The phylogenetic affinities of Otavipithecus namibiensis

Michelle Singleton Department of Vertebrate Paleontology, American Museum of Natural History, Central Park West and 79th Street, New York, NY 10034, U.S.A. Received 8 April 1999 Revision received 31 August 1999 and accepted 8 September 1999 Keywords: Otavipithecus, Miocene hominoids, Afropithecini, gap coding. The phylogenetic affinities of Otavipithecus namibiensis The middle Miocene hominoid Otavipithecus namibiensis is the first and most complete fossil ape from sub-equatorial Africa and represents a significant addition to the taxonomically sparse African middle Miocene hominoid fossil record. The Otavipithecus hypodigm comprises the holotype mandible, which presents a unique mosaic of dental and gnathic characters, and several attributed cranial and postcranial elements which resemble the stem hominoid Proconsul. Contrary to initial hopes that this discovery would provide new insights into hominoid morphological diversity and phylogenetic relationships, a variety of conflicting phylogenetic hypotheses have been advanced suggesting ties to virtually every major large-bodied hominoid group (Conroy et al., 1992; Andrews, 1992a; Conroy, 1994; Pickford et al., 1994; Begun, 1994a). Cladistic analysis of a matrix of 22 qualitative and ten quantitative characters of the mandible and mandibular dentition found no support for a close phylogenetic relationship between Otavipithecus and either the African ape or great ape clades, or with any of the Eurasian fossil hominoids with which it has previously been compared. A close relationship between Otavipithecus and Kenyapithecus cannot be ruled out, but is deemed unlikely on the basis both of morphological comparisons and the absence of support within a cladistic framework. The present analysis indicates that Otavipithecus is most closely related to Afropithecus, as previously suggested by Andrews (1992a) among others. Due to lack of statistical support for this result, a conservative interpretation, that these taxa represented related but divergent lineages of a late early Miocene hominoid radiation, is currently favored. Findings are consistent with the allocation of Otavipithecus to Andrews’ (1992a) tribe Afropithecini which represents the sister group to Kenyapithecus and the extant ape clade.  2000 Academic Press Journal of Human Evolution (2000) 38, 537–573 doi:10.1006/jhev.1999.0369 Available online at http://www.idealibrary.com on Introduction Otavipithecus namibiensis, the fossil hominoid from northern Namibia, is the first and, with the exception of a molar fragment from Ryskop, South Africa (Senut et al., 1997), only known Miocene ape from subequatorial Africa (Conroy et al., 1992). In addition to extending the known range of Address correspondence to: Department of Anatomy, New York College of Osteopathic Medicine, P.O. Box 8000, Old Westbury, NY 11568, U.S.A. E-mail: [email protected] 0047–2484/00/040537+37$35.00/0 Miocene hominoids, Otavipithecus represents a significant addition to the taxonomically sparse African middle Miocene hominoid record, and the holotype, a right mandibular corpus with partial dentition, presents a unique mosaic of dental and gnathic characters. The mandible was recovered from dolomitic bone breccia collected in a mine dump context at the Berg Aukas site in the Otavi Mountain region of Namibia (Conroy et al., 1992, 1993b). While the disturbed context prevents  2000 Academic Press 538 .  Figure 1. Holotype of Otavipithecus namibiensis. Right mandibular corpus in lateral (top) and medial (bottom) views. Scale in centimeters. (Photos courtesy of Glenn Conroy.) stratigraphic dating, comparative analyses of associated microfauna have given dates in the 131 Ma range, indicating a middle Miocene age (Conroy et al., 1992, 1996; Conroy, 1996; Pickford et al., 1997). Subsequent work at Berg Aukas has yielded several cranial and postcranial elements attributed to Otavipithecus, including a partial frontal bone, an atlas vertebra, a proximal ulna, and a middle manual phalanx (Conroy et al., 1993a, 1996; Pickford et al., 1997). Conroy et al. (1992) provide a detailed description and diagnosis of the holotype mandible (see Figure 1), and the remaining attributed material has been published in a variety of forums (Conroy et al., 1993a; Pickford et al., 1997; Senut & Gommery, 1997). The Otavipithecus holotype is a right mandibular corpus preserving the internal symphysis, incisor alveoli, canine and distal P3 roots and P4–M3. Table 1 lists its diagnostic traits (Conroy et al., 1992), and Figure 2 shows the specimen in greater detail. The mandibular corpus is only moderately robust, both deeper and more gracile than forms such as Kenyapithecus1 and Sivapithecus. A lingual alveolar buttress extends distally from the symphysis to the level of mesial M1. The buttress is more pronounced than that observed in Afropithecus or Sivapithecus, most closely resembling that of Proconsul. The lateral buttress (lateral eminence) extends from the ramus to reach the M1–M2 level anteriorly, but is weaker than that of Kenyapithecus, Sivapithecus, and the more robust Afropithecus specimens. The well-developed postcanine fossa is restricted posterosuperiorly by the lateral buttress and anteriorly by the pronounced canine-P3 eminence. A single mental foramen lies at the anteroinferior margin of the postcanine fossa below the distal P3. The retromolar region is broad and the M3 falls anterior to the root of the ascending ramus. The medial corpus preserves the mylohyoid groove and mylohyoid line, as well as a distinct submandibular fossa (Conroy et al., 1992). The latter is large and clearly demarcated, but does not extend posteriorly to form the ‘‘intertoral sulcus’’ (Brown, 1989) found in specimens of Sivapithecus, Dryopithecus, and possibly Kenyapithecus. The external symphyseal surface is largely missing; however, the right subincisal region is present and shows distinct hollowing. The symphysis appears 1 Discussion of Kenyapithecus is complicated by the current lack of consensus as to how many genera and species are represented among the various sites (Andrews, 1992a; McCrossin & Benetift, 1997; Nakatsukasa et al., 1998; Ward et al., 1999). The Kenyapithecus sample for this study includes material from Maboko Island, Fort Ternan, Majiwa, and Nyakach, representing a maximum of two genera (Ward et al., 1999). Unless otherwise noted, ‘‘Kenyapithecus’’ is used here in the broadest sense, including all material previously referred to Kenyapithecus and/or Equatorius. Except as noted, morphological comparisons are restricted to features common to all sub-samples.    Table 1 OTAVIPITHECUS 539 Diagnostic traits of Otavipithecus namibiensis (Conroy et al., 1992) 1 2 3 4 5 6 7 8 9 10 11 12 ‘‘Puffy’’ molar cusps forming distinctive Y-5 pattern in which the hypoconid lingual slope extends beyond mid-axis of the tooth Absence of beaded buccal cingulum Presence of protostylar ridges on M2 and M3 Molars ‘‘squared-off’’, not elongated M2 >M3 >M1 Moderate development of the inferior transverse torus Large retromolar space Wear pattern suggesting thin enamel and/or high dento-enamel relief M3 not obscured by anterior root of ascending ramus Little differential wear on molars Mandible depth does not decrease markedly mesiodistally Extremely narrow incisor region more gracile than that of Kenyapithecus or Sivapithecus, although this may be due in part to incomplete development of the canine root (Brown, 1997). The internal symphysis is exceptionally narrow, with the alveolar planum vertically inclined and laterally restricted, particularly inferiorly at the level of the superior transverse torus. The inferior transverse torus is delimited by a small but distinct genial fossa. It is slightly larger than the superior transverse torus, extending posteriorly to the mesial P4 level, but is neither thickened in the manner of Kenyapithecus and Sivapithecus (Conroy et al., 1992), nor posteriorly elongated to form a true ‘‘simian shelf’’. The incisor alveoli are small relative to the adjacent canine root, strongly mesiodistally compressed, and vertically oriented. The incisor region is both extremely narrow and strongly curved. The canine root is large and relatively stout. The P3 fragment preserves the distal fovea, which is small and mesiodistally restricted. The position of the mesial alveolus and orientation of the distal crown indicate the tooth was buccally rotated (Conroy et al., 1992). The P4 crown is ovoid in outline, lacking the mesiobuccal flare found in Sivapithecus and Dryopithecus, and is slightly buccally rotated (Conroy et al., 1992). The subequal protoconid and metaconid are joined by a poorly defined protocristid, setting off a small diamond-shaped mesial fovea. The talonid basin is only slightly depressed and lingually restricted by the presence of a distinct entoconid. As noted by Conroy et al. (1992), the molar teeth are characterized by inflated, bunodont cusps with poorly defined shearing crests. The lingual face of the hypoconid extends beyond the longitudinal axis of the crown, restricting the talonid basin. The crowns are short and broad with lingual margins straight and vertical, while buccal margins are broadly curved with marked basal flare, most pronounced on M2. Each molar exhibits a deflecting wrinkle, a median wrinkle of the metaconid joining the entoconid near the center of the occlusal surface (Swindler & Ward, 1988). The first molar shows a minute ectostylid remnant at the mesiobuccal developmental groove, while M2 and M3 possess distinct protostylid and ectostylid (M3 only) cingular elements. The cheek teeth are minimally worn; however, M1 displays small apical pits, a wear pattern consistent with thin cuspal enamel and/or high dentine horn penetrance (Conroy et al., 1992). This inference has been confirmed via CT imaging and confocal microscopy which have shown Otavipithecus to have thin dental enamel with Pattern 1 enamel predominant in the more superficial layers (Conroy et al., 1995). Of the remaining material attributed to Otavipithecus, the frontal fragment is the 540 .  Figure 2. Cast of Otavipithecus holotype mandible demonstrating the diagnostic morphology (see Table 1) in clearer detail. Top to bottom: occlusal, lateral, and medial views. only additional cranial specimen. The specimen preserves the right orbital margin, the glabellar region, portions of the temporal crests, and a significant portion of the frontal squama (Pickford et al., 1997). The specimen is described as having a narrow    superciliary ridge and lacking both supraorbital torus and supratoral sulcus, but possessing an extensive frontal sinus which excavates the frontal squama into orbital and endocranial plates (Pickford et al., 1997). The interorbital region is characterized as broad, with an estimated width of 24 mm. The temporal crests are reported to be sharp, overhanging the temporal fossa, and failing to converge posteriorly. Their position relative to the superciliary eminence leads the authors to suggest that Otavipithecus was mildly klinorynch. In all aspects, the morphology is described as comparable to that of Proconsul (Pickford et al., 1997). Analyses of the known postcranial elements also depict a primitive hominoid morphology. The original description of the Otavipithecus atlas provides comparisons with arboreal and terrestrial cercopithecoids and extant hominoids (Conroy et al., 1996). The orientation of the superior and inferior articular facets; the proportions and relative projection of the transverse processes; and the dorsal proportions of the vertebral canal are intermediate between the hominoid and cercopithecoid conditions, while the dimensions of the anterior arch and configuration of the anterior tubercle are more hominoidlike, suggesting a transitional morphology (Conroy et al., 1996). However, Senut & Gommery (1997) emphasize the latter features, and propose that Otavipithecus habitually assumed a more orthograde posture similar to that of the bonobo. The proximal ulna is damaged, lacking both olecranon and coronoid processes, however the preserved trochlear (semilunar) notch is described as distally- and inferiorly-oriented in the manner of pliopithecids and African early Miocene hominoids, such as Proconsul and Dendropithecus (Conroy et al., 1993a, b; Senut & Gommery, 1997). Although eroded, the radial notch is described as small and triangular, most closely resembling that of Proconsul nyanzae (Senut & Gommery, OTAVIPITHECUS 541 1997). The phalanx is described as long, slender, and moderately curved, with welldeveloped insertions for the m. flexor digitorum superficialis, and articular surfaces indicating limited potential for hyperextension (Conroy et al., 1993a). These traits are present in Proconsul (Leakey & Walker, 1997), as well as arboreal Old and New World monkeys (Senut & Gommery, 1997). Thus, all known postcranial evidence is consistent with the interpretation that Otavipithecus, like Proconsul, was a slow-moving arboreal quadruped (Conroy et al., 1993a, b; Leakey & Walker, 1997; Senut & Gommery, 1997). Dental estimates give a body weight for Otavipithecus of 14–20 kg (Conroy et al., 1992). This figure is supported independently by weight estimates based on the Otavipithecus phalanx (Conroy et al., 1993a) and atlas (Conroy et al., 1996). Biomechanical studies of the Otavipithecus mandible have suggested a capability to withstand torsional stresses comparable to that of Pongo or Australopithecus (Schwartz & Conroy, 1996). However, the combination of bunodont molar cusps lacking welldefined shearing crests, thin dental enamel, minimal differential molar wear, and a narrow incisor region indicates a relatively soft and non-abrasive diet requiring minimal incisal preparation, probably soft fruits, young leaves, and other soft plant parts (Conroy et al., 1992, 1995). Faunal analysis indicates the middle Miocene climate of the Otavi region was more humid than today, and that some degree of woodland habitat was present (Conroy et al., 1992; Conroy, 1996), an environment consistent with this species’ inferred locomotor pattern (Senut & Gommery, 1997). Thus, Otavipithecus is a Miocene Everyman: a medium-sized ‘‘hominoid of archaic aspect’’ (Pilbeam, 1996), lacking apparent locomotory or dietary specializations, sharing individual traits with specific taxa, but morphologically distinct. Given its temporal 542 .  