Reassessing Hominoid Phylogeny: Evaluating Congruence in the Morphological and Temporal Data

Paleobiology, 30(4), 2004, pp. 614–651

Reassessing hominoid phylogeny: evaluating congruence in the morphological and temporal data

John A. Finarelli and William C. Clyde

Abstract.—The phylogenetic relationships of fossil and extant members of the primate superfamily Hominoidea are reassessed by using both conventional (morphological) cladistic and stratocladistic (incorporating morphological and temporal data) techniques. The cladistic analysis recovers four most parsimonious cladograms that distinguish postcranially primitive (‘‘archaic’’) and derived (‘‘modern’’) hominoid clades in the earliest Miocene of East Africa and supports distinct hominine and pongine clades. However, the relationships among the pongines and hominine clades and other Eurasian hominoids remain ambiguous and there is weak support (Bremer decay indices, reduced consensus, and bootstrap proportions) for several other parts of the proposed phylogeny. An examination of the partitioning of homoplasy across the two major hominoid clades recovered in the cladistic analysis indicates that the majority of the observed homoplasy resides in the postcranially derived clade. An examination of the partitioning of homoplasy across anatomical regions indicates that dental characters display a significantly higher level of homoplasy than postcranial characters. A rarefaction analysis demonstrates that the higher homoplasy associated with the dental characters is not the result of sampling biases, indicating that postcranial skeletal characters are likely the more reliable phylogenetic indicators in the hominoids. The branching order of the most parsimonious cladograms shows better than average congruence with the observed ordering of first appearances in the fossil record, implying that the hominoid fossil record is surprisingly good. As with morphologic parsimony debt, most of the stratigraphic parsimony debt in these cladograms is associated with the ‘‘modern’’ hominoid clade. A stratocladistic analysis of the data recovers a single most parsimonious phylogenetic tree with a different cladistic topology from the morphological cladogram. The most striking difference is the elimination of the postcranially primitive clade of hominoids in the early Miocene in favor of a pectinate succession of taxa. The relative position of the late-appearing taxon Oreopithecus is also altered in the stratocladistic hypothesis. Topological differences between the cladistic and stratocladistic hypotheses highlight two intervals of significant discord between the morphological and temporal data—the early Miocene of eastern Africa and the late Miocene of Eurasia. The first discrepancy is likely the result of poor preservation and morphological homoplasy in Morotopithecus, as the fossil record in the early Miocene of eastern Africa for the ingroup is rather good. The second discrepancy is likely the result of the unusual preservation conditions associated with the late Miocene hominoid Oreopithecus.

John A. Finarelli. Committee on Evolutionary Biology, University of Chicago, 1025 East Fifty-seventh Street, Culver Hall 402, Chicago, Illinois 60637, and Department of Geology, The Field Museum of Natural History, 1400 South Lake Shore Drive, Chicago, Illinois 60605. E-mail: [email protected] William C. Clyde. Department of Earth Sciences, 56 College Road, University of New Hampshire, Durham, New Hampshire 03824. E-mail: [email protected]

Accepted: 18 February 2004

Introduction date no phylogenetic analysis of the Homi- Despite a considerable research effort to re- noidea has explicitly incorporated temporal ﬁne our understanding of the evolutionary re- data in hypothesis testing, and no quantitative lationships among known fossil and extant evaluation of the congruence of the morpho- hominoid taxa, little consensus has emerged logical and temporal data for the hominoid (e.g., Begun 1992a,b, 1994, 1995; Begun et al. fossil record has occurred. In an attempt to 1997a; de Bonis and Koufos 1993; Dean and clarify these evolutionary relationships and to Delson 1992; Andrews et al. 1996; MacLatchy evaluate congruence across data types, a phy- et al. 2000). Although their use in phylogenetic logenetic reassessment of the Hominoidea was analysis remains controversial, temporal data performed incorporating both morphological represent a potentially relevant class of infor- and stratigraphic data. mation that is often overlooked (Fisher 1991, Although cladistic hypotheses are evaluat- 1994; Huelsenbeck 1994; Wagner 1995). To ed with no explicit reference to time, all clad-

᭧ 2004 The Paleontological Society. All rights reserved. 0094-8373/04/3004-0007/$1.00 DATA CONGRUENCE IN HOMINOID PHYLOGENY 615 ograms make implicit statements about the ered under stratocladistic analysis, without si- relative time of divergence among taxa by the multaneously observing significant decreases order of branching events (Fisher 1991; Wag- in the fit of morphological data. In computer ner 1995). The sequence of branching events in simulations in which fossil records were cre- a morphological cladistic hypothesis is often ated for hypothetical taxa with known evolu- harmonized with the fossil record of the in- tionary histories, stratocladistics significantly group through the creation of ‘‘ghost lineag- outperformed conventional cladistic analysis es,’’ artificial extensions of a taxon’s range be- in recovering the true phylogenies (Fox et al. yond its observed first appearance in the fossil 1999). Additionally, in cases where neither record (Norell 1993). This approach essential- method was able to recover the correct phy- ly erases any discrepancy between the ob- logeny, the stratocladistic hypotheses more served order of appearance events and the or- closely matched the true phylogenies (Fox et der implied by the hypothesis. Insofar as al. 1999). ghost lineages explain away discrepancies between (stratigraphic) observation and (cladis- Methods tic) hypothesis, they may be considered ap- Morphological Data peals to ad hoc support, analogous to the way homoplasy is invoked to explain away mor- Thirteen fossil and five extant genera of the phological data that are incongruent with a primate superfamily Hominoidea constituted cladistic hypothesis (Fisher 1991, 1994). the ingroup of this analysis. The morpholog- Stratocladistics is a parsimony-based crite- ical data matrix was modified from the char- rion that evaluates competing phylogenetic acter-by-taxon matrix of Begun et al. (1997a). hypotheses relative to both morphological That study examined 240 morphological char- characters and a stratigraphic character de- acters for eight fossil and five extant genera. rived from the stratigraphic record of the in- Modifications to the Begun et al. (1997a) data group taxa (Fisher 1991, 1992, 1994). The mor- set are discussed below. The character state phological component is evaluated as in con- descriptions and the character by taxon matrix ventional cladistic analyses, where each in- used in this phylogenetic analysis have been stance of homoplasy imparts a unit of included as Appendices 1 and 2, respectively. ‘‘morphological parsimony debt’’ upon the Added Taxa. The Begun et al. (1997a) ma- hypothesis. In addition, each instance of in- trix was first expanded to include five addi- congruence between the stratigraphic data tional taxa (Turkanapithecus, Equatorius [sensu and the hypothesis, where a taxon is predicted Ward et al. 1999], Griphopithecus, Morotopithe- to exist by the branching order yet is not ob- cus [sensu Gebo et al. 1997], and Ankarapithe- served, imparts a unit of ‘‘stratigraphic par- cus). Morphological data for these taxa were simony debt’’ (Fisher 1992). Stratocladistics obtained from the literature. Character coding then sums the morphological and stratigraph- for the added taxa preferentially followed cla- ic parsimony debt values of each hypothesis, distic treatments of these taxa in other studies creating a single ‘‘total parsimony debt’’ val- by the authors of the Begun et al. (1997a) anal- ue. The minimum total parsimony debt value ysis to ensure consistency in character coding. over the set of possible phylogenetic hypoth- Additional character state information was eses determines the overall most parsimoni- also incorporated from descriptions of holo- ous phylogenetic hypothesis (Fisher 1992, types and additional fossil material. 1994). Morotopithecus was coded following Begun Two recent studies have shown that strato- and Gu¨leç 1998 and Ward 1997a, and with de- cladistics may outperform conventional cla- scriptions of new postcranial fossils from the distic analyses in recovering evolutionary his- Moroto II locality by Gebo et al. (1997) and tories for fossil taxa. Working with published MacLatchy et al. (2000). Turkanapithecus was cladistic analyses, Clyde and Fisher (1997) coded following Rae 1997, Rose 1997, and noted significant increases in the fit of strati- Ward 1997a. Additional character information graphic data to phylogenic hypotheses recov- was obtained from Leakey et al. 1988. Follow- 616 JOHN A. FINARELLI AND WILLIAM C. CLYDE ing the separation of Equatorius from Kenyapi- logical trait, characters were consolidated to thecus by Ward et al. (1999), coding of Equa- avoid unduly weighting the impact of these torius followed the character coding for Keny- characters on the cladistic analysis. Combined apithecus africanus by Rose (1997) and Ward pairs (again following the numbering scheme (1997a). Additional character information was of Begun et al. 1997) include: characters 5 and obtained from descriptions of fossil material 6 describing the morphology of the scapula, by Ward et al. (1999), Kelley et al. (2002), and 18 and 19 describing the trochlea of the hu- Sherwood et al. (2002). Begun (2000) has ar- merus, 39 and 40 describing the trapezoid fac- gued that Equatorius should be assigned to the ets, 47 and 48 describing ulnar articulation genus Griphopithecus on the basis of dental with carpal elements, characters 94 and 95 de- similarity, although Kelley et al. (2002) argued scribing the shape of the trochlea of the talus, that differences between Equatorius and Turk- characters 143 and 188 describing the shape of ish specimens attributed to Griphopithecus the nasal aperture (see above, ‘‘Added Char- support separation at the generic level. Recod- acters’’), characters 213 and 215 describing the ing of the OTU Kenyapithecus after the removal morphology of the nasal clivus, and charac- of Equatorius followed Rose 1997 and Begun et ters 232 and 233 describing the morphology of al. 1997a for those characters attributable to K. the upper molars. wickeri. Griphopithecus was coded following Character 88, describing the depth of the Begun and Kordos 1997 and Begun and Gu¨leç femoral condyle, character 96, describing the 1998, with additional character information angle of the talar neck, and characters 112 and from Andrews et al. 1996, Alpagut et al. 1990, 113, describing the morphology of the first and Begun 1992c. Ankarapithecus was coded metatarsal, were discarded from the matrix. following Begun and Gu¨leç 1998 and Alpagut The femoral, tarsal, and metatarsal morphol- et al. 1996. In addition to the taxa added to the ogies coded in these characters are believed to analysis, newly described fossil material (Ma- represent derived hominin (ϭ human and all dar et al. 2002) provided character state infor- direct ancestors [Tattersall et al. 1988]) char- mation for the taxon Sivapithecus. acters associated with bipedalism and there- Added Characters. Five characters, which fore are not homologous to the primitive pro- have been proposed as phylogenetically infor- pliopithecid condition (Jungers 1988; Ward mative but were not included in the Begun et 1997a). Additionally, characters 141, 142, and al. (1997a) analysis, were added to the data 176, describing the morphology of the maxil- matrix for this study. These were morphology lary sinuses, were excluded from the data ma- of the olecranon process and orientation of the trix, as variation in maxillary sinus morphol- radial notch (Rose 1997), morphology of the ogy in the hominoids has been demonstrated entocunieform facet on the navicular (Ward to be a function of isometric scaling (Rae and 1997a), torso shape (Schultz 1961; Ward Kopp 2000). 1997a), and a modification of the character coding nasal aperture shape (Begun and Gu¨- Temporal Data leç 1998). All five characters were coded in There is generally good stratigraphic con- studies using the same coding scheme as Be- trol for hominoid fossil localities suggesting gun et al. (1997a), thus ensuring consistency. that this information could be of potential val- Deleted Characters. The data matrix was ue in resolving phylogenetic relationships scanned for characters that were uninforma- within this group. For this study, stratigraphic tive for the ingroup, and these were deleted data in the form of first and last appearance from the matrix. Following the numbering events (FAEs and LAEs) for each hominoid scheme of Begun et al. (1997a) these were taxon were compiled from the literature. characters 1, 7, 8, 9, 10, 11, 12, 13, 14, 15, 35, Wherever available, radiometric ages or paleo- 36, 41, 46, 49, 50, 51, 64, 69, 70, 78, 97, 100, 161, magnetic data linking fossil localities to the 181, 225 and 240. Geomagnetic Polarity Timescale (GPTS) of In several instances where multiple charac- Cande and Kent (1995) were used. However, ters appeared to code for the same morpho- the temporal distributions of several taxa are DATA CONGRUENCE IN HOMINOID PHYLOGENY 617 known only in relation to biostratigraphy and LAE for this taxon is somewhat problem- (e.g., Ouranopithecus), and for these cases ap- atic. This is due to some uncertainty in attri- propriate regional faunal calibrations (e.g., buting the Meswa Bridge (FAE) and Ngorora European MN [Neogene Mammal] Zones) Formation (LAE) material to Proconsul (see were used to correlate these taxa into a uni- Appendix 3). In the case of Meswa Bridge, if versal temporal succession of first and last ap- the material were removed from the opera- pearance events (FAEs and LAEs) for the tional taxonomic unit of this study, the coding Hominoidea. of the stratigraphic character would not be al- The coding of the stratigraphic character for tered as the point occurrence of Morotopithecus this study follows the methodology outlined would be coded with the Kisingiri and Tin- in previous stratocladistic analyses by Clyde deret levels, as the earliest stratigraphic hori- and Fisher (1997) and Bloch et al. (2001). zon in the hominoid succession (‘‘a’’ in Fig. 1). Range-through assumptions were made for The LAE poses a more complicated problem. each taxon between its FAE and LAE, and With the assignment of the Ngorora Forma- those taxa observed (or inferred by the range- tion material to Proconsul, the young age of through assumption) to exist at multiple levels these localities (approximately 12.5 Ma [Deino were assigned multiple character states (Clyde et al. 1990; Hill et al. 1985, 2002]) creates a and Fisher 1997). Although it is possible to large extension for Proconsul with the range- code sampling gaps in the stratigraphic char- through assumption. However, if the Ngorora acter by skipping a letter or number in the material were not included in the OTU, then coding scheme (see Fisher 1992), a conserva- the coding of the stratigraphic character for tive coding scheme leaves such gaps uncoded Proconsul would comprise only a single char- and treats intervals where no ingroup taxa are acter state (‘‘a’’ in Fig. 1). This could poten- observed as ‘‘no data’’ (Clyde and Fisher 1997; tially have a considerable effect on the result Finarelli and Clyde 2002). Such intervals are of the stratocladistic analysis. As such, we per- therefore not factored into the calculation of formed the stratocladistic analysis twice, us- stratigraphic parsimony debt. ing both coding schemes for Proconsul—first, Hominoids are observed to range from the using a single stratigraphic character state (as- first appearance of Proconsul at Meswa Bridge suming that the Ngorora material is not Pro- (biostratigraphically constrained to ca. 23.5 consul), and second, spanning multiple strati- Ma [Pickford and Andrews 1981; Tassy and graphic character states (assuming that it be- Pickford 1983]) through the Recent. The strati- longs to Proconsul—as depicted in Fig. 1). The graphic data divide the range of the homi- results of both analyses were identical. There- noids into 11 distinct stratigraphic intervals fore, the coding of the stratigraphic character (Fig. 1). The detailed stratigraphic information here reflects the most inclusive set of fossils at- that was used to order the FAEs and LAEs is tributed to Proconsul; however, it should be reported in Appendix 3. All of the boundaries noted that additional material (especially for stratigraphic character states are defined from the Ngorora Formation) may significant- by FAEs and/or LAEs, but not all appearance ly alter the appearance events for this taxon. events were used in defining character state A recurring difficulty in coding the strati- boundaries (e.g., Turkanapithecus; Fig. 1). That graphic information for the hominoids is is to say, in some cases the coding scheme was ‘‘point occurrences’’ in the fossil record. Tur- coarsened such that the appearance events of kanapithecus provides a good example. At Ka- a particular taxon were subsumed within an- lodirr in Kenya, Turkanapithecus is demonstra- other stratigraphic character state, reflecting bly younger than some of the Afropithecus ma- areas where there is a lack of precision in the terial (Leakey and Leakey 1986b; Brochetto et stratigraphic data. Individual cases where this al. 1992), yet it is also found at the same level occurred are discussed below. as other Afropithecus fossils. It is easy to dem- Although the taxonomic assignment of the onstrate contemporaneity of the two genera; Kisingiri and Tinderet material to the genus however, demonstrating that Afropithecus ex- Proconsul is rather secure, the potential FAE isted prior to, and exclusive of, Turkanapithecus 618 JOHN A. FINARELLI AND WILLIAM C. CLYDE

