The Presence of a Large Cercopithecine (Cf. Theropithecus Sp.) in the ‘Ubeidiya Formation (Early Pleistocene, Israel)

ARTICLE IN PRESS

Journal of Human Evolution xxx (2009) 1–11

Contents lists available at ScienceDirect

Journal of Human Evolution

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

The presence of a large cercopithecine (cf. Theropithecus sp.) in the ‘Ubeidiya formation (Early Pleistocene, Israel)

Miriam Belmaker

Department of Anthropology, Harvard University, 11 Divinity Ave, Cambridge MA 02138, USA article info abstract

Article history: This study presents the discovery of a right cercopithecine calcaneus from the site of ‘Ubeidiya, Israel, Received 25 June 2008 dated to ca. 1.6 Ma. The fossil is described and statistically compared to bones of modern and fossil Accepted 20 August 2009 cercopithecids. The specimen can be attributed to a large-bodied cercopithecine and represents a new primate taxon previously unidentiﬁed in the Early Pleistocene of the Southern Levant. Among extant Keywords: genera, it is most clearly similar to calcanei of Theropithecus. However, it could also represent Para- Primate biogeography dolichopithecus, but this alternative is unlikely due to the morphological uniqueness of the latter taxon. Out of Africa I The ﬁnding of an African taxon in the Levant suggests a circum-Mediterranean dispersal route for the Cercopithecidae taxon out of Africa, and emphasizes the importance of the Levantine corridor as a biogeographic dispersal route between Africa and Eurasia during the Early Pleistocene. Evidence for the biogeography of large-bodied primates is essential for the understanding of the dispersal routes of ‘‘Out of Africa I’’ taxa and can help elucidate Homo dispersal patterns in the Early Pleistocene. Ó 2009 Elsevier Ltd. All rights reserved.

Introduction Pleistocene contexts (Delson, 1980). Paradolichopithecus is known from the late Ruscinian to the middle Villafranchian in Europe and This study reports the discovery of a large cercopithecid calca- Asia (Delson, 1974; Szalay and Delson, 1979; Ardito and Mottura, neus from the Early Pleistocene site of ‘Ubeidiya, Israel. The spec- 1987; Delson et al., 2000; Rook and Martıńez-Navarro, in press), imen, ‘Ubeidiya (UB) 330, was found in stratum III 12 during the and Theropithecus occurs sporadically in Eurasia in Early Pleistocene 1993 excavation season and is housed in the paleontological sediments, although it was widely distributed, ranging from the collection of the Hebrew University of Jerusalem, Israel (HUJI). The Iberian Peninsula in the west to the Indian sub-continent in the east only other recorded cercopithecid primate from the Levant during (Delson, 1993; Jablonski, 1993; Gibert et al., 1995; Delson et al., this period is the assemblage of Macaca sylavanus from ‘Ubeidiya 2000; Rook et al., 2004). In the Levant, the absence of large cerco- (Tchernov and Volokita, 1986). Readily observable size differences pithecids, often found sympatric with Macaca in other Eurasian between Macaca calcanei and UB 330 suggest that UB 330 repre- sites (Ardito and Mottura,1987), may be attributed to sampling bias. sents a different taxon, probably Theropithecus sp. (Belmaker, 2002). The goal of this study is to test the hypothesis that the specimen Taxonomic identification of primate foot bones remains chal- UB 330 cannot be attributed to Macaca sylvanus, and to evaluate the lenging. However, based on studies of modern taxa, primate calcanei alternative hypotheses that UB 330 represents one of the large have been shown to be morphologically distinct at the familial level cercopithecine genera present in Eurasia during the Early Pleisto- (Langdon, 1986; Strasser, 1988), and subfamilies and genera within cene, and, specifically, a species in the genus Theropithecus. The Old World monkeys can be distinguished based on linear mammalian fauna at the site of ‘Ubeidiya includes several African measurements and multivariate analyses of the calcaneus, which taxa such as Pelorovis oldwayensis and Kolpochoerus olduvaiensis probably relate to differences in locomotion and degree of terres- (Geraads, 1986). The presence of the African genus Theropithecus in triality vs. arboreality (Strasser, 1992; Yirga, 2002; Youlatos, 2003). ‘Ubeidiya would serve to further confirm an African-Asian dispersal At least three cercopithecine genera were present in Eurasia route along the Levantine corridor (Tchernov, 1981) and shed light during the Early Pleistocene: the Eurasian genera Macaca and Par- on possible hominin dispersal routes during this time period. adolichopithecus, and the African Theropithecus. Macaca sylvanus is the most common cercopithecine recovered from European Early Geological context

The ‘Ubeidiya Formation lies about 3 km south of the Sea of E-mail address: [email protected] Galilee in Israel, on the ﬂanks of the western escarpment of the

Please cite this article in press as: Belmaker, M., The presence of a large cercopithecine (cf. Theropithecus sp.) in..., J Hum Evol (2009), doi:10.1016/ j.jhevol.2009.08.004 ARTICLE IN PRESS

2 M. Belmaker / Journal of Human Evolution xxx (2009) 1–11

Jordan Rift (Fig. 1). The archaeological layers of the ‘Ubeidiya have been found in strata II 33 and II 23–24 in the Fi member and Formation have been systematically excavated since 1960 (Stekelis, have been assigned to the Cobb Mt. (1.215–1.190 Ma) and the Gilsa 1966a,b; Stekelis et al., 1969; Bar-Yosef and Goren-Inbar, 1993) (1.575–1.567 Ma), respectively (Sagi, 2005). The dating of these through the late 1990s (Stekelis,1966a; Stekelis et al.,1969; Bar-Yosef short polarity events is corroborated by local faunal turnovers. The and Goren-Inbar, 1993; Gue´rin et al., 1996,2003; Shea and Bar-Yosef, ‘Ubeidiya fauna can be assigned to a local mammalian fauna bio- 1998), and are known for rich faunal (Haas, 1966,1968; Tchernov, zone older than that in the sites of Bitzat Ruhama, Evron, and 1986; Belmaker, 2006) and lithic assemblages (Bar-Yosef and Goren- Latamne dated to ca. 1.0–1.2 Ma, suggesting that the ‘Ubeidiya Inbar, 1993; Shea and Bar-Yosef, 1998). The primate assemblage normal polarity events in strata II 23-24 and II 33 should both includes dental and postcranial material of Macaca sylvanus (Tcher- predate the Jaramillo (0.99–1.07 Ma). (For a detailed stratigraphic nov and Volokita,1986) as well as a small sample of Homo cf. ergaster/ correlation, see Supplementary Online Material Figure S1)(Bel- erectus dental material (Tobias, 1966a,b; Belmaker et al., 2002). maker, 2009). Furthermore, the large mammalian assemblage of Estimated dates for the fossil-bearing strata of the ‘Ubeidiya ‘Ubeidiya is similar to the Farneta faunal unit (the sites of Selvella Formation are between ca. 1.6–1.2 Ma. Paleomagnetic analysis of and Pietraﬁtta, Italy) (Belmaker, 2006; Martı´nez-Navarro et al., the ‘Ubeidiya Formation indicate that it overlies the ‘Erq el Ahmar 2009), which has been dated to ca. 1.6–1.2 Ma (Caloi and Palombo, Formation, which is dated at 1.96–1.78 Ma (Ron and Levi, 2001) and 1997, and references therein), and the lithic assemblage is similar to has a reversed polarity, suggesting that it predates the Brunhes– those from East African sites (Stekelis et al., 1969; Bar-Yosef and Matuyama reversal (Opdyke et al., 1983; Braun et al., 1991; Verosub Goren-Inbar, 1993) such as Olduvai Upper Bed II, dated to ca. 1.53– and Tchernov, 1991). Two short, normal paleomagnetic episodes 1.27 Ma (Gowlett, 1979; Cerling and Hay, 1986).

Figure 1. Location of the site of ‘Ubeidiya in the Southern Levant.

Please cite this article in press as: Belmaker, M., The presence of a large cercopithecine (cf. Theropithecus sp.) in..., J Hum Evol (2009), doi:10.1016/ j.jhevol.2009.08.004 ARTICLE IN PRESS

M. Belmaker / Journal of Human Evolution xxx (2009) 1–11 3

Figure 2. UB 330, a right calcaneus. A: Medial view, B: Lateral view, C: Plantar (inferior) view, D: Dorsal (superior) view. The scale bar represents 5 cm.

