, represented by threeteeth, one partial mandible, and one distal humerus.

Systematic paleontology

Order Primates Linnaeus, 1758Suborder Anthropoidea Mivart, 1864

Infraorder Catarrhini E. Geoffroy, 1812Superfamily Cercopithecoidea Gray, 1821

Family Victoriapithecidae von Koenigswald, 1969Victoriapithecidae gen. et sp. indet.

Table 1Miocene cercopithecoid specimens from the Tugen Hills.

Taxon Specimen number Anatomical element Geological unit BPRP site # Estimated age

Victoriapithecidae gen. et sp. indet. KNM-TH 27447 Left mandibular frag. with M1�M3 Muruyur 89B 15.83e15.59 MaVictoriapithecidae gen. et sp. indet. KNM-TH 31013 Left P4 Muruyur 89A 15.83e15.59 MaVictoriapithecidae gen. et sp. indet. KNM-TH 31014 Right P4 Muruyur 89 15.83e15.59 MaVictoriapithecidae gen. et sp. indet. KNM-TH 31015 Distal left humerus Muruyur 89A 15.83e15.59 MaVictoriapithecidae gen. et sp. indet. KNM-TH 45125 Left M1? Muruyur 89 15.83e15.59 Ma*Victoriapithecidae gen. et sp. indet. KNM-TH 23136 Right M3 Ngorora 38 12.49 Ma*Victoriapithecidae gen. et sp. indet. KNM-TH 23137 Right M2? Ngorora 38 12.49 Ma*Victoriapithecidae gen. et sp. indet. KNM-TH 23138 Right M1? Ngorora 38 12.49 Ma*Victoriapithecidae gen. et sp. indet. KNM-TH 23139 Right M3 Ngorora 38 12.49 Ma*Victoriapithecidae gen. et sp. indet. KNM-TH 23141 Right M2? Ngorora 38 12.49 Ma*Victoriapithecidae gen. et sp. indet. KNM-TH 23142 Left M1? Ngorora 38 12.49 Ma*Victoriapithecidae gen. et sp. indet. KNM-TH 23143 Right P4 Ngorora 38 12.49 Ma*aff. Colobinae KNM-BN 1251 Right P4 Ngorora 38 12.49 Ma*Microcolobus tugenensis KNM-BN 1740 Partial mandible preserving the symphysis,

both corpora and the entire dentitionexcept for right C1

Ngorora 25 9.5e9 Ma

Colobinae gen. et sp. indet A KNM-TH 23169 Right M3 Mpesida 85 6.7e7.2 Ma to 6.37 MaColobinae gen. et sp. indet A KNM-TH 30975 Left M2? Mpesida 85A 6.7e7.2 Ma to 6.37 MaColobinae gen. et sp. indet B KNM-TH 36742 Right M2? Lukeino 76A 5.88e5.7 MaColobinae gen. et sp. indet B KNM-LU 344 Right astragalus Lukeino 2/228 w6.1 Macf Colobinae KNM-TH 26706 Right P4? Lukeino 29 w6.1 Macf. Parapapio lothagamensis KNM-TH 36744 Right M3? Lukeino 76A 5.88e5.7 Macf. Parapapio lothagamensis KNM-LU 861 Right M1 or M3? Lukeino 29 w6.1 MaCercopithecidae gen et. sp. indet KNM-LU 940 Right C1 Lukeino 76A 5.88e5.7 MaCercopithecidae gen et. sp. indet KNM-LU 771 Left C1 Lukeino 29 w6.1 MaCercopithecidae gen et. sp. indet KNM-LU 860 Proximal left humerus with crushed

proximal shaftLukeino 2/225 w5.88 Ma

Cercopithecidae gen et. sp. indet KNM-LU 939 Left calcaneum Lukeino 2/225 w5.88 Ma

Notes: * indicates specimens that have been described previously (see Benefit and Pickford, 1986; Hill et al., 2002).

C.C. Gilbert et al. / Journal of Human Evolution 59 (2010) 465e483 467

Referred material

KNM-TH 27447, KNM-TH 45125, KNM-TH 31013, KNM-TH 31014,KNM-TH 31015 (Table 1).

Horizon

Muruyur Beds outcrops, Tugen Hills, Kenya.

Localities/Sites

Muruyur Beds, Kipsaramon site complex, site BPRP#89: from15.83 � 0.04 Ma to 15.59 � 0.03 Ma (Behrensmeyer et al., 2002).

Descriptions

KNM-TH 27447 This specimen is a left mandibular fragmentpreserving M1�3 with M1 heavily damaged (Fig. 2aec). Themandibular corpus is broken, so the full depth of the corpus underthe tooth row is unknown. The ascending ramus is not preserved.The extramolar sulcus is narrow, as is generally typical ofcercopithecines.

The M2 and M3 are in good condition and relatively unworn(Table 2 offers a complete list of dental measurements). Themolars are bilophodont, but both the M2 and M3 exhibit

Table 2Measurements of the permanent molars.

Taxon Specimen number M1 MD M1 MBL M1 DBL M1 N

Victoriapithecidae gen. et sp. indet. KNM-TH 45125 6.6 7.0 6.2 3cf. Parapapio lothagamensis KNM-TH 36744 X X Xcf. Parapapio lothagamensis KNM-LU 861 6.9 6.9 6.5

Taxon Specimen number M1 MD M1 MBL M1 DBL M1 N

Victoriapithecidae gen. et sp. indet. KNM-TH 27447 (5.9) X XColobine gen. et sp. indet. A KNM-TH 23169 X X XColobine gen. et sp. indet. A KNM-TH 30975 X X XColobine gen. et sp. indet. B KNM-TH 36742 X X X

Notes: Values in parentheses represent estimates. X ¼ unavailable measurement. MD ¼ mmesial loph/lophid; DBL ¼ max distal buccolingual width measured across the distal lop1993, 1999 for details.

a relatively weak distal lophid. M2 is a relatively square tooth, andif it is representative of the larger population, cercopithecid bodymass estimates derived from M2 area regressions suggest that thevictoriapithecid population averaged somewhere between w6and 7 kg (‘all cercopithecids equation’ from Delson et al., 2000).The mesial shelf on M2 is as buccolingually broad and as large, orlarger, than the distal shelf, which appears to be slightlycompressed by the M3 behind it (Fig. 2c). There is no hypoconulidon M2, but M3 has a large hypoconulid. No obvious tuberculumsextum or cuspules are present adjacent to the hypoconulid inthis specimen.

In general, the molars display relatively low and bunodontcusps, consistent with a frugivorous diet similar to that inferred forVictoriapithecus macinnesi (e.g., Benefit, 1987, 1999; Benefit andMcCrossin, 1990). Frugivory is also indicated by the notch height/crown height below the notch ratio (NH/NR) (Table 2), as the M2lingual notch height is relatively short and the tooth displays thetall crown/short cusp height ratio that is typical of later cercopi-thecinemonkeys. Themedian buccal clefts aremoderate in size anddevelopment, and not nearly as deep as seen in colobines and

H/NR M2 MD M2 MBL M2 DBL M2 NH/NR M3 MD M3 MBL M3 DBL M3 NH/NR

3.3 X X X X X X X XX X X X X 7.7 7.8 6.5 48.3X X X X X X X X X

H/NR M2 MD M2 MBL M2 DBL M2 NH/NR M3 MD M3 MBL M3 DBL M3 NH/NR

X 6.1 6 5.7 52.4 7.6 5.7 4.6 XX X X X X 10.2 6 5.8 158.8X 8.0 6.3 6.4 130.0 X X X XX 6.5 4.8 5.2 150.0 X X X X

ax mesiodistal length; MBL ¼ max mesial buccolingual width measured across theh/lophid; NH/NR ¼ notch height/crown height below the notch; see Benefit, 1987,

Figure 2. Victoriapithecidae gen. et sp. indet. mandibular and dental material from the Muruyur Beds, Tugen Hills, Kenya. KNM-TH 27447: left corpus fragment with M1�3; a) buccalview; b) lingual view; c) occlusal view; KNM-TH 45125: left upper molar; d) occlusal view. Scale bar ¼ 1 cm.

C.C. Gilbert et al. / Journal of Human Evolution 59 (2010) 465e483468

Theropithecus. Small mesial and distal buccal clefts are present onM2, but only the mesial buccal cleft is present on M3. Both molarsshow a moderate to high degree of basal flare.KNM-TH 45125 This specimen is represented by an upper leftmolar, most likely anM1 based on size and shape comparisons withthe sample from Maboko (Benefit, 1987). The tooth is broken justbelow the crown (Fig. 2d). Similar to KNM-TH 27447, the crown isrelatively high and the cusps are relatively low. The distal cusps, inparticular, are closely approximated, leading to a high degree ofbasal flare. The median lingual cleft is tall and narrow, as is typicalof Victoriapithecus upper molars, and there is a large mesial lingualcleft but no distal lingual cleft. Frugivory is indicated by the low,rounded cusps and the corresponding low NH/NR ratio (Table 2). Amesiodistally short mesial shelf is present at the front of the tooth,and a small and buccolingually compressed distal shelf is present inthe back (Fig. 2d). The mesial loph is complete but weakly devel-oped, and the distal loph appears to be incomplete or absent(Fig. 2d).KNM-TH 31013 This tooth, a relatively unworn left P4 crown, lackspreserved roots. The mesial fovea is moderate in size and slantslingually toward the metaconid. The metaconid and protoconid arenearly equal in height, similar to V. macinnesi, with the metaconidbeing slightly higher in buccal view. The protoconid and meta-conid are connected by a lophid running between them. Like othercercopithecoid P4s, there is a mesiobuccal extension of enameldown onto the beginning of the root of the tooth below the pro-toconid. A wide and shallow median cleft is present buccally. Thelingual side of the metaconid is damaged, but a slight lingual notchis present.

At the distal end of the tooth, both a hypoconid and ento-conid are present with a crest or lophid connecting them. Unlikemost V. macinnesi P4s, a well-defined distal fovea is presentbehind the distal lophid. Overall, the crown height appears tallcompared to the cusp height and the tooth displays a moderate

to high degree of flare for a P4. Measurements are provided inTable 3.KNM-TH 31014 Like the previous specimen, this right P4 crownlacks roots. Compared to KNM-TH 31013, this specimen is smallerand slightly more worn, but still falls well within the range ofvariation for V. macinnesi P4s (see Table 3 for measurements). Again,a well-defined mesial fovea is present and slants lingually towardthe metaconid. Both the metaconid and protoconid are prominentand relatively equal in height, with the metaconid only slightlytaller. There is a crest or lophid connecting the protoconid andmetaconid. Amesiobuccal extension of enamel is present below theprotoconid. Buccally, a well-defined V-shaped median buccal cleftis present.

In contrast to the larger KNM-TH 31013 but similar to manyVictoriapithecus P4s fromMaboko, KNM-TH 31014 lacks a definitivehypoconid or entoconid, resulting in a large distal basin or distalshelf behind the protolophid. There is a tiny distal fovea at the backof the tooth. Finally, KNM-TH 31014 appears to exhibit a moderatedegree of flare.KNM-TH 31015 This distal left humeral fragment preserves thearticular surface as well as approximately half of the shaft. The bonepresents several available measurements (Table 4). It is fairly large,exceeding theVictoriapithecushumeri described byHarrison (1989),but stillwithin the rangeof variation seen in theMaboko sample andcomparing favorably in size with KNM-MB 33513 (Fig. 3). On thebasis of humeral measurements and strong muscle markings, thisindividual was probably an adult male. On the posterior side of thehumerus, a deep olecranon fossa is present. The medial and lateralpillars surrounding the olecranon fossa are strong, with the lateralpillar appearing slightly taller (Fig. 3b and d). A large medial epi-condyle is present and projects medioposteriorly at an angle ofw45� (Fig. 3e). The medial epicondyle is larger and wider than thelateral epicondyle; the large size of the medial epicondyle impliesthat this individual possessed powerful wrist flexors, while the

Figure 3. KNM-TH 31015: Victoriapithecidae gen. et sp. indet. left distal humerus from the Muruyur Beds, Tugen Hills, Kenya; a) anterior view; b) posterior view; c-e) compared toV. macinnesi humeri from Maboko, left to right, KNM-MB 33513, KNM-TH 31015, KNM-MB 25314, KNM-MB 21207; c) anterior view; d) posterior view; e) distal view. Note theorientation of the medial epicondyle. Scale bar ¼ 1 cm.