FIGURE 1. Observed stratigraphic ranges for the hominoid taxa in this study. Stratigraphic ranges are based on information presented in Appendix 3. Coding of the stratigraphic character for the Hominoidea is shown at the right. The stratigraphic data for the ingroup produce 11 distinct stratigraphic intervals (coded ‘‘a’’ through ‘‘k’’). First and last appearance events (FAEs and LAEs) define all stratigraphic character states. In several cases where appearance events are poorly resolved, they are not used for defining character state boundaries. The stippled time interval between the LAE of Oreopithecus andtheFAEofAustralopithecus is not coded in the stratigraphic character. See text for discussion. is difficult. Instead of coding two separate in- an interval for the point occurrence of this tax- tervals for Afropithecus before the appearance on, and an interval following it. Instead, Mo- of Turkanapithecus and a geologically instan- rotopithecus is considered here to occupy the taneous interval comprising the overlap of Af- stratigraphic interval defined by the FAE of ropithecus and Turkanapithecus, they are con- Proconsul and the FAE of Afropithecus. Keny- servatively coded as occupying a single strati- apithecus is known from Fort Ternan, and its graphic level. occurrence is entirely subsumed within the Similar arguments are made for Morotopi- upper range of Griphopithecus (Ward et al. thecus, Kenyapithecus,andAnkarapithecus. Mo- 1999). Ankarapithecus is known from two lo- rotopithecus is known from the Moroto fossil calities in the Sinap Formation at Yassıo¨ren, sites in Uganda, which are radiometrically Turkey (Ozansoy 1965; Andrews and Tekkaya dated to be older than 20.6 Ma (Gebo et al. 1980; Alpagut et al. 1996) and correlates to the 1997), and is known only from specimens at interval defined by Sivapithecus and Dryopithe- this locality. If the coding of the stratigraphic cus. See Appendix 3 for details of the strati- character were to coincide with all FAE and graphic data. LAE data, then one would have to code an in- The taxonomic assignment of the fossil ma- terval before the occurrence of Morotopithecus, terial at Engelswies is uncertain. It is usually DATA CONGRUENCE IN HOMINOID PHYLOGENY 619 assigned to ‘‘cf. Griphopithecus,’’ although terminus of a faunal zone for which a range- Heizmann and Begun (2001) urge caution and through assumption is made, and coding of a suggest that the Engelswies hominoid may distinct character state is unwarranted. represent a distinct genus. The FAE of Gri- phopithecus is therefore documented either at Cladistic Analysis of the Hominoidea Engelswies or at Pas¸alar (Heizmann and Be- The Morphologically Most Parsimonious gun 2001). Using either fossil locality for the Cladograms FAE does not alter the coding of the stratigraphic character (see Appendix 3). The lower A cladistic analysis using the branch and MN5FAEofGriphopithecus in Eurasia dem- bound algorithm in PAUP* (version 4.0b10 onstrates that all of the western Kenyan Equa- [Swofford 2002]) recovered four most parsi- torius localities are contained within the range monious cladograms with respect to the mor- of Griphopithecus. However, the lower limit on phological data. Each had a tree length of 447 this appearance event is less well constrained. steps and a Retention Index (RI) of 0.69 (Fig. If the earliest possible calibration for the FAE 2A). This value for the RI is somewhat low of Griphopithecus is used (Base MN 5 [see Heiz- when compared to other published RIs for mann and Begun 2001]) and the latest possible studies that include fossil taxa. For example, calibration for the LAE of Afropithecus is also Clyde and Fisher (1997: Table 2) compared RI used (Ad Dabtiyah is biostratigraphically cor- values across 29 studies of varied taxonomic related to the base of Maboko [see Gentry resolution. These studies had a median RI of 1987a,b]), then some degree of overlap would 0.80, indicating that the degree of homoplasy be implied, which would also imply a separate observed in the Hominoidea is somewhat stratigraphic character state coded on the ba- higher than for other taxonomic groups. sis two indefinite biostratigraphic correla- The strict consensus cladogram for the mor- tions. Until either or both of these appearance phologically most parsimonious cladograms events are better constrained, the data do not (hereafter referred to as the MMPC) high- warrant the creation of this distinct character lights topological features common to all of state. As such, Afropithecus and Griphopithecus the cladograms as well as areas of ambiguity are coded here as non-overlapping (Fig. 1). (Fig. 2B). Proconsul is the sister taxon to all The FAE of Sivapithecus is documented other ingroup taxa. Unfortunately, this does within the Chron C5Ar.1, making it strati- not resolve the debate over the phyletic posi- graphically lower than the latest possible es- tion of Proconsul, because this position is con- timate for the LAE of Griphopithecus. A range- sistent with its being either a basal hominoid through assumption to the end of the MN 6 (Andrews and Martin 1987a; Andrews 1992) would place the LAE of Griphopithecus locally or an undifferentiated Miocene catarrhine at the Chron C5Ar.1r (Steininger et al. 1996). (Harrison 1987; Harrison and Sanders 1999). However, coding an overlap in this case would The most notable feature of the MMPC is the create a character state based on the correla- distinction between a postcranially primitive tion of the well-documented FAE of Sivapithe- clade of early to middle Miocene hominoids cus to the terminus of a faunal zone for which from East Africa and a postcranially derived a range-through assumption must be made. hominoid clade that includes both late Mio- Conservatively, such an assumption of con- cene Eurasian forms and all of the extant hom- temporaneity is not warranted. Similarly, Or- inoid genera. Morotopithecus is joined to the eopithecus overlaps with the terminal portion base of the derived clade, supporting Mac- of the Sivapithecus range. The V2 horizon at Latchy et al. (2000), who hypothesized the ex- Baccinello is correlated to earliest MN 13 istence of two evolutionarily distinct homi- (Rook et al. 2000). If a similar range-through noid lineages in East Africa by 20 Ma on the assumption to the end of MN 13 were made, basis of several derived features of the femur Oreopithecus’s LAE would extend beyond Si- and axial skeleton. However, it is not the in- vapithecus. Again, this correlation compares a clusion of Morotopithecus in this phylogenetic well-documented stratigraphic datum to the analysis that is responsible for the division of 620 JOHN A. FINARELLI AND WILLIAM C. CLYDE

FIGURE 2. A, The four morphologically most parsimonious cladograms (MMPC) recovered for the character by taxon matrix compiled for the 18 ingroup taxa (Appendix 2). Each cladogram has a tree length of 447 steps and an RI of 0.69. B, The strict consensus cladogram of the MMPC. The morphologically most parsimonious hypotheses distinguish two distinct clades of hominoids, a clade of ‘‘archaic’’ hominoids from the early Miocene of East Africa, and a clade of ‘‘modern’’ hominoids, including all extant hominoids and the early Miocene hominoid Morotopithecus. Numbers above branches indicate Bremer Decay indices for corresponding internal nodes. DATA CONGRUENCE IN HOMINOID PHYLOGENY 621 the hominoids into these two separate clades. hominines (Andrews 1992; Andrews et al. The phylogenetic analysis was repeated on the 1996; Dean and Delson 1992), to the pongines data set while excluding Morotopithecus,anda (Schwartz 1990; Moya`-SolaàndKo¨hler 1993), set of eight cladograms was recovered, each or to all extant great apes (Fig. 2A). However, distinguishing a monophyletic radiation of the Ouranopithecus/hominine/pongine clade early to middle Miocene hominoids seen in is consistently derived in the MMPC with re- the MMPC. The topologies of each of these spect to Dryopithecus (contra Begun 1992b, eight cladograms were identical to the 1994; Begun et al. 1997a; Begun and Gu¨leç MMPC, except that the relative position of 1998). Additionally, the interrelationships Proconsul varied across the set creating a po- among the chimpanzee, gorilla, and human lytomy between Proconsul and the two homi- lineages are not resolved in the strict consen- noid clades. Thus, early Miocene hominoids sus. A large suite of synapomorphies includ- from East Africa form a distinct clade in the ing postcranial, cranial, and dental characters MMPC, and the early appearing Morotopithe- unites the great ape clade (Table 1). Synapo- cus is allied with the postcranially derived morphies associated with their unique facial clade of hominoids. morphology and extreme heteromorphy of This hypothesis parallels, although is not the upper incisors unite the pongine clade, exactly identical to, Pilbeam’s (1997) concept whereas morphology primarily associated of distinct evolutionary histories for homi- with the hands and feet (e.g., fusion of os cen- noids of ‘‘archaic’’ and ‘‘modern’’ aspects. For trale) and cranium (e.g., increased robusticity ease of reference, Pilbeam’s terminology will of the supraorbital torus, broad supraorbital be adopted here, and the clade of postcrani- sulcus) unite the hominine clade (Table 1). ally primitive hominoids from the early and The MMPC have tree lengths of 447 steps middle Miocene of East Africa (Afropithecus, with 201 homoplasies. However, 17 clado- Turkanapithecus, Equatorius, Kenyapithecus,and grams were recovered with tree lengths of 448 Griphopithecus) will be informally referred to steps (202 homoplasies) and 67 cladograms as ‘‘archaic’’ hominoids, and the postcranially with 449 steps (203 homoplasies). By 455 steps derived clade (Morotopithecus, etc.) as ‘‘mod- (209 homoplasies) there are more than 16,000 ern’’ hominoids (Fig. 2B). Analysis of unam- recovered cladograms. This large number of biguous character state reconstructions (DEL- cladograms with only slightly higher tree TRAN optimization) demonstrates that the lengths raises concern over the strength of clade of ‘‘archaic’’ hominoids is supported by support for the topologies observed in the synapomorphies primarily centered on the strict consensus cladogram. Although this is morphology of the face (Table 1). Derived fea- certainly not a new debate, there is still no tures of the vertebrae and hindlimb, which are method to determine if the cladograms with thought to be associated with positional be- 201 homoplasies are significantly better than havior and locomotion (Ward 1997a), are re- those with 202 homoplasies (Wagner 1995). In constructed as synapomorphies for the clade the absence of a rigorous criterion for deter- of ‘‘modern’’ hominoids (Table 1). mining significance, we used additional anal- Within the clade of ‘‘modern’’ hominoids, yses to evaluate the strength of support for the there is a monophyletic clade of great apes phylogenetic statements in the most parsi- (Hominidae), which includes distinct clades of monious cladograms. South Asian great apes (Ponginae: Ankarapi- thecus, Sivapithecus, Lufengpithecus,andPongo) Consensus Analyses and African great apes (Homininae: Pan, Go- By evaluating strict consensus cladograms rilla,andAustralopithecus) (Fig. 2B). The inter- for successively higher tree lengths, the num- relationships between the pongines, the hom- ber of additional steps needed to ‘‘break’’ a inines, and the late Miocene European homi- node can be used as a measure of the relative noid Ouranopithecus are ambiguous in the strength of support for the phylogenetic state- strict consensus, as Ouranopithecus is alter- ments in a cladogram (Bremer 1988). Bremer nately hypothesized as the sister taxon to the Decay indices were calculated for the strict 622 JOHN A. FINARELLI AND WILLIAM C. CLYDE

TABLE 1. Unambiguous synapomorphies uniting major clades in the MMPC: character states reconstructed using the DELTRAN optimization.