and spatial position, it was originally hoped this discovery would provide new insights into hominoid morphological diversity and phylogenetic relationships. Instead, a variety of disparate phylogenetic hypotheses have been advanced advocating ties to virtually every major large-bodied hominoid group (Andrews, 1992a; Conroy et al., 1992; Conroy, 1994; Pickford et al., 1994; Begun, 1994a). This multiplication of contradictory reconstructions has led Pilbeam (1996:163) to conclude that ‘‘The [Otavipithecus] mandible is uninformative and there is no reason to believe that any of the currently available characters have any particular phylogenetic utility’’. By contrast, Begun proposed that, while ‘‘mandibles are poor indicators of phylogenetic relationships among hominoids’’ (Begun, 1994a:392), a cladistic analysis of ‘‘a carefully assembled set of characters for which ranges and pattern of variation . . . are well established’’ (1994a:385) would clarify the affinities of Otavipithecus. In fact, much of the debate regarding the affinities of Otavipithecus has surrounded the rigor of character analyses upon which the competing hypotheses are based (Conroy, 1994; Begun, 1994a). This emphasis of the assessment of characters reflects a growing awareness among primate systematists of the need for explicit methods of identifying and testing the character state distributions used in cladistic analysis. This need arises in part from the difficulties involved in coding quantitative morphological data into the discrete character states required by parsimony analysis, and is particularly pertinent to the study of fossil taxa as a large proportion of hard tissue characters show continuous distributions, often with considerable range overlap between groups, rather than qualitative variation (Trinkaus, 1990). As part of a broader examination of the treatment of such characters in hominoid phylogenetic studies (Singleton, 1998), a cladistic analysis of hominoid mandibular morphology was conducted in the hope of clarifying the phylogenetic affinities of Otavipithecus. Materials and methods The study sample comprised 597 mandibular specimens—complete mandibles, fragments, and isolated teeth—representing 25 extinct and extant catarrhine genera (see Table 2). Samples of extant taxa were sexbalanced and included only wild-shot individuals of known provenience. The fossil sample included the majority of recognized African and Eurasian fossil hominoid genera, and sample sizes were maximized within the limits of available material. With few exceptions, data were collected on original specimens. Where original material was unavailable, data were collected on good quality plaster or resin casts. Standard linear measurements of the mandible and dentition were taken emphasizing the anatomical regions preserved in the Otavipithecus holotype, namely the postcanine tooth row, the mandibular corpus, and the symphyseal region (see Appendix A). Molar teeth were sorted into wear categories according to Benefit’s (1993) criteria, and measurements of cusp height and proximity were made only where the original positions of cusp apices could be determined with a high degree of certainty (wear category 5 or less). Characters for cladistic analysis (Appendices B and C) were drawn from the hominoid phylogeny literature (Delson & Andrews, 1975; Fleagle & Kay, 1983; Andrews, 1985, 1987, 1988, 1992b; Groves, 1986, 1987; Martin, 1986; Andrews & Martin, 1987; Groves & Paterson, 1991; Begun, 1992, 1994a), creating a comprehensive list of mandibular characters, both qualitative and quantitative, traditionally considered to have phylogenetic value for Miocene hominoids.    543 OTAVIPITHECUS Table 2 Study sample Taxa n Mandibles/ isolated teeth Extant Gorilla gorilla Pan troglodytes Pan paniscus Pongo pygmaeus Hylobates lar European fossils Dryopithecus brancoi Dryopithecus crusafonti Dryopithecus fontani Dryopithecus laietanus Ouranopithecus macedoniensis Oreopithecus bambolii Asian fossils Sivapithecus indicus Sivapithecus sivalensis Sivapithecus simonsi Griphopithecus alpani Lufengpithecus lufengensis African fossils Aegyptopithecus zeuxis Afropithecus turkanensis Dendropithecus macinnesi Kalepithecus songhorensis Kenyapithecus africanus Kenyapithecus wickeri Limnopithecus evansi Limnopithecus legetet Micropithecus clarki Micropithecus leakeyorum Nyanzapithecus pickfordi Nyanzapithecus vancouveringorum Otavipithecus namibiensis Proconsul africanus Proconsul heseloni Proconsul major Proconsul nyanzae ‘‘Proconsul major’’ (Napak) Propliopithecus ankeli Propliopithecus chirobates Rangwapithecus gordoni Simiolus enjiessi Turkanapithecus kalakolensis Reference for specimen attribution 36/0 35/0 34/0 41/0 40/0 2/0 1/2 3/0 9/13 9/0 9/0 Begun & Kordos, 1993 Begun, 1992 Lartet, 1856 Golpe-Posse, 1993 Bonis & Melentis, 1977 Hürzeler, 1949 9/0 12/0 2/0 1/0 2/0 Brown, 1989 Brown, 1989 Brown, 1989 Alpagut et al., 1990 Wu, 1987 16/0 8/4 16/24 7/1 2/20 2/5 13/8 5/12 5/12 5/4 0/13 1/0 1/0 4/7 11/24 5/22 12/25 5/1 2/0 7/0 6/23 3/0 1/0 Simons & Rasmussen, 1991 Leakey & Leakey, 1986a Harrison, 1988 Harrison, 1988 Pickford, 1985 Pickford, 1985 Harrison, 1988 Harrison, 1988 Harrison, 1988 Harrison, 1989 Harrison, 1986 Harrison, 1986, 1989 Conroy et al., 1992 Walker et al., 1993 Walker et al., 1993 Martin, 1980 Walker et al., 1993 Bishop, 1958 Simons et al., 1987 Simons & Rasmussen, 1991 Harrison, 1988 Leakey & Leakey, 1987 Leakey & Leakey, 1986b Sample sizes are broken down by ‘‘mandibles’’ (including mandible fragments and associated tooth rows) and isolated teeth. References are to sources for fossil specimen attribution, incuding original descriptions and recent revisions and reviews. Qualitative character coding The coding of qualitative characters requires an assessment of the character states observed in individual specimens and the assignment of discrete character codes to groups based on the state(s) observed in their members (Forey et al., 1992; Thiele, 1993). Where a single character state is 544 .  present, coding simply involves assigning the value for that state, but the presence of multiple states in a single taxon makes coding more difficult. A common solution is to choose an arbitrary cut-off, usually a percentage representation, below which an observed character state is not represented in the code for that group. For example, Strait et al. (1997) used a 10% cut-off for large extant samples, but advocated the use of strict polymorphic coding for fossil groups, i.e., groups were coded for multiple states even if a given state appeared in only a single individual. Because of the large number of taxa sampled, as well as the observed variability of some dental traits, strict polymorphic coding was found to be impractical. Instead, a cut-off value was applied in the coding of all taxa, both extant and fossil. Qualitative morphological traits of the mandible and mandibular dentition were scored for each specimen, individual qualitative character scores were tallied by group, and, where more than one state was present within a taxon, character state percentages were calculated and character codes assigned based on a set of predetermined criteria. Appendix B summarizes the qualitative characters examined, the character states considered, and the specific criteria by which final codes were assigned. In general, for simple binary characters, a state occurring in fewer than 15% of individuals in a taxon was considered absent, and the character was coded for the presence of the remaining state. If both states were present in 15% or more of individuals, the character was coded as polymorphic, i.e., both states occurring coequally. The 15% cut-off was determined based upon the binomial distribution of alternate binary character states for the median taxon sample size of n=12 [see Singleton (1998) for a discussion of cut-off calculation]; however, multistate characters were coded using the same criterion. In the case of P3 Honing Facet Development (#4), a character whose expression is dependent on wear, coding was modified to reflect the possibility of sampling bias against older individuals as well as the observed character state distributions in the present sample. Appendix D provides the resulting qualitative character matrix. Quantitative character coding Quantitative character coding involves the computation of character variables, the delimitation of states from the resulting distribution, and the assignment of codes to the recognized states. Character variables traditionally have been computed from linear metric variables expressing ‘‘shape’’ as a proportional relationship—‘‘relative width’’ or ‘‘relative depth’’ of a feature compared to a chosen size surrogate (Jungers et al., 1995). The two most frequently employed such transformations, ratios and residuals from the allometric line, ‘‘represent competing expressions of shape within this relative framework’’ (Jungers et al., 1995:138). Both ratios and residuals are dimensionless variables connected for isometric size. However, it is well established that ratios do not correct for allometric effects (Gould, 1966; Corruccini, 1975, 1978; Atchley et al., 1976; Dodson, 1978; Preuschoft, 1989; O’Higgins, 1989; Smith, 1999), whereas residuals from the ordinary least squares (OLS) regression line may do so (Atchley et al., 1976; Atchley, 1978; Atchley & Anderson, 1978; Albrecht, 1978; Smith, 1984; 1999; Preuschoft, 1989; Albrecht et al., 1993). Because morphological similarities due to allometric effects are a potential source of homoplasy, phylogenetic analyses have favored residual values as ‘‘size-corrected’’ shape variables (Hartman, 1983; Jungers et al., 1995; Smith, 1999). Still, the statistical and geometric properties of these variables and the biological assumptions underlying their use remain sources of    dispute (Corruccini, 1987, 1995; Jungers et al., 1995). Given this uncertainty, Singleton (1998) compared the performance of log-ratio and log-residual representations of 23 hominoid mandibular and dental characters (see Appendix C). Characters were evaluated for the presence of allometric scaling, ratio and residual character values were converted to discrete character codes, and parsimony analyses were performed. Despite the prevalence of allometry (less than half the characters scaled isometrically), ratio and residual variables yielded equivalent taxon rankings, character state distributions, and cladogram topologies. However, residual-based analyses consistently resulted in more highly resolved cladograms, shorter tree lengths, and better goodness-of-fit indices, an apparent result of residuals’ statistical properties (Singleton, 1998; but see Smith, 1999). Therefore, in this study, quantitative characters were represented by individual residual values from the OLS regression of log-transformed variables. The fragmentary fossil sample precluded the use of Mosimann size variables such as the geometric means of all measurements (Mosimann, 1970; Mosimann & Malley, 1979). Instead, linear measurements such as M2 length or P4 corpus depth were chosen as independent, i.e., size, variables on a character by character basis. While less robust than Mosimann size variables, this approach permitted the inclusion of incomplete specimens and isolated teeth. Appendix C details the measurements used in the computation of residual values. A variety of approaches has been recommended for the delimitation of discrete character states from continuous shape variables (Mickevich & Johnson, 1976; Simon, 1983; Thorpe, 1984; Archie, 1985; Chappill, 1989; Goldman, 1988; Strait et al., 1996). In this study, residual character values were converted to discrete character states and coded using simple gap coding OTAVIPITHECUS 545 (Mickevich & Johnson, 1976; Thorpe, 1984), a method in which differences between successive pairs of rank-ordered taxon means are compared to a critical value to identify ‘‘gaps’’ in the character distribution. Where the difference exceeds this value, usually the pooled within-group standard deviation sp, the taxa are assigned to different states and coded accordingly. This approach is analogous to discriminate analysis (Archie, 1985), in so far as betweengroup differences are scaled against an estimate of average within-group variation; however it does not constitute a formal statistical test. Gap coding has been shown to be sensitive to sampling error and changes in sample composition (Thorpe, 1984), and often fails to recognize statistically significant differences between taxon means (Archie, 1985). For these reasons, methods based upon multiple pairwise comparisons tests— homogenous subset coding (Simon, 1983) and divergence coding (Thorpe, 1984), among others—have been recommended on theoretical grounds (Rae, 1998) and on the basis of actual empirical comparisons (Chappill, 1989). Unfortunately, these methods have demonstrated poor performance for small and/or unequal sample sizes (Archie, 1985), both key issues in fossil studies. Singleton (1998) demonstrated that gap coding provides a useful, albeit conservative, approach to analysis of fossil hominoid character distributions and showed that inclusion of gap coded characters can improve the resolution of estimated phylogenies. Only characters exhibiting low ratios of within-group to between-group variability are suitable for gap coding (Thorpe, 1984; Chappill, 1989); therefore, one-way analysis of variance (ANOVA) was conducted for character values across groups using SPSS 6.1.1 for the Power Macintosh. All analyses were significant at P<0·01; however, the assumptions of the ANOVA model were not 546 .  