C.C. Gilbert et al. / Journal of Human Evolution 59 (2010) 465e483 469

posterior direction of the medial epicondyle suggests a strength-ening of the medial side of the olecranon fossa against weight-bearing stress during forearm pronation (Harrison, 1989). There isalso a well-developed medial trochlear keel, a feature whichpromotes elbow stabilization during quadrupedal locomotion(Fig. 3a and c; Jenkins, 1973; Harrison, 1989). The lateral trochlearkeel is poorly developed and the capitulum is lowand rounded, as inother Victoriapithecus specimens, suggesting restricted rotation andelbow stabilization during full forearm extension (Fig. 3a and c;Harrison, 1989). At the distal articulation, the trochlea is narrowerthan the capitulum(Fig. 3a andc; Table3). Finally, the supinator crestis well defined and extends proximally, similar to Macaca, Cercoce-bus, Mandrillus and Parapapio lothagamensis (Fig. 3a and c; Fleagleand McGraw, 2002; CCG pers. obs.). The large and extended supi-nator crest implies a large and powerful brachioradialis, a majorforearm flexor, as well as strong wrist extensors, which are impor-tant in stabilizing the wrist and promoting grasping and digitalflexion. In total, the combination of features seen in the humerussuggest a significant terrestrial component to this animal’s loco-motor repertoire and an emphasis on activities that require

powerful forearm flexion and grasping, such as manual foragingand/or climbing.

Remarks

Overall, the Muruyur sample of victoriapithecids fits within therange of variation seen in the large V. macinnesi sample fromMaboko, as previously recognized by Behrensmeyer et al. (2002).However, Pickford and Kunimatsu (2005) recently argued thata larger sample of victoriapithecid specimens they described fromMuruyur sites at Kipsaramon and from a site they refer to as Keturo,belonged to a separate victoriapithecid species (Prohylobates kip-saramanensis). More recently, Miller et al. (2009) have examinedthese specimens and tentatively assigned them to a new genus ofvictoriapithecid as cf. Noropithecus kipsaramanensis. Consequently,given the taxonomic ambiguity of the victoriapithecids at this site,we have conservatively decided to assign our specimens to Vic-toriapithecidae gen. et sp. indet. until proper comparisons can bemade. Unfortunately, we have not been able to examine the cf. N.kipsaramanensis sample, as it has not been deposited at the

Table 3Measurements of the canines and 4th premolars.

Taxon Specimen number C1 MD C1 BL P4 MD P4 BL

cf. Colobinae KNM-TH 26706 X X 5.5 6.4Cercopithecidae gen.

et sp. indet.KNM-LU 771 9.5 7.1 X X

Cercopithecidae gen.et sp. indet.

KNM-LU 940 7.8 5.6 X X

Taxon Specimen number C1 MD C1 BL P4 MD P4 BL

Victoriapithecidae gen.et sp. indet.

KNM-TH 31013 X X 5.7 4.6

Victoriapithecidae gen.et sp. indet.

KNM-TH 31014 X X 4.8 4.0

Notes: X ¼ unavailable measurement. MD ¼ max mesiodistal length; BL ¼ maxbuccolingual width.

C.C. Gilbert et al. / Journal of Human Evolution 59 (2010) 465e483470

National Museums of Kenya as Kenyan law requires. We also notethat there is much taxonomic and stratigraphic confusion regardingthe previously described cf. N. kipsaramanensis specimens and theKipsaramon site complex itself.1

While Miller et al. (2009) have acknowledged a new species ofvictoriapithecid at this site (cf. N. kipsaramanensis), from our owninterpretation of published descriptions, we see very little basis forseparating most of these previously described specimens from thespecimens described here or from the large and quite variable V.macinnesi hypodigm fromMaboko. The published range of the cf.N.kipsaramanensis tooth measurements, at least those that can beconfidently placed in correct serial position, either extensivelyoverlap or lie completely within the range of variation seen in V.macinnesi at Maboko (see Pickford and Kunimatsu, 2005, as well asTable 5). The described morphology of the molars also conforms tothe species definition of V. macinnesi.

The only apparently diagnostic dental feature used to separatethe cf. N. kipsaramanensis material from V. macinnesi is the rela-tively narrower P3s of cf. N. kipsaramanensis; however, it is difficultto ascertain which premolar dimensions Pickford and Kunimatsu(2005) compared. We infer that Pickford and Kunimatsu (2005)measured the maximum length of the P3 (including the honingflange) and this measurement results in the appearance of a rela-tively narrower tooth when compared to Benefit’s (1993) length/

1 There are a number of discrete sites in the Kipsaramon site complex, not all ofthem the same age. Pickford and Kunimatsu (2005) are not generally very forth-coming about the exact sites or stratigraphic levels that their particular primatespecimens come from, nor do they seem to acknowledge temporal differenceswithin the local succession. We see no reason to accept their claim that the age ofthe Kipsaramon sites is around 14.5 Ma (see also Morales and Pickford, 2008). BPRPhas produced a detailed lithostratigraphy and chronostratigraphy of the unit atKipsaramon, and Behrensmeyer et al. (2002) present a finely documented accountof those results. For example, the main bone bed, BPRP#89, crops out at a numberof separate exposures, and is bracketed between 15.83 � 0.04 Ma and 15.59 � 0.03 Ma. Site BPRP#91, from which comes a large catarrhine talus (Hill and Ward,1988), is about the same age. The locality where an Equatorius partial skeletonwas found (Ward et al., 1999; BPRP#122) is dated between 15.58 Ma and 15.36 Ma.These determinations are consistent with biostratigraphic data, such as the pres-ence of Diamantomys at BPRP#89, which is not compatible with Pickford’ssuggestion of 14.5 Ma (Winkler, 2002). It is apparent from the text of Pickford andKunimatsu (2005), as well as from their Appendix 5, that their victoriapithecidspecimens were found in at least two regions, Kipsaramon and also the Keturo site.Although they specify that a few particular teeth come from Keturo, it is not clearexactly which specimens come from Keturo and which come from Kipsaramon. Nordo they give details of the stratigraphic relations between the two areas. And it isalso not clear whether some teeth come from additional sites in the Kipsaramonsite complex itself. Despite this general lack of detail about sites and stratigraphiclevel, Pickford and Kunimatsu (2005) do in fact present the geographic co-ordinatesof the type site of Prohylobates kisparamanensis. This is presumably the site fromwhich the mandible (BAR 219’02) derives, and these co-ordinates are consistentwith an outcrop of the bone bed, BPRP#89, the locale of our victoriapithecid teeth.

width ratios using occlusal length only. When compared toBenefit’s (1993) maximum length measurements for the P3 sampleat Maboko, the difference in P3 length (and shape) between V.macinnesi and cf. N. kipsaramanensis disappears (see Table 5).

In contrast to the isolated dental material, the type mandible ofcf. N. kipsaramanesis, BAR 219’02, displays symphyseal features thatare argued to be quite distinct from V. macinnesi and warranttaxonomic separation (Pickford and Kunimatsu, 2005; Miller et al.,2009; Benefit, pers. comm.). In particular, in their diagnosis of cf. N.kipsaramanensis, Miller et al. (2009) note that BAR 219’02 hasa more vertical mandibular symphysis than V. macinnesi. In theabsence of diagnostic dental features, under this taxonomicscheme, most victoriapithecid genera can only be diagnosed in thepresence of mandibular specimens preserving complete symphy-seal regions.

Thus, there are a few possibilities regarding the taxonomy of theKipsaramon victoriapithecids, including the BPRP specimensdescribed here. One view is to suggest that V. macinnesi is repre-sented by the specimens described in this paper as well as most ofPickford and Kunimatsu’s isolated dental specimens (2005). It isalso possible that the mandible BAR 219’02 represents V. macinnesi;comparisons of symphyseal measurements with the Mabokosample do not reveal any overwhelming differences to us. Theheight of the symphysis relative to AeP depth is slightly higher inthe Kipsaramon specimen compared to V. macinnesi from Maboko,but the difference is small and only three specimens from theMaboko sample preserve the relevant morphology (Table 6). Theposition of the genial pit of the Kipsaramon specimen falls withinthe range seen at Maboko (Table 6). Thus, given the low samplesizes and variation with which to compare the Kipsaramonmandible, we would assign this specimen to Victoriapithecidaegen. et sp. indet., pending additional material to corroborate themorphology observed in BAR 219’02 and to justify the recognitionof a new victoriapithecid genus and species at Kipsaramon on thebasis of this mandible.

An alternative view tomost of thematerial being assignable to V.macinnesi and the rest left as Victoriapithecidae gen. et sp. indet. isto suggest that multiple, identifiable taxa are present. For instance,it is important to point out that, in their initial description, Pickfordand Kunimatsu (2005) recognize that some of their isolated teethappear distinct from the others and raise the possibility that twospecies are present at the sample. Thus, it is quite possible that twovictoriapithecid taxa are indeed present, V. macinnesi and cf. N.kipsaramanensis or some other as yet unnamed taxon. We see noproblem with this scenario, in particular because multiple cerco-pithecine and colobine taxa, often in the same genus, are routinelyfound in extant Old World forests today. In fact, given that multiplecercopithecoid taxa are routinely found in extant forests as well asat most Plio-Pleistocene sites, it is somewhat curious that only onevictoriapithecid taxon is commonly found at most Miocene locali-ties. If two taxa are present at Kipsaramon, it is possible that someof the isolated dental specimens described by Pickford andKunimatsu (2005) belong with the specimens described here andothers may belong with BAR 219’02 as a second victoriapithecidtaxon such as cf. N. kipsaramanensis. Other taxonomic possibilitiesare possible as well, depending on howmany taxa can be identifiedconfidently. Increased taxonomic resolution in this casewill only bepossible with further study of the entire victoriapithecid samplefrom Kipsaramon, currently unavailable for study in a single loca-tion, as well as the collection of additional and more completespecimens.

A final possibility would be to suggest that a single taxon ispresent at Kipsaramon, but that this taxon is not V. macinnesi. Inthis case, all of the specimens from Kipsaramon, including the onesdescribed in this paper, would be included in cf. N.

Table 4Humeral measurements.

MeasurementTaxon(Specimen number)

Taxon(Specimen number)

Taxon(Specimen number)

Taxon(Specimen number)

Victoriapithecidaegen. et sp. indet.(KNM-TH 31015)

Victoriapithecusmacinnesi(Range from KNM-MB3, 19, 12044)

Parapapio lothagamensis(Range from KNM-LT 23067,23074, 23077, 24123, 28769)

Cercopithecidae gen.et sp. indet.(KNM-LU 860)

Max ML breadth of olecranon fossa 8.8 6.6e7.8 X XMax PD height of olecranon fossa 8.0 6.1e8.3 X XML breadth from medial trochlear

keel to lateral epicondyle18.6 13.5e15.6 X X

PD height of the capitulum 8.0 6.0e6.7 X XML breadth of capitulum 9.7 6.8e8.5 6.8e8.3 XBiepicondylar breadth 22.5 16.7e17.5 X XPD height of the trochlea 12.1 7.0e9.2 X XAP breadth of the trochlea 11.0 6.8e8.4 X XML breadth of distal articulation 16.4 10.7e13.7 18.4e20.7 XAP breadth of distal humerus 14.7 9e11.1 X XPosterior angulation of medial epicondyle w45 degrees 42e63 degrees 40e75 degrees XRelative breadth of medial epicondyle 17.3 10.9e19.2 X XAP length of humeral head X 13.3 17.3 14.8ML breadth of humeral head X 11.5 15.3 13.1Max diameter of greater tuberosity X 10.3 15.2 10.7Min distance between tuberosities X 3.5 X 5.4Max diameter of lesser tuberosity X 6.7 9.0 8.1Depth of bicipital groove X Shallow Shallow Shallow

Notes: X ¼ unavailable measurement. ML ¼ medial-lateral; PD ¼ proximal-distal; AP ¼ anterior-posterior. Measurements for Victoriapithecus and Parapapio lothagamensisfrom Harrison (1989) and Leakey et al. (2003).

C.C. Gilbert et al. / Journal of Human Evolution 59 (2010) 465e483 471

kipsaramanensis. However, there are morphological reasons todoubt that the sample described here belongs to the same vic-toriapithecid taxon as all of the other specimens currently includedin the cf. N. kipsaramanensis hypodigm by Miller et al. (2009) andPickford and Kunimatsu (2005). For instance, the P4s describedhere are within the range of the Maboko specimens in terms ofoverall shape and dimensions, but outside the described range ofcf. N. kipsaramanensis in terms of BL width (Table 5). In addition,KNM-TH 45125, an upper molar, lacks a clear distal loph, a featureshared with V. macinnesi specimens but not with the described cf.N. kipsaramanensis hypodigm (Fig. 2d). Finally, the KNM-TH 31015humerus is very similar to other humeri found at Maboko andsuggests either an assignment to Victoriapithecus or that cf. N.kipsaramanensis is nearly identical to V. macinnesi in its humeralmorphology. In summary, including the specimens here with thepreviously described specimens from Kipsaramon would require

Table 5Dental measurements of Victoriapithecus macinnesi vs. cf. Noropithecus kipsaramanensis.