The total accumulation between the two normal episodes is ca. was divided by the geometric mean1, producing 14 size-adjusted 30 m. The micromorphological analysis of the paleolake ‘Ubeidiya ratio variables (designated by the variable name followed by the delta system included periods of hiatus, probably in the range of subscript ‘‘/GM,’’ e.g., Cal 1/GM). Second, 16 ratios were calculated several thousands of years during which pedogenic processes following Langdon (1986), Strasser (1992), and Yirga (2002). Thus, occurred, suggesting that the estimated duration of 400 k.yr. is not a total of 30 variables (14 Mosimann shape-adjusted and 16 ratios) inconsistent with the geomorphology of the site (Mallol, 2006). were used in this analysis (Tables S1 and S2). The specimen UB 330 described here was found in stratum III 12 Size-adjusted values of UB 330 were compared to modern cer- in the Li member. It is stratigraphically below the Fi member that copithecid generic means using the single observation means t-test contains the short, normal polarity events, indicating that it most (Sokal and Rohlf, 1995). Multiple comparisons, such as the one probably predates the Gilsa (ca. 1.575 Ma). Thus, the estimated date performed here, require adjusting the probability values for the for stratum III 12 and specimen UB 330 is ca.1.6–1.58 Ma (Figure S2). number of simultaneous tests to avoid Type I errors. To increase the power of the test, the sequential Bonferroni method was applied Materials and methods (Rice, 1989). A P value of 0.00033 was set as the test criterion of the single sample t-test. The fossil specimen UB 330 (Fig. 2) was compared to calcanei of Discriminant Function Analysis (DFA) uses correlation metrics to adult extant and fossil Cercopithecidae. Calcanei were measured address weight combinations of variables and emphasizes between from specimens of extant Cercopithecidae species from the Amer- group variation while minimizing within group variation. In this ican Museum of Natural History, the Museum of Comparative study, a two-tier stepwise Linear Discriminant Function was Zoology, as well as one specimen from the personal collection of applied to the size-adjusted variables using stepwise insertion of Philip Rightmire. Comparative fossil material included an unpub- variables (maximizing the smallest F ratio) with UB 330 treated as lished Macaca sylvanus from ‘Ubeidiya (UB 101), as well as two casts a separate group (Sokal and Rohlf, 1995). First, a DFA was run at the of Theropithecus oswaldi (from Kanjera, Kenya), and a cast of Para- subfamilial level to test the hypothesis that UB 330 could be colobus chemeroni (Chemeron Fm., Kenya) generously provided by identified as a member of Colobinae or Cercopithecinae. Second, E. Delson. All extant comparative specimens (n ¼ 146) were adult, a DFA was run at the generic level confined to cercopithecines based on tooth eruption and fusion of limb elements (Table 1). (Cercopithecus, Macaca, Mandrillus, Papio, and Theropithecus). Measurements were taken with digital calipers on both modern (Methodological considerations of the DFA are presented in the and fossil specimens to an accuracy of two decimal places (Fig. 3). Supplementary Online Material.) Leave-one-out cross validation Measurements followed Langdon (1986) and Yirga (2002). Analysis was used to assess the overall error rate for the DFA. Furthermore, was performed on size-adjusted data. Two methods of size bias and standard error around the predicted posterior probabili- adjustments were used. First, raw variables were transformed to ties for UB 330 were estimated using the jackknife procedure. This size-adjusted Mosimann shape data (Mosimann, 1970; Falsetti was performed by running the DFA while randomly removing et al., 1993; Jungers et al., 1995). Each of the 14 raw measurements a single observation at a time and iterated for the total number of specimens (n ¼ 150). Linear regressions of cercopithecid indices of calcaneus pedal 1 The geometric mean is calculated as the nth root of the product of n power arm (Cal 20) and calcaneal load arm (Cal 21) with body mass, measurements. have shown significant correlations at the 0.001 significance level

Please cite this article in press as: Belmaker, M., The presence of a large cercopithecine (cf. Theropithecus sp.) in..., J Hum Evol (2009), doi:10.1016/ j.jhevol.2009.08.004 ARTICLE IN PRESS

4 M. Belmaker / Journal of Human Evolution xxx (2009) 1–11

Table 1 Comparative species measured by sex.

Family Genus Species \_U Total Colobinae Colobus Colobus guereza 221 5 Colobus polykomos 106 7 Colobus sp. 0 0 1 1 Procolobus Procolobus badius 211 4 Paracolobus Paracolobus chemeroni 001 1 Cercopithecinae Cercopithecus ascanius 130 4 Cercopithecus diana 012 3 Cercopithecus erythrotis 001 1 Cercopithecus hamlyni 011 2 Cercopithecus mitis 37111 Cercopithecus mona 020 2 Cercopithecus nictitans 001 1 Macaca Macaca arctoides 100 1 Macaca assamensis 021 3 Macaca fascicularis 18 19 0 37 Macaca fuscata 030 3 Macaca maura 010 1 Macaca mulatta 030 3 Macaca nemestrina 76013 Macaca nigra 001 1 Macaca ochreata 100 1 Macaca sp. 0 1 0 1 Macaca sylvanus 202 4 Macaca thibetana 010 1 Macaca tonkeana 150 6 Mandrillus Mandrillus leucophaeus 010 1 Mandrillus sphinx 450 9 Papio Papio hamadryasa 2160 18 Theropithecus Theropithecus gelada 120 3 Theropithecus oswaldi 002 2

Total 150

a Papio hamadryas includes individuals assigned to ﬁve subspecies: P. h. ursinus, P. h. hamadryas, P. h. anubis, P. h. cynocephalus, and P. h. papio.

with R2 above 0.9 (Strasser, 1992). Estimated body mass of UB 330, and a better leverage for foot plantar ﬂexion), a larger insertion for based on calcaneus body mass regression equations developed by a bulkier m. triceps surae, and a wider and shorter posterior talar Strasser (1992), were compared to Plio-Pleistocene fossil primate facet (reducing the sliding function, proximal inversion and ever- body mass estimates, as retrieved from the literature (Delson et al., sion, and helicoid movement in the joint). In comparison, the 2000). morphology of Colobinae exhibits a shorter proximal calcaneal All analyses were performed using SPSS (version 16.0) statistical region, a longer and narrower m. triceps surae insertion, and software. The jackknife DFA procedure was calculated using R. narrow and long posterior talar facets, which serve to increase the precision and power in the mobility of the foot during movement across difﬁcult terrain, such as branches, during arboreal walking Taxonomic comparisons and statistical results and climbing (Langdon, 1986; Strasser, 1988, 1992). Description

A complete right calcaneus, catalogue number ‘Ubeidiya (UB) 330, was found in stratum III 12 (Fig. 2). The calcaneal tuber, lateral process, and lateral edge are weathered. The calcaneal tuberosity is in advanced epiphyseal fusion corresponding to stage C in the development of appendicular bone in primates as defined by Gal- liari (1988), and the articular surfaces are well defined and angular. Age and bone fusion correlations for Old World monkeys (Papio and Macaca) suggest that fusion of the calcaneus begins between the ages of 3–4 years and is completed by the age of 7 (Bramblett, 1969; Kimura and Hamada,1990). This would suggest an ‘‘adolescent’’ age for UB 330 based on the age class of Kawai et al. (1983) and a nearly mature adult maximum length (Scheuer and Black, 2004). Within the Cercopithecidae, there are two subfamilial morphological patterns that typify adaptations for terrestrial and arboreal locomotion. The cercopithecine morphotype is adapted to increased stress during plantarflexion and inteversion as well as dorsiversion and eversion. In comparison, the colobine morphotype is adapted to an increased ability in grasping and the supi- Figure 3. Linear measurements taken on UB 330 and modern comparative calcanei. A: nation of the forefoot (Strasser, 1988). This results in a longer Anterior view, B: Medial view, C: Superior view, D: Posterior view. These measure- proximal calcaneal region in cercopithecines (a longer lever arm ments correspond to the descriptions of Cal 1–14 in Table S1.