satisfied. Levene tests run concurrently with each ANOVA showed that for only a minority of characters was the assumption of homogeneity of variance supported. This finding was notable both because groups must be statistically distinguishable for gap coding to be applied meaningfully and because the requisite critical value is derived within the ANOVA model. To explore the possible impact of this heterogeneity of variance, a Kruskal–Wallis test of character values was performed. This nonparametric test allowed an examination of group distinctiveness unhampered by distributional assumptions. For all characters, groups were significantly different at P<0·05, confirming the statistical distinctiveness of the groups within this sample. Having satisfied this criterion for quantitative character coding, all characters were retained for subsequent analysis. To examine the impact of heteroscedasticity on the formulation of the mean square (s2p) values from which the critical values are derived, ANOVAs were recomputed using only the larger extant samples. All analyses were significant at P<0·001, and Levene statistics showed the assumption of homogeneity of variance to be satisfied in all cases. Comparison of mean square values for the extants-only analyses with those for the entire sample showed that values for the extant sample were either identical to or smaller than those for the total sample in all cases. Under the gap coding criterion, larger critical values require that groups must be ‘‘more different’’ to be recognized. Thus, while the mean square values derived from the extants-only analyses are statistically more robust, the values based on the entire sample are actually more conservative, and these values were retained for use in subsequent analyses. Of the original 23 quantitative characters, 10 yielded parsimony-informative codings under gap coding with a critical value of 1/2sp. Appendix D provides the resulting quantitative character matrix. Signal tests The matrix combining parsimonyinformative qualitative and quantitative characters was evaluated for phylogenetic signal using both the random tree-length distribution (g1) test (Hillis & Huelsenbeck, 1992) and the permutation tail probability (PTP) test (Archie, 1989; Faith, 1991; Faith & Cranston, 1991). While these tests have been criticized as dependent upon tree topologies and lacking sufficient discriminatory power (Slowinski & Crother, 1998; Lyons-Weiler & Hoelzer, 1997), they do provide a minimum standard for phylogenetic signal content. The Random Trees option of PAUP 3.1.1 (Swofford, 1993) was used to construct the tree-length frequency distribution for 100,000 randomly sampled trees. The resulting distribution was leftskewed with g1 =
0.31. Interpolating from published critical values (Hillis & Huelsenbeck, 1992, Tables 1 and 2), this value is significant at P<0·01, indicating that the tree-length distribution is significantly more skewed than would be expected from random data (Hillis & Huelsenbeck, 1992). The permutation tail probability for 1000 random permutations of the character matrix was calculated using PAUP* 4.0b (Swofford, 1998). The resulting PTP value of 0·001 permits rejection of the null hypothesis of random character covariation and indicates the presence of significant ‘‘cladistic covariation’’ within the data set (Faith & Cranston, 1991). Parsimony analysis Parsimony analysis was performed upon the combined matrix of informative qualitative and quantitative characters using PAUP 3.1.1 (Swofford, 1993). Trees were rooted using Propliopithecus and Aegyptopithecus as outgroups. All characters were unordered and equally weighted. Although a consensus is emerging in favor of ordering multistate characters, particularly those derived from continuous morphological data (Lipscomb,    1992; Slowinski, 1993; Thiele, 1993), Slowinski (1993) has shown that, while ordering of multistate characters increases the resolution of cladistic analyses, it does not necessarily increase the accuracy of the estimated phylogenies. Further, he found no statistically significant differences in congruence between trees based upon ordered and unordered characters. Given the goals of the broader study, it was considered desirable to minimize the number of assumptions, and subsequent analyses confirmed that the use of ordered characters would not have altered the final conclusions. The size of the data matrix precluded the use of exhaustive or branch-and-bound search algorithms (Swofford, 1991). To maximize the chances of identifying all most parsimonious trees, a series of heuristic searches was performed varying stepwise addition and branch swapping methods to identify the approximate length of the shortest tree. This tree length was adopted as an upper boundary for a subsequent, more comprehensive heuristic search using ten repetitions of random stepwise addition followed by TBR branch-swapping on all trees in memory (Steepest Descent Option), saving all minimal trees at each repetition (MULPARS Option). This procedure was repeated to look for additional islands of trees potentially missed by earlier searches. This approach yielded 225 equally parsimonious trees (Length=332; CI=0·88; RI=0·61), 150 of which represented unique and fully-resolved solutions, with the remaining trees containing one or more polytomies. The strict consensus tree was computed and consensus and goodnessof-fit indices calculated. Bremer support indices (Bremer, 1988, 1994) were calculated for the strict consensus nodes, using TreeRot (Sorenson, 1996) to generate constraint statements for nodes of an arbitrarily selected most parsimonious tree and PAUP 3.1.1 (Swofford, 1993) to search for the shortest trees inconsistent those nodes. OTAVIPITHECUS 547 Bootstrap analysis (Felsenstein, 1985) to assess statistical robustness of the tree topology was performed using 1000 replicates of the previously outlined search protocol, performing a single random addition sequence and saving only a single most parsimonious tree (MULPARS off) per bootstrap replicate. The strict consensus tree was transferred to MacClade 3.07 (Maddison & Maddison, 1992) for further branch swapping and character analysis. Most parsimonious character reconstructions were produced using MacClade’s ‘‘Soft Polytomies’’ option, treating polytomies as uncertainties in resolution (rather than multiple speciation events) and minimizing individual character changes accordingly. Results A cladistic analysis of a single anatomical region cannot be expected to reproduce results developed from more extensive character sets. Nevertheless, results are broadly congruent with recent cladistic analyses of hominoid relationships (Begun et al., 1997). The strict consensus [Figure 3(a)] reconstructs a clade comprising the extant great ape and Eurasian fossil forms with Kenyapithecus, Afropithecus, and Proconsul as successive outgroups. Otavipithecus forms a clade with Afropithecus and the Napak large-bodied hominoid. This clade is rooted in a higher order polytomy, indicating uncertainty concerning the position of the Otavipithecus clade relative to Kenyapithecus and later forms. The consensus topology is moderately resolved. The consistency index of 0·87 indicates the presence of only minimal homoplasy, however the retention index value of 0·59 indicates that actual character support for this topology is considerably weaker (Farris, 1989a,b), as do the small decay index values (Bremer, 1988, 1994). The strict consensus tree and representative most parsimonious trees [Figure 3(b)] 548 .  (a) Strict consensus Length: 334 CI: 0.87 RI: 0.59 Rohlf 's CI1: 0.68 3 4 * 2 5 5 1 5 6 2 2 2 1 2 2 3 3 0 1 2 4 3 3 4 4 * 2 5 4 1 8 1 1 2 1 3 1 2 1 3 6 4 2 10 Ouranopithecus Griphopithecus Lufengpithecus Gorilla Dryopithecus Sivapithecus Pan Pongo Kenyapithecus Limnopithecus Kalepithecus Oreopithecus Turkanapithecus Nyanzapithecus Rangwapithecus Simiolus Napak material Afropithecus Otavipithecus Dendropithecus Hylobates Aegyptopithecus Proconsul Micropithecus Propliopithecus Figure 3. (a) Strict Consensus of 225 equally parsimonious trees. Numbers indicate approximate maximum branch lengths (MacClade 3.07 ‘‘Almost All Possible Changes’’ option for polytomous trees). Asterisks indicate nodes with Bremer Support Index (Decay Index) values of 2 (Bremer, 1994); all other nodes have values of 1. (b) Representative most parsimonious trees showing possible positions for the Otavipithecus clade and alternate resolutions of the great ape clade. The inconsistent placement of Aegyptopithecus results from the coding of relative incisor size (I1–I2). Incorrect resolution of relationships among the extant great apes and Sivapithecus is caused by homoplasy in relative incisor size (I1–I2) and molar cingulum development (CING). also illustrate the limitations of the restricted morphological sample and small character set. The presence of an OreopithecusNyanzapithecus clade and the position of Hylobates relative to the African large-bodied fossil taxa are inconsistent with recent findings based upon postcranial evidence (see Harrison & Rook, 1997; Rose, 1997; Begun et al., 1997 for reviews), but are predictable results of an analysis restricted to craniodental material. Prior analyses of dental morphology resulted in classifications uniting Oreopithecus and Nyanzapithecus (Harrison, 1986), and the cladistic position of the hylobatids relative to the large-bodied African fossils continues to turn upon differing interpretations of conflicting cranial and postcranial phylogenetic signals (Begun et al., 1997). With so few characters, homoplasies have disproportionate influence. The inconsistent placement of Aegyptopithecus, resolved incorrectly on all most parsimonious trees, appears to result from the coding of relative incisor size (I1–I2), which unites it with Hylobates and several of the Miocene small-bodied forms to the exclusion of Proconsul. Likewise, similarities between Pan and Pongo in relative incisor size (I1–I2) and molar cingulum development (CING) result in the incorrect resolution of relationships among the extant great apes and Sivapithecus on the most parsimonious and bootstrap consensus trees [Figure 3(b) and 4]. The small number of available mandibular characters contributes both to the poor resolution of some regions of the strict    549 OTAVIPITHECUS Griphopithecus Lufengpithecus Ouranopithecus Gorilla Dryopithecus Sivapithecus Pan Pongo Kenyapithecus Limnopithecus Kalepithecus Napak material Afropithecus Otavipithecus Oreopithecus Turkanapithecus Nyanzapithecus Rangwapithecus Simiolus Dendropithecus Hylobates Aegyptopithecus Proconsul Micropithecus Propliopithecus (b) Most parsimonious trees Length: 332 CI: 0.88 RI: 0.61 Dryopithecus Sivapithecus Pongo Pan Griphopithecus Lufengpithecus Ouranopithecus Gorilla Kenyapithecus Limnopithecus Kalepithecus Oreopithecus Turkanapithecus Nyanzapithecus Rangwapithecus Simiolus Napak material Afropithecus Otavipithecus Dendropithecus Hylobates Aegyptopithecus Proconsul Micropithecus Propliopithecus Figure 3. (b). 550 .  