TaxonI1 MD Mean(Range)

I1 BL Mean(Range)

I2 MD Mean(Range)

I2 BL Mean(Range)

P3

(Ra

V. macinnesi (Maboko) 5.1(4.2e6.2)n ¼ 42

4.1(3.1e4.7)n ¼ 46

3.6(3.2e4.2)n ¼ 15

4.0(3.1e4.7)n ¼ 17

4.5(4.n ¼

cf. N. kipsaramanensis 5.1(4.9e5.5)n ¼ 4

4.1(3.7e4.3)n ¼ 4

X X 4.5n ¼

TaxonI1 MD Mean(Range)

I1 BL Mean(Range)

I2 MD Mean(Range)

I2 BL Mean(Range)

P3(R

V. macinnesi (Maboko) 3.4(3.1e3.9)n ¼ 37

3.8(3.1e4.5)n ¼ 43

3.1(2.6e3.7)n ¼ 29

3.9(3.2e4.7)n ¼ 30

6.4(3n

Victoriapithecidae gen.et sp. indet. (Kipsaramon)

X X X X X

cf. N. kipsaramanensis 3.6(3.4e3.8)n ¼ 3

4.0(3.5e4.4)n ¼ 3

2.7(2.6e2.7)n ¼ 2

4.0(3.9e4.0)n ¼ 2

6.3(5n

Notes: X ¼ unavailable measurement. MD¼maxmesiodistal length; BL ¼max buccolinguand this study. Only teeth that could be confidently attributed to serial position are com

expanding the morphological diagnosis of cf. N. kipsaramanensis.Of the three hypotheses offered here, we suggest that this lasthypothesis is the least likely.

The only way that these taxonomic issues will be resolved isthrough further study of the entire victoriapithecid sample fromKipsaramon. In any case, the presence of victoriapithecids in theMuruyur Beds in addition to their presence in the later NgororaFormation suggests that the group existed in the Tugen Hills forat least 3 million years. In addition to suggesting a semi-terres-trial habitus, the distal humerus (KNM-TH 31015) is interestingin the context of recent analyses of cercopithecine and, inparticular, papionin postcranial morphology. The proximalextension of the supinator crest, a feature that Fleagle andMcGraw (2002) associated with an adaptational complex forterrestrial manual foraging and/or vertical climbing, is also seenin the earliest described papionin, Pp. lothagamensis, as well as in

MD Meannge)

P3 BL Mean(Range)

P4 MD Mean(Range)

P4 BL Mean(Range)

M3 MD Mean(Range)

M3 BL Mean(Range)

1e5.1)6

5.4(5.1e5.9)n ¼ 6

4.7(4.3e4.8)n ¼ 9

6.5(6.0e7.0)n ¼ 9

6.5(5.5e7.7)n ¼ 43

7.1(5.6e9.0)n ¼ 43

15.5n ¼ 1

4.4(3.9e4.6)n ¼ 4

6.7(4.9e7.8)n ¼ 4

6.0(5.6e6.3)n ¼ 8

7.2(6.3e7.5)n ¼ 8

MD Meanange)

P3 BL Mean(Range)

P4 MD Mean(Range)

P4 BL Mean(Range)

M3 MD Mean(Range)

M3 BL Mean(Range)

.7e8.6)¼ 29

4.1(3.2e5.8)n ¼ 37

5.6(3.7e6.8)n ¼ 41

4.7(3.8e6.5)n ¼ 49

8.9(6.8e10.9)n ¼ 75

6.3(5.2e7.7)n ¼ 72

X 5.3(4.8e5.7)n ¼ 2

4.3(4.0e4.6)n ¼ 2

7.6n ¼ 1

5.7n ¼ 1

.8e7.2)¼ 3

3.5(3.4e3.6)n ¼ 3

5.1(4.5e6.2)n ¼ 3

5.1(4.7e5.4)n ¼ 3

8.1 (7.3e8.8)n ¼ 7

6.1(5.7e6.6)n ¼ 7

al width. Measurements taken from Benefit (1993), Pickford and Kunimatsu (2005),pared.

Table 6Measurements of the mandibular symphysis.

TaxonSpecimennumber

Symphysealheight

Symphyseal APthickness

Symphysealheight/thickness

Genial pit heightbelow alveolus

Genial pit height aboveinferior border

Genial pit height above/height below

Victoriapithecus macinnesi(Maboko)

KNM-MB18993

15.0 14.5 1.0 9.1 4.6 0.5

Victoriapithecus macinnesi(Maboko)

KNM-MB31282

20.2 17.5 1.2 13.1 6.6 0.5

Victoriapithecus macinnesi(Maboko)

KNM-MB31283

17.5 13.2 1.3 10.8 7.3 0.7

Victoriapithecus macinnesi(Maboko)

KNM-MB11951

22.1 X X 14.05 8.55 0.6

Victoriapithecus Range(Maboko)

See above 15e22.1 13.2e17.5 1.0e1.3 9.1e14.05 4.6e8.55 0.5e0.7

Noropithecus bulukensis(Buluk)

KNM-WS123

24.6 20.5 1.2 10.55 10.74 1.0

Noropithecus bulukensis(Buluk)

KNM- WS12639

26.9 X X 15.5 11.76 0.8

cf. Noropithecuskipsaramanensis(Kipsaramon)

BAR 219’02 19.42 13.88 1.4 12.35 7.67 0.6

Notes: X ¼ unavailable measurement. AP ¼ anterior-posterior. All measurements taken from Miller et al. (2009).

C.C. Gilbert et al. / Journal of Human Evolution 59 (2010) 465e483472

the modern genera Macaca, Cercocebus, and Mandrillus. Theappearance of this feature in all of these taxa further supportsthe contention that it is a primitive feature (Fleagle and McGraw,2002), and that the primitive papionin morphotype may haveincluded both manual terrestrial foraging as well as verticalclimbing behaviors. It likewise supports the notion that themore distally restricted supinator crest seen in Papio, Lophocebus,and extant Theropithecus is a derived condition (Fleagle andMcGraw, 2002).

Ngorora Formation

The Ngorora Formation is defined as being those sedimentsbetween the Tiim Phonolites and the Ewalel Phonolites. There aredates on the Tiim Phonolites of 13.15 Ma and those for the cappingEwalel Phonolite are 7.6 Ma and 7.2 Ma. The type section of theFormation is in the north of the Tugen Hills, at Kabasero, wherebetween 370 and 400 m of volcaniclastic sediments are exposed.Here, they range in age from about 13 Ma to a little younger than10.5 Ma (Tauxe et al., 1985; Deino et al., 1990). About 40 km furthersouth are outcrops of the formation at Ngeringerowa which areyounger than this, where dates are as recent as 8.8 Ma (Deino, pers.comm.). Two horizons provide cercopithecoid fossils. BPRP#38,which has produced victoriapithecids, is in the Kabasero typesection, and this site is particularly well dated at 12.5 Ma. Inaddition, a small colobine comes from the younger exposures atNgeringerowa (BPRP#25; 1/1005), for which there is a date esti-mate of 9.5e9 Ma (Deino, pers. comm.). Here, we briefly reviewthese fossils and refer interested readers to Hill et al. (2002) andBenefit and Pickford (1986) for the systematic paleontology anddetailed descriptions.

Systematic paleontology

Family Victoriapithecidae von Koenigswald, 1969Victoriapithecidae gen. et sp. indet.

Referred material

KNM-TH 23136, KNM-TH 23137, KNM-TH 23138, KNM-TH23139, KNM-TH 23141, KNM-TH 23142, KNM-TH 23143.

Horizon

Ngorora Formation outcrops, Tugen Hills, Kenya.

Localities/Sites

Ngorora Formation, site BPRP#38; 12.5 Ma (Hill et al., 2002).

Remarks

Benefit and Pickford (1986) discussed a primate premolar fromsite BPRP#38 (2/1) in the Kabasero type section of the NgororaFormation which had been found, though not recorded, by theAguirre/Philip Leakey expedition (Aguirre and Leakey, 1974). Theydescribed it as an indeterminate cercopithecoid. The Baringo Pale-ontological Research Project later retrieved several more teeth fromthis same site, and Hill et al. (2002), described these, along withBenefit and Pickford’s tooth, as Victoriapithecus sp. indet. Thesespecimens are the latest occurrence of victoriapithecids in the fossilrecord; BPRP#38 is securely dated at 12.5 Ma. Hill et al. (2002)believed they may represent a new species of Victoriapithecus,largely based on the morphology of the P4. This is still possibly true,even excluding the P4, but the recent work of Miller et al. (2009)raises questions as to their generic attribution within Victor-iapithecidae on the basis of dental morphology. Under their schemeit is difficult to allocate victoriapithecid dental specimens to genus,as many generic distinctions involve symphyseal characters. Giventhe uncertainty surrounding these specimens as well as those fromthe Muruyur Beds, we conservatively reassign KNM-TH 23136,KNM-TH 23137, KNM-TH 23138, KNM-TH 23139, KNM-TH 23141,KNM-TH 23142, and KNM-TH 23143 to Victoriapithecidae gen. etsp. indet. The victoriapithecid specimens from Kabasero sharemany molar features with both Victoriapithecus and the victor-iapithecids described in this paper from the Muruyur Beds,including overall size and dimensions as well as the lack of a distalloph on the upper molars; thus, we refrain from assigning theMuruyur and Ngorora specimens to separate victoriapithecidspecies. However, we exclude the P4 originally described by Benefitand Pickford (1986), as we believe this specimen (KNM-BN 1251)has colobine affinities (Rossie et al., in prep).

Systematic paleontology

Family Cercopithecidae Gray, 1821Subfamily Colobinae Jerdon, 1867

Microcolobus tugenensis Benefit and Pickford, 1986

Referred material

KNM-BN 1470 is a fairly complete mandible preserving thesymphysis, both corpora, and the entire dentition except the rightC1; the ascending rami are not preserved.

Figure 4. Colobinae gen. et sp. indet. A from the Mpesida Beds, Tugen Hills, Kenya. a)KNM-TH 30975 (left) and KNM-TH 23169 (right) in occlusal view; b) KNM-TH 30975 inbuccal view; c) KNM-TH 30975 in lingual view; d) KNM-TH 23169 in buccal view; e)KNM-TH 23169 in lingual view. Scale bar ¼ 1 cm.

C.C. Gilbert et al. / Journal of Human Evolution 59 (2010) 465e483 473

Horizon

Ngorora Formation outcrops, Tugen Hills, Kenya.

Localities/Sites

Ngorora Formation, site 1/1005, (BPRP#25); 9.5e9 Ma.

Remarks

Benefit and Pickford (1986) described a mandible of a smallcolobine from the Ngeringerowa area of the Ngorora Formation,where younger exposures of the unit crop out to the south. This wasa new genus and species e M. tugenensis, and the earliest knowncolobine. The type (KNM-BN 1740) is the only specimen.

Mpesida Beds

Kingston et al. (2002) referred tomodern-looking colobine teethfrom the Mpesida Beds, but these specimens have remainedundescribed until now. The Mpesida Beds is a rather complicatedgeological unit that is not well understood. It is a member of theKabarnet Trachyte Formation and consists of a number of sedi-mentary lenses associated with, and interbedded with, the Kabar-net Trachyte volcanic flows, and there is not necessarily any lateralcontinuity between them. Outcrops are therefore probably ofseveral distinct ages, but all occurrences are limited to between ca.7.2 or 6.7 Ma and ca. 6.2 Ma (Hill et al., 1985, 1986; Kingston et al.,2002). The teeth described here are more precisely constrainedradiometrically, both coming from site BPRP#85, located nearRormuch, where local dating suggests an age of about 6.37 Ma(Kingston et al., 2002). Immediately underlying the site at thislocation is a coarse pyroclastic ash unit that incorporated thevegetation overwhich it flowed. Plant fossils provide evidence of anextensive wet lowland forest with similarities to those in West andCentral Africa, and tree heights are estimated to have been up to50 m (Kingston et al., 2002).

Systematic paleontology

Family Cercopithecidae Gray, 1821Subfamily Colobinae Jerdon, 1867Colobinae gen. et sp. indet. A

Referred material

KNM-TH 23169, KNM-TH 30975 (Table 1).

Horizon

Mpesida Beds outcrops, Tugen Hills, Kenya.

Localities/Sites

Mpesida Beds, site BPRP#85, 6.37 Ma (Kingston et al., 2002).