Please cite this article in press as: Belmaker, M., The presence of a large cercopithecine (cf. Theropithecus sp.) in..., J Hum Evol (2009), doi:10.1016/ j.jhevol.2009.08.004 ARTICLE IN PRESS

M. Belmaker / Journal of Human Evolution xxx (2009) 1–11 5

Table 2 from 88.9–89.6% for the two subfamilies. Leave-one-out cross Pooled within-group correlations between discriminating variables and standard- validation results indicate a robust classification, as nearly identical ized canonical discriminant functions*. classification rates were obtained with 89% of specimens correctly A. Subfamilial DFA classified. Misclassification rates of each of the subfamilies were Function 1 nearly identical with 10.4% of the cercopithecines identified as Cal 23 0.565 colobines and 11.1% of the colobines identified as cercopithecines. Cal 25 0.565 Cross validation results were similar indicating a highly robust Cal 4/GM 0.340 Eigenvalues 0.506 classification (Table 3). % of variance 100 Posterior probabilities of classification to each subfamily were Canonical Correlation 0.50 obtained for each quartile, and indicated that UB 330 should be P value <0.001 assigned to the subfamily Cercopithecinae with a median proba- B. Cercopithecine genus DFA bility of 97.6% (with an inter-quartile range of 97.2–98.2%), and Function 1 2 3 4 could be assigned to Colobinae with a median probability of 2.3% Cal 20 0.834 0.248 0.064 0.043 (with an inter-quartile range of 1.79–2.79%). Cal 22 0.092 0.552 0.020 0.020 The DFA produced a single eigenvector that explained 100% Cal 19 0.120 0.467 0.135 0.010 of the variance (Table 2A). The three variables, Cal 23 (the index of Cal 28 0.085 0.393 0.296 0.134 breadth of cuboid facet to distal length), Cal 25 (the index Cal 27 0.186 0.292 0.092 0.124 of breadth to height of calcaneal tuberosity), and Cal 4/ (posterior Cal 18 0.126 0.094 0.515 0.335 GM talar facet breadth), are consistent with observed morphological Cal 9/GM 0.200 0.343 0.514 0.213 Cal 3/GM 0.092 0.044 0.466 0.308 differences between cercopithecines and colobines. An increase in Cal 17 0.179 0.293 0.304 0.509 the size of the cuboid facet relative to the size of the calcaneus has Cal 23 0.359 0.320 0.069 0.450 been shown to secure the distal calcaneal joint (Langdon, 1986). Eigenvalues 2.616 0.846 0.281 0.079 % of variance 68.4 22.1 7.3 2.1 Similarly, a decrease in height-to-width ratio for the insertion of m. Canonical Correlation 0.851 0.677 0.468 0.270 triceps surae (Cal 25) and an increase in length of the posterior talar P value <0.001 <0.001 0.001 0.217 facet (Cal 4/GM) are indicative of a more flexible distal calcaneal *Variables ordered by absolute size of correlation within function. joint and calcaneal-astragalar joint. These are adaptive for leaping, bounding, and galloping performed during arboreal locomotion in comparison to terrestrial locomotion (Langdon, 1986; Strasser, The specimen UB 330 has proportions and morphology similar 1988, 1992). to that of Cercopithecidae in general, and terrestrial Cercopitheci- These results are consistent with the qualitative morphological nae in particular. The anterior articular facet and posterior articu- observations that support an identification of UB 330 as a terrestrial lation for the talus possess several derived features typical of cercopithecine. cercopithecines: the anterior articular facet is fused into a single facet and is not divided into two sub-facets; the posterior articu- Cercopithecine genera stepwise linear DFA lation for the talus is well defined, quadrilateral, medially deviated relative to the long axis of the bone, and strongly arched; the The variables which met the stepwise selection criterion (in calcaneal tuberosity process is dorso–ventrally thick with order of descending absolute size of correlation within function) a rounded dorsal margin indicative of a well-developed insertion of were: Cal 20, Cal 22, Cal 19, Cal 28, Cal 27, Cal 18, Cal 9/GM, Cal 3/GM, m. triceps surae; the posterior talar facet breadth is wide and short, Cal 25, Cal 17, and Cal 23. Stepwise linear discriminate function indicative of reduced movement along the calcaneo-astragalar analysis on size-adjusted ratios calculated four discriminate func- facet; and a weak peroneal process, which is a characteristic of tions. The first function accounted for roughly 68.4% of the variance, terrestrial Cercopithecinae as compared to the pronounced process the remaining functions accounting for 22.1%, 7.3%, and 2.1%, in in arboreal or semi-arboreal taxa (Harrison, 1989). descending order (Table 2B). Cercopithecine genera differed in 2 calcaneus morphology (Wilk’s l ¼ 0.108, c 40 ¼ 278.835, P value <0.001), and all variables significantly affected genus differences. One sample t-test The average correct classification rate for cercopithecine genera was 84.1%, ranging from 70–89% per genus (Table 3). Leave-one-out Table S3 presents the descriptive statistics (means and standard cross validation rate indicates an overall correct classification rate deviations) for the raw measurements for the comparative sample of 75% with a range of 50–79.2%. The genera were ordered roughly and UB 330, and Table S4 presents the descriptive statistics and the according to their level of arboreality, from Cercopithecus as the results for the one sample t-test. Despite overall similarities in the most arboreal, and following in descending order to the most majority of measurements, the one sample t-test indicates that UB terrestrial: Macaca, Mandrillus, Papio, and Theropithecus. Genera 330 differs from each of the modern genera by at least one size- tend to misclassify into genera that were most similar, or only ‘‘one adjusted measurement, with the exception of Theropithecus and step over,’’ in locomotion pattern. However, there were two Procolobus, which do not differ from UB 330 in any measurement. exceptions. In addition to Cercopithecus and Mandrillus, Macaca is However, both have the lowest sample sizes (5 and 4, respectively), misclassified (albeit in very low proportions) as the extremely which may have affected the results of the t-test. terrestrial Theropithecus. Similarly, Theropithecus is only misclassified as Macaca in the original reclassification and misclassified Subfamilial stepwise linear DFA as both Macaca and Papio in the leave-one-out cross validation reclassification. This result appears contradictory to the expectation Three variables met the criteria for the stepwise linear DFA derived from the behavioral position of the two genera, and criterion (in order of descending absolute size of correlation within suggests that some other behavioral similarities exist between the function): Cal 23, Cal 25, and Cal 4/GM were used in this analysis two genera. A closer examination reveals that the three Macaca (Table 2A). Calcaneus morphology correctly distinguished between specimens that were misclassified as Theropithecus are M. fas- Colobinae and Cercopithecinae in 89.5% of the specimens, ranging cicularis. It has been hypothesized by Rodman (1979) that foraging

Please cite this article in press as: Belmaker, M., The presence of a large cercopithecine (cf. Theropithecus sp.) in..., J Hum Evol (2009), doi:10.1016/ j.jhevol.2009.08.004 ARTICLE IN PRESS

6 M. Belmaker / Journal of Human Evolution xxx (2009) 1–11

Table 3 DFA classiﬁcation results.

A. Actual rows by predicted columns for subfamilial DFA classiﬁcation results: counts and (%), DFA correctly classiﬁed 89.5% of original grouped cases.

Cercopithecinae Colobinae Total Cercopithecinae 120 (89.6) 14 (10.4) 134 Colobinae 2 (11.1) 16 (88.9) 18

B. Actual rows by predicted columns for leave-one-out cross validation for subfamilial DFA classiﬁcation results: counts and (%), DFA correctly classiﬁed ca. 89% of cross-validated grouped cases. Cercopithecinae Colobinae Total Cercopithecinae 119 (88.8) 15 (11.2) 134 Colobinae 2 (11.1) 16 (88.9) 18

C. Actual rows by predicted columns for cercopithecine genera DFA classiﬁcation results: counts and (%), DFA correctly classiﬁed 84.1% of original grouped cases. Cercopithecus Macaca Mandrillus Papio Theropithecus Total Cercopithecus 21 (87.5) 3 (12.5) 0 0 0 24 Macaca 6 (12) 63 (84) 2 (2.7) 0 1 (1.3) 75 Mandrillus 0 1 (10) 7 (70) 2 (20) 0 10 Papio 0 0 0 16 (88.9) 2 (11.1) 18 Theropithecus 0 1 (20) 0 0 4 (80) 5