Bootstrap consensus 1000 replicates 27 13 4 29 7 20 2 17 8 2 27 63 3 22 21 12 4 11 12 1 34 17 Pan Pongo Gorilla Dryopithecus Sivapithecus Ouranopithecus Lufengpithecus Griphopithecus Otavipithecus Napak material Afropithecus Proconsul Oreopithecus Turkanapithecus Rangwapithecus Nyanzapihecus Micropithecus Simiolus Limnopithecus Kalepithecus Kenyapithecus Dendropithecus Hylobates Aegyptopithecus Propliopithecus Figure 4. Bootstrap consensus tree. Consensus of 1000 bootstrap replicates showing all groups compatible with majority rule consensus. Numbers indicate bootstrap percentage of the adjacent node. Only the Oreopithecus–Turkanapithecus sister pair achieves even 50% representation. consensus [Figure 3(a)] and the instability of the tree structure under bootstrap resampling. The bootstrap consensus (Figure 4) contains an Otavipithecus clade. However, this clade now includes Griphopithecus, and the overall topology is rather different from the strict consensus. While the relative positions of the great ape–Eurasian clade, the Otavipithecus clade, and Proconsul have been retained, they now form a monophyletic group to the exclusion of all other taxa, most notably Hylobates and Kenyapithecus. With the sole exception of the Oreopithecus– Turkanapithecus sister pair, no clade achieves even 50% representation. The Otavipithecus clade occurs on only 10% of the bootstrap trees, far below the threshold of statistical support. Expanding the dataset to include cranial and postcranial characters would be expected to increase resolution, correct discrepancies in the positions of key taxa, and increase overall support for the tree topology. However, this would be unlikely to contribute to our understanding of Otavipithecus, whose cranium and postcranium are largely unknown and, where known, uniformly primitive. Given that an Otavipithecus–Afropithecus relationship is found in both consensus topologies, and that the phylogenetic position of Afropithecus is reasonably well understood (Leakey & Walker, 1997; Begun et al., 1997; Ward, 1998), it seems more productive to evaluate the support for this relationship directly. Should it be supported, the phylogenetic position of Otavipithecus may then be inferred from that of its sister taxon. A concern which bears directly on the position of Otavipithecus is the potential influence of long-branch effects on the estimated phylogeny. A long branch may be defined as ‘‘a lineage that [has] evolved so much between nodes in the phylogeny that    551 OTAVIPITHECUS Table 3 Parsimony analyses excluding one or more Otavipithecus clade members Trial 1 2 3 4 5 6 7 Taxa excluded No. of trees Length CI RI Consensus length* Rohlf’s CI1 Otavipithecus, Afropithecus, Napak Afropithecus, Napak Otavipithecus, Napak Otavipithecus, Afropithecus Napak Otavipithecus Afropithecus 45 10,000+ 2985 798 10,000+ 10,000+ 10,000+ 292 299 315 306 321 329 312 0·90 0·89 0·88 0·90 0·88 0·88 0·89 0·68 0·65 0·62 0·66 0·60 0·60 0·64 292 299 331 306 321 335 315 0·83 0·73 0·24 0·81 0·83 0·64 0·52 *Strict consensus length is reported for Trial 1, while majority rule consensus lengths are reported for Trials 2–7. its character states have been effectively randomized with respect to other taxa’’ (LyonsWeiler & Hoelzer, 1997:375). It is well established that such long branches tend to ‘‘attract’’, potentially affecting the accuracy of a parsimony analysis and resulting in a false inference of cladistic affinity (Felsenstein, 1978; Hendy & Penny, 1989; Rohlf et al., 1990; Huelsenbeck, 1995). While long branch effects are sometimes difficult to detect, the branch lengths observed for the Otavipithecus clade [Figure 3(a)] immediately raise the possibility that this clade is a product of long branch attraction. Such long branch effects may be mitigated by the exclusion of problematic taxa, sampling of additional taxa to break up long branches, or the use of subsets of characters which reduce branch lengths (Lyons-Weiler & Hoelzer, 1997). Unfortunately, in this case the problematic long-branch taxa are precisely those whose relationships are of greatest interest. Given the limited morphological sample, neither inclusion of additional characters nor exclusion of character subsets is practical, and the middle Miocene hominoid record provides no additional taxa appropriately positioned to disrupt potential long branches. Thus, it is necessary to determine both the influence of long-branch effects on the analysis as a whole and the extent to which long-branch attraction ‘‘drives’’ the inferred Otavipithecus clade. To explore this issue, an additional series of parsimony analyses was run excluding each member of the Otavipithecus clade individually, each of the three possible taxon pairs, and all three taxa simultaneously (see Table 3). Figure 5 shows selected consensus trees produced by these trials. Trial 1, the analysis excluding all three taxa, yielded 45 equally parsimonious trees whose strict consensus is shown in Figure 5(a). As can be seen, other than excluding the Otavipithecus clade, this tree is virtually identical to the original analysis. This establishes that any long-branch effect is restricted to the placement of the Otavipithecus clade members and has no effect on the relationships of other taxa. The reintroduction, singly or in pairs, of the excluded taxa resulted in extremely high numbers of equally parsimonious trees (between 798 and 10,000+) and largely or completely unresolved strict consensus topologies. Of these, Trial 7 [Figure 5(d)] yields the most highly resolved consensus topology, including an OtavipithecusNapak sister pair which 99% of most parsimonious trees unite with Kenyapithecus. Similarly, Trial 5 [Figure 5(c)] unites Otavipithecus and Afropithecus on 99% of most parsimonious trees. Only when both Afropithecus and Napak are excluded [Trial 2, Figure 5(b)] is Otavipithecus paired with a non-clade member, namely Kenyapithecus. While not definitive, these results strongly suggest the inferred Otavipithecus clade is 552 .  Ouranopithecus Griphopithecus Lufengpithecus Gorilla Dryopithecus Sivapithecus Pan Pongo Kenyapithecus Limnopithecus Kalepithecus Oreopithecus Turkanapithecus Nyanzapithecus Rangwapithecus Simiolus Dendropithecus Hylobates Aegyptopithecus Proconsul Micropithecus Propliopithecus (a) Trial 1 Strict consensus (b) Trial 2 Majority rule consensus 54 99 60 94 94 94 89 86 84 57 100 68 63 87 60 Figure 5. (a) and (b). 100 Dryopithecus Sivapithecus Pan Pongo Ouranopithecus Griphopithecus Lufengpithecus Gorilla Kenyapithecus Otavipithecus Limnopithecus Oreopithecus Turkanapithecus Rangwapithecus Nyanzapithecus Simiolus Micropithecus Dendropithecus Hylobates Kalepithecus Proconsul Aegyptopithecus Propliopithecus    (c) Trial 5 Majority rule consensus 63 75 84 66 98 99 99 78 100 100 85 100 90 100 55 69 (d) Trial 7 Majority rule consensus 57 100 64 100 100 100 99 100 99 64 553 OTAVIPITHECUS 61 100 68 64 100 Pan Pongo Sivapithecus Dryopithecus Gorilla Griphopithecus Lufengpithecus Ouranopithecus Afropithecus Otavipithecus Proconsul Oreopithecus Turkanapithecus Rangwapithecus Nyanzapithecus Micropithecus Simiolus Limnopithecus Kenyapithecus Dendropithecus Kalepithecus Hylobates Aegyptopithecus Propliopithecus Dryopithecus Sivapithecus Pan Pongo Ouranopithecus Griphopithecus Lufengpithecus Gorilla Napak material Otavipithecus Kenyapithecus Limnopithecus Oreopithecus Turkanapithecus Rangwapithecus Nyanzapithecus Simiolus Micropithecus Hylobates Aegyptopithecus Dendropithecus Kalepithecus Proconsul Propliopithecus Figure 5. (c) and (d). Figure 5. Selected consensus trees based upon exclusion trials (see Table 3): (a) Trial 1 excluding Otavipithecus, Afropithecus, and Napak; (b) Trial 2 excluding Afropithecus and Napak; (c) Trial 5 excluding Napak; (d) Trial 7 excluding Afropithecus. Numbers indicate bootstrap percentage of the adjacent node. Only when both Afropithecus and Napak are excluded (Trial 2) is Otavipithecus united with Kenyapithecus. 554 P4ECD 21 steps 0 1 2 Polymorphic Equivocal Propliopithecus Micropithecus Proconsul Aegyptopithecus Hylobates Dendropithecus Otavipithecus Afropithecus Napak material Simiolus Rangwapithecus Nyanzapithecus Turkanapithecus Oreopithecus Limnopithecus Kalepithecus Kenyapithecus Sivapithecus Dryopithecus Pongo Lufengpithecus Griphopithecus Ouranopithecus Gorilla Pan .  (a) Figure 6. Most parsimonious character reconstructions of apparent synapomorphies supporting the Otavipithecus clade (MacClade 3.07 ‘‘Soft Polytomies’’ option for reconstructing character changes on polytomous trees). This clade is characterized by (a) possession of a P4 entoconid (P4ECD, State 2); (b) presence of a molar deflecting wrinkle (DEFWRIN, State 1); and (c) molar basal flare (MBF, States 1 and 2) which achieves its greatest expression (State 2) in Otavipithecus. not merely an artefact of long-branch attraction. What, then, is the basis for this finding? Andrews (1992a) listed six characters shared by Otavipithecus and Afropithecus, namely, inflated molar cusps, absence of the beaded buccal cingulum, square molars, moderate inferior transverse torus development, even mandibular depth, and a narrow incisor region. Neither inflated molar cusps nor incisor region breadth were addressed in this study, and coding of corpus shallowing (even mandibular depth) was uninformative under gap coding. The remaining characters—reduced buccal cingulum, relatively broad molars, and a moderate inferior transverse torus—are reconstructed on the strict consensus as shared characters of all post-Proconsul large-bodied Miocene hominoids. Ramus position, a trait discussed in the literature (Conroy et al., 1992; Andrews, 1992b) but excluded from Andrews’ (1992a) list, was likewise a shared character for this more inclusive group. Thus the presence of these traits within the Otavipithecus clade appears to reflect a common hominoid ancestry rather than specific cladistic affinities. Character reconstructions (see Figure 6) identified several additional potential synapomorphies of the Otavipithecus clade. These taxa share the presence of a P4 entoconid and a molar deflecting wrinkle. The clade is also characterized by molar basal flare, which achieves its most pronounced expression in Otavipithecus. However, the presence of this trait in both Aegyptopithecus and Propliopithecus admits the possibility that the Otavipithecus clade retains a more primitive catarrhine condition. Given the observed variability of the Gorilla Ouranopithecus Griphopithecus Lufengpithecus Pongo Pan Gorilla Ouranopithecus Griphopithecus Lufengpithecus Pongo Dryopithecus Sivapithecus Sivapithecus Kenyapithecus Kenyapithecus Limnopithecus Limnopithecus Kalepithecus Kalepithecus Oreopithecus Oreopithecus Turkanapithecus Turkanapithecus Nyanzapithecus (b) (c) Dryopithecus Nyanzapithecus Rangwapithecus Simiolus Simiolus Napak material Napak material Afropithecus Afropithecus Otavipithecus Otavipithecus Dendropithecus Dendropithecus Hylobates Hylobates Aegyptopithecus Aegyptopithecus Proconsul Proconsul Micropithecus Micropithecus Propliopithecus Propliopithecus OTAVIPITHECUS Rangwapithecus    DEFWRIN 8 steps 0 1 Polymorphic MBF 4 steps 0 1 2 Equivocal Pan 555 556 .  former traits—in each case a number of taxa received polymorphic codings—and the questionable polarity of the latter, their status as shared derived characters is uncertain. Discussion Previously stated hypotheses concerning the phylogenetic affinities of Otavipithecus namibiensis have included: (1) that it represents the sister group to the African ape–human clade (Conroy, 1994); (2) that it is a member of the great ape–human clade with no special relationship to the African apes (Begun, 1994a); (3) that it is a member of a middle Miocene ‘‘Kenyapithecus– Dryopithecus complex’’ predating the African ape–human divergence (Pickford et al., 1994); or, (4) that it is most closely related to the early Miocene form Afropithecus and a member of the tribe Afropithecini (Andrews, 1992a). Given the limited nature of the available sample, a decisive resolution of the ‘‘Otavipithecus problem’’ remains elusive. However, cladistic analysis is a useful tool for the exploration of character distributions and the separation of homoplasies from synapomorphies, issues which have clouded the Otavipithecus debate. While inconclusive in many respects, results do permit the exclusion of several hypotheses, and clarify the evidence for those remaining. No support was found for a close relationship to the African ape or great ape clades or to any of the late Miocene Eurasian hominoids with which Otavipithecus has been compared. While Otavipithecus and Dryopithecus both possess thin dental enamel (Andrews & Martin, 1991; Conroy et al., 1996), they differ in most other aspects of molar morphology. Dryopithecus molars lack the restricted talonid basins and pronounced basal flare of Otavipithecus, and possess expanded mesial foveae, more peripheral cusps, and tapering M3s. Otavipithecus shares with Sivapithecus a narrow symphy- seal region, the presence of accessory cusplets on the M3, and reduced molar cingula; however, cingulum reduction is considerably more pronounced in Sivapithecus, and Otavipithecus lacks its thickened dental enamel, robust mandibular corpus, and heavily buttressed symphysis (Conroy et al., 1992). Further, there is currently no evidence to indicate that Otavipithecus shares the derived elbow morphology which unites both Dryopithecus and Sivapithecus with the great ape clade (Pilbeam et al., 1990; MoyàSolà & Köhler, 1996; Ward, 1997; Begun & Kordos, 1997). While Otavipithecus clearly represents an outgroup to the ‘‘hominoids of modern aspect’’ (Pilbeam, 1996), this is equally true of Afropithecus and Kenyapithecus. Current evidence clearly favors one or both of the major Kenyapithecus groups as the most likely immediate sister group to (McCrossin & Benefit, 1997; Begun et al., 1997), if not an actual member of (McCrossin et al., 1998a), the extant ape clade, a position the present analysis also supports. Several sub-analyses do suggest a closer relationship between Otavipithecus and other middle Miocene hominoids. The bootstrap consensus reconstructs Otavipithecus and Griphopithecus as sister taxa on 17% of the bootstrap trees, and Trials 2 and 7 reconstruct relationships between Otavipithecus and Kenyapithecus. Although most parsimonious character reconstructions fail to identify even a single trait supporting an Otavipithecus–Kenyapithecus sister relationship, less parsimonious trees uniting Kenyapithecus with various Otavipithecus clade members are not significantly different from the most parsimonious solution at