Descriptions

KNM-TH 23169 This specimen is a modern-looking colobine M3exhibiting typical colobine characteristics such as a low crowned/high cusped appearance, a well-developed deep median buccalcleft with tall columnar-like cusps, and strong lophids that areobliquely oriented relative to the mesiodistal axis of the tooth(Fig. 4a, d and e). Body size estimates suggest that this individualwas derived from a population averaging between w11 and 15 kg,depending on whether the specimen is male or female, or slightlylarger than modern Colobus guereza (Fleagle, 1999; Delson et al.,2000). The hypoconulid lobe is strongly curved buccally, andmesiolingual to the hypoconulid there are noticeable wrinkles orfolds in the enamel that do not quite form a distinct cusp, but arereminiscent of a tuberculum sextum (Fig. 4a). These accessorycuspules are clearly visible in both occlusal and buccal views

(Fig. 4). The mesial shelf is mesiodistally short, and there is a mesialbuccal cleft present on the tooth as well as a distal buccal cleftbetween the hypoconid and hypoconulid (Fig. 4a).KNM-TH 30975 This specimen is a lower left molar, most likely anM2. Similar to KNM-TH 23169, this appears to be a modern colobinemolar, comparing very well in size (estimated population bodymassw10 kg if female,w13 kg if male, fromDelson et al., 2000) andoverall morphology with modern Colobus. The lophids are strongand obliquely oriented (Fig. 4a). The tooth exhibits tall cusps andlow crowns, as indicated by the high NH/NR ratio (Table 2; see alsoFig. 4b and c). The high NH/NR ratio also suggests a degree offolivory similar to modern colobines. Moderately-sized mesial anddistal shelves are present on the tooth, with the distal shelf perhapsslightly larger and more mesiodistally extended than the mesialshelf, but overall the shelves are similar in size (Fig. 4a). Themedianbuccal cleft is wide and relatively deep, andmesial buccal and distalbuccal clefts are also present on the tooth.

Remarks

While we refrain from naming a new genus and species on thebasis of two isolated teeth in this case, the two specimens describedhere likely represent a single population from the Mpesida Bedsand are important in the context of colobine evolution. Given theage of this site, at about 6.37 Ma, these teeth represent some of theearliest evidence of the modern colobine morphotype, possibly themodern African colobine morphotype, and accord well withmolecular estimates of the split betweenmodern African and Asiancolobines w12.3e8.2 Ma (Raaum et al., 2005; Ting, 2008). Incontrast to early colobines such as Microcolobus, Libypithecus, andto some extent, Mesopithecus, the colobine teeth described herehave the full complement of features characterizing modern colo-bines, particularly modern African colobines, including the highNH/NR ratio and the implicated adaptations for a high degree offolivory. The overall proportions of the teeth (wider distal lophidson the M2 but wider mesial lophids on the M3) most likely repre-sents the primitive condition for crown colobines and are typical ofextant Asian colobines and African colobines except Colobus,whichdisplays wider distal lophids on the M3 (Szalay and Delson, 1979;

Figure 5. Cercopithecoid dental remains from the Lukeino Formation, Tugen Hills,Kenya. a) Occlusal view of cercopithecoid specimens. From left to right: Colobinae gen.et sp. indet. B KNM-TH 36742, cf. Parapapio lothagamensis KNM-TH 36744, cf. Pp.lothagamensis KNM-LU 861. b) Occlusal view of cf. Colobinae KNM-LU 26706. c) Lateralview of cercopithecoid specimens. From left to right: cf. Cercopithecidae gen. et sp.indet. KNM-LU 940, Cercopithecidae gen. et sp. indet. KNM-LU 771. Scale bar ¼ 1 cm.

C.C. Gilbert et al. / Journal of Human Evolution 59 (2010) 465e483474

Frost, 2001). These teeth are similar in length but much narrowerthan the smallest species of Paracolobus (Paracolobus enkorikae),and they are larger than the recently described Late MioceneColobinae sp. B from Lemudong’o atw6.0Ma (Hlusko, 2007).Whilethese taxa from Lemudong’o are also quite modern in terms ofdental morphology, they appear to be separate populations fromthe specimens found in the Mpesida Beds of the Tugen Hills.

With regards to other described colobine material, Kuseracolobusaramisi from theMiddle Awash (AramisMember, KuseraleeMember,and Adu-Asa Formation) is larger, has relatively wider molars, andoften displays a well-developed, distinct tuberculum sextumcombined with a well-developed distal lingual notch on the M3 thatare not possessed by theMpesida specimens (Frost, 2001; Frost et al.,2009). Compared to the Late Miocene colobine specimens fromMenacer, the colobine from Nkondo, as well as the indeterminatelarge colobine from theKuseraleeMember andAdu-Asa Formation inthe Middle Awash, the Mpesida specimens are much smaller in size(Arambourg, 1959; Senut, 1994; Frost et al., 2009).

Colobinae gen. et sp. indet. A from the Mpesida Beds comparesmost favorably in morphological features to Colobinae sp. C fromthe Apak Member at Lothagam (w4.2 Ma), but the Tugen Hillspopulation appears to be slightly smaller in size. Overall, themorphology of the Mpesida M3, in particular, with its multipleaccessory cuspules adjacent to the hypoconulid, seems distinctamong known extinct or extant colobine taxa. If this condition isconfirmed as a typical feature rather than a rare variant in futurecolobine M3 specimens from the Mpesida Beds, it would suggestthat this is indeed a distinct taxon; however, it is also conceivablethat future specimens may bridge the morphological gap betweenthe current material and another named colobine taxon.

Lukeino Formation

The Lukeino Formation sediments lie between the underlyingKabarnet Trachyte Formation and the overlying Kaparaina Basalts.They are therefore bracketed temporally by rocks dated at around6.2 Ma and 5.72 Ma (Deino et al., 2002).

There are four taxonomic categories of cercopithecoid describedhere, which come from a number of sites exposed within severaldiscrete exposures of the Formation. The internal chronology of thesuccession of strata exposed at Kapcheberek is particularly wellunderstood through radiometric and paleomagnetic calibration(Deino et al., 2002). Site BPRP#76, within that sequence, fallsbetween 5.88 and 5.70 Ma, and has produced specimens attributedto Colobinae gen. et sp. indet. B, to Cercopithecidae gen. et sp. indet.,and to one of the specimens allocated as cf. Pp. lothagamensis. Also inthat local sequence is site 2/225, in the EAGRU system (a site not yetformally cataloged by BPRP). Two specimens belonging to Cercopi-thecidae gen. et sp. indet. come from this locality. They werecollected by Pickford in the 1970s, and he describes this site as beingsituatedwithin the lowest fossiliferous horizon of three in the upperpart of the local succession (Pickford,1975). It is probable that this isat the older end of the range given for our dated level at siteBPRP#76, 5.88e5.70 Ma (Deino et al., 2002).

Another Lukeino outcrop, Aragai, contains another site that hasnot been cataloged in the BPRP locality database, site 2/228. Thislies near the base of the Formation and therefore may date toaround 6.1 Ma. A specimen from this locality was collected byPickford and is here attributed to Colobinae gen. et sp. indet. B.

Cheboit, BPRP#29, is another site lying in a separate Lukeinoarea of exposures. Again it is conventionally regarded as being nearthe base of the Lukeino succession, and therefore dated to around6.1 Ma. However, our work in the area of Rormuch (Kingston et al.,2002) has cast doubt on this, so that its position has not yet beenresolved. It seems quite possible that sediments in this area in fact

should be allocated to the Mpesida Beds; this would date the tootharound 0.5 M yr older. One specimen attributed to cf. Colobinae andone attributed to Cercopithecidae gen. et sp. indet. derive from thissite.Andwhileoneof the specimensattributedtocf.Pp. lothagamensiscomes fromBPRP#76, and is confidently dated, the other comes fromBPRP#29 at Cheboit, so the same reservations applywith regard to itsage. Pp. lothagamensis at Lothagam, its type site, is found in both theLower and Upper parts of the Nawata Formation, and hence spansabout 7.4 Ma to younger than 6.5 Ma (McDougall and Feibel, 2003).

Systematic paleontology

Family Cercopithecidae Gray, 1821Subfamily Colobinae Jerdon, 1867Colobinae gen. et sp. indet. B

Referred material

KNM-TH 36742, KNM-LU 344 (Table 1).

Horizon

Lukeino Formation outcrops, Tugen Hills, Kenya.

Localities/Sites

Lukeino Formation KNM-TH 36742 comes from Kapcheberek,site BPRP#76A, a horizon which is very well limited by radiometricestimations and paleomagnetic calibration to between 5.88 and5.70Ma (Deino et al., 2002). KNM-LU 344 comes fromAragai, site 2/228, near the base of the Formation, maybe around 6.1 Ma.

Descriptions

KNM-TH 36742 This tooth is a lower right M1 or M2 (most likely anM2), preserving only the crown. The specimen is clearly smallerthan the Colobine sp. A described from the Mpesida Beds (Table 2),which necessitates a separate species designation. The mean esti-mated bodymass of the population represented by this specimen isw6e7 kg based on female and male M2 area regressions for colo-bine taxa, respectively (Delson et al., 2000). Similar to the speci-mens from the Mpesida Beds, the molar is modern in appearance,with a low crown, a very tall median lingual notch, and very talllingual cusps (Figs. 5a and 6a). The NH/NR ratio is again very high,

Figure 7. KNM-LU 344, Colobinae gen. et sp. indet. B, right astragalus from the Lukeino Formleft to right, Papio hamadryas cynocephalus, C. albogularis, Colobinae gen. et sp. indet. B KNlateral suppression of the facet for the lateral calcaneonavicular ligament in KNM-LU 344 aspecimens. Also note the swelling distal to the groove for the m. flexor tibialis in the colobigroove for the tendon of the m. flexor tibialis in KNM-LU 344 and the Colobus specimen an

Figure 6. Cercopithecoid dental remains from the Lukeino Formation, Tugen Hills,Kenya. a) Colobinae gen. et sp. indet. B KNM-TH 36742, left: buccal view; right: lingualview. b) cf. Colobinae KNM-TH 26706, left: buccal view; right: lingual view. c) occlusalview of cercopithecoid specimens. left: Cercopithecidae gen. et sp. indet. KNM-LU 940;right: Cercopithecidae gen. et sp. indet. KNM-LU 771. Scale bar ¼ 1 cm.

C.C. Gilbert et al. / Journal of Human Evolution 59 (2010) 465e483 475

similar to modern colobines, and suggests a high degree of folivory(Table 2; Fig. 6a). Mesial and distal shelves are both present andmoderate in size, with the distal shelf appearing slightly moreexpanded mesiodistally than the mesial shelf (Fig. 5). The medianbuccal cleft is moderately deep, and the cusps are not stronglycolumnar for a colobine. Mesial and distal buccal clefts are present,but these are small.KNM-LU 344 This specimen is a well-preserved right astragaluswith some damage on the dorsal portion of its head (Fig. 7a). It isslightly smaller than a modern Colobus female, or approximatelythe size of the modern Sykes’ monkey (Cercopithecus albogularis).The bone is easily identified as a colobine by at least two diagnosticfeatures. First, as is typical for colobines, KNM-LU 344 displayslateral suppression of the facet for the lateral calcaneonavicularligament (Fig. 7b; see Strasser, 1988). Second, on the proximalportion of the astragalus, there is a distinctive groove for the m.flexor tibialis (Fig. 7c; see Strasser and Delson, 1987; Strasser, 1988).Compared to modern colobines, the development of the groove, aswell as the “swelling” distal to the groove, on KNM-LU 344 is mostsimilar to that seen in African rather than Asian colobines (Fig. 7band c; see Strasser, 1988).

In addition to diagnostic features such as the m. flexor tibialisgroove and the lateral suppression of the facet for the lateral calca-neonavicular ligament, there are other features of the astragaluswhichsuggest it belonged to an arboreal animal. For example, the proximaltalo-calcaneal facet is strongly curved, a feature found in both arboreal

ation, Tugen Hills, Kenya, compared to modern African cercopithecoid astragali. FromM-LU 344, Colobus abyssinicus kikuyensis. a) superior view; b) inferior view; note thend the Colobus specimen and the continuous articulation found in the cercopithecinene specimens, indicated by an asterisk; c) medial view; note the presence of a narrowd the absence of this groove in the two cercopithecine specimens. Scale bar ¼ 1 cm.

Table 7Astragalar measurements.

MeasurementTaxon (Specimen number)

Colobinae gen. et sp. indet.(KNM-LU 344)

Max PD length parallel to trochlear groove 23.2PD length of trochlea at its midpoint 11.4ML width of trochlea at its midpoint 8.6Max diameter of head XMin diameter of head XDepth of the trochlear groove ShallowMax astragalar width (14.3)Min width of the neck 6.9Max length of posterior calcaneal facet 11.1Max breadth at the center of the posterior

calcaneal facet5.9

Notes: Values in parentheses represent estimates. X ¼ unavailable measurement.ML ¼ medial-lateral; PD ¼ proximal-distal.

C.C. Gilbert et al. / Journal of Human Evolution 59 (2010) 465e483476

colobines aswell as arboreal guenons amongmodern cercopithecines.Standard astragalar measurements are given in Table 7.