D. Actual rows by predicted columns for leave-one-out cross validation for cercopithecine genera DFA classiﬁcation results: counts and (%), DFA correctly classiﬁed ca. 75% of original grouped cases. Cercopithecus Macaca Mandrillus Papio Theropithecus Total Cercopithecus 19 (79.2) 5 (20.8) 0 0 0 24 Macaca 11 (14.7) 59 (78.7) 2 (2.7) 0 3 (4.0) 75 Mandrillus 1 (20) 1 (10) 5(50) 2 (20) 0 10 Papio 0 0 1 (5.6) 14 (77.8) 3 (16.8) 18 Theropithecus 0 1 (20) 0 1 (20) 3 (60) 5

efficiency may be an additional contributing factor in long limb Cercopithecus), which score negative values on the first function. skeletal differences between the arboreal M. fascicularis and the While UB 330 scores clearly in the large terrestrial cercopithecine terrestrial M. nemestrina (Rodman, 1979). While he did not consider group, it occupies a unique position on the plot, specifically in the osteological morphology of the foot in his original study, it is relation to Theropithecus, and will be discussed later. intriguing to speculate if this hypothesis can be applied to the The first function is affected primarily by Cal 20 (pedal power similarities observed here between Theropithecus and Macaca. arm). Arboreal taxa such as Macaca have a shorter pedal power arm Theropithecus forage over much shorter distances while feeding compared to more terrestrial taxa, since the ‘‘high gear’’ ratio than the sympatric anubis baboons (Iwamoto, 1993), and perhaps contributes to increasing take off velocity required for leaping provide an analogy to the M. fascicularis–M. nemestrina study by locomotion in comparison to a ‘‘lower gear’’ ratio, which is found in Rodman (1979). If this is correct, the misclassification between Theropithecus and Macaca fascicularis may be due to similarities in foraging efficiencies, which may confound the distinction in calcaneal morphology observed among arboreal and terrestrial cercopithecines. Posterior probabilities of classification to each taxon were obtained for each quartile. The results indicate that UB 330 should be assigned to the genus Theropithecus with a median probability of 98.5% (with an inter-quartile range of 98.1–99.2%). Results indicated that UB 330 can be assigned to Cercopithecus with a median probability of 1.1% (with an inter-quartile range of 0.06–1.67%), and could be assigned to Papio with a median probability of 0.22% (with an inter-quartile range of 0.01–0.35%). Probability of assignment to Macaca and Mandrillus was less than 0.001 percent. The high posterior probability in assignment of UB 330 to Theropithecus with very narrow inter-quartile ranges provides strong support for the identification of UB 330 as Theropithecus. In order to understand the morphological differences that may be driving the distinction between the genera, a scatter plot of the two first functions can be observed (Fig. 4) and analyzed in relation to the results of the stepwise linear DFA (Table 2B). The DFA plots indicate that the major separation along the first function, which explains 68.4% of the variance, is between the Figure 4. Bivariate plot of the first two axes of the Discriminant Function Analysis (DFA) separating the five cercopithecine genera and UB 330. The unique position of UB larger and terrestrial cercopithecine genera (Papio, Mandrillus, and 330 is discussed in the text. Discriminant function 1 explains 68.4% of the total Theropithecus), which score positive values on the first function, variance, and discriminant function 2 explains 21.1% of the total variance. Theropithecus and the smaller and more arboreal genera (Macaca and gelada and T. oswaldi were grouped together in the analysis.

Please cite this article in press as: Belmaker, M., The presence of a large cercopithecine (cf. Theropithecus sp.) in..., J Hum Evol (2009), doi:10.1016/ j.jhevol.2009.08.004 ARTICLE IN PRESS

M. Belmaker / Journal of Human Evolution xxx (2009) 1–11 7 the more terrestrial locomotion of baboons (Strasser, 1992). shown to have occupied a more arboreal habitat than modern However, pedal power arm is also positively correlated with body T. gelada (Elton, 2002), and ‘‘shuffling forward bipedally’’ is an early mass (Langdon, 1986; Strasser, 1992). Thus, we cannot exclude locomotor adaptation in the Theropithecus lineage (Krentz, 1993), a body mass component in the distinction between groups of cer- both of which support the former hypothesis. copithecine genera along the first function. The average body weight calculated for Macaca and Cercopithecus is ca. 5.5 kg Body mass estimates compared to the average body weight of 21.3 kg for Papio, Man- drillus, and Theropithecus. The body mass for UB 330 was 25.56 kg based on the pedal The distinction along the second dimension, which accounts for power arm regression equation and 22.83 kg based on the calcaneal only 21.1% of the variance, is more difficult to explain. In the load arm regression equation (Strasser, 1992). Since DFA results smaller-bodied cercopithecine group, there is a gradient from indicated that UB 330 could be classified as a cercopithecine, the negative to positive scores, from Macaca (a more terrestrial genus) body size of UB 330 was compared to fossil and extant cercopi- to Cercopithecus (a more arboreal genus on average). In the larger thecines only. The body size estimates for UB 330 are higher than taxa, a reverse trend is observed from negative to positive scores, modern Cercopithecus and Macaca (both sexes). It is also higher from Mandrillus (which is the most arboreal), to Papio, and then than modern Mandrillus, Papio, and Theropithecus females but Theropithecus, which is the most terrestrial. This similarity in within the range of the males of these genera. It is only slightly positive scores between Cercopithecus and Theropithecus is above the range of modern Theropithecus males (Delson et al., surprising since both have different behavioral positions. Most of 2000). This value is higher than Plio-Pleistocene fossil populations the Cercopithecus species sampled in this study are arboreal, and of Macaca in general (10–16 kg for males and 6.5–12.5 kg for the two that score the highest on the second function of the DFA are females) and the ‘Ubeidiya Macaca sylvanus in particular C. mitis and C. ascanius, while Theropithecus is an extremely (6.5–9.5 kg for females) (Delson et al., 2000). The body mass esti- terrestrial species with a ‘‘shuffling forward bipedally’’ type of mate is within the estimated range for T. oswaldi oswaldi, dated locomotion (Wrangham, 1980). between 2.5–1.2 Ma, and obtained from both dental and postcranial The variable that most affected the second dimension is Cal 22, measurements (13–36 kg for females and 20–86 kg for males) anterior height to anterior length of the calcaneus. An increase in (Delson et al., 2000), and is consistent with the identification of UB height relative to length provides additional strength and robus- 330 as Theropithecus (Table 4). However, it is also just within the ticity to the calcaneus and suggests an adaptation to stresses in the estimated body size range for Paradolichopithecus avernensis sagittal plane produced by m. triceps surae (Langdon, 1986; Youla- females (12–23 kg) and males (25–41 kg) (Delson et al., 2000). tos, 2003). A well-developed m. triceps surae provides the necessary force required for terrestrial locomotion, as opposed to the more Discussion slender m. triceps surae observed in more arboreal taxa (Strasser, 1988; Youlatos, 2003). Thus, it is not surprising that we observe the The identification of primate fossil material to genus based on trend in the smaller-bodied cercopithecines. The more terrestrial postcranial material alone is a difficult task. This study indicates Macaca score negatively (i.e., have higher anterior calcanei relative that cercopithecid calcanei can be used to distinguish between to length) while the more arboreal Cercopithecus have more posi- subfamily and genera using both univariate and multivariate tive scores (i.e., have shorter anterior calcanei relative to length). methods. The measurements of specimen UB 330 may have been However, the results of the larger-bodied cercopithecines are more a slight underestimation of the full adult size due to its adolescent difficult to explain as it may have been expected that Theropithecus age. However, the estimated adult length was probably only an would score negatively on the second function. However, unlike additional 1–2 mm due to incomplete ossification of the epiphysis. other terrestrial cercopithecines, Theropithecus invert their feet Primate calcanei attain their adult shape by adolescence. Since much of the time while feeding, so a more agile and flexible ankle is Macaca sylvanus are smaller than other terrestrial Cercopithecinae, advantageous, and perhaps convergent in some of its morphology the comparisons with adult Macaca provide a conservative size to more arboreal forms (Krentz, 1993). Theropithecus shares with comparison with UB 330. the arboreal Colobus an angulated medial malleolus and Based on the results in this study, UB 330 can be identified as a pronounced notch for m. tibialis posterior, which aids in inverting a cercopithecine with a very high level of confidence. DFA analysis the foot and increased flexibility of the ankle (Krentz, 1993). identified UB 330 as a cercopithecine with a posterior probability It is interesting to note that the score for UB 330 on the second of nearly 90%. Of the cercopithecine genera analyzed, UB 330 function falls above any of the observed values for the comparative differed significantly from Cercopithecus, Mandrillus, Papio, and samples. This is supported by the fact that the two Theropithecus Macaca. One sample t-test differed in one variable or more from oswaldi from Kanjera score the highest on the second dimension each of these genera, and stepwise linear DFA classification results within the Theropithecus sample (T. oswaldi mean ¼ 2.53, indicated that UB 330 could be classified as Cercopithecus, S.D. ¼1.279; T. gelada mean ¼ 0.84, S.D. ¼ 0.645). However, the two Mandrillus,orMacaca with a probability of less than 0.0001. species do not differ significantly along the second dimension (two- Moreover, UB 330 fell above the estimated size range for tailed student t-test P value >0.1), probably due to the low sample Cercopithecus and Macaca as well as Mandrillus females. The size of the comparative Theropithecus sample, making it difficult to distinction from Cercopithecus and Mandrillus is not surprising as evaluate the significance of the extreme value of UB 330. While their current and fossil biogeographic distribution is confined to probabilistically UB 330 is more similar to Theropithecus than other sub-Sahara Africa (Pickford, 1987). genera, the unique position of UB 330 along the second function The distinction from Macaca is of particular importance. Macaca may indicate either an undocumented high variability in has been previously found in ‘Ubeidiya (Tchernov and Volokita, Theropithecus calcaneal morphology reflective of variable locomo- 1986). Given the difference in size between UB 330 and other tion patterns (Elton, 2002), or an otherwise unknown calcaneus of Macaca specimens at the site, specifically the Macaca calcaneus UB a fossil taxon, such as Paradolichopithecus. However, current 101, the null hypothesis was that UB 330 represents a large male assessments of the positional behavior of the latter taxon point to Macaca sylvanus. The DFA provides the most significant distinction an increased terrestrial locomotion (Sondaar et al., 2006; E. Delson along the first function, which separates the smaller-bodied taxa and W. Harcourt-Smith, pers. comm.), whereas T. oswaldi has been (Macaca and Cercopithecus) from the larger and more terrestrial