Clade Description Characters ‘‘Hominoids of Archaic Aspect’’ Reduction of molar cusps and lower canine 111, 117 Deepening of the zygomatic and more vertical orientation of zygomatic 161, 169 Lowering of the nasal aperture relative to the orbits 171 Increased length of nasal bones 174 Deep canine fossa 179 Lengthening of nasoalveolar clivus 197 ‘‘Hominoids of Modern Aspect’’ Increased vertebral body height 57 More dorsally positioned transverse process 59 Increased asymmetry of the femoral condyle 70 Hominidae More medially asymmetric trochlear keel 15 Loss of articulation of ulna with pisiform/triquetral 36 Broadening of sternebrae 60 Increased size of plantar tubercle 80 Increased cuboid wedging 83 Loss of prehallux facet (MT 1) 90 Increased robusticity of MT 2 and MT 5 91, 95 Shift in the foot axis through the second digit 94 Reduction of male canine size 100 Upper P3: reduced cusp heteromorphy, reduced para- cone 103, 104 Lower P3 metaconid present 107 Lengthening of lower P4 108 Loss of metacones, increase in upper molar crown size 110, 131 Relative increase in lower M1 size, lengthening of upper M1 and M2 119, 128 Increased heteromorphy of the lateral incisors 121 Inflation of glabella 143 Presence of supraorbital torus and superciliary ridges 144, 145 Reduction of frontal sinus size 148 Increased length of neurocranium 155 Reduction of articular tubercle and articular/tympanic fused to temporal 157, 158 Raising of nasal aperture relative to alveolar plane 172 Presence of deep incisive fossa 195 Lengthening of nasoalveolar clivus 197 Large incisive canal present 199 Ponginae Greatly increased lateral incisor heteromorphy 121 Greatly increased orbital breadth 141 Zygomatics positioned more anteriorly and vertically oriented 162, 169 Decreased distance from nasal aperture to orbits (primitive reversal) 164 Deep canine fossa 179 Piriform nasal aperture 200 Homininae Fused os centrale 28 Concavoconvex centrale facet on capitate 42 Angled posterior talar facet 78 Decreased cuboid peg 82 Increase in size of phalangeal flexor ridges 97 Longer upper P4 124 Increased robusticity of supraorbital torus, broad supraorbital sulcus 144, 146 Inferior orientation of the nuchal plane 154 More horizontal orientation of the zygomatics 169 Inflated maxillary alveolar process 182 Larger lesser palatine foramina 186 Pyramidal process inferiorly broad 190 Pterygoid process compressed 191 DATA CONGRUENCE IN HOMINOID PHYLOGENY 623 consensus of the MMPC using the program ic’’ hominoid clade upon exclusion of mutu- AutoDecay (Version 5.0 [Eriksson 2001]) and ally segregating units (i.e., if Turkanapithecus is PAUP* (heuristic searches with 100 random excluded, then a Kenyapithecus/Equatorius/ sequence additions) (Fig. 2B). The Bremer De- Griphopithecus clade is recovered; if Kenyapithe- cay indices calculated for the MMPC are very cus/Equatorius/Griphopithecus are excluded, low; no internal node in the strict consensus then Turkanapithecus/Afropithecus is recovered persists past three additional steps. A distinct through three additional steps, and through clade of ‘‘modern’’ hominoids is recoverable four steps if Kenyapithecus is removed. only through two additional steps, and the From the reduced consensus analysis, it ‘‘archaic’’ clade is recoverable through a single was apparent that much of the poor nodal additional step. Several clades are not recov- support was due to Kenyapithecus, which was erable through even one additional step (Bre- repeatedly excluded from reduced consensus mer Decay Index ϭ 0; Fig. 2B), such as the Af- partitions at and beyond three additional rican great apes, as one of the cladograms steps. The great ape clade is supported in re- with a tree length of 448 allies Ouranopithecus duced consensus through four additional with Australopithecus. This low Bremer sup- steps upon removal of only Kenyapithecus,and port is likely the result of the relatively high it can be recovered through six additional proportion of homoplasy and highly incom- steps if Lufengpithecus is also excluded. When plete preservation for several of the fossil taxa. Kenyapithecus is excluded, a clade uniting Pon- However, strict consensus cladograms are go and Sivapithecus is recovered through six highly insensitive to repeated clustering of additional steps. Although these results do taxa when the phylogenetic positions of a few imply a good deal of ambiguity in the rela- taxa in the ingroup are highly mobile (Wilkin- tionships of late Miocene European apes and son 1999), and such mobility can be greatly the hominines, they argue for distinct homi- enhanced by incomplete representation in the nine and pongine lineages. In addition, the re- data matrix, and therefore highly incomplete peated exclusion of Kenyapithecus from re- fossil taxa. Because of this insensitivity, rela- duced consensus cladograms demonstrates tionships preserved in strict consensus clad- that the phylogenetic position of this taxon is ograms can be inferred as strong statements, highly unstable at slightly longer tree lengths. but important phylogenetic statements can be It is likely that this instability is due to incom- lost when relying solely upon strict consensus plete preservation. Kenyapithecus is coded for methods (Wilkinson 1994, 1996). We therefore only 25% of the characters in the data matrix used reduced consensus analysis (Wilkinson (Appendix 2). 1994, 1996, 1999) to uncover any such ob- scured statements, highlighting groupings Bootstrap Analysis that are consistently implied by a set of clad- Bootstrapping was also used to evaluate ograms if one or more taxa with variable po- support for MMPC topologies (Felsenstein sitions are removed from consideration. It is 1985). Two hundred characters were random- important to note that this method does not ly resampled with replacement for 1000 heu- ignore data. The entire data matrix is used to ristic search replicates (ten random sequence recover the set of cladograms; taxa are only additions per replicate) and a majority-rule subsequently pruned to uncover repeated to- consensus cladogram was produced. Most of pologies lost in the strict consensus (Wilkin- the nodes observed in the MMPC are left am- son 1994, 1999). biguous in the 50% majority-rule cladogram For the reduced consensus analyses, we for the bootstrap (Fig. 3A). Additionally, those used the program ‘‘strict.exe’’ in the REDCON nodes that were resolved have low bootstrap software package (Version 3.0 [Wilkinson proportions, indicating only marginal sup- 2001]). Reduced consensus analyses demon- port. It is likely that this, too, is a function of strate that the morphological support for the the incompleteness of certain taxa in the data ‘‘archaic’’ lineage is limited. These analyses matrix, which increases ambiguity under re- were only able to recover parts of the ‘‘archa- sampling techniques such as the bootstrap. We 624 JOHN A. FINARELLI AND WILLIAM C. CLYDE

FIGURE 3. A, The majority-rule bootstrap consensus of the full character by taxon matrix (heuristic search ten random sequence additions; 1000 replicates). Many of the internal nodes in the bootstrap topology are ambiguous, and bootstrap proportions are generally low. These ambiguities are introduced by the highly incomplete taxon Kenyapithecus. B, The majority-rule bootstrap consensus for the same character data, excluding Kenyapithecus. Num- bers above branches indicate bootstrap proportions for corresponding internal nodes. Bootstrap proportions for the internal nodes resolved in A are either maintained or increased in B. DATA CONGRUENCE IN HOMINOID PHYLOGENY 625 noted above that Kenyapithecus played a large strap proportions. It is likely that much of the role in the low Bremer support for many of the uncertainty in the position of Lufengpithecus nodes in the MMPC. Therefore, a second boot- and Ankarapithecus is caused by the relatively strap was performed on the data set excluding poor preservation of a few key taxa. Kenyapithecus (Fig. 3B). The topology of this Two additional bootstraps were performed second bootstrap is fully consistent with that removing either Ankarapithecus or Lufengpithe- of the first, only more resolved, and the boot- cus in addition to Kenyapithecus. The bootstrap strap proportions indicate stronger support topologies produced for the remaining taxa for several of the internal nodes. Upon remov- for these additional analyses were identical to al of Kenyapithecus,anEquatorius/Griphopithe- the bootstrap excluding Kenyapithecus (Fig. 3B) cus clade is also recovered. Removal of Keny- with several notable exceptions. The removal apithecus also serves to separate the clade of of the more incomplete Lufengpithecus causes ‘‘modern’’ hominoids (albeit without Moroto- the bootstrap proportion for Sivapithecus/Pon- pithecus) relative to early Miocene hominoids go to rise to 79%, and a monophyletic Pongi- from East Africa with strong support, as well nae is recovered with 73% support. However, as increasing the support for a distinct clade if Ankarapithecus is excluded and Lufengpithe- of great apes, including Oreopithecus. cus retained, then the bootstrap proportion for The bootstrap topologies do differ signifi- Sivapithecus/Pongo rises to 88%, and a mono- cantly from the strict consensus cladogram in phyletic Ponginae is recovered 52% of the several respects. In the strict consensus of the time. The removal of additional incompletely MMPC, both the pongine clade and Ourano- coded taxa decreases some of the ambiguity in pithecus are hypothesized as derived relative the bootstrap results in other regions of the to Dryopithecus and Oreopithecus. In both boot- cladogram. The bootstrap proportions for the straps (with and without Kenyapithecus), all Hominidae increase in both of these analyses European hominoids are joined in an unre- to about 90%, and there is an increase support solved polytomy with the hominines (Fig. for the Equatorius/Griphopithecus clade (ap- 3A,B). The support for this group as a clade proximately 70%), and a weakly supported that is both distinct from and derived in re- (about 55%) clade of ‘‘modern’’ hominoids is lation to the South Asian hominoids is ex- recovered. None of the bootstrap analyses re- tremely weak (52% in both bootstraps), indi- covered a distinct clade of ‘‘archaic’’ homi- cating that the relative positions of the Hom- noids. ininae and Ponginae with respect to late Mio- Judging from these various analyses, it is cene European hominoids is rather weakly clear that poorly preserved fossil taxa and rel- supported by the morphological data, despite atively high levels of homoplasy infuse con- its apparently clear resolution in the MMPC. siderable uncertainty into the cladistic analy- The polytomy in the bootstrap topologies sis of hominoids. In fact only a few partitions formed by the South Asian hominoids does of the MMPC (e.g., Homininae, Sivapithecus/ distinguish a Pongo/Sivapithecus clade with Pongo clade) are consistently supported and reasonable support. The lack of resolution of the results are highly sensitive to the choice of a monophyletic Ponginae here is likely also a ingroup taxa. function of incompleteness of Ankarapithecus and Lufengpithecus. These taxa are 46% and Homoplasy Partitioning 38% coded, respectively, whereas Sivapithecus To understand better how homoplasy is dis- and Pongo are 72% and 100% coded, respec- tributed across the hominoid clade, we per- tively. As the incompleteness of a taxon in- formed additional character analyses. The increases, it becomes more likely that resam- group was divided into two subsets of taxa pling techniques, such as the bootstrap, will based upon the split between the ‘‘archaic’’ sample heavily from missing character data and ‘‘modern’’ hominoid clades observed in for that taxon. This leads to an increase in the the MMPC and reevaluated to examine the ambiguity of its placement within and among partitioning of homoplasy across the ingroup. bootstrap replicates, and serves to lower boot- The ‘‘archaic’’ group here included Proconsul 626 JOHN A. FINARELLI AND WILLIAM C. CLYDE with Afropithecus, Turkanapithecus, Equatorius, sy observed when comparing the fit of these Kenyapithecus,andGriphopithecus. The ‘‘mod- character partitions to the MMPC. For in- ern’’ group was composed of the remainder of stance, convergence may be masked by the rel- the ingroup taxa. The ‘‘modern’’ hominoid atively poorer preservation of postcranial subset produced eight most parsimonious characters, or the larger number of postcranial cladograms. The topology of their strict con- characters may implicitly weight the outcome sensus is identical to the analogous partition of a morphology-based analysis, generating a of the strict consensus of the MMPC, except cladogram that is fundamentally biased in fa- Morotopithecus and Hylobates form a polytomy vor of the postcrania. at the base of the consensus cladogram. A sin- To eliminate these potential sources of bias, gle most parsimonious cladogram was recov- a weighted rarefaction analysis was per- ered for the ‘‘archaic’’ subset with a topology formed. Dental character data were randomly identical to that seen in all of the MMPC. deleted from the fossil taxa in the original data Interestingly, the degree of observed ho- matrix until the completeness (proportion of moplasy is markedly different between these cells coded in the data matrix) equaled that of taxonomic partitions. The RI for the ‘‘modern’’ the postcranial character data (29% across fos- hominoid clade is 0.54, whereas the RI for the sil taxa). This procedure was repeated 100 ‘‘archaic’’ hominoid clade is 0.90. The mor- times creating a set of replicate data sets that phological data for the clade of ‘‘archaic’’ exhibit equal preservation in both the dental hominoids is therefore internally consistent, and postcranial character partitions for the indicating that the resolution problem noted fossil taxa in this analysis. Weighting the den- in the consensus and bootstrap analyses for tal characters by 2.62 equalized the relative in- these taxa is not the result of excessive ho- put of the postcranial and dental character moplasy, but rather of highly incomplete pres- partitions (97 vs. 37 characters, respectively) ervation (with the exception of Proconsul). The in determining the resulting most parsimo- elevated rate of homoplasy for the Hominoi- nious cladogram for each replicate. The set of dea in general is apparently confined almost optimal cladograms was then recovered for entirely to the clade of ‘‘modern’’ hominoids. each rarefaction with the branch-and-bound We further examined homoplasy across algorithm, allowing a null hypothesis (H0: no dental and postcranial character partitions to significant difference in level of homoplasy determine the relative proportion of homopla- between dental and postcranial characters ex- sy in these anatomical regions. Average char- ists) to be tested. acter RI values obtained for the set of postcra- Mean character RI scores for the dental and nial skeletal characters (1–97) and dental char- postcranial character blocks were calculated acters (characters 98–134) were calculated separately over the returned cladograms for across the strict consensus of the MMPC. Re- each rarefaction of the dental data. The distri- sults show that postcranial characters (mean bution of rarefaction mean character RIs for RI ϭ 0.75) exhibit lower levels of homoplasy postcranial characters is centered on an aver- than dental characters (mean RI ϭ 0.59). How- age of 0.78, compared with a distribution ever, several aspects of the dental character mean of 0.73 for average dental character RI subset raise concern about simply accepting scores (Fig. 4 top). A paired t-test on the rar- this pattern as a reflection of the degree of ho- efaction data demonstrates that postcranial RI moplasy in the hominoid dentition and po- scores are significantly higher than the corre- stcrania. First, the dental character subset is sponding RI scores for dental characters (t ϭ characterized by more complete preservation. 5.965, p K 0.01; one-tailed). Examination of The dental subset is 90% coded in the data each rarefaction on a case-by-case basis re- matrix, as compared with only 49% for post- veals that 72% of the cases are characterized cranial characters. Second, the dental charac- by higher mean postcranial RI values (Fig. 4 ter data are represented by a smaller number bottom). of characters in the data matrix. Both of these These results support earlier studies on factors may influence the amount of homopla- hominoid data sets, which concluded that DATA CONGRUENCE IN HOMINOID PHYLOGENY 627

FIGURE 4. Top, Histogram of RI values for postcranial and dental characters after rarefaction. The rarefaction analysis controlled for differential preservation and implicit weighting. Arrows plot mean values for each distribution. The mean for the postcranial character distribution is signiﬁcantly higher than the mean for the dental characters (Paired t-test: t ϭ 5.965; p K 0.01). Bottom, Case-by-case comparison of each rarefaction showing higher RI values for the dental characters than the postcranial characters for the majority of cases (72/100).

dental character data may not reliably recover and therefore assumed the correctness of the phylogenetic relationships (Hartman 1988; molecular hypothesis. Regardless of any ar- Collard and Wood 2000). However, these ear- gument for preferring a cladogram derived lier approaches compared the performance of from molecular evidence (Collard and Wood anatomical subsets in reproducing the cladis- 2000: p. 5003), it must remain an assumption tic topologies recovered in molecular studies, that the molecular data recovered ‘‘true’’ hom- 628 JOHN A. FINARELLI AND WILLIAM C. CLYDE

TABLE 2. Total parsimony scores for the phylogenetic hypotheses proposed in this study.