Remarks

The tooth KNM-TH 36742 bears similarities to a modern colo-bine tooth, but its small size and lack of distinctive characteristicspreclude its assignment to any known colobine taxon. Comparisonswith the Late Miocene small colobine taxon (Colobinae sp. B) fromLemudong’o (Hlusko, 2007) indicate that KHM-TH 36742 is slightlysmaller and narrower than this taxon. Colobinae sp. B from Lukeinois relatively narrower and completely out of the known size rangeof Kuseracolobus from the Middle Awash (Frost, 2001; Frost et al.,2009). Likewise, it is much smaller than colobine specimens fromMenacer and Nkondo. KNM-TH 36742 is most similar in size andoverall measurements to Colobinae sp. A from the Lower NawataFormation at Lothagam (w7.4e6.7 Ma) (Leakey et al., 2003;McDougall and Feibel, 2003), but the specimen from Lukeinodisplays small mesial and distal buccal clefts while the lonemandibular specimen from Lothagam does not. Further material isneeded to determine if the material from Lothagam is conspecificwith Colobinae gen. et sp. B from Lukeino.

The right astragalus KNM-LU 344 is not associated with anycraniodental material, precluding any assignment to previouslynamed colobine taxa. However, its size aligns with the estimatedbody mass of KNM-TH 36742, and given that both are found in theLukeino formation, it is most parsimonious at this time to assignthem to the same taxon. In the absence of any definitive evidence ofmultiple colobine taxa, we therefore assign both specimens toColobinae gen. et sp. B. While this arrangement avoids the problemof taxonomic inflation, we admit that the lack of association makesit difficult to exclude the possibility that we are instead maskingtaxonomic diversity. Future fossils from the Lukeino formation offera basis for separating these specimens.

The distinctive arboreal colobine features and, more specifically,African colobine features make the KNM-LU 344 astragalus animportant specimen for understanding colobine evolution.Previous authors have argued that the earliest colobines weresemi-terrestrial or terrestrial, in contrast to the modern taxawhich are almost exclusively arboreal (e.g., Andrews, 1982; Harriset al., 2003; Leakey et al., 2003). Based on paleoecological recon-structions as well as fragmentary postcranial remains at Lemu-dong’o, Hlusko (2007) suggested that the most parsimoniousinterpretation of the available evidence is to assume that thecolobine radiation was mostly arboreal by the Late Miocene. Theastragalus from Lukeino bolsters this last suggestion with moredefinitive anatomical evidence of arboreality in a Late Miocenecolobine. In addition, the distinctive m. flexor tibialis groove found

on the proximal portion of KNM-LU 344 and the associated“swelling” distal to the groove may provide some of the earliestdefinitive evidence of the modern African colobine radiation aswell as the modern African colobine arboreal morphotype. The LateMiocene age of KNM-LU 344 is in-line with molecular estimates forthe divergence of the modern African and Asian colobine radiationsw12.3e8.2 Ma (Raaum et al., 2005; Ting, 2008), as well as thedivergence of the modern African colobine clades at w8.7e6.3 Ma(Ting, 2008). Therefore, beyond being an African colobine, KNM-LU344 could conceivably belong to either the Colobus or Procolobus/Piliocolobus clades. In any case, KNM-TH 36742 and KNM-LU 344further document a radiation of Late Miocene colobines that appearto contain some of the earliest relatives of modern taxa.

Systematic paleontology

cf. Colobinae

Referred material

KNM-TH 26706 (Table 1).

Horizon

Lukeino Formation outcrops, Tugen Hills, Kenya.

Localities/Sites

Lukeino Formation, Cheboit, site BPRP#29, near the base of theformation, probably around 6.1 Ma (but see introductory remarksabove).

Description

KNM-TH 26706 This specimen, an upper right premolar, possiblya P4, is relatively unworn, and exhibits both a paracone and pro-tocone. The specimen is relatively small, being slightly smaller thanmodern C. guereza but similar in size to Colobus polykomos (seeTable 3 for measurements). The protocone is approximately thesame height as the paracone, a feature found most often amongmodern Asian colobine premolars, and both the paracone andprotocone are relatively tall cusps. A small mesial shelf is present,and it appears strongly angled or sloping mesially in lateral view(Figs. 5b and 6b). A larger distal shelf is also present and slopesdistally in lateral view (Figs. 5b and 6b).

Remarks

The right premolar KNM-TH 26706 lacks diagnostic features thatwould allow confident taxonomic placement. It appears to comparemost favorablywith colobines rather than cercopithecines, but this isuncertain. While cercopithecine premolars occasionally displaysimilarly sized cusps, this condition is most likely to be seen amongcolobine taxa and among extant Asian colobine taxa in particular.However, some extant African colobine teeth also show a similarlysized paracone and protocone. If KNM-TH 26706 is a colobine, it isalmost certainly a different taxon than Colobinae gen. et sp. indet. BfromtheLukeinoFormation sinceKNM-TH26706 represents aP4 thatis larger in size than the lower molar representing Colobinae sp. B.

Systematic paleontology

Subfamily Cercopithecidae Gray, 1821Tribe Papionini Burnett, 1828Genus Parapapio Jones, 1937

(¼ or including: Papio Erxleben, 1777: Leakey and Leakey, 1976, inpart. Cercocebus Geoffroy, 1812: Hopwood, 1936, in part. Papio(Simopithecus) Andrews, 1916: Dietrich, 1942. Brachygnathopithecus

C.C. Gilbert et al. / Journal of Human Evolution 59 (2010) 465e483 477

Kitching, 1952, in part. Papio (Parapapio): Delson, 1975, in part.Papionini gen. et. sp. indet. B. Leakey and Leakey, 1976).

Type species

Parapapio broomi Jones, 1937

Other included species

Pp. ado Hopwood, 1936; Pp. jonesi Broom, 1940; Pp. whiteiBroom, 1940; Pp. lothagamensis Leakey et al., 2003.

Generic diagnosis

This diagnosis largely follows that of Freedman (1957), Szalayand Delson (1979), Frost and Delson (2002), and Leakey et al.(2003), with the addition of features discussed by Gilbert (2007,2008). Parapapio is a medium to large sized papionin, ranging inestimated body mass from the smallest and oldest species, Pp.lothagamensis (w9 kg; Delson et al., 2000; Leakey et al., 2003), tothe largest species Pp. whitei (w19e25 kg; Delson et al., 2000).Parapapio is distinguished cranially from Papio, Mandrillus, Ther-opithecus, Pliopapio, Dinopithecus, and Gorgopithecus by the lack ofan anteorbital drop, a feature it shares with Procercocebus, Cerco-cebus, Lophocebus and most specimens of Macaca. Also in lateralprofile, Parapapio exhibits straight to slightly upturned nasal bones.Parapapio is distinguished most easily from other similar sizedAfrican papionins (Subtribe: Papionina, sensu Strasser and Delson,1987), except Pliopapio, by the lack of any definitive facial fossae,both on the maxilla as well as the mandible. Likewise, maxillaryridges on Parapapiomales are typically poorly developed or absent.Maxillary and mandibular premolars are relatively small, similar tomany African papionin taxa but distinct from Cercocebus, Man-drillus, and Procercocebus.

Where the frontal is preserved, Parapapio specimens usuallyexhibit a weak brow ridge and a glabella that is not prominent,although male Pp. jonesi specimens may be an exception. Anophryonic groove is not usually present in individuals of Parapapio,and it is shallow when it exists. The temporal lines of Parapapiomales are typically pinched, a feature it shares with most Africanpapionins with the exception of Cercocebus, Mandrillus, and Pro-cercocebus. The nuchal crest of Parapapio males is variable inmorphology, similar to the condition seen among the variousspecies of Macaca. Also similar to Macaca, the inferior petrousprocess, when preserved, appears variable in position.

Pp. lothagamensis Leakey et al., 2003

Holotype

KNM-LT 23091, a male mandible missing both rami butpreserving both corpora as well as left I2, C1, P4-M3 and right I2, C1crown, P4-M3.

Referred material

See Leakey et al. (2003).

Horizon

Late Miocene outcrops of the Nawata Formation, Lothagam,Kenya.

Specific diagnosis

As described by Leakey et al. (2003), Pp. lothagamensis isdistinguished from other species of Parapapio by its small size, longand sloping mandibular symphysis, relatively broad P3s, and dP4sthat are not fully bilophodont and lack a distal transverse crest.There is slight sexual dimorphism in the size of the teeth. Pp.

lothagamensis also differs from other species of Parapapio and issimilar to Victoriapithecus in a number of features such as thecommon occurrence of ametaconid on the P3, an obliquely orientedP4, a buccolingually wide M1, distally constricted M3s with thevariable absence of a distal shelf, and dp4s with variable expressionof both a crista obliqua and a hypoconulid.

cf. Pp. lothagamensis Leakey et al., 2003

Referred material

KNM-TH 36744, KNM-LU 861 (Table 1).

Horizon

Lukeino Formation outcrops, Tugen Hills, Kenya.

Localities/Sites

Lukeino Formation KNM-TH 36744 comes from Kapcheberek,site BPRP#76, and is dated to 5.88e5.70 Ma (Deino et al., 2002);KNM-LU 861 comes from Cheboit, BPRP#29, near the base of theFormation and therefore dated to around 6.1 Ma (but see intro-ductory remarks above).

Descriptions

KNM-TH 36744 This upper right M3 is slightly smaller in size, butvery similar in shape and morphology to Pp. lothagamensis M3sfrom Lothagam (Figs. 5a and 8). The distal loph is buccolinguallyrestricted and narrow compared to the mesial loph, as is often seenin Pp. lothagamensis and other small-to-medium sized papionins(Figs. 5a and 8c). KNM-TH 36744 exhibits a relatively compressedmesial shelf and no real distal shelf, consistent with other Pp.lothagamensis specimens (Fig. 8c). The tooth is moderately worn,with low, rounded cusps, a low NH/NR index, and a moderate tohigh degree of basal flare (Table 2; Fig. 8a and c). Themedian lingualcleft is moderately developed and narrow. A shallowmesial lingualcleft and a small distal lingual cleft appear to be present as well.KNM-LU 861 This specimen is an upper right molar crown, rep-resenting either an M1 or an M3. Based on the molar dimensions ofPp. lothagamensis from Lothagam, it is most likely an M1 (Table 2;Figs. 5a and 8b). The cusps are low and rounded and the toothdisplays a moderate to high degree of basal flare (Figs. 5a and 8b).Similar to KNM-TH 36744, the tooth is slightly smaller than thesample of M1s from Lothagam. KNM-LU 861 displays a short,compressed mesial shelf and very short distal shelf as well(Fig. 5a). Because the molar is heavily worn, it is not clear if thesmall size of the mesial and distal shelves is due to wear, as Pp.lothagamensis M1s and M2s typically exhibit larger mesial anddistal shelves (Leakey et al., 2003). There is a moderate to shallowmedian lingual cleft.

Remarks

Both KNM-TH 36744 and KNM-LU 861 resemble Pp. loth-agamensis in the overall dimensions and general morphology of themolars, particularly KNM-TH 36744 (see Fig. 8e). In addition, thesespecimens differ from contemporaneous papionin taxa such asPliopapio in Ethiopia and Macaca sp. in North Africa (Macaca libyca,Menacer, and As Sahabi). Compared to Pliopapio and cf. Pliopapiofrom the Sagantole and Adu-Asa Formations in the Middle Awash(Frost, 2001; Frost et al., 2009), KNM-TH 36744 and KNM-LU 861are lower-crowned, smaller in size, relatively wider compared totheir length, display greater levels of flare, and display reduceddistal shelves (particularly for the M3). Compared to specimensassigned to Macaca sp. from northern Africa (Arambourg, 1959;Benefit et al., 2008), the Tugen Hills specimens are smaller inoverall size, display M1s that are just as long as they are wide, anddisplay M3s with a greater reduction of the distal loph. In total,

Figure 8. cf. Parapapio lothagamensis dental remains from the Lukeino Formation, Tugen Hills, Kenya. a) KNM-TH 36744, left: lingual view; right: buccal view; b) KNM-LU 861, left:lingual view; right: buccal view; c) Comparison of Pp. lothagamensis from Lothagam to cf. Parapapio lothagamensis from the Lukeino Formation in occlusal view, left: Pp. loth-agamensis KNM-LT 419; right: cf. Pp. lothagamensis KNM-TH 36744. Scale bar ¼ 1 cm.

C.C. Gilbert et al. / Journal of Human Evolution 59 (2010) 465e483478

a tentative assignment to Pp. lothagamensis is best supported giventhe available evidence. However, there are a couple differencesfrom the Pp. lothagamensis population from Lothagam. First, thetwomolars are slightly smaller than the population from Lothagam.Given the current specimens, the estimated bodymass of the TugenHills population (from Delson et al., 2000 cercopithecine regres-sions) was between w7.5 and w10 kg. Second, KNM-LU 861appears to have smaller mesial and distal shelves than is typical forPp. lothagamensis M1s, but again, this could simply be due to wear.More material is necessary to confirm that the population atLukeino is indeed Pp. lothagamensis, or perhaps demonstrate thatthe Lukeino fossils represent a slightly smaller and closely relatedtaxon.