Please cite this article in press as: Belmaker, M., The presence of a large cercopithecine (cf. Theropithecus sp.) in..., J Hum Evol (2009), doi:10.1016/ j.jhevol.2009.08.004 ARTICLE IN PRESS

8 M. Belmaker / Journal of Human Evolution xxx (2009) 1–11

Table 4 Body mass (kg) (fossil estimates) for modern and fossil Cercopithecinae.

Species Sex Body mass (kg) (fossil estimates) Known age range Modern genera Cercopithecus _ 1.8–8.0 Extant \ 1.8–5 Macaca _ 4.9–17.5 Extant \ 3.05–14.1 Mandrillus _ 27–45 Extant \ 10–17 Papio _ 15–37.2 Extant \ 8.8–20.5 Theropithecus _ 16.5–20.25 Extant \ 9–13.8 Fossil species Paradolichopithecus avernensis _ 25–41 2.5–1.6 Ma \ 12–23 Theropithecus oswaldi oswaldia _ 20–86 2.5–1.2 Ma \ 13–36 Macaca sylvanus (fossil) _ 10–17 Late Miocene – Early Pleistocene \ 6.5–12.5 Macaca sylvanus (from ‘Ubeidiya) \ 6.5–9.5 1.6–1.2 Ma

UB 330 ? 22.83–25.56 1.6–.2 Ma

Values for fossils are maximum estimated body mass ranges derived from cranial and postcranial material and include 20% estimated mass from Delson et al. (2000), as are comparative extant data. a Measured specimens in this study are from Kanjera, Kenya, and dated to ca. 2 Ma.

ones. Misclassifications of the DFA were very low between these 2003), which postdate the Early Pleistocene dispersal of African two groups. Therefore, the null hypothesis was rejected with a high species from East Africa into Eurasia ca. 1.8–1.6 Ma, which is degree of probability, and UB 330 was assigned to a large Cerco- apparent in ‘Ubeidiya. pithecinae previously unidentified in ‘Ubeidiya. All previous Pleis- tocene cercopithecid material in the Levant has been attributed to Do the data support an assignment of UB 330 to the genus the small-bodied Macaca sylvanus, and UB 330 represents the Paradolichopithecus? finding of a new taxon in the Early Pleistocene of the Southern Levant. Comparison between UB 330 and fossil calcanei of Para- Assignment to genera within the large-bodied Cercopithecinae, dolichopithecus could not be done, because no Paradolichopithecus Papio, Paradolichopithecus,orTheropithecus, is more difficult based calcanei are yet known; however, the estimated body size range on the current data set but some taxa are more probable, based on obtained for UB 330 falls within the estimated range for Para- morphology, body size, and biogeography. dolichopithecus males and females. Therefore, the possibility that UB 330 may be attributed to this Eurasian taxon cannot be Do the data support an assignment of UB 330 to the genus Papio? excluded. Nonetheless, based on current information, this identification would appear less probable. A talus (PO 157F) from the site The question of the identification of UB 330 is very interesting of Vatera on the island of Lesvos, Greece, assigned to P. arvernensis from a biogeographic point of view. Papio hamadryas hamadryas is and dated to MN17 (St. Vallier faunal Unit) ca. 2.4–1.8 Ma, allows for the only Papio to disperse beyond the African continent. It is an indirect comparison of foot bone morphology. The morphology currently found in the south of the Arabian Peninsula in the of the talus is said to be unique among cercopithecine tali in that republic of Yemen and Saudi Arabia (Harrison and Bates,1991). Two the facet for the tibial malleolus in plantar view is flattened rather alternative (although not mutually exclusive) routes have been than bowl shaped; furthermore, the trochlea is only slightly suggested for the dispersal of Papio from Africa to Arabia: 1) wedge–shaped, and the dorsal aspect of the fibular articulation a longer route, which includes a dispersal northward though the protrudes laterally (Sondaar and Van der Geer, 2002; Van der Geer Nile valley and Sinai Peninsula into the Levant and then southward and Sondaar, 2002; Sondaar et al., 2006). This would suggest that to the Arabian Peninsula, and 2) crossing the Bab el Mandeb strait UB 330, with its typical cercopithecine calcaneal morphology and during periods of low sea levels (Kummer, 1995). The finding of similarity to Theropithecus calcanei, could not articulate with the Papio in the Early Pleistocene Levant would provide support for the unique talus of Paradolichopithecus. It is worth noting that this former northern route. morphological interpretation, while intriguing, is highly contro- Evidence presented in this study does not provide strong versial and a comparison by Delson and colleagues of the distal support for the assignment of UB 330 to the genus Papio. While UB tibia and talus of the Vatera specimen indicates no significant 330 falls within the body size range of both modern and fossil distinction between the postcranial elements of PO 157F and other Papio species (Delson et al., 2000), DFA could assign UB 330 to large papionins (E. Delson and W. Harcourt-Smith, pers. comm.), Papio with a median probability of <1%, and it differed significantly suggesting that such conclusions may be premature. Nonetheless, from Papio in the variable Cal 4/GM. Therefore, the probability that since the DFA results presented here suggest that cercopithecine UB 330 may be assigned to Papio is very low. This is concordant calcanei can be used to identify genera, and given the high posterior with analyses of the levels of genetic variation in P. h. hamadryas in probability obtained in this study of the assignment of UB 330 to Southern Arabia compared to the population in the Horn of Africa, Theropithecus, while the possibility that UB 330 is a Para- which suggests a colonization date of ca. 200–400 ka (Wildman dolichopithecus cannot be excluded, it would seem less probable. et al., 2004; Winney et al., 2004). This coincides with the formation However, if future fossils were to confirm that UB 330 is indeed of a land bridge in the Bab el Mandeb straits at 18 ka, 130 ka, Paradolichopithecus, its recovery in ‘Ubeidiya would provide 270 ka, 370 ka, and 440 ka (Rohling et al., 1998; Siddall et al., a geographic expansion of this taxon westward and southward into

Please cite this article in press as: Belmaker, M., The presence of a large cercopithecine (cf. Theropithecus sp.) in..., J Hum Evol (2009), doi:10.1016/ j.jhevol.2009.08.004 ARTICLE IN PRESS