Morphologic Stratigraphic Stratigraphically

tree length parsimony debt augmented tree length RImorph RIstrat Phylogenetic hypothesis MMPC 1 447 37 484 0.69 0.61 MMPC 2 447 35 482 0.69 0.63 MMPC 3 447 37 484 0.69 0.61 MMPC 4 447 38 485 0.69 0.60 Stratocladistic hypothesis 455 25 480 0.68 0.73

inoid phylogenetic relationships. Thus, these than as cladograms, it is always possible (giv- studies are actually comparing the goodness- en the coding scheme for the stratigraphic of-ﬁt of various anatomical character parti- character used here) to hypothesize an ana- tions with a molecular topology (or the con- genetic lineage without ghost ranges (Clyde gruence between the molecular and various and Fisher 1997). Under stratocladistics, when morphological data partitions). This approach cladograms (such as the MMPC) are evaluated is different in that the two anatomical subsets with respect to the stratigraphic character, are equalized with rarefaction and weighting they are actually treated as their isomorphic and then evaluated relative to one another us- phylogenetic trees (identical branching pat- ing a phylogenetic hypothesis that is derived tern and no observed ancestors). Therefore, from their combined input. This analysis pro- when evaluating stratigraphically augmented vides a quantitative measure of relative per- tree lengths, a cladogram must have some formance of postcranial and dental character non-zero stratigraphic debt value associated data subsets under the parsimony criterion with the ranges connecting taxa at terminal and suggests that, when using this optimiza- nodes to hypothetical ancestors at internal tion criterion, postcranial characters are more nodes (Finarelli and Clyde 2002). The RIstrat reliable phylogenetic indicators in the Homi- values for the MMPC (0.60 to 0.63; Table 2) are noidea. higher than the median stratigraphic debt value (0.53) reported in Clyde and Fisher (1997: Incorporation of Stratigraphic Data Table 2) across 29 cladistic analyses in the lit- Stratigraphic Debt in the Morphologically erature. Thus, the hominoids show a better Most Parsimonious Cladograms than average congruence between the stratigraphic order of observed appearance events The coded stratigraphic character was ap- and cladistic branching order. pended to the morphological data matrix and stratigraphically augmented tree lengths (i.e., Partitioning of Stratigraphic Parsimony Debt morphological tree length ϩ stratigraphic parsimony debt) were calculated for the In order to evaluate the partitioning of MMPC. The stratigraphically augmented tree stratigraphic parsimony debt, the ingroup lengths, stratigraphic debt values, and the RI was again divided into the ‘‘archaic’’ and scores relative to the stratigraphic character ‘‘modern’’ hominoid subsets and the strati-

(RIstrat [Clyde and Fisher 1997]) for each of the graphic character was recoded for each subset

MMPCs have been compiled in Table 2. independently. The RIstrat of the most parsi- The maximum possible stratigraphic debt monious cladogram for the ‘‘archaic’’ subset is value for the observed stratigraphic ranges is 0.67, indicating a much better fit of the strati- 94 steps and the minimum possible value is 0, graphic data to this part of the phylogeny than corresponding to an anagenetic lineage in to the strict consensus for the cladograms re- which the ingroup taxa evolve in the order of covered for the ‘‘modern’’ subset (RIstrat ϭ their first appearances. It should be noted that 0.38). This difference in RIstrat values mirrors because the stratigraphic character evaluates the pattern observed for morphology, indicat- all hypotheses as phylogenetic trees rather ing that both the morphological and strati- DATA CONGRUENCE IN HOMINOID PHYLOGENY 629 graphic data for the ‘‘modern’’ hominoids dis- sible to guarantee the return of the most par- play a poorer fit to the proposed cladistic re- simonious phylogenetic tree in a stratocladis- lationships. tic analysis using the ‘‘debt ceiling’’ approach This poorer fit of stratigraphic data to the of Fisher (1992), in practice this is usually not cladistic ordering of the ‘‘modern’’ hominoids feasible because of the large number of poten- is largely due to long ghost lineages for Hy- tial hypotheses and the lack of an automated lobates and Pongo caused by their essentially algorithm. The manual heuristic search of to- nonexistent fossil records. Another problem pologies and ancestor configurations is not lies in the early appearance of Sivapithecus. guaranteed to uncover the most parsimonious The relatively primitive morphology of Hylo- tree(s). However, for a relatively small number bates requires a ghost lineage extending back of taxa such as in this analysis, a researcher to the FAE of Sivapithecus. In addition, signif- can reliably minimize total parsimony debt icant accumulation of stratigraphic debt is in- with the manual search in MacClade (Fisher curred by the implied close relationships of 1992). Ankarapithecus, Lufengpithecus, and especially Following the methodology detailed above, Pongo. Because of the active geological setting an arbitrary cutoff of 455 steps was used in a of the Siwaliks, it is possible that the observed branch-and-bound search, recovering 28,093 FAE of Sivapithecus more closely approximates cladograms. Because the stratigraphic data do its true FAE than is the case for coeval homi- fit the MMPC relatively well and the number noid taxa. Counterintuitively then, a higher of stratigraphic character states is small rela- preservation probability for Sivapithecus,as tive to the total number of morphological compared with its hypothesized close rela- character states, reductions of stratigraphic tives in the MMPC, may contribute to the low debt are minor compared with the corre-

RIstrat observed in the ‘‘modern’’ hominoids. If sponding morphological debt that is incurred. this is true then the assumption of uniform Evaluation by manual branch swapping and preservation probabilities (Fisher 1992; Riep- incorporation of explicit statements of ances- pel and Grande 1994) may be violated for try recovered a single most parsimonious some of the stratigraphic levels deﬁned above. phylogenetic tree (Fig. 5). This tree has a strat- However, to test this rigorously would require igraphically augmented tree length of 480 incorporating quantitative range extensions steps, a morphological tree length of 455 steps that require signiﬁcantly more stratigraphic (RImorph ϭ 0.68), and 25 units of stratigraphic information than is presently available. parsimony debt (RIstrat ϭ 0.73; see Table 2). As in Clyde and Fisher (1997), a substantial in- Stratocladistic Analysis of the Hominoidea crease in the value of RIstrat was observed,

For the stratocladistic analysis, the set of all without a corresponding decrease in RImorph. cladograms with tree lengths below a thresh- This stratocladistic hypothesis shows a pri- old value was imported into MacClade (ver- marily pectinate succession of hominoids sion 3.08a [Maddison and Maddison 1999]). through the early Miocene (Fig. 5). Kenyapithe- As there is currently no automated search al- cus is hypothesized as the ancestor to all ex- gorithm for performing a stratocladistic anal- tant hominoids and their fossil relatives (An- ysis, this threshold value was chosen on the drews and Martin 1987a). This statement of basis of the number of returned cladograms, ancestry involves no increase in the morpho- seeking to maintain manageability of the re- logical debt for the hypothesis. Rather it sim- turned set. The imported trees were evaluated ply reduces the stratigraphic debt relative to relative to total parsimony debt (morpholog- the corresponding hypothesis without an ex- ical ϩ stratigraphic parsimony debt). Manual plicitly hypothesized ancestor. This implies branch swapping was performed to uncover that no autapomorphic character-state transi- additional topologies with lower total debt, tions occur in Kenyapithecus, which is likely a and hypotheses of explicit ancestry were test- function of the original character coding (see ed (Fisher 1992; Clyde and Fisher 1997; Bloch Bloch et al. 2001). In addition, this proposed et al. 2001). Although it is theoretically pos- ancestral position of Kenyapithecus is at odds 630 JOHN A. FINARELLI AND WILLIAM C. CLYDE

FIGURE 5. The most parsimonious phylogenetic tree recovered in the stratocladistic analysis. The stratigraphically augmented tree length of this stratocladistic hypothesis is 480 steps, with 25 steps of stratigraphic parsimony debt

(RImorph ϭ 0.68, RIstrat ϭ 0.73). Kenyapithecus is recognized as an ancestral taxon in this hypothesis. with recent phylogenetic analyses, which gen- Comparison of Morphological and erally consider Kenyapithecus as too primitive Stratocladistic Hypotheses to be a close relative of the extant hominoids As stratocladistic hypotheses are phyloge- (Begun et al. 1997a; McCrossin and Beneﬁt netic trees, it is desirable to compare the to- 1997), although these studies did incorporate ‘‘Kenyapithecus’’ prior to its partition into the pologies of the MMPC with the cladistic to- two OTUs, Equatorius and Kenyapithecus, used pology that is consistent with the phylogenet- in this study (Ward et al. 1999). Given the in- ic tree recovered by the stratocladistic analy- completeness of coding for Kenyapithecus, sup- sis. Such a topological comparison between port for this explicit statement of ancestry is the cladistic and stratocladistic hypotheses re- rather weak. However, it is important to note veals several important differences (Fig. 6). that the array of ‘‘archaic’’ taxa in a pectinate Most signiﬁcantly, the division of the Homi- manner is strongly supported by the strati- noidea into ‘‘archaic’’ and ‘‘modern’’ clades is graphic data, as it minimizes the amount of not recovered in the stratocladistic analysis. stratigraphic debt incurred by the hypothesis Although Afropithecus and Turkanapithecus re- by eliminating the long ghost lineage implied main sister taxa, the remainder of the ‘‘archa- by Morotopithecus in the MMPC. ic’’ taxa are arranged in pectinate succession DATA CONGRUENCE IN HOMINOID PHYLOGENY 631

FIGURE 6. Comparison of the MMPC (left) and the cladogram associated with the stratocladistic hypothesis. Al- though Afropithecus and Turkanapithecus remain grouped as sister taxa in the stratocladistic hypothesis, the remainder of the ‘‘archaic’’ clade is not recognized. These taxa are arranged in a pectinate manner, eliminating the long ghost lineage implied by the position of Morotopithecus in the MMPC. Two arrows point to taxa whose relative positions in the cladogram are signiﬁcantly changed. Morotopithecus (an early-appearing hominoid with derived postcranial morphology) is displaced baseward in the stratocladistic hypothesis, whereas the late-appearing taxon (Oreopithecus) is displaced crownward. See text for discussion. following their order of appearance, essen- acters in the data matrix. Closer examination tially grafting the ‘‘archaic’’ hominoids into revealed that the repositioning of Morotopithe- the implied gap between the appearances of cus baseward and the inverted relative posi- Morotopithecus and Sivapithecus. tions of Dryopithecus and Ouranopithecus facil- The displacement of the early-appearing itated the repositioning of Oreopithecus in the Morotopithecus baseward in the stratocladistic stratocladistic hypothesis. Without both of hypothesis is accompanied by the crownward these movements, displacing Oreopithecus movement of the late-appearing Oreopithecus. crownward is associated with large increases The savings in stratigraphic parsimony debt in morphological debt that are not compen- associated with these movements are obvious. sated for by the decreases in the stratigraphic In the case of Morotopithecus (27.5% coded in debt. Therefore, the position of other taxa in- the data matrix), incomplete coding of mor- cluding Morotopithecus has a large impact on phology makes its position relatively poorly the pattern of character polarity, making the constrained. Although incomplete fossil pres- switch in the relative position of Ouranopithe- ervation facilitates uncertainty in the phylo- cus and Dryopithecus possible, which in turn genetic position of some taxa it is not a nec- allows the crownward movement of Oreopi- essary condition for it. For instance, the dis- thecus to dramatically lower stratigraphic placement of Oreopithecus (in this case crown- debt. ward) eliminates the stratigraphic debt The change in the relative position of Dry- associated with its long ghost lineage in the opithecus and Ouranopithecus highlights anoth- MMPC (Fig. 7D). Unlike Morotopithecus how- er signiﬁcant difference between the MMPC ever, Oreopithecus is coded for 70% of the char- and the stratocladistic hypothesis—the pon- 632 JOHN A. FINARELLI AND WILLIAM C. CLYDE