Systematic paleontology

Cercopithecidae gen. et sp. indet.

Referred material

KNM-LU 771, KNM-LU 860, KNM-LU 939, KNM-LU 940 (Tables 1and 4).

Horizon

Lukeino Formation outcrops, Tugen Hills, Kenya.

Localities/Sites

Lukeino Formation KNM-LU 771 comes from Cheboit, siteBPRP#29, near the base of the formation, probably around 6.1 Ma(but see introductory remarks above). KNM-LU 860 and KNM-LU939 come from Kapcheberek, 2/225 in the EAGRU style scheme,stratigraphically lower than BPRP#76 and probably about 5.8 Ma.KNM-LU 940 comes from Kapcheberek, site BPRP#76,5.88e5.70 Ma (Deino et al., 2002).

Descriptions

KNM-LU 771 This specimen appears to be an upper left malecanine. As cercopithecid canines cannot be confidently attributedto subfamily (Hill and Gundling, 1999), we assign the specimen toCercopithecidae gen. et sp. indet. The base of the tooth is expandedand triangular in cross-section, with a strong distal shelf or flange(Figs. 5c and 6c). A mesial groove is present and extends onto theroot of the tooth.

Figure 9. KNM-LU 860, Cercopithecidae gen. et sp. indet. proximal left humerus fromthe Lukeino Formation, Tugen Hills, Kenya. a) anterior view; b) medial view; c)posterior view; d) lateral view; e-f) superior views. Scale bar ¼ 1 cm.

C.C. Gilbert et al. / Journal of Human Evolution 59 (2010) 465e483 479

KNM-LU 940 This is an upper right male canine tooth. Similar to themolars assigned to cf. Pp. lothagamensis from Lukeino, KNM-LU 940is smaller than the described male canine from Lothagam. A mesialsulcus is present and extends down onto the root. Compared withKNM-LU 771, there is no large expansion at the base of the toothand no distal shelf or flange (Figs. 5c and 6c; see Table 3 formeasurements).KNM-LU 860 This proximal left humerus preserves the humeralhead and neck along with a portion of the proximal humeral shaft(see Table 4 for standard humeral measurements). In pure sizecomparisons, KNM-LU 860 exceeds the size of most Victoriapithecushumeri, but is smaller than most Pp. lothagamensis humeri (Table4). The proximal humeral shaft is crushed approximately 14 mmbelow the head. The humeral head is positioned approximatelyeven with or slightly above the greater tuberosity. The greatertuberosity is larger than the lesser tuberosity (Table 4; Fig. 9), butthe lesser tuberosity is relatively large compared to that reportedfor Victoriapithecus and Pp. lothagamensis. There is a shallowbicipital groove between the tuberosities. Overall, the generalmorphology of this proximal humerus suggests a semi-terrestrialmonkey that still retains arboreal capabilities.KNM-LU 939 This specimen is a complete left calcaneus with smallareas of wear, damage, or both wear and damage. As Strasser (1988)

and Leakey et al. (2003) point out, cercopithecoid calcanei lackobvious diagnostic features among subfamilies. Therefore, weattribute this specimen to Cercopithecidae gen. et sp. indet. KNM-LU 939 exhibits a relatively narrow heel, and the overall outline ofthe specimen is most similar to a colobine (Fig. 10). Inferior to thesustentaculum tali, the groove for the tendon of the flexor hallucislongus is quite deep and obvious; the groove is relatively long andcurves around the sustentaculum tali to the anterior portion of thecalcaneus (see Fig. 10b). A deep groove appears to be morepronounced in arboreal guenons and colobines than in terrestrialcercopithecids such as Papio (Fig. 10b). Standard calcanealmeasurements of KNM-LU 939 compared to cercopithecids fromLothagam are given in Table 8.

Remarks

Both KNM-LU 860 and KNM-LU 939 are unassociated post-cranial specimens lacking any documented derived features to linkthem definitively with the modern cercopithecid subfamilies.Therefore, we conservatively and tentatively assign them Cercopi-thecidae gen. et sp. indet. However, both specimens are completeenough to infer something about their size and locomotor adap-tations, and these adaptations are at least suggestive. Specifically,KNM-LU 860 includes an interesting mixture of features typicallyseen in both arboreal and terrestrial cercopithecids. The position ofthe humeral head, evenwith, or slightly above, the greater tubercle,may reveal a significant arboreal component to the animal’s loco-motor repertoire. The relatively large size of the lesser tuberosity,although smaller than the greater tuberosity, is a feature moreoften associated in terrestrial cercopithecines (Fleagle and Simons,1982), although Harrison (1989) notes that there is a great degree ofoverlap. Among cercopithecoids, the shallow bicipital groove ismost typical of arboreal quadrupeds. These features suggesta quadrupedal animal with semi-terrestrial capabilities, retainingan arboreal component to the locomotor repertoire as well. Thesefeatures and their inferred locomotor repertoire are seen in bothVictoriapithecus and Pp. lothagamensis. Given that KNM-LU 860 issmaller than humeri described for Pp. lothagamensis but retainsfeatures seen in this species, it is possible that it belongs with KNM-TH 36744 and KNM-LU 861 in cf. Pp. lothagamensis. In fact, this isprobably the most likely scenario given the available evidence.However, since KNM-LU 860 is unassociated with any craniodentalremains, and because semi-terrestrial colobine taxa are known toexist in the Late Miocene and Early Pliocene, it is impossible toattribute this proximal humerus with certainty.

Similarly, KNM-LU 939 is unassociated and calcanei are sug-gested to be insufficiently diagnostic as to their more specific attri-bution (Strasser, 1988; Leakey et al., 2003). However, the overallshape and suite of features compares most favorably to an arborealcolobine, approximately the size of a modern Colobus female orslightly smaller. The size and locomotor repertoire suggested for thisspecimenwould be consistent with the inferred size and locomotorcapabilities of colobine specimensKNM-TH36742 andKNM-LU344,both also from the Lukeino Formation (see above). More associatedmaterial is needed to confirm the affinities of this specimen.

As mentioned earlier, the range of variation in cercopithecineand colobine canines makes it difficult to confidently assign iso-lated cercopithecid canines below the family level. While nodefinitive subfamily assignments can be made regarding KNM-LU771 and KNM-940, we note that the base of the canine KNM-LU 771appears most similar to the modern colobines we examined. Inaddition, given that the size of KNM-LU 940 is consistent with theteeth assigned to cf. Pp. lothagamensis, it is possible that KNM-LU940 is a part of the same population. The recovery of additionalspecimens in association with other cercopithecoid material isnecessary before assigning these canines to lower taxonomic levels.

Table 8Calcaneal measurements.

MeasurementTaxon(Specimen number)

Taxon(Specimen number)

Taxon(Specimen number)

Taxon(Specimen number)

Cercopithecidae gen. et sp.indet. (KNM-LU 939)

Parapapio lothagamensis (Range fromKNM-LT 24125, 26402, 28575)

Colobinae sp.(KNM-LT 30610)

Colobinae gen. et sp.indet (KNM-LT 26392)

Max PD length 34.3 34.5e36.7 X XAnterior PD length from the center of the posterior

astragalar facet to the cuboid facet15.8 12e14.2 X X

Posterior PD length from the center of the posteriorastragalar facet to the heel process

16.6 21.2e23.3 16.6 17.1

Posterior astragalar facet PD length 9.1 9.6e11.1 7.9 7.6Width of posterior astragalar facet (7.6) X X XMax width from the middle of the articular facet 16.5 X X XMax height of the heel process 14.8 X X XMax width of the heel process 10.2 X X X

Notes: Values in parentheses represent estimates. X ¼ unavailable measurement. PD ¼ proximal-distal. Measurements for Parapapio lothagamensis and Lothagam colobinesfrom Leakey et al. (2003).

Figure 10. KNM-LU 939, Cercopithecidae gen. et sp. indet. left calcaneus from the Lukeino Formation, Tugen Hills, Kenya, compared to modern African cercopithecoid calcanei.From left to right, Papio hamadryas cynocephalus, Cercopithecus albogularis, Cercopithecidae gen. et sp. indet. KNM-LU 939, Colobus abyssinicus kikuyensis.a) superior view; b) medial view; note the deep groove for the tendon of the flexor hallucis longus in the arboreal taxa, C. albogularis, KNM-LU 344 and C. abyssinicus kikuyensis; c)posterior view; d) anterior view. Scale bar ¼ 1 cm.

C.C. Gilbert et al. / Journal of Human Evolution 59 (2010) 465e483480

Discussion and conclusions

In total, the Tugen Hills records a minimum of 8 cercopithecoidtaxa in the time period represented between the Muruyur Bedsand the Lukeino Formation (Table 1). In the Muruyur Beds, at leastone victoriapithecid species is present and possibly two. From theNgorora Formation, there appears to be a victoripithecid along

with two colobine taxa: the previously described M. tugenensis(Benefit and Pickford, 1986) and at least one other taxon that is notrepresented well enough to be formally named (Benefit andPickford, 1986; Rossie et al., in prep). The Mpesida Beds appearto contain a single colobine taxon. Finally, the Lukeino Formationcontains at least two colobines and a cercopithecine tentativelyassigned here to Pp. lothagamensis. Overall, this section of the

C.C. Gilbert et al. / Journal of Human Evolution 59 (2010) 465e483 481

Tugen Hills seems notable for the prevalence of colobine taxa,which seems correlated with the presence of forests and/orwoodlands in the Ngorora Formation (Jacobs and Kabuye, 1987;Jacobs and Winkler, 1992; Hill et al., 2002), Mpesida Beds(Jacobs and Deino, 1996; Kingston et al., 2002), and LukeinoFormation (see above).

The cercopithecoids found in the Tugen Hills stratigraphicsection represented by the Muruyur Beds to the Lukeino Forma-tion document an important time period in cercopithecoidevolution, some of it not represented elsewhere. The major cer-copithecid subfamilies (Colobinae, Cercopithecinae) are thoughtto have diverged somewhere between 20 and 13.5 Ma (e.g.,Raaum et al., 2005; Ting, 2008) and the split between the majortribes within each subfamily are estimated to have taken placesoon after, between 12.3 and 8.2 Ma (Raaum et al., 2005; Ting,2008), within the time encompassed by the deposits in theTugen Hills. As such, it might be expected that members of themajor tribes be represented in the fossil material described here.These expectations appear to be confirmed by the morphologicalevidence.

The colobines found in the Mpesida Beds and LukeinoFormation, in particular, perhaps represent some of the earliestmembers of the African colobine (Tribe: Colobina) radiation. Thecolobine taxa found in the Mpesida Beds and the LukeinoFormation are associated with forested environments and exhibitfully modern, highly folivorous dental morphotypes. Thiscontrasts with other early colobine taxa such as Microcolobus,Libypithecus, and Mesopithecus, which display molars with rela-tively lower cusps than modern colobine taxa (Benefit andPickford, 1986). All of the colobine molars described from theMpesida Beds and Lukeino Formation are fully modern in theheight of their cusps, as indicated by the NH/NR ratios (Table 2).While all of the colobine molars and premolars described herepossess features characteristic of modern colobines and modernAfrican colobines, it should also be pointed out that there are nodefinitive dental synapomorphies linking the Tugen Hills speci-mens with modern African colobines to the exclusion of modernAsian colobines. It is entirely possible that they simply representa morphotype close to the common ancestor of all crown colo-bines. However, if Mesopithecus is related to modern Asian colo-bines, as hypothesized by some (e.g., Jablonski, 1998, 2002;Benefit, 1999; Groves, 2000), then this hypothesis becomes lesslikely as the Tugen Hills dental specimens are more modern/derived in appearance than Mesopithecus.

In contrast to the dental material, the colobine astragalus foundin the Lukeino Formation is significant in that it displays derivedfeatures linking it exclusively with the Colobina. Moreover, some ofthe same derived features suggest an arboreal locomotor reper-toire. Comparisons with other fossil colobine astragali, particularlyParacolobus chemeroni (KNM-BC 3), Paracolobus mutiwa (KNM-WT16827), and cf. Cercopithecoides (KNM-ER 30320) reveal that KNM-LU 344 is more derived in these aspects of its postcranialmorphology than these later and better known taxa. None of thesePlio-Pleistocene specimens appears to display the restricted groovefor the m. flexor tibialis or the “swelling” distal to the groove whichcharacterize living African colobines. Instead, these taxa seem tohave a shallow and wide groove (if any) with no distal swelling,most similar to the morphology seen among living Asian colobineand/or cercopithecine taxa. Among fossil colobine astragali exam-ined, only the newly named Colobus freedmani (KNM-ER 5896;Jablonski and Leakey, 2008) appears to display similar features asthose found in KNM-LU 344 and modern African colobines, andeven in this case the groove and the distal swelling are not as well-preserved or are simply less pronounced. Given that the estimatedage of KNM-LU 344 (w6.1 Ma) places it subsequent to the

hypothesized split within the Colobina between the Colobus andProcolobus/Piliocolobus clades (w8.7e6.3 Ma), it is possible that thisspecimen represents one of the earliest members of the modernAfrican colobine clades.