M. Belmaker / Journal of Human Evolution xxx (2009) 1–11 9 the Levant. Of more interest would be the implication for the role of the Bab el Mandeb straits during this time period (Derricourt, 2006; African vs. Eurasian taxa in Early Pleistocene dispersal events. Rook Fernandes et al., 2006; Turner and O’Regan, 2007). et al. (2004) proposed that the Early Pleistocene dispersal of Homo The presence of Theropithecus in ‘Ubeidiya suggests a possible from Africa to Eurasia occurred in parallel with a suite of four other circum Mediterranean dispersal route via ‘Ubeidiya (ca. 1.6 Ma), African taxa that included Megantereon whitei, Kolpochoerus old- Pirro Nord (ca. 1.6–1.3 Ma), and Cueva Victoria (ca. 1.2 Ma), which uvaiensis, Hippopotamus antiquus, and Theropithecus oswaldi. The may have allowed for the dispersal of Theropithecus from Africa presence of an ‘‘African assemblage’’ in Eurasia has been suggested without requiring Theropithecus (and other taxa including Homo)to as a faunal marker for the presence of Homo and has been ascribed cross open bodies of water such as the Gibraltar straits. This route to global climatic change (Martıńez-Navarro and Palmqvist, 1995; has been shown to be the most probable based on computer Martıńez-Navarro, 2004; Rook et al., 2004). This was based, among simulation of vegetation and climate models (Holmes, 2007), as other things, on the presence of Theropithecus in Cueva Victoria well as biogeographic models (O’Regan, 2008). Computer simula- (Gibert et al., 1995), Pirro Nord (Rook et al., 2004; Rook and tion models for the dispersal of Theropithecus’ such as ‘‘Stepping Martıńez Navaro, in press), Mirzapur (Gupta and Sahni, 1981), and Out,’’’ did not include the presence of this taxon in the Levantine ‘Ubeidiya (Belmaker, 2002). However, if future analysis of speci- corridor (Hughes et al., 2008); it would be of interest to rerun mens assigned to Theropithecus were to support a reassignment to additional computer simulation programs with the inclusion of an the Eurasian Paradolichopithecus, as has been suggested for Pirro ‘Ubeidiya Theropithecus to test the probability of such routes. Nord (Patel et al., 2007), the dispersal hypothesis would require The high proportion of African taxa in the mammalian faunal reevaluation. assemblage of ‘Ubeidiya, such as Pelorovis oldwayensis, Oryx sp., and Kolpochoerus olduvaiensis has suggested the presence of a well established dispersal route between East Africa and the Central Do the data support an assignment of UB 330 to the genus Jordan Valley (Haas, 1966; Tchernov, 1986; Martıńez-Navarro, Theropithecus? 2004). This route would have supported the dispersal of early hominin taxa as well. The presence of an additional African taxon Of the genera studied here, UB 330 is most similar to the modern (Theropithecus sp.) in the Central Jordan Valley provides additional genus Theropithecus. This is the only cercopithecine taxon that does support for this dispersal route during the Early Pleistocene. not differ from UB 330 in any size-adjusted measurement. DFA classification results indicate that the probability that UB 330 could Conclusions be classified as Theropithecus is 98.5%. The body size estimates are consistent with those for T. oswaldi females dating between Cercopithecid calcaneal morphology can be used to distinguish 2.5–1.2 Ma, and it has a high positive score along the first discrim- genera based on body size and degree of terrestrial vs. arboreal inant function, which is consistent with a high value of pedal power locomotion. A new specimen, UB 330, a right calcaneus from arm for large-bodied and terrestrial cercopithecines, as well as stratum III 12, ‘Ubeidiya, Israel, which has been dated to ca. 1.6 Ma, a high value for the ratio of anterior calcaneus height to length, can be attributed to a large-bodied cercopithecine and represents which is consistent with ankle inversion and flexibility. Within the a new primate taxon previously unidentified in the Early Pleisto- genus Theropithecus, UB 330 cannot be identified to species level. cene of the Southern Levant. At the genus level, it can be attributed However, T. oswaldi is the only Theropithecus species to have been to Theropithecus sp. with the highest probability, and represents the found in Eurasia during the time frame of ‘Ubeidiya (1.6–1.2 Ma) and only member of its genus in the Southern Levant and perhaps the is therefore the most probable species identification of UB 330. earliest in Eurasia. It is assigned to Theropithecus based on the high If this is correct, UB 330 represents the only occurrence value of its pedal power arm (indicative of terrestrial locomotion), of Theropithecus sp. in the eastern Mediterranean Levant and one of large body size, and high value of the ratio of anterior height to the oldest members of its genus out of Africa. The presence of length consistent with the inversion and ankle flexibility observed Theropithecus in Eurasia during the Early Pleistocene was sporadic in modern Theropithecus. While UB 330 could potentially also be although widely distributed, ranging from the Iberian Peninsula in attributed to Paradolichopithecus, this alternative is less probable. the west to the Indian sub-continent in the east (Jablonski, 1993). The presence of an African taxon in the Central Jordan Valley at T. oswaldi has been found at the site of Cueva Victoria, Spain (Gibert this date suggests a circum Mediterranean dispersal route for et al., 1995; Martıńez-Navarro et al., 2005), dated to ca. 1.2 Ma; at Theropithecus during the earlier part of the Pleistocene and the site of Mirzapur, India, dated to ca. 1.0–0.1 Ma (Gupta and Sahni, supports the presence of the Levantine corridor as a biogeographic 1981; Delson, 1993; Pickford,1993); and at the sites of Ternifine and route between Africa and Eurasia. The finding of Theropithecus sp. Thomas Quarry, Algeria (Delson, 1993), ranging in date from 1.0– in ‘Ubeidiya expands our knowledge of primate dispersals during 0.4 Ma (Alemseged and Geraads, 1998; Raynal et al., 2001). The the Early Pleistocene. The understanding and interpretation of the species has also been identified at the site of Pirro Nord, Italy, (Rook biogeography of large-bodied primates and the dispersal route of et al., 2004; Rook and Martıńez-Navarro, in press) dated to ca. 1.6– other ‘‘Out of Africa I’’ taxa is important in elucidating hominin 1.3 Ma, but this identification has recently been questioned (Patel dispersal patterns. et al., 2007). While Theropithecus fossils are rare in Eurasia, their finds Acknowledgments document the dispersal routes of large-bodied primates from Africa into Eurasia during the Early Pleistocene and mirror the possible The research was made possible by generous grants from the dispersal routes used by early Homo. To date, the presence of Irene Levy Sala CARE Foundation, the L.S.B. Leakey Foundation, the Theropithecus in Cueva Victoria in the west and in India in the east Richard Carley Hunt Wenner-Gren postdoctoral Fellowship, and have suggested two parallel, although not mutually exclusive, the American School of Prehistoric Research postdoctoral research dispersal routes: a westerly dispersal route via the Gibraltar straits, fellowship at Harvard University. The photographs of UB 330 were and a northern dispersal route along the Nile valley and Levantine taken by M. Barazani. I am indebted to O. Bar-Yosef, David Pilbeam, corridor (Tchernov, 1988; Petraglia, 2003; Turner and O’Regan, and the late E. Tchernov for their support and help throughout this 2007; O’Regan, 2008). An additional route across the Bab el Man- research, as well as to Navot Morag and Alon Barash for their deb strait is not probable as there is no evidence for a land bridge at technical assistance. I would like to thank Rivka Rabinovich, Judy

Please cite this article in press as: Belmaker, M., The presence of a large cercopithecine (cf. Theropithecus sp.) in..., J Hum Evol (2009), doi:10.1016/ j.jhevol.2009.08.004 ARTICLE IN PRESS