P=0·05 (Templeton’s Wilcoxon signed-ranks test; Templeton, 1983). Numerous similarities between Otavipithecus and Kenyapithecus have been previously noted (Andrews, 1992a; Conroy et al., 1992; Pickford et al., 1994). Certainly, the Otavipithecus inferior transverse torus    OTAVIPITHECUS 557 Figure 7. Comparison of Otavipithecus M1 (left) and KNM-NC 9740 (right), an isolated right molar from Nyakach attributed to Kenyapithecus africanus (Pickford, 1985). While the Otavipithecus M1 is more worn and its crown is slightly less ‘‘flexed’’ than KNM-NC 9740, these teeth share similar crown dimensions, cusp size and position, cingulum configuration, and degree of basal flare. more closely resembles that of Kenyapithecus than Afropithecus, and their molar teeth share, in varying degrees, bunodont and inflated cusps, restricted talonid basins, and buccal cingulum reduction. The resemblance between the Otavipithecus M1 and KNM-NC 9740 (see Figure 7), an isolated molar from Nyakach attributed to Kenyapithecus africanus (Pickford, 1985), is particularly striking. Despite different levels of wear, these teeth are clearly similar in crown dimensions, cusp size and position, cingulum configuration, and degree of basal flare. Whether this is evidence for Otavipithecus– Kenyapithecus affinities or cause for reassessment of the Nyakach material and further revision of the Kenyapithecus hypodigm (Ward et al., 1999) remains an open question. At the same time, Otavipithecus lacks the thickened dental enamel, enlarged and procumbent incisors, proclined symphyseal axis, and robust mandibular corpora which have led to suggestions that Kenyapithecus had a pitheciine-like adaptation (McCrossin & Benefit, 1997). While both show reduced buccal cingula, in Kenyapithecus this reduction is less marked and cingular remnants are generally restricted to the bases of the buccal transverse developmental grooves (McCrossin & Benefit, 1997), whereas in Otavipithecus the cingulum remnants manifest as stylar shelves. Finally, the Otavipithecus ulna and phalanx are not consistent with the unique postcranial features of Kenyapithecus, which ally the latter taxon more closely with the ‘‘hominoids of modern aspect’’ (Begun et al., 1997; Rose, 1997; McCrossin, 1997; McCrossin & Benefit, 1997; McCrossin et al., 1998a). The results of this analysis invariably indicate cladistic affinities among Otavipithecus, Afropithecus, and the Napak large-bodied hominoid, giving partial support to Andrews’ (1992a) Afropithecini hypothesis. This group is generally characterized by primitive hominoid morphology. In addition to the observed predominance of primitive mandibular and dental traits, both Otavipithecus and Afropithecus exhibit primitive hominoid skeletal morphology, in most cases indistinguishable from Proconsul (Conroy et al., 1993a,b; Pickford et al., 1997; Leakey & Walker, 1997). However, Afropithecus possesses specializations of the jaws and dentition, most notably thickened dental enamel and procumbent incisors, which have been interpreted as specializations for sclerocarp feeding (Leakey & 558 .  Walker, 1997). These characters are not shared by Otavipithecus, which has thin dental enamel and, based on the preserved alveoli, is thought to have had narrow, vertically implanted incisors (Conroy et al., 1992, 1995). Conversely, Afropithecus lacks the more strongly developed inferior transverse torus and pronounced molar basal flare observed in Otavipithecus. These facts, in combination with the primitive status of the majority of traits characterizing this group, the uncertain significance of the few putative synapomorphies supporting it, and the absence of statistical support under bootstrap analysis, clearly warrant cautious reading of this result. This analysis admits several possible interpretations: ( 1) The synapomorphies uniting Otavipithecus, Afropithecus and the Napak large-bodied hominoid are valid and the strict consensus is an accurate reconstruction of the phylogenetic position of Otavipithecus. This interpretation is given some weight by the relative stability of this relationship in the various elimination trials, even as the exact position of this group relative to the great ape clade remains unresolved. ( 2) The observed pattern of uniquely derived traits superimposed on a broader suite of shared primitive similarities is evidence that Otavipithecus, Afropithecus, and the Napak largebodied hominoid represent divergent lineages of an early Miocene hominoid radiation similar to that which gave rise to the ‘‘proconsulids’’ (Harrison, 1993). In this case, these taxa have no resolvable sister group relationship (Harrison, 1993) but, lacking the derived traits of other middle and late Miocene taxa, are united under the parsimony criterion. This would explain the long branches observed for Otavipithecus and Afropithecus. It might also account for the dramatic difference in resolution between the original analysis and Trial 1 (excluding all Otavipithecus clade members), on the one hand, and subsequent trials excluding subsets of this group, on the other. Under this scenario, the character distributions of the constituent taxa are most effectively explained in terms of traits shared by the group as a whole, whereas the individual taxa, sharing relatively few traits with the more derived forms, occupy any number of positions equally well, or poorly, as the case may be. ( 3) The strict consensus is incorrect, and Kenyapithecus does, in fact, have cladistic affinities with members of the Otavipithecus clade, as implied by Andrews (1992a) and supported by Trials 2 and 7. In fact, Begun et al. (1997) proposed that the dental complex of thickened molar enamel and enlarged procumbent incisors might represent a synapomorphy of an Afropithecus–Kenyapithecus clade, in which case Otavipithecus would be best interpreted as a less derived basal member of a group which subsequently underwent specialization for hard object feeding (Leakey & Walker, 1997; McCrossin & Benefit, 1994, 1997). Of these possibilities, the last is considered least probable. Branch-swapping showed all trees uniting Kenyapithecus with one or more members of the Otavipithecus clade to be at least two steps longer than the most parsimonious solution. Acceptance of such a clade would require homoplasy in the form of postcranial convergences between Kenyapithecus and later Eurasian hominoids and reversion to a more primitive inferior transverse torus morphology in Afropithecus. An exclusive Otavipithecus–Kenyapithecus sister relationship is deemed even less likely. While the similarities between these taxa are    well documented, they appear to result from a more distant common heritage. In fact, character reconstructions based on Trial 2 failed to identify even a single dental or mandibular synapomorphy to support such a relationship. Nevertheless, an Otavipithecus–Kenyapithecus relationship cannot be rejected statistically, and pending the recovery of additional Otavipithecus material, this possibility cannot be excluded. In particular, any evidence that Otavipithecus shared more derived postcranial traits with Kenyapithecus would alter the current picture drastically. Of the original hypotheses, the Otavipithecus–Afropithecus clade is clearly the strongest, however the lack of statistical support for this group and the weakness of the putative synapomorphies defining it cannot be overlooked. As with Kenyapithecus, more definitive answers must await the recovery of additional material. Should the facial morphology of Otavipithecus prove to resemble Afropithecus (Leakey et al., 1991), the ‘‘Afropithecini Hypothesis’’ (Andrews, 1992a) would be considerably strengthened. At present, the most conservative interpretation, that Otavipithecus and Afropithecus represent divergent lineages of an earlier hominoid radiation, is favored. This interpretation may be considered a null hypothesis, subject to revision as new material of the relevant taxa is recovered. However, it may also be an accurate depiction of later early to middle Miocene hominoid diversification. It is unclear what, if any, positive evidence could be marshaled in favor of this hypothesis. But, should future additions to the African Miocene hominoid record fail to resolve this ‘‘muddle in the middle Miocene’’, this scenario may gain greater currency. Conclusions Counter to initial expectations, the addition of Otavipithecus namibiensis to the African OTAVIPITHECUS 559 fossil hominoid record has failed to clarify significantly our understanding of stem hominoid diversification. Rather, the present analysis highlights the complexity of early hominoid evolutionary relationships and the limitations of the cladistic methods we employ to infer them. While this analysis permits the rejection of several prior hypotheses concerning the affinities of Otavipithecus, a strictly cladistic framework provides no solid basis for distinguishing among the remaining possibilities. A phylogenetic relationship between Otavipithecus and Kenyapithecus is neither supported by the present analysis, nor can it be rejected with any statistical confidence, and parsimony analysis is ill-suited to differentiate between a true Otavipithecus–Afropithecus sister relationship and a star phylogeny indicative of rapid radiation. If one takes a wider perspective, however, several general conclusions can be drawn. While the precise details of the ecological adaptation of Otavipithecus are unknown, the general pattern is fairly clear. The few known postcranial elements give no indication of locomotor behaviors outside the primitive hominoid repertoire (Leakey & Walker, 1997; Ward, 1998). The mandible, reportedly designed to withstand considerable masticatory loads (Schwartz & Conroy, 1996), seems over-engineered for a postcanine dentition lacking the extreme bunodonty and thick molar enamel associated with hard-object specialists. As noted by McCrossin & Benefit (1997) thin dental enamel need not preclude hard object feeding, however Otavipithecus lacks the compensatory anatomical adaptations reported in thin-enameled seed predators, and the observed pattern of gross molar wear is more consistent with a soft, probably frugivorous, diet (Conroy et al., 1992, 1995). This combination of a unique dentognathic mosaic and primitive postcranium is consistent with the established pattern for stem hominoid taxa (Ward, 1998). Whether true sister taxa 560 .  or not, Otavipithecus and Afropithecus appear to represent a similar grade of hominoid evolution. Comparisons between Otavipithecus and Kenyapithecus are hampered by the current diversity of opinion as to how many genera and species are represented among the various sites (Andrews, 1992a; McCrossin & Benefit, 1997; Nakatsukasa et al., 1998; Ward et al., 1999). McCrossin and Benefit have consistently maintained that Kenyapithecus africanus from Maboko Island and Kenyapithecus wickeri from Ft. Ternan fall within the range of variation for extant hominoid species, although they have not formally referred the Maboko material to the type species (McCrossin & Benefit, 1997; Benefit, 1999). By contrast, Andrews (1992a) considers Kenyapithecus africanus, as represented by the Maboko Island and Nachola samples, to be generically distinct from Kenyapithecus wickeri. Ward et al. (1999) concur, and have recently referred all material previously attributed to Kenyapithecus africanus—i.e., Maboko Island, Nachola, and Tugen Hills, as well as the limited samples from Ombo, Majiwa, Kaloma, and Nyakach—to the new genus Equatorius. Finally, based on significant differences in postcranial anatomy, Nakatsukasa et al. (1998) have suggested that the Maboko Island and Nachola samples represent separate species. However, Kenyapithecus samples from all these sites share characters indicating a significant adaptive shift involving more varied locomotor behaviors and increased dietary specialization. Of these, traits related to increased range of forearm pronation/supination—a narrower humero–radial articulation and laterally oriented radial notch—are shared exclusively with the extant ape clade (McCrossin & Benefit, 1997; Rose, 1997; Nakatsukasa et al., 1998). Regardless of whether Kenyapithecus sensu lato pre- or postdates the hylobatid divergence (Begun et al., 1997; McCrossin & Benefit, 1997), these features ally it with the extant hominoids to the exclusion of Otavipithecus and the early Miocene hominoids. All results, regardless of choice of interpretation, are broadly consistent with classifications in which Otavipithecus, Afropithecus, Heliopithecus, and Kenyapithecus/ Equatorius africanus are united in the tribe Afropithecini (Andrews, 1992a; Ward et al., 1999; Delson & Andrews, 1999). However, inclusion of the latter taxon is problematic because of conflicting opinions concerning the (con)generic status of Kenyapithecus africanus and Kenyapithecus wickeri (Andrews, 1992a; McCrossin & Benefit, 1997) and uncertainty concerning the phylogenetic position of the various Kenyapithecus taxa relative to the extant ape clade (Begun et al., 1997; McCrossin et al., 1998a; Ward et al., 1999). Both the present analysis and the postcranial synapomorphies noted above recommend the placement of all Kenyapithecus material in Andrews’ (1992a) Kenyapithecini (cf. Kenyapithecinae: Harrison & Rook, 1997). So revised, the tribe Afropithecini is wholly consistent with the phylogenetic analysis results, reflecting the cladistic