Previous authors have hypothesized that early colobines weresemi-terrestrial, and that modern colobine arboreal adaptationsare a recent development in both the Colobina and Presbytina(Andrews, 1982; Harris et al., 2003; Leakey et al., 2003). Incontrast, the evidence here suggests that early colobine taxa, atleast early African colobine taxa, were committed to an arborealway of life. An analysis by Hlusko (2007) on the Late Miocenemonkeys from Lemudong’o also suggested that the colobineradiation has long been adapted to life in the trees. Frost et al.(2009) recently described cercopithecid postcrania tentativelyattributable to a Late Miocene population of Kuseracolobus anddetermined that these specimens indicated a relatively arborealmode of locomotion for this colobine taxon as well. Perhaps mostconvincingly, new postcrania discovered by Nakatsukasa et al.(2010) strongly suggest that the earliest colobine taxon known,Microcolobus, was also an arboreal animal. In total, the mostparsimonious hypothesis based on the current evidence is thatearly colobines and early African colobines were mainly arborealand that the semi-terrestrial Late Miocene and Plio-Pleistocenecolobine taxa were secondarily derived in their locomotoradaptations.

The papionin specimens found in the Lukeino Formation arealso important in that they potentially document the presence ofa taxon previously known only from Lothagam. Benefit et al. (2008)hinted that papionin specimens found at As Sahabi might alsobelong to a species similar to Pp. lothagamensis, but they consideredan assignment to the genus Macaca equally likely. Given theconsistent difference in size between the Tugen Hills papioninpopulation and the Lothagam papionin population, it is entirelypossible that the fossils here attributed to cf. Pp. lothagamensisrepresent a distinct species. However, the evidence is too frag-mentary, at this point, to be certain.

What is interesting, however, is that the Lukeino cercopithecidrepresented by the proximal humerus appears to have beena semi-terrestrial animal with a significant arboreal component toits locomotor repertoire. Therefore, both the indeterminate cer-copithecid and colobine postcranial specimens from the LukeinoFormation suggest the presence of trees and would therefore seemto indicate a wooded environment in the Tugen Hills and RiftValley during this time period. In addition to evidence for forestedenvironments from macro-flora alluded to above (see MpesidaBeds), an independent body of information from carbon isotopessuggests forested or wooded, rather than open conditionsthroughout the Tugen Hills sequence (Kingston et al., 1994, 2002;Kingston, 1999). Although there is some evidence, particularlyafter about 7 Ma, of more or less exclusive C4 grazers (Morganet al., 1994), it appears that open grasslands never dominatedthis part of the Rift Valley. Suitable habitats for both arborealfrugivorous and folivorous primates would have been availablethroughout the time period discussed here, a time when it isbelieved that hominins were diverging from the lineage leading tochimpanzees.

Finally, the distal humerus of V. macinnesi from Muruyurpossesses a proximally extended supinator crest, a feature alsofound in papionin monkeys (Fleagle andMcGraw, 2002). Therefore,a semi-terrestrial morphotype including both manual terrestrialforaging as well as vertical climbing behaviors appears to have beenprimitive for papionin monkeys. The more distally restrictedsupinator crest seen in taxa such as Papio, Lophocebus, and extantTheropithecus is supported as a derived condition (Fleagle andMcGraw, 2002).

C.C. Gilbert et al. / Journal of Human Evolution 59 (2010) 465e483482

Acknowledgements

This research forms part of the work of the Baringo Paleonto-logical Research Project (BPRP), based at Yale University, andcarried out in collaboration with the National Museums of Kenya.We thank the Government of the Republic of Kenya for permissionto carry out research (Permit MOHEST/13/001/C 1391/issued toAH), and permission to excavate from the Minister for Home Affairsand National Heritage. While in Kenya, this study was improvedenormously from discussions with Brenda Benefit and, upon returnto the US, the RHOI Cercopithecoid Analytical Working Group alsoprovided helpful comments and insights. Steve Leigh, Eric Delson,and three anonymous reviewers provided comments thatimproved this manuscript greatly. Eric Delson kindly providedimportant papers during the study of the Tugen Hills cercopithe-coid material. We are also most grateful to John Kingston, whoprovided us with Fig. 1. This study was kindly supported by a Don-nelley Fellowship from the Yale Institute for Biospheric Studies toCCG, and a Yale University Enders Fellowship and the Brian Pat-terson Award for Field Research from the Society of VertebratePaleontology to EDG. BPRP has been supported by grants to AHfrom NSF (currently #BSC-071137 with John Kingston and AlanDeino), and also from Clayton Stephenson, and the Schwartz FamilyFoundation. We are grateful to Bonface Kimeu for his considerablehelp with field and laboratory work. The National Museums ofKenya provided valuable logistical support, and we thank particu-larly Emma Mbua, the Head of the Division of Earth Sciences.Finally, CCG thanks David McKenzie and Sarah Band for theirwonderful and enduring hospitality while in Nairobi.

References

Aguirre, E., Leakey, P., 1974. Nakali: nueva fauna de Hipparion del Rift Valley deKenya. Estudios Geol. Inst. Invest. Geol. Lucas Mallada 30, 219e227.

Andrews, C.W., 1916. Note on a new baboon (Simopithecus oswaldi, gen. et sp. nov.)from the Pliocene of British East Africa. Ann. Mag. Nat. Hist. 18, 410e419.

Andrews, P., 1982. Ecological polarity in primate evolution. Zool. J. Linn. Soc. Lond.74, 233e244.

Arambourg, C., 1959. Vertébres Continentaux du Miocène Supérieur de l’Afrique duNord. Publications Service Carte Geol. Algérie (Nouvelle Série). Paléontologie 4,5e159.

Behrensmeyer, A.K., Deino, A.L., Hill, A., Kingston, J.D., Saunders, J.J., 2002. Geologyand geochronology of the Middle Miocene Kipsaramon site complex, MuruyurBeds, Tugen Hills, Kenya. J. Hum. Evol. 42, 11e38.

Benefit, B.R., 1987. The Molar Morphology, Natural History, and Phylogenetic Posi-tion of the Middle Miocene Monkey Victoriapithecus. Ph.D. Dissertation, NewYork University.

Benefit, B.R., 1993. The permanent dentition and phylogenetic position of Victor-iapithecus from Maboko Island, Kenya. J. Hum. Evol. 25, 83e172.

Benefit, B.R., 1999. Victoriapithecus, the key to Old World monkey and catarrhineorigins. Evol. Anthropol. 7, 155e174.

Benefit, B.R., McCrossin, M.L., Boaz, N.T., Pavlakis, P., 2008. New fossil cercopithe-coids from the Late Miocene of As Sahabi, Libya. Garyounis Sci. Bull. Spec. Issue5, 265e282.

Benefit, B.R., McCrossin, M.L., 1990. Diet, species diversity and distribution ofAfrican fossil baboons. Kroeber Anthropol. Soc. Pap. 71e72, 77e93.

Benefit, B.R., Pickford, M., 1986. Miocene fossil cercopithecoids from Kenya. Am. J.Phys. Anthropol. 69, 441e464.

Birchette, M.G., 1982. The Postcranial Skeleton of Paracolobus chemeroni. Ph.D.Dissertation, Harvard University.

Broom, R., 1940. The South African Pleistocene cercopithecoid apes. Ann. Transv.Mus. 20, 89e100.

Burnett, G.T., 1828. Illustrations of the Manupeda of apes and their allies: being thearrangement of the Quadrumana or Anthropomorphous beasts indicated in theoutline. Quart. J. Sci. Lit. Art 26, 300e307.

Deino, A.L., Tauxe, L., Monaghan, M., Drake, R., 1990. 40Ar/39Ar age calibration of thelitho- and paleomagnetic stratigraphies of the Ngorora Formation, Kenya. J.Geol. 98, 567e587.

Deino, A.L., Tauxe, L., Monaghan, M., Hill, A., 2002. 40Ar/39Ar geochronology andpaleomagnetic stratigraphy of the Lukeino and lower Chemeron succession atTabarin and Kapcheberek, Tugen Hills, Kenya. J. Hum. Evol. 42, 117e140.

Delson, E., 1975. Evolutionary history of the Cercopithecidae. In: Szalay, F.S. (Ed.),Approaches to Primate Paleobiology. vol. 5: Contributions to Primatology.Karger, Basel, pp. 167e217.

Delson, E., Dean, D., 1993. Are Papio baringensis R. Leakey, 1969, and P. quad-ratirostris Iwamoto, 1982, species of Papio or Theropithecus? In: Jablonski, N.G.(Ed.), Theropithecus: The Rise and Fall of a Primate Genus. Cambridge Univer-sity Press, Cambridge, pp. 157e189.

Delson, E., Terranova, C.J., Jungers, W.L., Sargis, E.J., Jablonski, N.G., Dechow, P.C.,2000. Body mass in Cercopithecidae (Primates, Mammalia): estimation andscaling in extinct and extant taxa. Anthropol. Pap. Am. Mus. Nat. Hist. 83, 1e159.

Dietrich, W.O., 1942. Altestquartare Saugetiere aus der Sudlichen Serengti, Deutsch-Ostafrika. Palaeontographica 94, 44e75.

Eck, G.G., Jablonski, N.G., 1984. A reassessment of the taxonomic status and phyleticrelationships of Papio baringensis and Papio quadratirostris. Am. J. Phys.Anthropol. 65, 109e134.

Eck, G.G., Jablonski, N.G., 1987. The skull of Theropithecus brumpti compared withthose of other species of the genus Theropithecus. In: Coppens, Y., Howell, F.C.(Eds.), Les Faunes Plio-Pleistocenes de la Basse Vallée de l’Omo (Ethiopie), Tome3, Cercopithecidae de la Formation Shungura. Cahiers de Paléontologie, Editionsdu Centre National de la Récherche Scientifique, Paris, pp. 11e122.

Erxleben, J.C.P.,1777. SystemaRegni Animalis. Classis 1,Mammalia.Weygand, Lipsiae.Fleagle, J.G., 1999. Primate Adaptation and Evolution. Academic Press, New York.Fleagle, J.G., McGraw, W.S., 2002. Skeletal and dental morphology of African

papionins: unmasking a cryptic clade. J. Hum. Evol. 42, 267e292.Fleagle, J.G., Simons, E.L., 1982. Skeletal remains of Propliopithecus chirobates from

the Egyptian Oligocene. Folia Primatol. 39, 161e177.Freedman, L., 1957. The fossil Cercopithecoidea of South Africa. Ann. Transvaal Mus.

23, 121e262.Frost, S.R., 2001. New Early Pliocene Cercopithecidae (Mammalia: Primates) from

Aramis, Middle Awash Valley, Ethiopia. Am. Mus. Nov. 3350, 1e36.Frost, S.R., Delson, E., 2002. Fossil Cercopithecidae from the Hadar Formation and

surrounding areas of the Afar Depression, Ethiopia. J. Hum. Evol. 43, 687e748.Frost, S.R., Haile-Selassie, Y., Hlusko, L., 2009. Cercopithecidae. In: Haile-Selassie, Y.,

Woldegabriel, G. (Eds.), Ardipithecus kadabba: Late Miocene Evidence from theMiddle Awash, Ethiopia. University of California Press, Berkeley, pp. 135e158.

Geoffroy, E., 1812. Tableau des quadrumanes, 1. Ord. Quadrumanes. Ann. Mus. Hist.Nat. Paris 19, 85e122.

Gilbert, C.C., 2007. Craniomandibular morphology supporting the diphyletic originof mangabeys and a new genus of the Cercocebus/Mandrillus clade, Procerco-cebus. J. Hum. Evol. 53, 69e102.

Gilbert, C.C., 2008. African papionin Phylogenetic History and Plio-PleistoceneBiogeography. Ph.D. Dissertation, Stony Brook University.

Gray, J.E., 1821. On the natural arrangement of vertebrate animals. Lond. Med.Repos. Rec. 15, 296e310.

Groves, C.P., 2000. The phylogeny of the Cercopithecoidea. In: Whitehead, P.F.,Jolly, C.J. (Eds.), Old World Monkeys. Cambridge University Press, Cambridge,pp. 180e213.

Gundling, T., Hill, A., 2000. Geological context of fossil Cercopithecoidea fromeastern Africa. In: Whitehead, P.F., Jolly, C.J. (Eds.), Old World Monkeys. Cam-bridge University Press, Cambridge, pp. 180e213.

Harris, J.M., Leakey, M.G., Cerling, T.E., 2003. Early Pliocene tetrapod remains fromKanapoi, Lake Turkana Basin, Kenya. In: Harris, J.M., Leakey, M.G. (Eds.), Geologyand Vertebrate Paleontology of the Early Pliocene Site of Kanapoi, NorthernKenya. Contributions in Science, vol. 498. Nat. Hist. Mus. Los Angeles County,Los Angeles, pp. 39e113.