10 M. Belmaker / Journal of Human Evolution xxx (2009) 1–11

Chupasko, and Eileen Westwig for access to collections. I am pre´histoire Me´diterraneenne. E´ ditions du Comite´ des Travaux Historiques et grateful to Philip Rightmire, Eric Delson, W.E.H. Harcourt-Smith, Scientifiques, Paris, pp. 131–146. Gupta, V.L., Sahni, A., 1981. Theropithecus delsoni, a new cercopithecine species from and Tina Warinner for useful comments during the process of the Upper Siwliks of India. Bull. Geol. Soc. India. 14, 69–71. identification of the specimen, and for suggestions during the Haas, G., 1966. On the Vertebrate Fauna of the Lower Pleistocene Site of ‘Ubeidiya. preparation of this manuscript, and to Sarah Elton and two anon- Israel Academy of Sciences and Humanities, Jerusalem. Haas, G., 1968. On the Fauna of ‘Ubeidiya. Proc. Isr. Acad. Sci. Humanit. Sci. 7, 1–14. ymous reviewers for suggestions on earlier drafts of the Harrison, D.L., Bates, P.J.J., 1991. The Mammals of Arabia. Harrison Zoological manuscripts. Museum, Sevenoaks, Kent. Harrison, T., 1989. New postcranial remains of Victoriapithecus from the middle Miocene of Kenya. J. Hum. Evol. 18, 3–54. Supplementary data Holmes, K.M., 2007. Using Pliocene paleoclimatic data to postulate dispersal pathways of early hominins. Palaeogeogr. Palaeoclimatol. Palaeoecol. 248, 96–108. Supplementary data associated with this article can be found in Hughes, J.K., Elton, S., O’Regan, H.J., 2008. Theropithecus and ‘Out of Africa’ dispersal online version, at doi:10.1016/j.jhevol.2009.08.004. in the Plio-Pleistocene. J. Hum. Evol. 54, 43–77. Iwamoto, M., 1993. The ecology of Theropithecus gelada. In: Jablonski, N. (Ed.), Theropithecus: The Rise and Fall of a Primate Genus. Cambridge University Press, References Cambridge, pp. 441–452. Jablonski, N.G. (Ed.), 1993. Theropithecus: The Rise and Fall of a Primate Genus. Alemseged, Z., Geraads, D., 1998. Theropithecus atlanticus (Thomas 1984) (Primates: Cambridge University Press, Cambridge. Cercopithecidae) from the late Pliocene of Ahl al Oughlam, Casablanca, Jungers, W.L., Falsetti, A.B., Wall, C.E., 1995. Shape, relative size, and size-adjust- Morocco. J. Hum. Evol. 34, 609–621. ments in morphometrics. Yrbk. Phys. Anthropol. 38, 137–161. Ardito, G., Mottura, A., 1987. An overview of the geographic and chronological Kawai, M., Dunbar, R., Ohsawa, H., Mori, U., 1983. Social organization of Gelada distribution of West European Cercopithecoids. Hum. Biol. 2, 29–45. Baboons: social units and definitions. Primates 24, 13–24. Bar-Yosef, O., Goren-Inbar, N., 1993. The Lithic Assemblages of ‘Ubeidiya, a Lower Kimura, T., Hamada, Y., 1990. Development of epiphyseal union in Japanese Paleolithic Site in the Jordan Valley. The Institute of Archaeology, The Hebrew macaques of known chronological age. Primates 31, 79–93. University of Jerusalem, Jerusalem. Krentz, H.B., 1993. Postcranial anatomy of extant and extinct species of Belmaker, M., 2002. The first presence of Theropithecus sp. in the Southern Levant. Theropithecus. In: Jablonski, N.G. (Ed.), Theropithecus: The Rise and Fall of Isr. J. Zool. 48, 165. a Primate Genus. Cambridge University Press, Cambridge, pp. 383–422. Belmaker, M., 2006. Community structure through time: ‘Ubeidiya, a Lower Pleis- Kummer, H., 1995. Quest of the Scared Baboon: A Scientist’s Journey. Princeton tocene site as a case study. Ph.D. Dissertation. The Hebrew University of University Press, Princeton. Jerusalem. Langdon, J.H., 1986. Functional Morphology of the Miocene Hominoid Foot. Karger, Belmaker, M., 2009. Hominin adaptability and patterns of faunal turnover in the Basel. Lower-Middle Pleistocene transition in the Levant. In: Camps, M., Chauhan, P.R. Mallol, C., 2006. What’s in a beach? Soil micromorphology of sediments from the (Eds.), A Sourcebook of Paleolithic Transitions: Methods, Theories and Lower Palaeolithic site of ‘Ubeidiya, Israel. J. Hum. Evol. 51, 185–206. Interpretations. Springer, pp. 211–228. Martıńez-Navarro, B., 2004. Hippos, pigs, bovids, saber-toothed tigers, monkeys, Belmaker, M., Tchernov, E., Condemi, S., Bar-Yosef, J., 2002. New evidence for and hominids: dispersals through the Levantine corridor during late Pliocene hominid presence in the Lower Pleistocene in the Southern Levant. J. Hum. Evol. and early Pleistocene times. In: Goren-Inbar, N., Speth, J.D. (Eds.), Human 43, 43–56. Paleoecology in the Levantine Corridor. Oxbow Books, Oxford, pp. 37–52. Bramblett, C.A., 1969. Non-metric skeletal age changes in the Drajani Baboon. Am. Martıńez-Navarro, B., Palmqvist, P., 1995. Presence of the African Machairodont J. Phys. Anthropol. 30, 161–172. Megantereon whitei (Broom, 1937) (Felidae, Carnivora, Mammalia) in the Lower Braun, D., Ron, H., Marco, S., 1991. Magnetostratigraphy of the hominid tool-bearing Pleistocene site of Venta Micena (Orce, Granada, Spain), with some considerations Erk el Ahmar formation in the northern Dead Sea Rift. Isr. J. Earth Sci. 40, on the origin, evolution and dispersal of the genus. J. Archaeol. Sci. 22, 569–582. 191–197. Martıńez-Navarro, B., Belmaker, M., Bar-Yosef, O., 2009. The large carnivores from Caloi, L., Palombo, M.R., 1997. Biochronology of large mammals in the early and ’Ubeidiya (early Pleistocene, Israel): biochronological and biogeographical middle Pleistocene on the Italian peninsula. Hystrix 9, 3–12. implications. J. Hum. Evol. 56, 514–524. Cerling, T.E., Hay, R.L., 1986. An isotopic study of paleosol carbonates from Olduvai Martıńez-Navarro, B., Claret, A., Shabel, A.B., Pe´rez-Claros, J.A., Lorenzo, C., Palmqvist, P., Gorge. Quatern. Res. 25, 63–78. 2005. Early Pleistocene ‘‘hominid remains’’ from southern Spain and the taxonomic Delson, E., 1974. Preliminary review of Cercopithecid distribution in the Circum assignment of the Cueva Victoria phalanx. J. Hum. Evol. 48, 517–523. Mediterranean region. Me´moire du B.R.G.M. 78, 131–135. Mosimann, J.E., 1970. Size allometry: size and shape variables with characteristics of Delson, E., 1980. Fossil macaques, phyletic relationships and a scenario of deploy- the log normal and generalized gamma distribution. J. Am. Stats. Assoc. 665, ment. In: Lindburg, D.G. (Ed.), The Macaques: Studies in Ecology, Behavior and 930–945. Evolution. Van Nostrand Reinhold, New York, pp. 11–30. O’Regan, H.J., 2008. The Iberian Peninsula - corridor or cul-de sac? Mammalian Delson, E., 1993. Theropithecus fossils from Africa and India and the taxonomy of the faunal change and possible routes of dispersal in the last 2 million years. genus. In: Jablonski, N.G. (Ed.), Theropithecus: The Rise and Fall of a Primate Quatern. Sci. Rev. 27, 2136–2144. Genus. Cambridge University Press, Cambridge, pp. 157–189. Opdyke, N.D., Lindsay, E., Kukla, G., 1983. Evidence for earlier date of ‘Ubeidiya, Delson, E., Terranova, C.J., Jungers, W.L., Sargis, E.J., Jablonski, N.G., Dechow, P.C., Israel hominid site. Nature 304, 375. 2000. Body mass in Cercopithecidae (Primates, Mammalia): estimation and Patel, B.A., Gilbert, C.C., Ericson, K.E., 2007. Cercopithecoid cervical vertebral scaling in extinct and extant taxa. Anthropol. Paper Am. Mus. Nat. Hist. 83, morphology and implications for the presence of Theropithecus in early Pleis- 1–159. tocene Europe. J. Hum. Evol. 52, 113–129. Derricourt, R., 2006. Getting ‘‘Out of Africa’’: sea crossing, land crossings and culture Petraglia, M.D., 2003. The Lower Paleolithic of the Arabian Peninsula: occupations, in the hominin migrations. J. World Prehist. 19, 119–132. adaptations, and dispersals. J. World Prehist. 17, 141–179. Elton, S., 2002. A reappraisal of the locomotion and habitat preference of Pickford, M., 1987. The chronology of the Cercopithecoidea of East Africa. Hum. Evol. Theropithecus oswaldi. Folia Primatol. 73, 252–280. 2, 1–17. Falsetti, A.B., Jungers, W.L., Cote, T.M., 1993. Morphometrics of the callitrichid Pickford, M., 1993. Climatic change, biogeography, and Theropithecus. In: forelimb: a case study of size and shape. Int. J. Primatol. 14, 551–572. Jablonski, N.G. (Ed.), Theropithecus: The Rise and Fall of a Primate Genus. Fernandes, C.A., Rohling, E.J., Siddall, M., 2006. Absence of post-Miocene Red Sea Cambridge University Press, Cambridge, pp. 227–243. land bridges: biogeographic implications. J. Biogeogr. 33, 961–966. Raynal, J.P., Sbihi Alaoui, F.Z., Geraads, D., Magoga, L., Mohi, A., 2001. The earliest Galliari, C.A., 1988. A study of postnatal appendicular skeletal maturation in captive occupation of North Africa: the Moroccan perspective. Quatern. Int. 75, 65–75. born squirrel monkeys (Saimiri boliviensis). Am. J. Primatol. 16, 51–61. Rice, W.R., 1989. Analyzing tables of statistical tests. Evolution 43, 223–225. Geraads, D., 1986. Ruminants Pleístoce`ne d’Oubeídiyeh. In: Tchernov, E. (Ed.), Les Rodman, P.S., 1979. Skeletal difference of Macaca fascicularis and Macaca nemestrina Mammife`res du Pleístoce`ne Infe´rieur, de la Valleé du Jourdain a Oubeídiyeh. in relation to arboreal and terrestrial quadrupedalism. Am. J. Phys. Anthropol. Association Paleórient, Paris, pp. 143–182. 51, 51–62. Gibert, J., Ribot, F., Gilbert, L., Leakey, M.G., Arribas, A., Martıńez-Navarro, B., 1995. Rohling, E.J., Fenton, M., Jorissen, F.J., Bertrand, P., Ganssen, G., Caulet, J.P., 1998. Presence of the Cercopithecid genus Theropithecus in Cueva Victoria (Murcia, Magnitude of sea-level lowstands of the past 500,000 years. Nature 394, Spain). J. Hum. Evol. 28, 487–493. 162–164. Gowlett, G.A.J., 1979. Complexities of cultural evidence in the Lower and Middle Ron, H., Levi, S., 2001. When did hominids first leave Africa? New high-resolution Pleistocene. Nature 278, 14–17. magnetostratigraphy from Erk-el-Ahmar Formation. Isr. Geol. 29, 887–890. Gue´rin, C., Bar-Yosef, O., Debard, E., Faure, M., Shea, J., Tchernov, E., 1996. Mission Rook, L., Martıńez-Navarro, B., The large sized cercopithecid from Pirro Nord and the archeólogique et paleóntologique dans le Pleístoce`ne ancien d’Oubeídiyeh importance of Theropithecus in the early Pleistocene of Europe: Faunal markers for (Israe¨l): Re´sultats 1992–1994, 322. C.R. Acad. Sci., Paris, pp. 709–712. hominins outside Europe. Palaeontographica Abteilung A. (in press). Gue´rin, C., Bar-Yosef, O., Debard, E., Faure, M., Shea, J., Tchernov, E., 2003. Oubei- Rook, L., Martıńez-Navarro, B., Clark Howell, F., 2004. Occurrence of Theropithecus diyeh, carrefour biogeógraphique et culturel entre l’Afrique et l’Eurasie au sp. in the Late Villafranchian of Southern Italy and implication for Early Pleis- Paleólithique ancien. In: Vandermeersch, B. (Ed.), E´ changes et Diffusion dans la tocene ‘‘out of Africa’’ dispersals. J. Hum. Evol. 47, 267–277.