FIGURE 7. A, Graphic representation of the stratigraphic debt incurred upon the strict consensus of the MMPC. B, Graphic representation of the stratigraphic debt incurred upon the stratocladistic hypothesis. Note that Kenyapi- thecus is recognized as an ancestral taxon in this hypothesis. As in Figure 1, the stippled region represents a sampling gap that is not coded in the stratigraphic character. C, Arrow points to the removal of large amounts of stratigraphic debt by displacing Morotopithecus baseward in the stratocladistic hypothesis and thus eliminating distinct DATA CONGRUENCE IN HOMINOID PHYLOGENY 633 gines. Associated with the repositioning of cene of East Africa. The position of Morotopi- these two European hominoids is the dis- thecus as the sister taxon to all extant and late placement of the pongine clade, such that Dry- middle to late Miocene Eurasian hominoids in opithecus is consistently derived relative to the the MMPC implies a large gap in this group’s South Asian hominoids (Fig. 6). The strato- fossil record spanning the early to middle cladistic hypothesis produced a topology sim- Miocene of East Africa. Numerous hominoid ilar to the results of a cladistic analysis per- localities exist in this region that span this pe- formed by Begun (2002) in this respect. How- riod of time (approximately 21–14 Ma). The ever, in that analysis Lufengpithecus and An- long ghost lineage in the morphological clad- karapithecus remained in a monophyletic ogram is eliminated in the stratocladistic hy- pongine clade with Pongo and Sivapithecus (as pothesis by baseward displacement of Moro- in the MMPC of this study). In contrast, al- topithecus and elimination of distinct clades of though Sivapithecus and Pongo remain united hominoids during this interval. The MMPC as sister taxa, the positions of both Lufengpi- accrues stratigraphic parsimony debt by im- thecus and Ankarapithecus change, placing plying a long missing lineage of hominoids in them as sister taxa to European and African a region that has been well sampled and from hominoids in the stratocladistic hypothesis which many hominoid localities are known. If (Fig. 6). Unfortunately, Begun (2002) only in- the morphological hypothesis is correct, then cluded Griphopithecus and Afropithecus from the ‘‘modern’’ lineage is completely unrepre- the ‘‘archaic’’ clade, and therefore that analy- sented in the fossil record, despite the pres- sis does not address the question of support ence of fossil localities in the geographic re- for distinct early Miocene hominoid lineages. gion and in strata of the correct age where It is important to note here that the strato- these hominoids should exist. If the stratocladistic hypothesis is not simply arranging cladistic hypothesis is correct, then the de- the ingroup taxa by their order of appearance. rived postcranial morphology that Morotopi- For instance, the relative positions of Pongo thecus shares with later hominoids of ‘‘modern and Hylobates are not altered between the two aspect’’ would be homoplasious. Although it hypotheses (Figs. 6, 7) even though these hy- was demonstrated that the ‘‘archaic’’ clade pothesized positions imply long ghost line- formed a monophyletic group unto itself even ages. Repositioning either taxon crownward upon the exclusion of Morotopithecus, the mor- significantly lowers stratigraphic debt, but phological support of a distinct clade of ‘‘ar- any such movement simultaneously creates chaic’’ hominoids is weak (low Bremer sup- large amounts of morphological debt, and the port and low bootstrap proportions). We favor total parsimony criterion rejects these hypoth- the stratocladistic result here, as it is unlikely eses. that an entire lineage of early Miocene homi- The topological differences between the cla- noids is completely unrepresented in the rel- distic and stratocladistic hypotheses highlight atively good fossil record of the early Miocene two intervals in the fossil record of the Hom- of eastern Africa. New fossils from this region inoidea where significant incongruence exists will help to resolve this dispute not only be- between the morphological and temporal cause they would increase the chance of re- data. The first interval occurs in the early Mio- covering such a lineage, but also because any

← clades of ‘‘archaic’’ and ‘‘modern’’ hominoids. The previously recognized clade of ‘‘archaic’’ hominoids is now essentially grafted between the early appearance of Morotopithecus and the later FAE of Sivapithecus. D, Arrow points to the elimination of stratigraphic debt by the displacement of Oreopithecus crownward. These differences in topology highlight areas of inconsistency between stratigraphic and morphological data. For instance, either there is a lineage of ‘‘modern’’ hominoids unrepresented in the fossil record of eastern Africa, despite productive fossil localities in the region, or the morphology observed in Morotopithecus linking this taxon to a clade of ‘‘modern’’ hominoids is homoplasious. Note also that the long ghost lineages between Hylobates and Pongo and their sister groups persist in the stratocladistic hypothesis. These are not overturned by stratigraphic information because of the strong morphological evidence for their respective positions. 634 JOHN A. FINARELLI AND WILLIAM C. CLYDE new morphological information (especially in 1. The morphological evidence points to the the postcrania of such taxa as Morotopithecus presence of two distinct evolutionary line- and Kenyapithecus) should bear heavily on ages of hominoids in the earliest Miocene, phylogenetic hypotheses. a postcranially primitive clade of ‘‘archaic’’ The second interval of incongruence be- hominoids including early Miocene taxa tween morphology and stratigraphy occurs in from eastern Africa, and a postcranially the late Miocene across Eurasia. For instance, derived clade of ‘‘modern’’ hominoids that Oreopithecus is displaced crownward in the includes middle to late Miocene Eurasian cladogram, removing stratigraphic parsimo- hominoids and all extant taxa. Morotopithe- ny debt incurred by the MMPC. This late-ap- cus is positioned at the base of the ‘‘mod- pearing taxon is joined at the base of the hom- ern’’ hominoid clade consistent with the inid clade in the MMPC (Fig. 7A). If the mor- hypothesis of MacLatchy et al. (2000). How- phological hypothesis is correct, then a large ever, the support for the separation of ‘‘ar- sampling gap exists during this interval for chaic’’ hominoids from ‘‘modern’’ homi- the lineage leading to Oreopithecus. There is noids is shown to be rather weak by both some evidence to believe that this may be the consensus and bootstrap methods. Evalu- case. Oreopithecus is associated with a fauna ation of reduced consensus cladograms that is a unique mixture of African and Eu- and several bootstrapping analyses indi- ropean taxa, and the localities from which this cate that the cladistic hypothesis is very taxon is known are thought to sample the rem- sensitive to choice of ingroup taxa and that nant of a now-submerged land corridor be- the poor preservation of a few taxa (e.g., tween northern African and southern Europe Kenyapithecus) is responsible for much of the (Andrews et al. 1996; Harrison and Rook observed cladistic uncertainty. 1997). On the other hand, if the stratocladistic 2. The distribution of homoplasy across tax- result is correct, then Oreopithecus represents the sister group to the Homininae. This im- onomic and character partitions is highly plies a European origin for Oreopithecus (An- asymmetric. The ‘‘archaic’’ hominoid clade drews et al. 1996), and a southward migration displays very little homoplasy compared for the hominines into Africa, with Oreopithe- with the ‘‘modern’’ hominoid clade. This cus representing an island isolate that re- suggests that the elevated level of homopla- mained on the land bridge. The highly di- sy observed in the Hominoidea is confined verged morphology of Oreopithecus presents a largely to the clade of ‘‘modern’’ homi- problem in the evaluation of this hypothesis. noids. In addition, the level of homoplasy Oreopithecus is placed at the base of all great observed in dental characters relative to the apes in the MMPC, sharing only primitive MMPC is significantly higher than that ob- characters with the remainder of the Homin- served for postcranial characters. This re- idae, and bootstrap analyses could find no sult was upheld even when the analysis strong support for resolving the polytomy at was performed on a set of randomly gen- the base of this clade. We suspect that the in- erated rarefactions that were created to congruence in this case results from the un- equalize both preservation and implicit usual stratigraphic setting of Oreopithecus, weighting of the two character partitions. which may greatly underestimate the FAE of Postcranial characters appear to be the this lineage compared with coeval hominoid more reliable indicators of hominoid phy- taxa, in particular Sivapithecus, which likely logeny under the parsimony criterion, in- has a higher preservation potential given the dicating that they should become a focus of geologic setting in which the fossil material future collection efforts. has been recovered, and therefore likely has a 3. A better-than-average match is observed more accurate estimate of its FAE. between the ordering of hominoid FAEs Conclusions and the branching order in the MMPC, in- Several important conclusions can be drawn dicating that the hominoid fossil record is from the phylogenetic analyses presented relatively good. Most of the observed mis- here. match is within the group of ‘‘modern’’ DATA CONGRUENCE IN HOMINOID PHYLOGENY 635

hominoids. However, when stratigraphy is Sciences. This research was part of a Master’s incorporated into the analysis there are sig- Thesis that J.A.F. submitted to the Graduate nificant topological differences between School at the University of New Hampshire. the stratocladistic hypothesis and the MMPC. These differences highlight two in- Literature Cited tervals in the stratigraphic record where Agustı´, J., L. Cabrera, and M. Garce´s. 1998. Age of late Miocene significant incongruence exists between the hominoids from Europe. Phylogeny of European Neogene hominoid primates. Programme and Abstracts European Sci- morphological and stratigraphic data for ence Foundation: Network on Hominoid Evolution and En- the Hominoidea. The first interval in the vironmental Change in the Neogene of Europe 3rd Workshop. early Miocene of eastern Africa is the result Agustı´, J., L. Cabrera, M. Garce´s, W. Krijgsman, O. Oms, and J. M. Pares. 2001. A calibrated mammal scale for the Neogene of the morphological support for distinct of Western Europe: state of the art. Paleogeography, Paleocli- clades of ‘‘archaic’’ and ‘‘modern’’ homi- matology, and Paleoecology 52:247–260. Aiello, L., and M. Collard. 2001. Our newest oldest ancestor? noids. The early appearance of Morotopithe- Nature 410:526–527. cus implies an unobserved lineage of Allbrook, D., and W. W. Bishop. 1963. New fossil hominoid ma- ‘‘modern’’ hominoids spanning an interval terial from Uganda. Nature 197:1187–1190. Alpagut, B., P. Andrews, and L. Martin. 1990. A new hominoid that is otherwise characterized by good in- species from Pas¸alar, Turkey. Journal of Human Evolution 19: group preservation. In light of the rather 397–422. weak morphological support for the posi- Alpagut, B., P. Andrews, M. Fortelius, J. Kappelman, I. Tem, H. Celebi, and W. Lindsay. 1996. A new specimen of Ankarapi- tion of Morotopithecus, the reality of two dis- thecus meteai from the Sinap Formation of central Anatolia. tinct evolutionary lineages in the early Nature 382:349–351. Andrews, P. 1978. A revision of the Miocene Hominoidea of East Miocene is questionable. Rather it is likely Africa. Bulletin of the British Museum of Natural History (Ge- that the taxa in the ‘‘archaic’’ group here ology) 30:85–224. represent a stem lineage of early homi- ———. 1981. The Miocene fossil beds of Maboko island, Kenya: geology, age, taphonomy and palaeontology. Journal of Hu- noids. Additional material (especially po- man Evolution 10:35–48. stcrania) will certainly help to resolve the ———. 1992. Evolution and environment in the Hominoidea. evolutionary relationships of these early Nature 360:641–646. Andrews, P., and B. Alpalgut. 1990. Description of the fossilif- hominoid taxa. The second interval in the erous units at Pas¸alar, Turkey. Journal of Human Evolution 19: late Miocene of Eurasia is highlighted by 343–361. Andrews, P., and L. Martin. 1987a. Cladistic relationships of ex- the variable position of Oreopithecus. In this tant and fossil hominoids. Journal of Human Evolution 16: case, its morphology is well known but it 101–118. comes from a unique stratigraphic setting ———. 1987b. The phylogenetic position of the Ad Dabtiyah hominoid. Bulletin of the British Museum of Natural History in southern coastal Europe that makes the (Geology) 41:383–393. temporal range of Oreopithecus and/or its Andrews, P., and I. Tekkaya. 1980. A revision of the Turkish Miocene hominoid Sivapithecus meteai. Palaeontology 23:85– immediate ancestors poorly resolved com- 95. pared to coeval hominoid taxa. Andrews, P., T. Harrison, L. Martin, and M. Pickford. 1981. Hominoid primates from a new Miocene locality named Mes- Acknowledgments wa Bridge in Kenya. Journal of Human Evolution 10:123–128. Andrews, P., T. Harrison, E. Delson, L. Bernor, and L. Martin. We would like to thank W. Sanders for help- 1996. Distribution and biochronology of European and south- ful insights in the course of our analysis, and west Asian Miocene catarrhines. Pp. 168–206 in Bernor et al. 1996. D. Fisher and D. Begun for thoughtful and Anyonge, W. 1991. Fauna from a new Lower Miocene locality critical reviews of our manuscript, which have west of Lake Turkana, Kenya. Journal of Vertebrate Paleon- tology 11:378–390. greatly improved the quality of this work. In Azzaroli, A., M. Boccaletti, E. Delson, G. Moratti, and D. Torre. addition, the work presented here would not 1986. Chronological and paleogeographical background to have been possible without the support of R. the study of Oreopithecus bambolii. Journal of Human Evolu- tion 15:533–540. Potts and J. Clark at the Human Origins Pro- Badgley, C., G. Qi, W. Chen, and D. Han. 1988. Paleoecology of gram, Smithsonian Institution and A. Rosen- a Miocene, tropical upland fauna: Lufeng, China. National berger then at the Smithsonian Institution. Geographic Research 4:178–195. Begun, D. R. 1992a. Dryopithecus crusafonti sp.nov.AnewMio- Partial funding for this work was provided by cene hominoid species from Can Ponsic (northeastern Spain). the University of New Hampshire Graduate American Journal of Physical Anthropology 87:291–309. ———. 1992b. Miocene hominoids and the chimp-human clade. Teaching Assistant Summer Fellowship and Science 257:1929–1933. the Jonathon W. Herndon Scholarship in Earth ———. 1992c. Phyletic diversity and locomotion in European 636 JOHN A. FINARELLI AND WILLIAM C. CLYDE