Harrison, T., 1989. New postcranial remains of Victoriapithecus from the MiddleMiocene of Kenya. J. Hum. Evol. 18, 3e54.

Hill, A., 1985. Early hominid from Baringo, Kenya. Nature 315, 222e224.Hill, A., 1995. Faunal and environmental change in the Neogene of east Africa:

evidence from the Tugen Hills sequence, Baringo District, Kenya. In: Vrba, E.S.,Denton, G.H., Partridge, T.C., Burckle, L.H. (Eds.), Paleoclimate and Evolution, withEmphasis on Human Origins. Yale University Press, New Haven, pp. 178e193.

Hill, A., 1999. The Baringo Basin, Kenya: from Bill Bishop to BPRP. In: Andrews, P.,Banham, P. (Eds.), Late Cenozoic Environments and Hominid Evolution:a Tribute to Bill Bishop. Geological Society of London, London, pp. 85e97.

Hill, A., 2002. Paleoanthropological research in the Tugen Hills, Kenya. J. Hum. Evol.42, 1e10.

Hill, A., Drake, R., Tauxe, L., Monaghan, M., Barry, J.C., Behrensmeyer, A.K., Curtis, G.,Fine Jacobs, B., Jacobs, L., Johnson, N., Pilbeam, D., 1985. Neogene paleontologyand geochronology of the Baringo basin, Kenya. J. Hum. Evol. 16, 749e773.

Hill, A., Curtis, G., Drake, R., 1986. Sedimentary stratigraphy of the Tugen Hills,Baringo District, Kenya. In: Frostick, L.E., Renaut, R.W., Reid, I., Tiercelin, J.-J.(Eds.), Sedimentation in the African Rifts. Geological Society of London SpecialPublication, vol. 25. Blackwell, Oxford, pp. 285e295.

Hill, A., Behrensmeyer, A.K., Brown, B., Deino, A.L., Rose, M., Saunders, J.J., Ward, S.C.,Winkler, A.J., 1991. Kipsaramon: a lower Miocene hominoid site in the TugenHills, Baringo District, Kenya. J. Hum. Evol. 20, 67e75.

Hill, A., Ward, S.C., Deino, A.L., Curtis, G., Drake, R., 1992. Earliest Homo. Nature 355,719e722.

Hill, A., Leakey,M.,Kingston, J.D.,Ward,S.C., 2002.Newcercopithecoids andahominoidfrom 12.5 Ma in the Tugen Hills succession, Kenya. J. Hum. Evol. 42, 75e93.

Hill, A., Gundling, T., 1999. A monkey (Primates; Cercopithecidae) from the LateMiocene of Abu Dhabi, United Arab Emirates. In: Whybrow, P.J., Hill, A. (Eds.),Fossil Vertebrates of Arabia. Yale University Press, New Haven and London, pp.198e202.

Hill, A., Ward, S.C., 1988. Origin of the Hominidae: the record of African largehominoid evolution between 14 My and 4 My. Yearb. Phys. Anthropol. 31,49e83.

C.C. Gilbert et al. / Journal of Human Evolution 59 (2010) 465e483 483

Hlusko, L.J., 2007.AnewlateMiocenespeciesofParacolobusandotherCercopithecoidea(Mammalia: Primates) fossils from Lemudong’o, Kenya. Kirtlandia 56, 72e85.

Hopwood, A.T., 1936. New fossil mammals from Olduvai, Tanganyika territory. Ann.Mag. Nat. Hist. 17, 636e638.

Jablonski, N.G., 1998. The evolution of the doucs and snub-nosed monkeys and thequestion of the phyletic unity of the odd-nosed colobines. In: Jablonski, N.G.(Ed.), The Natural History of the Doucs and Snub-Nosed Monkeys. WorldScientific, Singapore, pp. 13e52.

Jablonski, N.G., 2002. Late Neogene cercopithecoids. In: Hartwig, W.C. (Ed.), ThePrimate Fossil Record. Cambridge University Press, Cambridge, pp. 255e299.

Jablonski, N.G., Leakey, M.G., 2008. Systematic paleontology of the small colobines.In: Jablonski, N.G., Leakey, M.G. (Eds.), Koobi Fora Research Project, vol. 6: TheFossil Monkeys. California Academy of Sciences, San Francisco, pp. 12e30.

Jacobs, B.F., Deino, A.L., 1996. Test of climate-leaf physiognomy regression models,their application to two Miocene floras from Kenya, and 40Ar/39Ar dating of theLateMioceneKapturo site. Palaeogeogr. Palaeoclimatol. Palaeoecol.123, 259e271.

Jacobs, B.F., Kabuye, C.H.S., 1987. A Middle Miocene (12.2 My old) forest in the EastAfrican Rift Valley, Kenya. J. Hum. Evol. 16, 147e155.

Jacobs, B.F., Winkler, D.A., 1992. Taphonomy of a Middle Miocene autochthonousforest assemblage, Ngorora Formation, central Kenya. Palaeogeogr. Palae-oclimatol. Palaeoecol. 99, 31e40.

Jenkins, F.A., 1973. The functional anatomy and evolution of the mammalianhumero-ulnar articulation. Am. J. Anat. 137, 281e298.

Jerdon, T.C., 1867. The Mammals of India: a Natural History of all the AnimalsKnown to Inhabit Continental India. Thomason College Press, Roorkee.

Jones, T.R., 1937. A new fossil primate from Sterkfontein, Krugersdorp, Transvaal. S.Afr. J. Sci. 33, 709e728.

Kelley, J., Ward, S.C., Brown, B., Hill, A., Duren, D., 2002. Dental remains of Equatoriusafricanus from Kipsaramon, Tugen Hills, Baringo District, Kenya. J. Hum. Evol.42, 39e62.

Kingston, J.D., 1999. Environmental determinants in early hominid evolution; issuesand evidence from the Tugen Hills, Kenya. In: Andrews, P., Banham, P. (Eds.),Late Cenozoic Environments and Hominid Evolution: a Tribute to Bill Bishop.Geological Society of London, London, pp. 69e84.

Kingston, J.D., Marino, B.D., Hill, A., 1994. Isotopic evidence for Neogene hominidpaleoenvironments in the Kenya Rift Valley. Science 264, 955e959.

Kingston, J.D., Jacobs, B.F., Hill, A., Deino, A., 2002. Stratigraphy, age, and environ-ments of the late Miocene Mpesida Beds, Tugen Hills, Kenya. J. Hum. Evol. 42,95e116.

Kitching, J.W., 1952. A new type of fossil baboon: Brachygnathopithecus peppercorni,gen. et sp. nov. S. Afr. J. Sci. 49, 15e17.

von Koenigswald, G.H.R., 1969. Miocene Cercopithecoidea and Oreopithecoideafrom the Miocene of East Africa. In: Leakey, L.S.B. (Ed.), Fossil Vertebrates ofAfrica, vol. 1. Academic Press, London and New York, pp. 39e51.

Leakey, R.E.F., 1969. New Cercopithecidae from the Chemeron Beds of Lake Baringo,Kenya. In: Leakey, L.S.B. (Ed.), Fossil Vertebrates of Africa, vol. 1. Academic Press,London and New York, pp. 53e70.

Leakey, M.G., Teaford, M.F., Ward, C.V., 2003. Cercopithecidae from Lothagam. In:Leakey, M.G., Harris, J.M. (Eds.), Lothagam: The Dawn of Humanity in EasternAfrica. Columbia University Press, New York, pp. 201e248.

Leakey,M.G., Leakey, R.E.F.,1976. Further Cercopithecinae (Mammalia: Primates) fromthe Plio-Pleistocene of East Africa. In: Savage, R.J.G., Coryndon, S.C. (Eds.), FossilVertebrates of Africa, vol. 4. Academic Press, London and New York, pp. 121e146.

Linnaeus, C., 1758. Systema Naturae per Regna Tria Naturae, Secundum Classes,Ordines Genera, Species cum Characteribus, Differentris, Synonymis, Locis. In:Editis Decima, Reformata, Tomus 1. Laurentii Salvii, Stockholm.

McDougall, I., Feibel, C.S., 2003. Numerical age control for the Miocene-Pliocenesuccession at Lothagam, a hominoid-bearing sequence in the northern KenyaRift. In: Leakey, M.G., Harris, J.M. (Eds.), Lothagam: the Dawn of Humanity inEastern Africa. Columbia University Press, New York, pp. 43e64.

Miller, E.R., Benefit, B.R.,McCrossin,M.L., Plavcan, J.M., Leakey,M.G., El-Barkooky,A.N.,Hamdan, M.A., Abdel Gawad, M.K., Ali, S.M., Simons, E.L., 2009. Systematics ofEarly and Middle Miocene Old World monkeys. J. Hum. Evol. 57, 195e334.

Mivart, St.G., 1864. Notes on the crania and dentition of the Lemuridae. Proc. Zool.Soc. Lond. 1864, 611e648.

Morales, J., Pickford, M., 2008. Creodonts and carnivores from the Middle MioceneMuruyur Formation at Kipsaraman and Cheparawa, Baringo District, Kenya. C.R.Palevol. 7, 487e497.

Morgan, M.E., Kingston, J.D., Marino, B.D., 1994. Carbon isotopic evidence for theemergence of C4 plants in the Neogene from Pakistan and Kenya. Nature 367,162e165.

Nakatsukasa, M., Mbua, E., Yoshihiro, S., Sakai, T., Nakaya, H., Yano, W., Kunimatsu, Y.,2010. Earliest colobine skeletons fromNakali, Kenya. Am. J. Phys. Anthropol. doi:10.1002/ajpa.21327.

Patel, B., Susman, R., Rossie, J., Hill, A., 2009. Terrestrial adaptations in the hands ofEquatorius africanus revisited. J. Hum. Evol. 57, 763e772.

Pickford, M., 1975. Stratigraphy and Paleoecology of Five Late Cainozoic Formationsin the Kenya Rift Valley. Ph.D. Dissertation, University of London.

Pickford, M.H.L., Kunimatsu, Y., 2005. Catarrhines from the Middle Miocene (ca.14.5 Ma) of Kipsaraman, Tugen Hills, Kenya. Anthropol. Sci. 113, 189e224.

Raaum, R.L., Sterner, K.N., Noviello, C.M., Stewart, C.B., Disotell, T.R., 2005. Catar-rhine primate divergence dates estimated from complete mitochondrialgenomes: concordance with fossil and nuclear DNA evidence. J. Hum. Evol. 48,237e257.

Rossie, J.B., Gilbert, C.C., Hill, A., in prep. Early colobines from the Tugen Hills,Kenya.

Senut, B., 1994. Cercopithecoidea Néogène et Quaternaires du Rift Occidental(Ouganda). Geol. Paleontol. Albertine Rift Valley, Uganda-Zaire 2, 195e205.

Sherwood, R.J., Ward, S.C., Hill, A., 2002a. The taxonomic status of the Chemerontemporal (KNM-BC 1). J. Hum. Evol. 42, 153e185.

Sherwood,R.J.,Ward, S.C., Duren,D., Hill, A., Brown, B., Downs,W., 2002b. Preliminarydescription of the Equatorius africanus partial skeleton (KNM-TH 28860) fromKipsaramon, Tugen Hills, Baringo District, Kenya. J. Hum. Evol. 42, 63e73.

Strasser, E., 1988. Pedal evidence for the origin and diversification of cercopithecidclades. J. Hum. Evol. 17, 225e245.

Strasser, E., Delson, E., 1987. Cladistic analysis of cercopithecid relationships. J. Hum.Evol. 16, 81e99.

Szalay, F.S., Delson, E., 1979. Evolutionary History of the Primates. Academic Press,New York.

Tauxe, L., Monoghan, M., Drake, R., Curtis, G., Staudigel, H., 1985. Paleomagnetism ofMiocene east African Rift sediments and the calibration of the geomagneticreversal timescale. J. Geophys. Res. 90, 4639e4646.

Ting, N., 2008. Mitochondrial relationships and divergence dates of the Africancolobines: evidence of Miocene origins for the living colobus monkeys. J. Hum.Evol. 55, 312e325.

Ward, S.C., Brown, B., Hill, A., Kelley, J., Downs, W., 1999. Equatorius: a new homi-noid genus from the Middle Miocene of Kenya. Science 285 (5432), 1382e1386.

Ward, S.C., Hill, A., 1987. Pliocene hominid partial mandible from Tabarin, Baringo,Kenya. Am. J. Phys. Anthropol. 72, 21e37.

Winkler, A.J., 2002. Neogene paleobiogeography and East African paleoenviron-ments: contributions from the Tugen Hills rodents and lagomorphs. J. Hum.Evol. 42, 237e256.