Please cite this article in press as: Belmaker, M., The presence of a large cercopithecine (cf. Theropithecus sp.) in..., J Hum Evol (2009), doi:10.1016/ j.jhevol.2009.08.004 ARTICLE IN PRESS

M. Belmaker / Journal of Human Evolution xxx (2009) 1–11 11

Sagi, A., 2005. Magnetostratigraphy of ‘Ubeidiya Formation, Northern Dead Sea Tchernov, E., 1988. The biogeographical history of the southern Levant. In: Yom- Transform, Israel. M.Sc. Dissertation. The Hebrew University. Tov, Y., Tchernov, E. (Eds.), The Zoogeography of Israel. Dr. Junk Publishers, Scheuer, L., Black, S., 2004. The Juvenile Skeleton. Elsevier Academic Press, London. Dordrecht, pp. 159–250. Shea, J.J., Bar-Yosef, O., 1998. Lithic assemblages from new (1988–1994) excava- Tchernov, E., Volokita, M., 1986. Insectivores and Primates from the early Pleisto- tions at ‘Ubeidiya: a preliminary report. Mitekufat Haeven J. Isr. Prehist. Soc. cene of ‘Ubeidiya Formation. In: Tchernov, E. (Ed.), Les Mammife`res du Pleís- 28, 5–19. toce`ne Infe´rieur, de la Valleé du Jourdain a Oubeídiyeh. Association Paleórient, Siddall, M., Rohling, E.J., Almogi-Labin, A., Hemleben, C., Meischner, D., Schmelzer, I., Paris, pp. 54–62. Smeed, D.A., 2003. Sea-level fluctuations during the last glacial cycle. Nature Tobias, P., 1966a. A Member of the Genus Homo from ‘Ubeidiya. Publications of the 423, 853–858. Israel Academy of Sciences and Humanities, Jerusalem. Sokal, R.R., Rohlf, J.F., 1995. Biometry: The Principles and Practice of Statistics in Tobias, P.V., 1966b. Fossil hominid remains from ‘Ubeidiya, Israel. Nature 211, Biological Research. W.H. Freeman and Company, New York. 130–133. Sondaar, P., Van der Geer, A.A.E., 2002. Arboreal and terrestrial traits as revealed by Turner, A., O’Regan, H.J., 2007. Zoogeography: primate and early hominin distri- the primate ankle joint. Annales Geólogiques des Pays Helleńiques 1e Se´rie 39A, bution and migration patterns. In: Henke, W., Tattersall, I. (Eds.), Handbook of 88–98. Paleoanthropology. Principles, Methods and Approaches. Springer-Verlag, Ber- Sondaar, P., Van der Geer, A.A.E., Dermitzakis, M.D., 2006. The unique postcranial of lin, pp. 421–502. the Old World Monkey Paradolichopithecus: more similar to Australopithecus Van der Geer, A.A.E., Sondaar, P., 2002. The postcranial elements of Para- than to Baboons. Hellenic J. Geosci. 41, 19–28. dolichopithecus arvernesis (Primates, Cercopithecidae, Papionini) from Lesvos, Stekelis, M., 1966a. Archaeological Excavations at ‘Ubeidiya, 1962–1964. Israel Greece. Annales Geólogiques des Pays Helleńiques 1e Se´rie 39A, 71–86. Academy of Sciences and Humanities, Jerusalem. Verosub, K., Tchernov, E., 1991. Resultats pre´liminaires de l’e´tude magneto- Stekelis, M., 1966b. Archaeological Excavations at ‘Ubeidiya: 1960–1963. The Israel stratigraphique d’une se´quence se´dimentaire a` l’industrie humaine en Israe¨l. In: Academy of Sciences and Humanities, Jerusalem. Vandermeersch, B. (Ed.), Les Premiers Peuplements de l’Europe. C.N.R.S., Paris, Stekelis, M., Bar-Yosef, O., Schick, T., 1969. Archaeological Excavations at Ubeidiya, pp. 237–242. 1964–1966. Israel Academy of Sciences and Humanities, Jerusalem. Wildman, D.E., Bergman, T.E., al-Aghbari, A., Sterner, K.N., Newman, T.K., Philips- Strasser, E., 1988. Pedal evidence for the origin and diversification of cercopithecid Conroy, J.E., Jolly, C.J., Disotell, T.R., 2004. Mitochondrial evidence for the origin clades. J. Hum. Evol. 17, 225–245. of hamadryas baboons. Mol. Phylogent. Evol. 32, 287–296. Strasser, E., 1992. Hindlimb proportions, allometry, and biomechanics in Old World Winney, B.J., Hammond, R.L., Macasero, W., Flores, B., Boug, A., Biquand, V., Monkeys (Primates, Cercopithecidae). Am. J. Phys. Anthropol. 87, 187–213. Biquand, S., Bruford, W., 2004. Crossing the Red Sea: Phylogeography of the Szalay, F.S., Delson, E., 1979. Evolutionary History of the Primates. Academic Press, hamadryas baboon, Papio hamadryas hamadryas. Mol. Ecol. 13, 2819–2827. New York. Wrangham, R., 1980. Bipedal locomotion as a feeding adaptation in gelada baboons, Tchernov, E., 1981. The biostratigraphy of the Levant. In: Cauvin, J., Sanlaville, P. and its implications for hominid evolution. J. Hum. Evol. 9, 329–331. (Eds.), Pre´histoire du Levant. Chronologie et Organisation de l’Espace depuis les Yirga, S., 2002. Hind limb bones and locomotion in the Old World monkeys. Sinet Origines jusqu’au VIe Milleńaire. C.N.R.S., Paris, pp. 67–97. 25, 205–226. Tchernov, E. (Ed.), 1986. Les Mammife`res du Pleístoce`ne Infe´rieur de la Valleédu Youlatos, D., 2003. Calcaneal features of the Greek Miocene primate Mesopithecus Jourdain a Oubeídiyeh. Association Paleórient, Paris. pentelicus (Cercopithecoidea: Colobinae). Geobios 36, 229–239.

Please cite this article in press as: Belmaker, M., The presence of a large cercopithecine (cf. Theropithecus sp.) in..., J Hum Evol (2009), doi:10.1016/ j.jhevol.2009.08.004