affinities of Afropithecus and Otavipithecus as well as their inferred gradistic similarities. One or both tribes are certainly paraphyletic (Delson & Andrews, 1999), but the Afropithecini may be broadly considered the out group to both the Kenyapithecini and the extant ape clade. Biostratigraphic concerns should not influence phylogenetic analyses, at least in their initial stages. Still, the grouping of Otavipithecus, a southern African species which may be as young as 12 Ma, with east African fossils 4–5 million years older is somewhat troubling. However, given the tremendous geographic gap between Otavipithecus and all other known Miocene hominoids, we should not be surprised that its closest apparent relatives are temporally as well as spatially remote. By contrast, the paleoecological implications of this finding    are quite unanticipated. If the present picture of Otavipithecus is correct, it represents the persistence in southern Africa of a primitive hominoid during a period when east African forms were expanding into new locomotor and dietary niches in response to increasing seasonality, the expansion of open country habitats, and intensified competition from the newly emergent cercopithecoid monkeys (McCrossin & Benefit, 1997; McCrossin et al., 1998b; Nakatsukasa et al., 1998). The extent to which these factors were prevalent in the Otavi Mountain region is unclear. Although the earliest recorded presence of cercopithecoids in southern Africa—isolated teeth and postcrania dated to approximately 9 Ma from the northern Namibian site of Harasib—is considerably later (Conroy et al., 1996), it would be imprudent to attribute the continued presence of a generalized stem hominoid in the region to absence of Old World monkey competitors. Similarly, palynological studies (Scott, 1995) show progressive climatic aridification across southern Africa from the late middle Miocene onward comparable to that seen in contemporary East African sites (McCrossin et al., 1998b). However, the faunal unit which includes the Otavipithecus type specimen contains a mix of riparian forest and savanna woodland species, suggesting the presence of riverine forest in close proximity to drier, more open habitats (Conroy et al., 1992, 1993b). Such forests might have provided refugia for generalized hominoids, at least during the initial period of climate change. Clearly, more detailed paleoenvironmental studies are needed to evaluate this possibility. In conclusion, rather than resolving old questions concerning Miocene hominoid evolutionary relationships, the discovery of Otavipithecus has raised a series of new questions concerning the patterns of stem hominoid diversification and the spatiotemporal OTAVIPITHECUS 561 distribution of primitive hominoid morphotypes. Future discoveries will almost certainly amend the present assessment of Otavipithecus. But, barring total revision and reinstatement of Otavipithecus as an unequivocal extant ape ancestor, the issues raised by the present interpretation should focus renewed attention on the complexities of stem hominoid evolution. Ultimately, this may prove the most significant contribution of Otavipithecus to Miocene hominoid studies. Summary Otavipithecus namibiensis is a medium-sized ‘‘hominoid of archaic aspect’’ (Pilbeam, 1996), derived relative to the stemhominoid Proconsul, but lacking apparent locomotor or dietary specializations. Based on a cladistic analysis of hominoid mandibular morphology, there is no support for a close cladistic relationship between Otavipithecus and either the African ape or great ape clades, or with any of the Eurasian fossil hominoids with which it has previously been compared (Conroy et al., 1992; Pickford et al., 1994). While suggested by several analyses, a relationship between Otavipithecus and Kenyapithecus is deemed unlikely on the basis both of morphological comparisons and the absence of support within a cladistic framework. The present analysis indicates Otavipithecus is most closely related to Afropithecus and the largebodied Napak hominoid. Given the lack of statistical support for this result, a conservative interpretation, that these taxa represent related but divergent lineages of a later early Miocene hominoid radiation, is currently favored. All results are consistent with the taxonomic allocation of Otavipithecus to Andrews’ (1992a) tribe Afropithecini, which may be considered the sister group to Kenyapithecus and the ‘‘hominoids of modern aspect’’. 562 .  Acknowledgements For access to materials and curatorial assistance, thanks are due to the following individuals and institutions: Glenn C. Conroy; David Begun; Stephen Ward; Barbara Brown; John Alexander and the American Museum of Natural History; Bruce Latimer, Lyman Jellema and the Cleveland Museum of Natural History; Elwyn Simons, Friderun Ankel-Simons, Prithijit Chatrath and the Duke University Primate Center Paleontology Laboratory; Maria Rutzmoser and the Harvard Museum of Comparative Zoology; Richard Thorington, Linda Gordon, Karie Darrow and the National Museum of Natural History; Mary Ann Turner and the Yale Peabody Museum; Peter Andrews, Louise Humphrey and the British Museum (Natural History); Brigitte Senut, Martin Pickford and the Muséum National d’Histoire Naturelle; Paul Mazza, Lorenzo Rook and the Museo di Geologia e Paleontologia; S. Schüffler and the Institut für Geologie und Minerologie, Universität Erlangen; Meike Köhler, Salvador Moyà-Solà and the Institut Paleontològic M. Crusafont; Antonio Abad and the Museo y Laboratoria de Geologia, Seminario de Barcelona; Wim Van Neer and the Royal Central African Museum; George D. Koufos and the Geological Laboratory, Aristotle University of Thessaloniki; Ezra Musiime and the Uganda Museum (Paleontology); Meave Leakey, William Anyonge, Emma Mbua, the Board of Governors of the Kenya National Museum, and the Office of the President of Kenya. I owe a tremendous debt to my doctoral advisor, Glenn Conroy, and members of my thesis committee—Tab Rasmussen, Richard Smith, Jim Cheverud, Allan Larson, and Jonathan Losos—for their assistance and feedback at various stages of this project. Many thanks to Eric Delson for his helpful comments on several drafts of this manuscript and to Terry Harrison and an anonymous reviewer whose suggestions improved the final version considerably. Any errors of fact or interpretation are, of course, my own. I wish to thank Glenn Conroy for providing photos of the Otavipithecus holotype specimen: Meave Leakey and the Board of Governors of the Kenya National Museum for permission to use the photo of KNM-NC 9740; and Barbara Brown for permission to reproduce the line drawings in Appendix A. My immense gratitude to Chester Tarka and Lorraine Meeker for photographic and layout assistance. This research was supported by the National Science Foundation, the WennerGren Foundation for Anthropological Research, the Boise Fund, the Foundation for Science and Disability, and the New York Consortium in Evolution Primatology. Appendix A. Measurements of the mandible and mandibular dentition Mandibular measurements (see Figure 8) ( 1) Symphyseal chord (CORD)—length from infradentale to most posteroinferior point on symphysis. ( 2) Symphyseal breadth (SYMB)— maximum breadth measured perpendicular to CORD. ( 3) Corpus depth at P3/P4 (P3D)— perpendicular to the alveolar plane. ( 4) Corpus depth at P4/M1 (P4D)— perpendicular to the alveolar plane. ( 5) Corpus depth at M2/M3 (M2D)— perpendicular to the alveolar plane. ( 6) Corpus depth at mental foramen (FORD)—perpendicular to the alveolar plane. ( 7) Mental foramen depth below alveolar margin (MFD)—perpendicular to the alveolar plane. ( 8) Bicanine breadth (BCB) ( 9) Bimolar breadth (BMB) (10) Corpus breadth at P4 (P4B)    OTAVIPITHECUS 563 7 2 36 64 5 1 8 9 10 11 12 Figure 8. Measurements of the mandible as described in Appendix A. Diagrams adapted from Brown (1989, Figure 5.1). (11) Lateral eminence breadth (LAT) (12) M1–M3 Length (M1M3) Dental measurements ( 1) Incisor length (IL)—mesiodistal length at incisive edge. ( 2) Incisor breadth (IB)—maximum buccolingual breadth normal to length. ( 3) Incisor height (IHT)—maximum crown height measured buccally from the cemento-enamel junction (CEJ) to the incisive edge. ( 4) Canine length (CL)—maximum crown length measured along the long axis. ( 5) Canine breadth (CB)—maximum crown breadth measured normal to length. ( 6) Canine maximum crown height (CMAX)—maximum crown height measured buccally from the CEJ to the crown tip. ( 7) Canine mesial crown height (CMES)—crown height measured from mesial CEJ to the crown tip. ( 8) Canine mesial ridge length (CMRL)—measured along mesial ridge from crown tip to union with lingual cingulum. 564 .  ( 9) P3 Maximum crown length (P3ML)— measured along the long axis. (10) P3 Maximum crown breadth (P3MB)—measured normal to the maximum length. Mesial height (P3MHT)— (11) P3 perpendicular distance between the protocone apex and the most inferior point on the mesiobuccal CEJ. Distal height (P3DHT)— (12) P3 perpendicular distance between the distobuccal CEJ and the distal marginal ridge. (13) P4 Maximum mesiodistal length (P4L) (14) P4 Mesial breadth (P4MB)— maximum buccolingual breadth measured across the trigonid basin. (15) P4 Distal breadth (P4DB)—maximum buccolingual breadth measured across the talonid basin. (16) P4 Metaconid height (P4MHT)— perpendicular height measured lingually from the CEJ to the metaconid apex. (17) P4 Talonid height (P4DHT)— perpendicular height measured lingually from the CEJ to the lowest point of the lingual notch or talonid rim. (18) Molar maximum mesiodistal length (MxL) (19) Molar mesial breadth (MxMB)— maximum buccolingual breadth measured across the trigonid basin. (20) Molar distal breadth (MxDB)— maximum buccolingual breadth measured across the talonid basin. (21) Molar cuspal breadth (MxDB)—linear distance between the protoconid and metaconid apices. (22) Molar metaconid height (MxMHT)— perpendicular height measured lingually from the CEJ to the metaconid apex. (23) Molar lingual notch height height (MxDHT)—perpendicular measured lingually from the CEJ to the lowest point of the lingual notch. Appendix B. Qualitative character definition and coding With the exception of P3 honing facet development (see P3HF below), all characters were coded using a 15% cut-off, i.e. a state occurring in fewer than 15% of individuals was considered absent, and the character was coded for the presence of the remaining state. Where multiple states were each present in 15% or more of individuals, the character was coded as polymorphic. CBH Canine basal heel: 0=the crown cross-section is ovoid; 1=the distolingual crown is elaborated to form a heel (often continuous with the lingual cingulum) producing a waisted or triangular cross-section. CMG Canine mesial groove: 0=the mesial ridge is weakly defined and the crown surface is continuous across it; 1=a vertical groove is present distal to a sharply defined and elevated mesial ridge. P3HF Prevalence and extent of P3 honing facet: 0=present in <33% of individuals; 1=present d33% of individuals; 2=extends to mesial root in one or more individuals. P3MLB P3 crown shows distinct mesiolingual projection or ‘‘beak’’ (Begun, 1994a): 0=absent; 1=present. P3MCD P3 metaconid development: 0=absent; 1=present as enamel tubercle on the distolingual ridge; 2=present as cusp(ule) connected to the protoconid by a transverse cristid distinct in orientation from the distolingual ridge. P4ECD P4 entoconid development: 0=absent; 1=present as enamel tubercle on the lingual talonid    P4HCD CING DEFW HCLD HPOS ACC SEXT LMR rim; 2=present as cusp(ule) demarcated from talonid rim by developmental groove. P4 hypoconid development: 0=absent; 1=present as enamel tubercle on the buccal talonid rim; 2=present as cusp(ule) demarcated from talonid rim by developmental groove. Molar buccal cingulum is: 0=absent; 1=reduced with discontinuous cingular remnants largely restricted to buccal developmental grooves; 2=largely continuous around buccal margin, with or without ‘‘beading’’. Molar deflecting wrinkle, a median wrinkle of the metaconid joining the entoconid near the center of the occlusal surface (Swindler & Ward, 1988), is: 0=absent; 1=present on one or more molars. Molar hypoconulid development: 0=absent; 1=present with crown area less than hypoconid or entoconid; 2=present with crown area comparable to hypoconid or entoconid. Molar hypoconulid is: 0=centrally positioned, impinging on the longitudinal axis; 1=positioned buccal of the longitudinal axis. Multiple accessory cusps on the distal talonid rim are: 0=absent; 1=present on one or more molars. Tuberculum sextum (Swindler, 1976): 0=absent; 1=present. Lingual marginal ridges (Swindler, 1976) are: 0=restricted to lingual fissure; 1=extend along lingual metaconid face to create a shelf OTAVIPITHECUS WRIN RAMA RAMP PCF OBL ITT STT 565 continuous with the talonid;2=extend as (1) approximating mesial marginal ridge. Unworn molar occlusal surfaces show: 0=absence of enamel wrinkling; 1=presence of enamel wrinkling; 2=dense enamel wrinkling. Mandibular ramus angle: 0=ramus is raked back, meeting corpus at ]>110; 1=ramus is vertical, meeting corpus at ]90. In lateral view, mandibular ramus: 0=partially or wholly obscures the M3; 1=does not obscure the M3. Postcanine fossa development: 0=no discernible hollowing of the lateral corpus posterior to the C-P3 eminence; 1=hollowing is present but restricted to basal corpus immediately posterior to the C–P3 eminence; 2=hollowing is more pronounced, extending both distally and superiorly. Oblique line (Brown, 1989): 0=extends inferiorly and is continuous with the basal line; 1=terminates superior to the basal line. Inferior transverse torus development: 0=no discernible torus with median symphyseal surface running inferoanteriorly below the genial fossa; 1=present as distinct posterior swelling inferior to the genial fossa; 2=extends posteriorly to mid-P4 level or beyond. Superior transverse torus development: 0=no discernible torus with superior symphyseal surface continuous with genial fossa; 1=present as distinct 566 RTS ITS .  