hominoids. American Journal of Physical Anthropology 87: inid from the Upper Miocene of Chad, Central Africa. Nature 311–340. 418:145–151. ———. 1994. Relationships among the great apes and humans: Cande, S. C., and D. V. Kent. 1995. Revised calibration of the new interpretations based on the fossil great ape Dryopithecus. geomagnetic polarity timescale for the Late Cretaceous and Yearbook of Physical Anthropology 37:11–63. Cenozoic. Journal of Geophysical Research 100:6093–6095. ———. 1995. Late Miocene European orang-utan, gorilla, hu- Clyde, W. C., and D. C. Fisher. 1997. Comparing the fit of strati- man or none of the above? Journal of Human Evolution 29: graphic and morphologic data in phylogenetic analysis. Pa- 169–180. leobiology 23:1–19. ———. 2000. Technical comments: Middle Miocene hominoid Collard, M., and B. Wood. 2000. How reliable are human phy- origins. Science 287, http://www.sciencemag.org/cgi/ logenetic hypotheses? Proceedings of the National Academy content/full/287–5462/ 2375a. of Sciences USA 97:5003–5006. ———. 2002. European hominoids. Pp. 339–368 in W. C. Har- Dean, D., and E. Delson. 1992. Second gorilla or third chimp? twig, ed. The primate fossil record. Cambridge University Nature 359:676–677. Press, Cambridge. de Bonis, L. de, and G. D. Koufos. 1993. The face and the man- Begun, D. R., and E. Gu¨leç. 1995. Restoration and reinterpreta- dible of Ouranopithecus macedoniensis: description of new spec- tion of the facial specimen attributed to Sivapithecus meteai imens and comparisons. Journal of Human Evolution 24:469– from Kayincak (Yassyó¨ren), Central Anatolia, Turkey. Amer- 491. ican Journal of Physical Anthropology 20(Suppl.):63–64. Deino, A., L. Tauxe, M. Monaghan, and R. Drake. 1990. 40 AR/ ———. 1998. Restoration of the type and palate of Ankarapithecus 39 AR age calibration of the litho- and paleomagnetic stratig- meteai: taxonomic and phylogenetic implications. American raphies of the Ngorora Formation, Kenya. Journal of Geology Journal of Physical Anthropology 105:279–314. 98:567–587. Begun, D. R., and L. Kordos. 1993. Revision of Dryopithecus bran- Delson, E., and F. Szalay. 1985. Reconstruction of the 1958 cra- coi SCHLOSSER, 1901 based on the fossil hominoid from Ru- nium of Oreopithecus bambolii and its comparison to other cat- dabanya. Journal of Human Evolution 25:271–285. arrhines. American Journal of Physical Anthropology 66:163. ———. 1997. Phyletic affinities and functional convergence in Drake, R. L., J. A. VanCouvering, M. Pickford, G. H. Curtis, and Dryopithecus and other Miocene and living hominids. Pp. 291– J. A. Harris. 1988. New chronology for early Miocene mam- 317 in Begun et al. 1997b. malian faunas of Kisingiri, western Kenya. Journal of the Geo- Begun, D. R., S. Moya`-Sola`, and M. Kohler. 1990. New Miocene logical Society, London 145:479–491. hominoid specimens from Can Llobateres (Valles Penedes, Eriksson, T. 2001. AutoDecay, Version 5.0 (program distributed Spain) and their geological and paleoecological context. Jour- by author). Bergius Foundation, Royal Swedish Academy of nal of Human Evolution 19:255–268. Sciences, Stockholm. Begun, D. R., C. V. Ward, and M. D. Rose. 1997a. Events in hom- Felsenstein, J. 1985. Confidence limits on phylogenies: an ap- inoid evolution. Pp. 389–416 in Begun et al. 1997b. proach using the bootstrap. Evolution 39:783–791. ———. 1997b. Function, phylogeny, and fossils: Miocene Hom- Finarelli, J. A., and W. C. Clyde. 2002. Comparing the gap excess inoid evolution and adaptation. Plenum, New York. ratio and the retention index of the stratigraphic character. Behrensmeyer, A. K., A. L. Deino, A. Hill, J. D. Kingston, and J. Systematic Biology 51:166–176. J. Saunders. 2002. Geology and geochronology of the middle Fisher, D. C. 1991. Phylogenetic analysis and its application in Miocene Kipsaramon site complex, Muruyur Beds, Tugen evolutionary analysis. In N. L. Gilinsky and P. W. Signor, eds. Hills. Kenya Journal of Human Evolution 42:11–38. Analytical paleobiology. Short courses in paleontology 4:103– Bernor, R. L., and H. Tobien. 1990. Mammalian geochronology 121. Paleontological Society, Knoxville, Tenn. and biogeography of Pas¸alar (Middle Miocene, Turkey). Jour- ———. 1992. Stratigraphic parsimony. Pp. 124–129 in W. P. Mad - nal of Human Evolution 19:551–568. dison and D. R. Maddison, eds. MacClade: analysis of phy- Bernor, R. L., V. Fahlbusch, and H.-W. Mittmann, eds. 1996. The logeny and character evolution, version 3. Sinauer, Sunder- evolution of western Eurasian Neogene mammal faunas. Co- land, Mass. lumbia University Press, New York. ———. 1994. Stratocladistics: morphological and temporal pat- Bestland, E. 1990. Sedimentology and paleopedology of Mio- terns and their relation to phylogenetic process. Pp. 133–171 cene alluvial deposits at the Pas¸alar hominoid site, western in L. Grande and O. Rieppel, eds. Investigating the hierarchy Turkey. Journal of Human Evolution 19:363–378. of nature. Academic Press, New York. Bishop, W. W., J. A. Miller, and F. J. Fitch. 1969. New Potassium- Fox, D. L., D. C. Fisher, and L. R. Leighton. 1999. Reconstructing Argon age determinations relevant to the Miocene fossil se- phylogeny with and without temporal data. Science 284:1816– quence of East Africa. American Journal of Science 267:669– 1819. 699. Gebo, D., L. MacLatchy, R. Kityo, A. Deino, J. Kingston, and D. Bloch, J. I., D. C. Fisher, K. D. Rose, and P. D. Gingerich. 2001. Pilbeam. 1997. A hominoid genus from the early Miocene of Stratocladistic analysis of Paleocene Carpolestidae (Mam- Uganda. Science 276:401–404. malia, Plesiadapiformes) with description of a new late Tif- Gentry, A. W. 1987a. Mastodons from the Miocene of Saudi Ara- fanian genus. Journal of Vertebrate Paleontology 21:119–131. bia. Bulletin of the British Museum of Natural History (Ge- Bremer, K. 1988. The limits of amino acid sequence data in an- ology) 41:395–407. giosperm phylogenetic reconstruction. Evolution 42:795–803. ———. 1987b. Rhinoceroses from the Miocene of Saudi Arabia. Brochetto, H. B., F. H. Brown, and I. McDougall. 1992. Stratig- Bulletin of the British Museum of Natural History (Geology) raphy of the Lothidok Range, northern Kenya and K/Ar ages 41:409–432. of its Miocene primates. Journal of Human Evolution 22:47– Harrison, T. 1987. The phylogenetic relationships of the early 71. catarrhine primates: a review of the current evidence. Journal Brunet, M., F. Guy, D. Pilbeam, H. T. Mackaye, A. Likius, D. of Human Evolution 16:41–80. Ahounta, A. Beauvilain, C. Blondel, H. Bocherens, J. Boisserie, Harrison, T., and L. Rook. 1997. Enigmatic anthropoid of mis- L. de Bonis, Y. Coppens, J. Dejax, C. Denys, P. Duringer, V. understood ape? The phyletic affinities of Oreopithecus bam- Eisenmann, F. Fanone, P. Fronty, D. Geraads, T. Lehmann, F. bolii revisited. Pp. 327–362 in Begun et al. 1997b. Lihoreau, A. Louchart, A. Mahamat, G. Merceron, G. Mouch- Harrison, T., and W. Sanders. 1999. Scaling of lumbar vertebrae elin, O. Otero, P. P. Campomanes, M. P. De Leon, J. Rage, M. in anthropoid primates: its implications for positional behav- Sapanet, M. Schuster, J. Sudre, P. Tassy, X. Valentin, P. Vig- ior and phylogenetic affinities of Proconsul. American Journal naud, L. Viriot, A. Zazzo, and C. Zollikofer. 2002. A new hom- of Physical Anthropology 28(Suppl.):146. DATA CONGRUENCE IN HOMINOID PHYLOGENY 637

Harrison, T., X. P. Ji, and D. Su. 2002. On the systematic status Dam, A. J. van der Meulen, J. Agustı´, and L. Cabrera. 1996. A of the late Neogene hominoids from Yunnan Province, China. new chronology for the middle to late Miocene continental re- Journal of Human Evolution 43:207–227. cord in Spain. Earth and Planetary Science Letters 142:367– Hartman, S. E. 1988. A cladistic analysis of hominoid molars. 380. Journal of Human Evolution 17:489–502. Leakey, M. G., and A. Walker. 1997. Afropithecus: function and Heizmann, E. J. P., and D. R. Begun. 2001. The oldest Eurasian phylogeny. Pp. 225–240 in Begun et al. 1997b. hominoid. Journal of Human Evolution 41:463–481. Leakey, M. G., C. S. Feibel, I. McDougall, and A. Walker. 1995a. Hill, A., R. Drake, L. Tauxe, M. Monaghan, J. C. Barry, A. K. Beh- New four-million-year-old hominid species from Kanapoi, rensmeyer, G. Curtis, B. Fine-Jacobs, L. Jacobs, N. Johnson, and Allia, Bay Kenya. Nature 376:565–571. and D. Pilbeam. 1985. Neogene paleontology and geochro- Leakey, M. G., P. S. Ungar, and A. Walker. 1995b. A new genus nology of the Baringo Basin, Kenya. Journal of Human Evo- of large primate from the late Oligocene of Lothidok, Turkana lution 14:749–773. District, Kenya. Journal of Human Evolution 28:519–531. Hill, A., M. Leakey, J. D. Kingston, and S. Ward. 2002. New cer- Leakey, M. G., C. S. Feibel, I. McDougall, and A. Walker. 1999. copithecoids and a hominoid from 12.5 Ma in the Tugen Hills New specimens and confirmation of an early age for Austral- succession, Kenya. Journal of Human Evolution 42:75–93. opithecus anamensis. Nature 393:62–66. Huelsenbeck, J. P. 1994. Comparing the stratigraphic record to Leakey, R. E. F., and M. G. Leakey. 1986a. A new Miocene hom- estimates of phylogeny. Paleobiology 20:470–483. inoid from Kenya. Nature 324:143–146. Huerzeler, J., and B. Engesser. 1976. Les faunes mammife`res ———. 1986b. A second new Miocene hominoid from Kenya. neóge`ne du Bassin de Baccinello (Grosseto, Italie). Comptes Nature 324:146–148. Rendus de l’Acade´mie des Sciences de Paris, se´rie III, 296:497– Leakey, R. E. F., and A. Walker. 1985. New higher primates from 503. the early Miocene of Buluk, Kenya. Nature 318:173–175. Ishida, H., Y. Kunimatsu, M. Nakatsukasa, and Y. Nakano. 1999. Leakey, R. E. F., M. G. Leakey, and A. Walker. 1988. The mor- New hominoid genus from the middle Miocene of Nachola, phology of Turkanapithecus kalakolensis from Kenya. Ameri- Kenya. Anthropological Science 107:189–191. can Journal of Physical Anthropology 76:277–288. Johnson, G. D., P. Zeitler, C. W. Naeser, N. M. Johnson, D. M. MacLatchy, L., D. Gebo, R. Kityo, and D. Pilbeam. 2000. Post- Summers, C. D. Frost, N. D. Opdyke, and R. A. K. Tahirkheli. cranial functional morphology of Morotopithecus bishopi, with 1982a. The occurrence and fission track ages of late Neogene implications for the evolution of modern ape locomotion. volcanic sediments, Siwalik Group, northern Pakistan. Paleo- Journal of Human Evolution 39:159–183. geography, Paleoclimatology, and Paleoecology 37:63–93. Madar, S. I., M. D. Rose, J. Kelley, L. MacLatchy, and D. Pilbeam. Johnson, G. D., N. D. Opdyke, S. K. Tandon, and A. C. Nanda. 2002. New Sivapithecus postcranial specimens from the Si- 1983. The magnetic polarity stratigraphy of the Siwalik Group waliks of Pakistan. Journal of Human Evolution 42:705–752. at Haritalyangar (India) and a last appearance datum for Ra- Maddison, M. J., and D. R. Maddison. 1999. MacClade: inter- mapithecus and Sivapithecus. Paleogeography, Paleoclimatolo- active analysis of phylogeny and character evolution, Version gy, and Paleoecology 44:223–249. 3.08a. Sinauer, Sunderland, Mass. Johnson, N. M., N. D. Opdyke, G. D. Johnson, E. H. Lindsay, and McCrossin, M. L., and B. R. Benefit. 1997. On the relationships R. A. K. Tahirkheli. 1982. Magnetic polarity stratigraphy and and adaptations of Kenyapithecus a large-bodied hominoid ages of Siwalik Group rocks of the Potwar Plateau, Pakistan. from the Middle Miocene of eastern Africa. Pp. 241–268 in Be- Paleogeography, Paleoclimatology, and Paleoecology 37:17– 42. gun et al. 1997b. Jungers, W. L. 1988. Relative joint size and hominoid locomo- Moya`-Sola`, S., and M. Ko¨hler. 1993. Recent discoveries of Dry- tory adaptations with implications for the evolution of biped- opithecus shed new light on evolution of great-apes. Nature alism. Journal of Human Evolution 17:247–265. 365:543–545. Kappelman, J., J. Kelly, D. Pilbeam, K. A. Sheikh, M. A. Anwar, Nakatsukasa, M., A. Yamanaka, Y. Kunimatsa, D. Shimizu, and J. C. Barry, B. Brown, P. Hak, N. M. Johnson, S. M. Raza, and H. Ishida. 1998. A newly discovered Kenyapithecus skeleton S. M. I. Shah. 1991. The earliest occurrence of Sivapithecus and its implications for the evolution of postural behavior in from the middle Miocene Chinji Formation of Pakistan. Jour- Miocene East African hominoids. Journal of Human Evolution nal of Human Evolution 21:61–73. 34:657–664. Kelley, J., and D. Pilbeam. 1986. The dryopithecines: taxonomy, Norell, M. 1993. Tree-based approaches to understanding his- comparative anatomy, and phylogeny of Miocene large hom- tory: Comments on ranks, rules and the quality of the fossil inoids. Pp. 361–411 in D. R. Swindler and J. Irwin, eds. Com- record. American Journal of Science 293-A:407–417. parative primate biology, Vol. 1. Systematics, evolution, and Opdyke, N. D., N. M. Johnson, G. D. Johnson, E. H. Lindsay, and anatomy. Liss, New York. R. A. K. Tahirkheli. 1982. Paleomagnetism of the Middle Si- Kelley, J., and J. M. Plavcan. 1998. A simulation test of hominoid walik formations of northern Pakistan, and rotation of the Salt species number at Lufeng, China: implications for the use of Range decollement. Paleogeography, Paleoclimatology, and the coefficient of variation in paleotaxonomy. Journal of Hu- Paleoecology 37:1–15. man Evolution 35:572–596. Ozansoy, F. 1965. E´ tude des gisent continenteaux et des mam- Kelley, J., S. Ward, B. Brown, A. Hill, and D. L. Duren. 2002. Den- mife`res du Cenezoı¨que de Turquie. Me´moires de la Socie´te´ tal remains of Equatorius africanus from Kipsaramon, Tugen Geologique de France 44:5–89. Hills, Baringo District, Kenya. Journal of Human Evolution 42: Patnaik, R., and D. Cameron. 1997. New Miocene fossil ape lo- 39–62. cality, Dangar, Hari-Talyangar region, Siwaliks, Northern In- Kondopulou, D., S. Sen, G. Koufos, and L. de Bonis. 1992. Mag- dia. Journal of Human Evolution 32:93–97. neto and biostratigraphy of the Late Miocene mammalian lo- Pelaéz-Campomanes, P., and R. Daams. 2002. Middle Miocene calities of Prochome (Macedonia, Greece). Paleontologica i rodents from Pas¸alar, Anatolia, Turkey. Acta Palaeontologica Evolucion (Barcelona) 24– 25:138–139. Polonica 47:125–132. Kordos, L. 2000. New results of Hominoid research in the Car- Pickford, M. 1984. Kenya paleontology gazetteer, Vol. 1. Western pathian Basin. Acta Biologica Szegediensis 44:71–74. Kenya. National Museums of Kenya, Department of Sites and Koufos, G. D. 1995. The first female maxilla of the hominoid Monuments Documentation, Nairobi. Ouranopithecus macedoniensis from the Late Miocene of Mac- ———. 1985. A new look at Kenyapithecus based on recent dis- edonia, Greece. Journal of Human Evolution 29:385–399. coveries in western Kenya. Journal of Human Evolution 14: Krijgsman, W., M. Garce´s, G. C. Langereis, R. Daams, J. van 113–143. 638 JOHN A. FINARELLI AND WILLIAM C. CLYDE

Pickford, M., and P. Andrews. 1981. The Tinderet Miocene se- Mountain roots to mountain tops. In A. Macfarlene, R. B. quence in Kenya. Journal of Human Evolution 10:11–33. Sorkhabi, and J. Quade, eds. Himalaya and Tibet: mountain Pickford, M., B. Senut, D. Hadoto, J. Musisi, and C. Kariira. 1986. roots to mountain tops. Geological Society of America Special Dećouvertes recentes dans le sites mioce`nes de Moroto (Oug- Paper 328:1–7. anda Oriental): aspects biostratigraphiques et paleoecolo- Steininger, F. F., W. A. Berggren, D. V. Kent, R. L. Bernor, S. Sen, giques. Comptes Rendus de l’Acade´mie des Sciences, se´rie II, and J. Agustı´. 1996. Circum Mediterranean Neogene (Mio- 302:681–686. cene and Pliocene) marine and continental chronological cor- Pilbeam, D. 1997. Research on Miocene hominoids and hominid relations of European Mammal Units. Pp. 7–46 in Bernor et al. origins: the last three decades. Pp. 13–28 in Begun et al. 1997b. 1996. Pilbeam, D., M. Morgan, J. C. Barry, and L. Flynn. 1996. Euro- Steininger, F. F. 1999. Chronostratigraphy, geochronology and pean MN Units and the Siwalik faunal sequence of Pakistan. biochronology of the Miocene ‘‘European Land Mammal Pp. 96–105 in Bernor et al. 1996. Mega Zones’’ (ELMMZ) and the Miocene ‘‘Mammal Zones’’ Rae, T. C. 1997. The early evolution of the hominoid face. Pp. 59– (MN-Zones). Pp. 9–24 in G. E. Rossner and K. Hessig, eds. The 78 in Begun et al. 1997b. Miocene land mammals of Europe. Pfeil, Munich. Rae, T. C., and T. Kopp. 2000. Isometric scaling of maxillary si- Swofford, D. L. 2002. PAUP*. Phylogenetic Analysis Using Par- nus volume in hominoids. Journal of Human Evolution 38: simony (*and Other Methods), Version 4. Sinauer, Sunder- 411–423. land, Mass. Raza, S. M., J. C. Barry, D. Pilbeam, M. D. Rose, S. M. I. Shah, Tassy, P., and M. Pickford. 1983. Un nouveau Mastodonte zyg- and S. C. Ward. 1983. New hominoid primates from the Mid- olophodonte (Proboscidea, Mammalia) dans le Mioce`ne In- dle Miocene Chinji Formation, Potwar Plateau, Pakistan. Na- fe´rieur d’Afrique Orientale: Syste´matique et Paleóenviron- ture 406:52–54. nement. Geobios 16:53–77. Retallack, G. J. 1991. Miocene paleosols and Ape Habitats of Tattersall, I., E. Delson, and J. VanCouvering. 1988. Pp. 245–255 Pakistan and Kenya. Oxford Monographs on Geology and in Encyclopedia of human evolution and prehistory. Garland, Geophysics No. 19. Oxford University Press, New York. New York. Retallack, G. J., E. A. Bestland, and D. P. Degas. 1995. Miocene Wagner, P.J. 1995. Stratigraphic tests of cladistic hypotheses. Pa- paleosols and habitats of Proconsul on Rising Island. Journal leobiology 21:153–178. of Human Evolution 29:53–91. Walker, A., and M. F. Teaford. 1988. The Kaswanga Primate Re- Rieppel, O., and L. Grande. 1994. Summary and comments on search Site; and early Miocene hominoid site on Rusinga Is- systematic pattern and evolutionary process. Pp. 227–255 in land, Kenya. Journal of Human Evolution 17:539–544. L. Grande and O. Rieppel, eds. Investigating the hierarchy of Walker, A., M. F. Teaford, L. Martin, and P. Andrews. 1993. A nature. Academic Press, NY. new species of Proconsul from the early Miocene of Rusinga/ Rook, L., T. Harrison, and B. Engesser. 1996. The taxonomic sta- Mfangano Islands, Kenya. Journal of Human Evolution 25:43– tus and biochronological implications of Oreopithecus from 56. Baccinello (Tuscany, Italy). Journal of Human Evolution 30:3– Ward, C. V. 1997a. Functional anatomy and phyletic implica- 27. tions of the hominoid trunk and hindlimb. Pp. 101–130 in Be- Rook, L., P. Rene, M. Benvenuti, and M. Papini. 2000. Geochro- gun et al. 1997b. nology of Oreopithecus-bearing succession at Baccinello (Italy) Ward, C. V., A. Walker, M. F. Teaford, and I. Odhiambo. 1993. and the extinction pattern of European Miocene hominoids. Partial skeleton of Proconsul nyanzae from Mfangano Island, Journal of Human Evolution 39:577–582. Kenya. American Journal of Physical Anthropology 90:77– Rose, M. D. 1997. Functional and phylogenetic features of the 111. forelimb in Miocene hominoids. Pp. 79–100 in Begun et al. Ward, S. 1997b. The taxonomy and phylogenetic relationships 1997b. of Sivapithecus revisited. Pp. 269–290 in Begun et al. 1997b. Schultz, A. H. 1961. Vertebral column and thorax. Primatologia Ward, S., B. Brown, A. Hill, J. Kelley, and W. Downs. 1999. Equa- 4:1–66. torius: a new hominoid genus from the Miocene of Kenya. Sci- Schwartz, J. H. 1990. Lufengpithecus and its potential relation- ence 285:1382–1386. ship to an orang-utan clade. Journal of Human Evolution 19: White, T. D., G. Suwa, and B. Asfaw. 1994. Australopithecus ram- 591–605. idus, a new species of early hominid from Aramis, Ethiopia. ———. 1997. Lufengpithecus and hominoid phylogeny: Problems Nature 371:306–312. in delineating and evaluating phylogenetically relevant char- ———. 1995. Australopithecus ramidus, a new species of early acters. Pp. 363–388 in Begun et al. 1997b. hominid from Aramis, Ethiopia. Nature 375:88. Senut, B., M. Pickford, D. Gommery, and Y. Kunimatsu. 2000 A new genus of Early Miocene hominoid from East Africa: Wilkinson, M. 1994. Common cladistic information and its con- Ugandapithecus major (Le Gros Clark & Leakey, 1950). Comptes sensus representation: reduced Adams and reduced consen- Rendus des Seánces de l’Acade´mie des Sciences 331:227–233. sus trees and profiles. Systematic Biology 43:343–368. Senut, B., D. Gommery, P. Mein, C. Cheboi, and Y. Coppens. ———. 1996. Majority-rule reduced consensus methods and 2001. First hominid from the Miocene (Lukeino Formation, their use in bootstrapping. Molecular Biology and Evolution Kenya). Comptes Rendus des Seánces de l’Acade´mie des Sci- 13:437–444. ences 332:137–144. ———. 1999. Choosing and interpreting a tree for the oldest Sherwood, R. J., S. Ward, A. Hill, D. L. Duren, B. Brown, and W. mammal. Journal of Vertebrate Paleontology 19:187–190. Downs. 2002. Preliminary description of the Equatorius afri- ———. 2001. REDCON 3.0: software and documentation. De- canus partial skeleton (KNM-TH 28860) from Kipsaramon, partment of Zoology, The Natural History Museum, London. Tugen Hills, Baringo District, Kenya. Journal of Human Evo- WoldeGabriel, G., T. D. White, G. Suwa, P.Renne, J. de Heinzelin, lution 42:63–73. W. K. Hart, and G. Heiken. 1994. Ecological and temporal Shipman, P., A. Walker, J. A. VanCouvering, P. J. Hooker, and J. placement of early Pliocene hominids at Aramis, Ethiopia. A. Miller. 1981. The Fort Ternan hominoid site, Kenya: geol- Nature 371:330–333. ogy, age, taphonomy, and paleoecology. Journal of Human Wu, R., Q. Xu, and Q. Lu. 1986. Relationship between Lufeng Evolution 10:49–72. Ramapithecus and Sivapithecus and their phylogenetic position. Sorkhabi, R. B., and A. Macfarlene. 1999. Himalaya and Tibet: Acta Anthropologica Sinica 5:1–30. DATA CONGRUENCE IN HOMINOID PHYLOGENY 639 Ј Ј 2 3 4 5 34 37 38 16 17 18 20 21 22 23 24 25 26 27 28 29 30 31 32 33 168 170 171 y/y 169 z/z citation original Number in citation Original Begun et al. 1997a Begun et al. 1997a Begun et al. 1997a Begun et al. 1997a Begun et al. 1997a Begun et al. 1997a Rose 1997 Begun et al. 1997a Begun et al. 1997a Begun et al. 1997a Begun et al. 1997a Begun et al. 1997a Begun et al. 1997a Begun et al. 1997a Begun et al. 1997a Begun et al. 1997a Begun et al. 1997a Begun et al. 1997a Begun et al. 1997a Begun et al. 1997a Begun et al. 1997a Begun et al. 1997a Begun et al. 1997a Begun et al. 1997a Begun et al. 1997a Begun et al. 1997a Begun et al. 1997a Begun et al. 1997a Rose 1997 Begun et al. 1997a w/strong medial keel 90 deg. 15 deg. medially narrow anterior gracile smaller Ͼ Ͼ Ͼ broad shallow abbreviated fused large palmar spherical narrow rounded medial merged notched large rounded present circular concave laterally reduced larger deep trochleiform straight Appendix 1 Character state definitions 0123 oflexed oflexed weak medial keel 90 deg. 15 deg. broad axillary robust larger smaller shallow cylindrical w/ retr Ͻ Ͻ narrow deep not abbreviated separate small lateral ovoid broad angular retr subequal distinct straight small oval absent oval flat anterolaterally prominent tubercle xillary margin haracter descriptions. angle curvature Infraglenoid Teres minor attachment Spinous process root Glenoid/a Coronoid/radial fossae size Coronoid fossa Trochlea Anteroposterior shaft Intertuberosity angle Humeral head torsion Trochlear notch Ulnar shaft Olecranon process Os centrale and scaphoid Scaphoid tuberosity Scaphoid tubercle Humeral head shape Bicipital groove Proximal shaft shape Medial epicondylar projection Trochlear keel symmetry Medial epicondyle/keel Superior trochlear border Lateral trochlear keel Radial head shape Radial head beveling Radial neck shape Distal radial articular surface Radial notch of ulna faces Deltopectoral plane Character number and description 1 2 3 4 5 6 7 8 9 10 25 26 27 28 29 30 11 12 13 14 15 16 17 18 19 20 21 22 23 24 Morphological c Scapula Humerus Carpo-metacarpals Radius/Ulna 640 JOHN A. FINARELLI AND WILLIAM C. CLYDE 39 42 43 44 45 47 52 53 54 55 56 57 58 59 60 61 62 63 65 66 67 68 71 72 73 74 75 76 77 79 In text citation original Number in citation Original Begun et al. 1997a Begun et al. 1997a Begun et al. 1997a Begun et al. 1997a Begun et al. 1997a Begun et al. 1997a Begun et al. 1997a Begun et al. 1997a Begun et al. 1997a Begun et al. 1997a Begun et al. 1997a Begun et al. 1997a Begun et al. 1997a Begun et al. 1997a Begun et al. 1997a Begun et al. 1997a Begun et al. 1997a Begun et al. 1997a Begun et al. 1997a Begun et al. 1997a Begun et al. 1997a Begun et al. 1997a Begun et al. 1997a Begun et al. 1997a Begun et al. 1997a Begun et al. 1997a Begun et al. 1997a Begun et al. 1997a Begun et al. 1997a Begun et al. 1997a Ward, C.V. 1997 short absent dorsal broad laterally broad absent broad short extensive high-angle absent irregular absent divided present narrow concavoconvex large lateral shallow short bowed proximal absent angulated absent dorsally square small ridged high medium small inter. inter. short/medio- Character state definitions 0123 long/narrow present narrow long restricted low-angle present flat present continuous absent broad flat small proximal deep long narrow dorsal present confluent present palmarly oval large convex low tall large ventral narrow craniocaudally trapezium pisiform and/or triquetrum surface of proximal phalanges tubercles articulation surface Os centrale articulation with Lunate: mediolateral width Lunate: dorsopalmar length Scaphoid facet on lunate Lunate, scaphoid radial facet Ulnar styloid articulation with MC3 facet on capitate Palmar MC4 facets on capitate MC2 facet on capitate Trapezoid facet on capitate Capitate head mediolaterally Central facet on capitate Hamulus size Triquetral face on hamate MC4 facet on hamate MC1 head proximodistally MC1 head dorsal part MC1/trapezium articular MC4 palmar capitate