posterior swelling superior to the genial fossa; 2=extends posteriorly to mid-P4 level or beyond. Relative transverse torus size: 0=STT>ITT; 1=STT<ITT; 2=STT]ITT. Intertoral sulcus (Brown, 1989), a longitudinal depression of the medial corpus continuous with the sublingual fossa between the superior and inferior transverse tori, is: 0=absent: 1=present. DIG IDG GGF Impressions for the mandibular attachment of the right and left digastric muscles are: 0=absent; 1=present as discrete paired structures; 2=continuous across the midline. Interdigastric tubercle is: 0=absent; 1=present. Genioglossal fossa is oriented: 0=posteriorly and inferiorly; 1=directly posteriorly. Appendix C. Quantitative character variable calculation Character Abbreviation 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 ISHP I1–I2 CSHP CHT CMRL P3SHP P3HT P4SHP P4HT P3–P4 MSHP M3TAP MBF MHT M1–M2 BUCC CORPD CORPS CORPR LAT DIV SYMR MFD Incisor shape Relative incisor size Canine shape Canine crown height Canine mesial ridge length P3 Shape P3 Crown projection P4 Shape P4 Talonid height Relative premolar size Molar shape M3 Taper Molar basal flare Molar cusp relief Relative molar size Buccinator groove width Relative corpus depth Corpus shallows posteriorly Corpus robustness Lateral mass development Tooth row divergence Symphysis robustness Mental foramen depth OLS residual transformation lnI1L on lnI1HT ln(I1L*I1HT) on ln(I2L*I2HT) lnCB on lnCL lnCMAX on lnCL lnCMRL on lnCMES lnP3B on lnP3L lnP3MHT on lnP3DHT lnP4MB on lnP4L lnP4MB on lnP4DHT ln(P4MB*P4L) on ln(P3MB*P3ML) lnM2MB on lnM2L lnM3MB on lnM3DB lnM2CB on lnM2MB lnM2MHT on lnM2DHT ln(M1MB*M1L) on ln(M2MB*M2L) lnLAT on lnM3MB lnP3D on lnM1M3 lnP3D on lnM2D lnP4B on lnP4D lnLAT on lnM2B lnBMB on lnBCB lnSYMB on lnCORD lnMFD on lnFORD Quantitative character variables were calculated as individual (rather than group mean) residual values from the ordinary least squares regression line and coded by simple gap coding (Mickevich & Johnson, 1976). Appendix D. Matrix of parsimony-informative qualitative and quantitative characters CMG P3HF P3MLB P3MCD P4ECD P4HCD CING DEFW HCLD HPOS ACC Pan Gorilla Pongo Hylobates Propliopithecus Aegyptopithecus Ouranopithecus Oreopithecus Dryopithecus Griphopithecus Lufengpithecus Sivapithecus Napak Material Afropithecus Turkanapithecus Dendropithecus Micropithecus Kalepithecus Limnopithecus Kenyapithecus Rangwapithecus Proconsul Nyanzapithecus Simiolus Otavipithecus 0 0 0&1 1 1 0&1 1 ? 0&1 0 0&1 0&1 ? 0&1 ? 1 0&1 0&1 0&1 0 0 0&1 ? 0&1 ? 0&1 0&1 0&1 0&1 0&1 0&1 0&1 ? 0&1 0 0&1 ? ? 1 ? 0&1 0&1 0&1 0&1 0&1 0&1 0&1 ? 0&1 ? 1 1 1 2 1 2 0 0 2 0 0 2 0 2 0 2 1 2 1 1 2 2 0 1 ? 1 1 1 0&1 0 0 1 0 1 1 1 0&1 0 0 0 0 0 0 0 0 0 0&1 0 0 ? 0&1&2 0&1 0&1&2 0 0 0&1 0 2 0&1 ? 2 0&2 ? 0 ? 0 0 0 0&1 0&2 0&2 0 0 0 ? 0&1&2 0&1&2 0&2 0&1 0 0 0 0 0&1&2 2 0&2 0&1&2 0&2 2 ? 0&1&2 0&1 0 0&1 0&2 0&1 0&1 0 0 2 0&1 0&1&2 0&2 0&1 0 0 0 0 0&1 ? 0&2 0&2 1&2 2 ? 0&1&2 0&1 0&1 0&1 0&1&2 0 0&1&2 0 0&1 0 0 1&2 0&1 0 2 1&2 1 1 0&1 1 1 0 1&2 1 1 1&2 1&2 1&2 1&2 0&1 1 1&2 0&1 1&2 1 0 0 0&1 0 0 0 0&1 0 0 0 0&1 0 1 0&1 ? 0 0 0 0 0&1 0 0&1 0 0&1 1 1&2 2 1 1&2 1 1 1&2 1&2 2 2 2 1 1&2 1 1 1&2 1 1&2 1 1&2 1 1&2 1 1&2 1 1 0&1 1 0 0 0 1 0&1 1 0 1 0&1 1 0&1 1 1 0&1 0&1 0&1 1 0&1 0&1 0&1 1 1 0 0 0 0 0 0 0&1 0&1 0&1 1 0&1 0&1 0 0&1 ? 0 0 0 0 0&1 0 0&1 0 0 1 OTAVIPITHECUS CBH    Genus 567 568 Appendix D. (Continued). Genus LMR WRIN RAMA RAMP PCF OBL ITT STT RTS ITS DIG 0&1 0 0&1 0 0 0 0&1 1 0 0 0&1 0 0&1 1 ? 0 0 0 0 0&1 0 0&1 0 0 0 0 1&2 0 0 0 0 0 0 0 0 0 0 0&1 0 ? 0 0 0 0 0 0 0 0 0 0 1 1 2 0 0 0 1 0 0&1&2 ? 1&2 0&1 0&1&2 1&2 ? 0 0 0 0 0&1 2 1&2 0&1 0 1 0&1 1 0&1 0&1 0&1 1 1 1 0&1 ? 1 1 ? 0 0 1 1 1 1 ? ? 0&1 ? ? 1 0&1 0&1 0&1 1 1 1 0 0&1 0&1 ? 0 0 ? 0&1 0 0&1 1 1 1 ? 0 0&1 1 0 1 1 1&2 1&2 1&2 1&2 1&2 1&2 1&2 1&2 ? 2 1&2 ? 2 1 1&2 0&1 1 0&1 1&2 1&2 1&2 2 1 1 0&1 0 0&1 0 0 0 ? 0 0 ? ? 0 0 0 0 0 0 0 0&1 0 0 0&1 0 ? 0 1&2 2 2 1 1 1 1&2 0 1&2 ? 2 1&2 1 0&1 0 0&1 0&1 0&1 0&1 1&2 0 0&1 0 0&1 1 1 1&2 1 1 1 1 1&2 0 1 ? 1 1 1 0&1 0 1 1 1 1 1 1 1&2 1 1 1 1 1 1 0&1 0&1&2 0&2 1 ? 1 ? 1 1 0&2 1 ? 0&2 0&2 0&2 0&2 1 0 0&2 0 0&2 2 0&1 1 1 0&1 0&1 0&1 0&1 ? 0&1 ? 1 1 ? 0 0 0 0 1 0&1 0&1 0 0 0 ? 0 1&2 1&2 0 1 0&1 1 1 1 0&1 ? 1 0&1 1 0&1 1 1 1 1 1 1 1 1 1 1 1 .  Pan Gorilla Pongo Hylobates Propliopithecus Aegyptopithecus Ouranopithecus Oreopithecus Dryopithecus Griphopithecus Lufengpithecus Sivapithecus Napak Material Afropithecus Turkanapithecus Dendropithecus Micropithecus Kalepithecus Limnopithecus Kenyapithecus Rangwapithecus Proconsul Nyanzapithecus Simiolus Otavipithecus SEXT Appendix D. (Continued). GGF ISHP I1_I2 CSHP CHT P3SHP P3HT P3_P4 MSHP MBF CORPR Pan Gorilla Pongo Hylobates Propliopithecus Aegyptopithecus Ouranopithecus Oreopithecus Dryopithecus Griphopithecus Lufengpithecus Sivapithecus Napak Material Afropithecus Turkanapithecus Dendropithecus Micropithecus Kalepithecus Limnopithecus Kenyapithecus Rangwapithecus Proconsul Nyanzapithecus Simiolus Otavipithecus 0&1 0&1 0 0&1 0&1 0&1 1 0 1 ? 1 0&1 1 0&1 1 1 1 0&1 0&1 0&1 0 1 0 1 0 1 1 1 1 1 0&1 1 1 1 ? 1 1 1 1 1 0&1 1 0&1 0&1 1 1 0&1 0 1 1 3 3 3 3 0 0 2 3 1 ? 0 3 1 2 ? 1 0 ? 1 1 1 0 ? 1 ? 4 2 4 3 2 3 2 3 1 ? 3 1 ? 0 ? 3 ? ? 4 ? 4 2 ? 3 ? 1 1 1 1 1 1 1 2 1 1 0 1 ? 1 ? 1 1 1 1 1 2 1 ? 1 ? 2 2 2 2 2 2 2 2 2 0 2 ? ? 0 ? 2 2 2 2 1 2 2 ? 2 ? 0 0 0 0 0 0 0 1 0 0 0 0 0 0 1 0 0 0 0 0 0 0 0 0 ? 0 1 1 0 1 1 1 1 1 ? 1 1 3 1 ? 1 1 2 1 2 1 1 1 1 ? 0 0 0 0 0 0 0 0 0 0 0 0 ? 2 ? 0 0 ? 0 0 2 1 ? 0 ? 2 2 2 2 3 2 2 1 2 2 2 2 2 2 1 2 1 2 2 2 1 2 0 1 2 2 2 2 2 1 1 2 2 2 ? 2 2 1 1 ? 2 2 2 2 2 2 2 2 2 0 1 1 1 1 1 1 1 3 1 ? 1 1 ? 0 1 1 1 1 1 3 1 1 2 ? 1 OTAVIPITHECUS IDG    Genus 569 570 .  References Albrecht, G. H. (1978). Some comments on the uses of ratios. Syst. Zool. 27, 67–71. Albrecht, G. H., Gelvin, B. R. & Hartman, S. E. (1993). Ratios as a size adjustment in morphometrics. Am. J. phys. Anthrop. 91, 441–468. Alpagut, B., Andrews, P. & Martin, L. (1990). New hominoid specimens from the middle Miocene site at Paşalar, Turkey. J. hum. Evol. 19, 397–422. Andrews, P. (1985). Family group systematics and evolution among catarrhine primates. In (E. Delson, Ed.) Ancestors: The Hard Evidence, pp. 14–22. New York: Alan R. Liss. Andrews, P. (1987). Aspects of hominoid phylogeny. In (C. Patterson, Ed.) Molecules and Morphology in Evolution: Conflict or Compromise?, pp. 21–53. Cambridge: Cambridge University Press. Andrews, P. (1988). Review of Comparative Primate Biology (D. R. Swindler & J. Erwin, Eds). Cladistics 4, 297–304. Andrews, P. (1992a). Evolution and environment in the Hominoidea. Nature 360, 641–646. Andrews, P. (1992b). An ape from the south. Nature 356, 106. Andrews, P. & Martin, L. (1987). Cladistic relationships of extant and fossil hominoids. J. hum. Evol. 16, 101–118. Andrews, P. & Martin, L. (1991). Hominoid dietary evolution. Phil. Trans. Roy. Soc. London B 334, 199–209. Archie, J. W. (1985). Methods for coding variable morphological features for numerical taxonomic analysis. Syst. Zool. 34, 325–345. Archie, J. W. (1989). A randomization test for phylogenetic information in systematic data. Syst. Zool. 38, 219–252. Atchley, W. R. (1978). Ratios, regression intercepts, and the scaling of data. Syst. Zool. 27, 78–83. Atchley, W. R. & Anderson, D. (1978). Ratios and the statistical analysis of biological data. Syst. Zool. 27, 71–78. Atchley, W. R., Gaskins, C. T. & Anderson, D. (1976). Statistical properties of ratios. I. Empirical results. Syst. Zool. 25, 137–148. Begun, D. R. (1992). Miocene fossil hominids and the chimp-human clade. Science 257, 1929–1933. Begun, D. R. (1994a). The significance of Otavipithecus namibiensis to interpretations of hominoid evolution. J. hum. Evol. 27, 385–394. Begun, D. R. (1994b). Relations among the great apes and humans: New interpretations based on the fossil great ape Drypithecus. Yearb. phys. Anthropol. 37, 11–63. Begun, D. R. & Kordos, L. (1993). Revision of Dryopithecus brancoi Schlosser, 1901, based on the fossil hominoid material from Rudabánya. J. hum. Evol. 25, 271–285. Begun, D. R. & Kordos, L. (1997). Phyletic affinities and functional convergence in Dryopithecus and other Miocene and living hominids. In (D. R. Begun, C. V. Ward & M. D. Rose, Eds) Function, Phylogeny, and Fossils—Miocene Hominoid Evolution and Adaptations, pp. 291–316. New York: Plenum Press. Begun, D. R., Ward, C. V. & Rose, M. D. (1997). Events in hominoid evolution. In (D. R. Begun, C. V. Ward & M. D. Rose, Eds) Function, Phylogeny, and Fossils—Miocene Hominoid Evolution and Adaptations, pp. 389–416. New York: Plenum Press. Benefit, B. R. (1993). The permanent dentition and phylogenetic position of Victoriapithecus from Maboko Island, Kenya. J. hum. Evol. 25, 83–172. Benefit, B. R. (1999). The dentition of Kenyapithecus africanus. In (H. Ishida, Ed.) Abstracts—International Symposium on the Evolution of Middle and Late Miocene Hominoids in Africa. Kyoto: Kyoto University. Bishop, W. W. (1958). Miocene Mammalia from the Napak volcanics, Karamoja, Uganda. Nature 182, 1480–1482. Bonis, L. de & Melentis, J. (1977). Les primates hominoı̈des du Vallésien de Macédoine (Grèce). Etude de la mâchoire inférieure. Geobios 10, 849–885. Bremer, K. (1988). The limits of amino acid sequence data in angiosperm phylogenetic reconstruction. Evolution 42, 795–803. Bremer, K. (1994). Branch support and tree stability. Cladistics 10, 295–304. Brown, B. (1989). The mandibles of Sivapithecus. Ph.D. Dissertation, Kent State University. Brown, B. (1997). Miocene hominoid mandibles— functional and phylogenetic perspectives. In (D. R. Begun, C. V. Ward & M. D. Rose, Eds) Function, Phylogeny, and Fossils—Miocene Hominoid Evolution and Adaptations, pp. 153–171. New York: Plenum Press. Chappill, J. A. (1989). Quantitative characters in phylogenetic analysis. Cladistics 5, 217–234. Conroy, G. C. (1994). Otavipithecus: Or how to build a better hominid—not. J. hum. Evol. 27, 373–383. Conroy, G. C. (1996). The cave breccias of Berg Aukas, Namibia: A clustering approach to mine dump paleontology. J. hum. Evol. 30, 349–355. Conroy, G. C., Pickford, M., Senut, B., Van Couvering, J. & Mein, P. (1992). Otavipithecus namibiensis, first Miocene hominoid from southern Africa. Nature 356, 144–148. Conroy, G. C., Pickford, M., Senut, B. & Mein, P. (1993a). Additional Miocene primates from the Otavi Mountains, Namibia. C.R. Acad. Sci. (Paris), Serie II 317, 987–990. Conroy, G. C., Pickford, M., Senut, B. & Mein, P. (1993b). Diamonds in the desert: the discovery of Otavipithecus namibiensis. Evol. Anthropol. 2, 46–52. Conroy, G. C., Lichtman, J. W. & Martin, L. B. (1995). Brief communication: some observations on enamel thickness and enamel prism packing in the Miocene hominoid Otavipithecus nambiensis. Am. J. phys. Anthrop. 98, 595–600. Conroy, G. C., Senut, B., Gommery, D., Pickford, M. & Mein, P. (1996). Brief communication: new primate remains from the Miocene of Namibia, southern Africa. Am. J. phys. Anthrop. 99, 487–492.    Corruccini, R. S. (1975). Multivariate analysis in biological anthropology: some considerations. J. hum. Evol. 4, 1–19. Corruccini, R. S. (1978). Morphometric analysis: uses and abuses. Yearb. phys. Anthropol. 21, 134–150. Corruccini, R. S. (1987). Shape in morphometrics: comparative analyses. Am. J. phys. Anthrop. 73, 289–303. Corruccini, R. S. (1995). Of ratios and rationality. Am. J. phys. Anthrop. 96, 189–191. Delson, E. & Andrews, P. (1975). Evolution and interrelationships of the catarrhine primates. In (W. P. Luckett & F. S. Szalay, Eds) Phylogeny of the Primates, pp. 405–446. New York: Plenum Press. Delson, E. & Andrews, P. (1999). ‘‘Kenyapithecinae’’. In (E. Delson, I. Tattersall, J. A. Van Couvering & A. S. Brooks, Eds) Encyclopedia of Human Evolution and Prehistory, 2nd edn. New York: Garland Publishing. Dodson, P. (1978). On the use of ratios in growth studies. Syst. Zool. 27, 62–67. Faith, D. P. (1991). Cladistic permutation tests for monophyly and nonmonophyly. Syst. Zool. 40, 366–375. Faith, D. P. & Cranston, P. S. (1991). Could a cladogram this short have arisen by chance alone? On permutation tests for cladistic structure. Cladistics 7, 1–28. Farris, J. S. (1989a). The retention index and homoplasy excess. Syst. Zool. 38, 406–407. Farris, J. S. (1989b). The retention index and the rescaled consistency index. Cladistics 5, 417–419. Felsenstein, J. (1978). Cases in which parsimony or compatibility methods will be positively misleading. Syst. Zool. 27, 401–410. Felsenstein, J. (1985). Confidence limits on phylogenies: An approach using the bootstrap. Evolution 39(4), 783–791. Fleagle, J. G. & Kay, R. F. (1983). New interpretations of the phyletic position of Oligocene hominoids. In (R. L. Ciochon & R. S. Corruccini, Eds) New Interpretations of Ape and Human Ancestry, pp. 181–211. New York: Plenum Press. Forey, P. L., Humphries, C. J., Kitching, I. J., Scotland, R. W., Siebert, D. J. & Williams, D. M. (1992). Cladistics: A Practical Course in Systematics. Systematics Association Publication No. 10. Oxford: Clarendon Press. Goldman, N. (1988). Methods for discrete coding of morphological characters for numerical analysis. Cladistics 4, 59–71. Golpe-Posse, J. M. (1993). Los Hispanopitecos (Primates, Pongidae) de los yacimientos del VallèsPenedes (Cataluña, España). II: Descripción del material existente en el Instituto de Paleontologı́a de Sabadell. Paleontol. Evol. 26–27, 151–224. Gould, S. J. (1966). Allometry and size in ontogeny and phylogeny. Biol. Rev. 41, 487–640. Groves, C. P. (1986). Systematics of the great apes. In (D. R. Swindler & J. Erwin, Eds) Comparative Primate Biology, Volume 1: Systematics, Evolution, and Anatomy, pp. 187–217. New York: Alan R. Liss. OTAVIPITHECUS 571 Groves, C. P. (1987). Monkey business with red apes. J. hum. Evol. 16, 537–542. Groves, C. P. & Paterson, J. D. (1991). Testing hominoid phylogeny with the PHYLIP programs. J. hum. Evol. 20, 167–183. Hartman, S. E. (1983). A critique of some regression adjustments used in allometric size ‘‘correction’’ in numerical taxonomy. Am. J. phys. Anthrop. 62, 305–309. Harrison, T. (1986). New fossil anthropoids from the middle Miocene of east Africa and their bearing on the origin of the Oreopithecidae. Am. J. phys. Anthrop. 71, 265–284. Harrison, T. (1988). A taxonomic revision of the small catarrhine primates from the early Miocene of East Africa. Folia primatol. 50, 59–108. Harrison, T. (1989). A new species of Micropithecus from the middle Miocene of Kenya. J. hum. Evol. 18, 537–557. Harrison, T. (1992). A reassessment of the taxonomic and phylogenetic affinities of the fossil catarrhines from Fort Ternan, Kenya. Primates 33, 501–522. Harrison, T. (1993). Cladistic concepts and the species problem in hominoid evolution. In (W. H. Kimbel & L. B. Martin, Eds) Species, Species Concepts, and Primate Evolution, pp. 345–371. New York: Plenum Press. Harrison, T. & Rook, L. (1997). Enigmatic anthropoid or misunderstood ape? The phylogenetic status of Oreopithecus bambolii reconsidered. In (D. R. Begun, C. V. Ward & M. D. Rose, Eds) Function, Phylogeny, and Fossils—Miocene Hominoid Evolution and Adaptations, pp. 327–362. New York: Plenum Press. Hendy, M. D. & Penny, D. (1989). A framework for the quantitative study of evolutionary trees. Syst. Zool. 38, 297–309. Hillis, D. M. & Huelsenbeck, J. P. (1992). Signal, noise, and reliability in molecular phylogenetic analyses. J. Hered. 83, 189–195. Huelsenbeck, J. P. (1995). The performance of phylogenetic methods in simulation. Syst. Biol. 44, 17–48. Hürzeler, J. (1949). Neubeschreibung von Oreopithecus bambolii Gervais. Schweiz. Palaeontol. Abh. 66, 1–46. Jungers, W. L., Falsetti, A. B. & Wall, C. E. (1995). Shape, relative size, and size-adjustments in morphometrics. Yearb. phys. Anthropol. 38, 137–161. Kay, R. F. (1982). Sivapithecus simonsi, a new species of Miocene hominoid, with comments on the phylogenetic status of the Ramapithecinae. Int. J. Primatol. 3, 113–173. Lartet, E. (1856). Note sur un grand singe fossile qui de rattache au groupe des singes supérieures. C.R. Acad. Sci. (Paris) 43, 219–223. Leakey, M. G. & Walker, A. (1997). Afropithecus— function and phylogeny. In (D. R. Begun, C. V. Ward & M. D. Rose, Eds) Function, Phylogeny, and Fossils—Miocene Hominoid Evolution and Adaptations, pp. 225–259. New York: Plenum Press. Leakey, M. G., Leakey, R. E., Richtsmeier, J. T., Simons, E. L. & Walker, A. C. (1991). Similarities in Aegyptopithecus and Afropithecus facial morphology. Folia primatol. 56, 65–85. 572 .  Leakey, R. E. & Leakey, M. G. (1986a). A second new Miocene hominoid from Kenya. Nature 324, 146–148. Leakey, R. E. & Leakey, M. G. (1986b). A new Miocene hominoid from Kenya. Nature 324, 143–146. Leakey, R. E. & Leakey, M. G. (1987). A new Miocene small-bodied ape from Kenya. J. hum. Evol. 16, 369–387. Lipscomb, D. L. (1992). Parsimony, homology and the analysis of multistate characters. Cladistics 8, 45–65. Lyons-Weiler, J. & Hoelzer, G. A. (1997). Escaping from the Felsenstein zone by detecting long branches in phylogenetic data. Mol. Phylogenes. Evol. 8, 376–384. Maddison, W. P. & Maddison, D. R. (1989). Interactive analysis of phylogeny and character evolution using the computer program MacClade. Folia primatol. 53, 190–202. Maddison, W. P. & Maddison, D. R. (1992). MacClade: Analysis of Phylogeny and Character Evolution, Version 3.0. Sunderland, MA: Sinauer Associates. Martin, L. (1980). New specimens of Proconsul from Koru, Kenya. J. hum. Evol. 10, 139–150. Martin, L. (1986). Relationships among extant and extinct great apes and humans. In (B. Wood, L. Martin & P. Andrews, Eds) Major Topics in Primate and Human Evolution, pp. 161–187. New York: Cambridge University Press. McCrossin, M. L. (1997). New postcranial remains of Kenyapithecus and their implications for understanding the origins of hominoid terrestriality. Am. J. phys. Anthrop. Supplement 24, 164. McCrossin, M. L. & Benefit, B. R. (1994). Maboko Island and the evolutionary history of Old World monkeys and apes. In (R. S. Corruccini & R. L. Ciochon, Eds) Integrative Paths to the Past: Paleoanthropological Advances in Honor of F. Clark Howell, pp. 95–122. Englewood Cliffs, NJ: Prentice Hall. McCrossin, M. L. & Benefit, B. R. (1997). On the relationships and adaptations of Kenyapithecus, a large-bodied hominoid from the middle Miocene of Eastern Africa. In (D. R. Begun, C. V. Ward & M. D. Rose, Eds) Function, Phylogeny, and Fossils— Miocene Hominoid Evolution and Adaptations, pp. 241–267. New York: Plenum Press. McCrossin, M. L., Benefit, B. R. & Gitau, S. N. (1998a). Functional and phylogenetic analysis of the distal radius of Kenyapithecus, with comments on the origin of the African great ape and human clade. Am. J. phys. Anthrop. Supplement 26, 158. McCrossin, M. L., Benefit, B. R., Gitau, S. N., Palmer, A. K. & Blue, K. T. (1998b). Fossil evidence for the origins of terrestriality among Old World higher primates. In (E. Strasser, J. Fleagle, A. Rosenberger & H. McHenry, Eds) Primate Locomotion—Recent Advances, pp. 337–352. New York: Plenum Press. Mickevich, M. F. & Johnson, M. F. (1976). Congruence between morphological and allozyme data in evolutionary inference and character evolution. Syst. Zool. 25, 260–270. Mosimann, J. E. (1970). Size allometry: size and shape variables, with characterizations of the log-normal and gamma distributions. J. Am. Stat. Assoc. 56, 930–945. Mosimann, J. E. & Malley, J. D. (1979). Size and shape variables. In (L. Orloci, C. R. Rao & W. M. Stiteler, Eds) Multivariate Methods in Ecological Work, pp. 175–189. Fairland, MD: International Cooperative Publishing House. Moyà-Solà, S. & Köhler, M. (1996). A Dryopithecus skeleton and the origins of great-ape locomotion. Nature 379, 156–159. Nakatsukasa, M., Yamanaka, A., Kunimatsu, Y., Shimizu, D. & Ishida, H. (1998). A newly discovered Kenyapithecus skeleton and its implications for the evolution of positional behavior in Miocene East African hominoids. J. hum. Evol. 34, 657–664. O’Higgins, P. (1989). Developments in cranial morphometrics. Folia primatol. 53, 101–124. Pickford, M. (1985). A new look at Kenyapithecus based on recent discoveries in Western Kenya. J. hum. Evol. 14, 113–143. Pickford, M., Senut, B., Conroy, G. C. & Mein, P. (1994). Phylogenetic position of Otavipithecus: questions of methodology and approach. In (B. Thierry, J. R. Anderson, J. J. Roeder & N. Herrenschmidt, Eds) Current Primatology, Volume 1, Ecology and Evolution, pp. 265–272. Strasbourg: Université Louis Pasteur. Pickford, M., Moyà-Solà, S. & Köhler, M. (1997). Phylogenetic implications of the first African middle Miocene hominoid frontal bone from Otavi, Namibia. C.R. Acad. Sci. (Paris) Serie II 325, 459–466. Pilbeam, D. (1996). Genetic and morphological records of the Hominoidea and hominid origins: A synthesis. Mol. Phylogenet. Evol. 5, 155–168. Pilbeam, D. (1997). Research on Miocene hominoids and hominoid origins: the last three decades. In (D. R. Begun, C. V. Ward & M. D. Rose, Eds) Function, Phylogeny, and Fossils—Miocene Hominoid Evolution and Adaptations, pp. 241–267. New York: Plenum Press. Pilbeam, D., Rose, M. D., Barry, J. C. & Shah, S. M. (1990). New Sivapithecus humeri from Pakistan and the relationship of Sivapithecus and Pongo. Nature 348, 237–239. Preuschoft, H. (1989). Quantitative approaches to primate morphology. Folia primatol. 53, 82–100. Rae, T. C. (1998). The logical basis for the use of continuous characters in phylogenetic systematics. Cladistics 14, 221–228. Rohlf, F. J. (1982). Consensus indices for comparing classifications. Math. Biosci. 59, 131–144. Rohlf, F. J., Chang, W. S., Sokal, R. R. & Kim, J. (1990). Accuracy of estimated phylogenies: Effects of tree topology and evolutionary model. Evolution 44, 1671–1684. Rose, M. D. (1997). Functional and phylogenetic features of the forelimb in Miocene hominoids. In (D. R. Begun, C. V. Ward & M. D. Rose, Eds) Function, Phylogeny, and Fossils—Miocene Hominoid    Evolution and Adaptations, pp. 79–100. New York: Plenum Press. Schwartz, G. T. & Conroy, G. C. (1996). Crosssectional geometric properties of the Otavipithecus mandible. Am. J. phys. Anthrop. 99, 613–623. Scott, L. (1995). Pollen evidence for vegetational and climatic change in southern Africa during the Neogene and Quaternary. In (E. S. Vrba, G. H. Denton, T. C. Partridge & L. H. Burckle, Eds) Paleoclimate and Evolution, With Emphasis on Human Origins, pp. 65–76. New Haven, CT: Yale University Press. Senut, B. & Gommery, D. (1997). Squelette postcrânien d’ Otavipithecus, Hominoidea du Miocène Moyen de Namibie. Ann. Paléontol. 83, 267–284. Senut, B., Pickford, M. & Wessels, D. (1997). Panafrican distribution of Lower Miocene Hominoidea. C.R. Acad. Sci. (Paris) Serie II 325, 741–746. Simon, C. (1983). A new coding procedure for morphometric data with an example from periodical cicada wing veins. In (J. Felsenstein, Ed.) Numerical Taxonomy, pp. 378–383. Berlin: Springer-Verlag. Simons, E. L. & Rasmussen, D. T. (1991). The generic classification of Fayum Anthropoidea. Int. J. Primatol. 12, 163–177. Simons, E. L., Rasmussen, D. T. & Gebo, D. L. (1987). A new species of Propliopithecus from the Fayum, Egypt. Am. J. phys. Anthrop. 73, 139–147. Singleton, M. (1998). The phylogenetic affinities of Otavipithecus namibiensis: quantitative character coding in hominoid phylogeny reconstruction. Ph.D. Dissertation. Washington University. Slowinski, J. B. (1993). ‘‘Unordered’’ versus ‘‘ordered’’ characters. Syst. Biol. 42, 155–165. Slowinski, J. B. & Crother, B. I. (1998). Is the PTP Test Useful? Cladistics 14, 297–302. Smith, R. J. (1984). Determination of relative size: the ‘‘criterion of subtraction’’ problem in allometry. J. Theor. Biol. 108, 131–142. Smith, R. J. (1999). Statistics of sexual size dimorphism. J. hum. Evol. 36, 423–459. Sorenson, M. D. (1996). TreeRot. University of Michigan, Ann Arbor. Strait, D. S., Moniz, M. A. & Strait, P. T. (1996). Finite mixture coding: a new approach to coding continuous characters. Syst. Biol. 45, 67–78. Strait, D. S., Grine, F. E. & Moniz, M. A. (1997). A reappraisal of early hominid phylogeny. J. hum. Evol. 32, 17–82. OTAVIPITHECUS 573 Swindler, D. R. (1976). Dentition of Living Primates. New York: Academic Press. Swindler, D. R. & Ward, S. (1988). Evolutionary significance of the deflecting wrinkle in the lower molars of the Hominoidea. Am. J. phys. Anthrop. 75, 405–411. Swofford, D. L. (1991). PAUP: Phylogenetic Analysis Using Parsimony, Version 3.1. Computer program distributed by the Illinois Natural History Survey, Champaign, Illinois. Swofford, D. L. (1993). PAUP: Phylogenetic Analysis Using Parsimony, Version 3.1.1. (MacIntosh). Computer program distributed by the Smithsonian Institution, Washington, DC. Swofford, D. L. (1998). PAUP*: Phylogenetic Analysis Using Parsimony (*and Other Methods), Version 4.0b1. Sunderland, Massachusetts: Sinauer Associates. Templeton, A. R. (1983). Phylogenetic inference from restriction endonuclease cleavage site maps with particular reference to the evolution of humans and the apes. Evolution 37, 221–244. Thiele, K. (1993). The holy grail of the perfect character: the cladistic treatment of morphometric data. Cladistics 9, 275–304. Thorpe, R. S. (1984). Coding morphometric characters for constructing distance Wagner networks. Evolution 38, 244–255. Trinkaus, E. (1990). Cladistics and the hominoid fossil record. Am. J. phys. Anthrop. 83, 1–11. Walker, A., Teaford, M. F., Martin, L. & Andrews, P. (1993). A new species of Proconsul from the early Miocene of Rusinga/Mfangano Islands, Kenya. J. hum. Evol. 25, 43–56. Ward, C. (1998). Afropithecus, Proconsul, and the primitive hominoid skeleton. In (E. Strasser, J. Fleagle, A. Rosenberger & H. McHenry, Eds) Primate Locomotion—Recent Advances, pp. 337–352. New York: Plenum Press. Ward, S. C. (1997). The taxonomy and phylogenetic relationships of Sivapithecus revisited. In (D. R. Begun, C. V. Ward & M. D. Rose, Eds) Function, Phylogeny, and Fossils—Miocene Hominoid Evolution and Adaptations, pp. 269–289. New York: Plenum Press. Ward, S. C., et al. (1999). Equatorius: a new hominoid genus from the middle Miocene of Kenya. Science 285, 1382–1386. Wu, R. (1987). A revision of the classification of the Lufeng great apes. Acta Anthrop. Sin. 6, 263–271.