facet MC facets on palmar hamate MC2-4 sesamoids MC heads broadest Proximal articulation surface Proximal phalangeal palmar Ray 1 terminal phalanx Costal angle Vertebral body height Accessory processes Transverse processes Sternebrae Torso shape . Continued. Character number and description 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 Appendix 1 Phalanges: Manus Trunk DATA CONGRUENCE IN HOMINOID PHYLOGENY 641 80 81 82 83 84 85 86 87 89 90 91 92 93 94 98 99 41 101 102 103 104 105 106 107 108 109 110 111 114 115 116 117 citation original Number in citation Original Begun et al. 1997a Begun et al. 1997a Begun et al. 1997a Begun et al. 1997a Begun et al. 1997a Begun et al. 1997a Begun et al. 1997a Begun et al. 1997a Begun et al. 1997a Begun et al. 1997a Begun et al. 1997a Begun et al. 1997a Begun et al. 1997a Begun et al. 1997a Begun et al. 1997a Begun et al. 1997a Begun et al. 1997a Begun et al. 1997a Begun et al. 1997a Begun et al. 1997a Begun et al. 1997a Begun et al. 1997a Begun et al. 1997a Begun et al. 1997a Begun et al. 1997a Begun et al. 1997a Begun et al. 1997a Begun et al. 1997a Begun et al. 1997a Begun et al. 1997a Ward, C. V. 1997 unique Begun et al. 1997a trical wide wide long deep sphere asymme- shallow short large flat large robust inter. high long inter. medium inter. inter. absent inter. short flared robust inter.inter. inter. inter. angled large inter. large small inter. strong short medial short inter. large short absent robust absent inter. distal Character state definitions 0123 nder narrow low short narrow short open cyli present symmetrical square distal thin small deep long small aligned curved small large small slight long distal long gracile small long present gracile present dorsal lunate surface navicular Iliac blade width Iliac blade height Lower iliac height Cranial aspect of acetabulum Pubic length Trochanteric fossa Femoral head Femoral neck tubercle Femoral condyle shape Distal tibia facet Medial malleolus projection Fibular robusticity Lateral malleolus Talar trochlear depth Distal calcaneus Flexor hallucis longus groove Posterior talar facet long axis Anterior talar facet Plantar calcaneal tubercle Calcaneo-navicular facet Cuboid peg Cuboid wedging Cuboid length Entocuneiform/MT 1 joint Cuneiform length MT 1 robusticity MT 1 sesamoid grooves MT 1 length MT 1 prehallux facet MT 2-5 robusticity Transverse arch Entocuneiform facet on . Continued. Character number and description 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 Appendix 1 Pelvis Femur Tibia/Fibula Tarso-metatarsals 642 JOHN A. FINARELLI AND WILLIAM C. CLYDE 221 222 223 224 226 227 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132 133 134 135 136 137 217 218 219 220 228 229 230 231 232 234 citation original Number in citation Original Begun et al. 1997a Begun et al. 1997a Begun et al. 1997a Begun et al. 1997a Begun et al. 1997a Begun et al. 1997a Begun et al. 1997a Begun et al. 1997a Begun et al. 1997a Begun et al. 1997a Begun et al. 1997a Begun et al. 1997a Begun et al. 1997a Begun et al. 1997a Begun et al. 1997a Begun et al. 1997a Begun et al. 1997a Begun et al. 1997a Begun et al. 1997a Begun et al. 1997a Begun et al. 1997a Begun et al. 1997a Begun et al. 1997a Begun et al. 1997a Begun et al. 1997a Begun et al. 1997a Begun et al. 1997a Begun et al. 1997a Begun et al. 1997a Begun et al. 1997a Begun et al. 1997a Begun et al. 1997a Begun et al. 1997a Begun et al. 1997a Begun et al. 1997a Begun et al. 1997a v. high v. low v. reduced M2 ϭ / Ͻ canine rotated high low/rounded strong digit 2 robust curved inter.narrow absent reduced compressed strong thick, rounded reduced low/rounded weak/absent large M3 low/rounded larger post- internally small large short broad present present long high reduced low/rounded long rectangle strong strong long weak/absent Character state definitions 0123 M2 Ͼ oriented digit 3 gracile straight weak broad present large robust narrow strong tall/narrow strong small M3 low tall weak high subequal buccolingually large small tall narrow absent absent short low large tall broad triangle weak weak broad strong lingually omorphy omorphy size ratio size ratio 2 2 size heter metacones shape 2 /M /M shape shape lingual flare posterior dentition 2–3 1 3 1–2 /I 4 3 4 1 M I P Foot axis runs through Phalangeal robusticity Phalangeal curvature Phalangeal flexor ridges Incisor breadth I2 cingulum Relative male canine size Male canines Canine cingula P3 cusp heter P3 paracones P3 shape P3 mesiolingual beak P3 metaconid P4 shape P4 talonid M Upper canine height Upper canine cervical flare P Premolar buccal flare P M Dentine penetrance Anterior dentition relative to Mandibular canine roots Lower central incisors Lower incisors labio Lower canine crown height Lower molar cingula M Molar cusps Upper molar cingula . Continued. Character number and description 94 95 96 97 98 99 124 125 126 127 128 129 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 Appendix 1 Phalanges: Pes Dentition DATA CONGRUENCE IN HOMINOID PHYLOGENY 643 235 236 237 238 239 138 139 140 144 145 146 147 148 149 150 151 152 153 154 155 156 157 158 159 160 162 163 164 167 172 173 178 175 177 citation original Number in citation Original Begun et al. 1997a Begun et al. 1997a Begun et al. 1997a Begun et al. 1997a Begun et al. 1997a Begun et al. 1997a Begun et al. 1997a Begun et al. 1997a Begun et al. 1997a Begun et al. 1997a Begun et al. 1997a Begun et al. 1997a Begun et al. 1997a Begun et al. 1997a Begun et al. 1997a Begun et al. 1997a Begun et al. 1997a Begun et al. 1997a Begun et al. 1997a Begun et al. 1997a Begun et al. 1997a Begun et al. 1997a Begun et al. 1997a Begun et al. 1997a Begun et al. 1997a Begun et al. 1997a Begun et al. 1997a Begun et al. 1997a Begun et al. 1997a Begun et al. 1997a Begun et al. 1997a Begun et al. 1997a Begun et al. 1997a v. large 3 nificant- ly lower distal to P v. elongate sunken strong absent absent inion sig- inferior Begun et al. 1997a large high vertical crenulated reduced thick biconvex distal to C deep flat thicker elongate narrow inflated present present broad above & below small horizontal concave broad inion below reduced posterior elongated deep small fused hollow below apex deep anterior hollow Character state definitions 0123 small low bulging smooth large thin flat opposite to C shallow curved narrow broad broad indistinct absent absent shallow/absent above large vertical convex/flat narrow inion above strong superiorly broad short shallow large unfused solid near apex shallow lateral robust ossa size ratio 2 size /M nasion protuberance temporal bones ization relative to nasal aperture 1 3 Palatine process Alveolar process Incisive fossa Maxillary depth Lateral malar surface Frontozygomatic breadth Upper molar crowns Upper molar sides Molar enamel M M Orbital breadth Nasal bones at nasion Glabella Supraorbital torus Supraciliary ridges Supraorbital sulcus Frontal sinus size relative to Frontal sinus size Frontal squama Facial profile Temporal fossa Inion relative to glabella External occipital Nuchal plane orientation Neurocranial length Glenoid f Articular tubercle Articular/tympanic and Zygomatic root pneumat- Inferior orbital foramen Zygomatic depth Zygomatic orientation Maxillary nasal process . Continued. Character number and description 135 136 137 138 139 140 130 131 132 133 134 141 142 143 144 145 146 147 148 149 150 151 152 153 154 155 156 157 158 159 160 161 162 163 Cranium Appendix 1 644 JOHN A. FINARELLI AND WILLIAM C. CLYDE 18 178 179 180 182 183 184 185 186 187 189 190 191 192 193 194 195 196 197 198 199 200 201 202 203 204 205 206 207 208 209 210 211 212 214 215 216 citation original Number in ¸ 1998 ¨lec citation Original Begun et al. 1997a Begun et al. 1997a Begun et al. 1997a Begun et al. 1997a Begun et al. 1997a Begun et al. 1997a Begun et al. 1997a Begun et al. 1997a Begun et al. 1997a Begun et al. 1997a Begun et al. 1997a Begun et al. 1997a Begun et al. 1997a Begun et al. 1997a Begun et al. 1997a Begun et al. 1997a Begun et al. 1997a Begun et al. 1997a Begun et al. 1997a Begun et al. 1997a Begun et al. 1997a Begun et al. 1997a Begun et al. 1997a Begun et al. 1997a Begun et al. 1997a Begun et al. 1997a Begun et al. 1997a Begun et al. 1997a Begun et al. 1997a Begun et al. 1997a Begun et al. 1997a Begun et al. 1997a Begun et al. 1997a Begun et al. 1997a Begun and Gu v. small v. largehorizontal high Begun et al. 1997a v. long flatinflated Begun et al. 1997a large deep smooth small piriform rotated small flat narrow below deep vertical broad v. low high yes long inter. posterior posterior anterior deeper medial externally collapsed horizontal elongate posterior small narrow weak deep inferior compressed deep broad high shallow stepped long horizontal large ovid Character state definitions 0123 large concave broad same level shallow inclined narrow low low none short anterior above above lateral shallow vertical in line solid vertical round anterior none broad strong shallow superiorly broad robust shallow compressed low absent fenestrated short vertical absent superiorly broad 2 1 distance position alveolar plane malar surface to M orientation Orbital/nasal aperture Orbital/nasal surface Interorbital distance Zygomatic arch relative orbit Zygomatic temporal process Zygomatic arch angle Nasal aperture breadth Nasal aperture relative to orbit Nasal aperture relative to Lacrimal fossa visible Nasal bone length Nasal aperture relative to Inferior orbital margin relative Nasal apex relative to M Maxillary surface Canine fossa Canine foot angulation Canine root rotation Maxillary alveolar process Incisor orientation Greater palatine foramen Greater palatine position Lesser palatine foramina Horizontal palatine Palatine crest Posterior palate Pyramidal process position Pterygoid process Alveolar process depth Zygomatico-alveolar crest Zygomatic root height Incisive fossa Subnasal floor Nasoalveolar clivus length Nasoalveolar clivus Incisive canal caliber Nasal aperture shape . Continued. Character number and description 164 165 166 167 168 169 170 171 172 173 174 175 176 177 178 179 180 181 182 183 184 185 186 187 188 189 190 191 192 193 194 195 196 197 198 199 200 Appendix 1 DATA CONGRUENCE IN HOMINOID PHYLOGENY 645 1 2 0 1 1 1 0 1 1 1 0 1 1 1 1 2 1 0 0 0 0 0 0 0 1 1 0 2 1 1 1 2 1 1 1 2 1 1 0 3 1 0 0 0 0 0 0 0 1 0 3 0 1 0 3 0 1 0 3 0 1 0 2 1 1 0 0 0 0 0 0 0 1 2 1 2 1 2 1 2 1 2 1 1 1 2 2 2 1 2 1 0 0 1 0 0 1 0 1 1 1 1 1 1 1 1 1 1 0 1 0 2 1 1 0 0 0 0 0 0 1 1 1 2 1 1 1 2 0 1 1 2 1 1 1 1 1 0 0 0 0 0 0 0 1 1 2 1 1 1 2 1 0 1 2 1 1 1 0 1 1 0 0 0 1 0 0 0 1 2 1 1 1 1 1 1 1 1 1 1 1 1 0 0 1 0 0 0 0 0 0 — 1 0 1 1 1 1 1 1 1 1 1 1 0 1 1 1 0 0 0 0 0 0 0 — 0 1 1 1 0 1 1 1 0 1 0 1 1 1 2 0 1 0 0 0 0 0 0 0 0 1 0 1 0 1 0 1 0 1 1 1 0 1 1 1 0 0 0 0 1 0 0 0 0 0 0 1 1 1 0 1 1 1 0 1 1 1 1 1 1 1 0 1 0 0 0 0 1 1 1 0 1 0 1 1 1 0 1 1 1 0 1 0 1 0 0 0 0 1 0 0 1 2 2 1 1 2 2 1 1 2 1 1 1 2 1 1 0 1 0 0 0 1 0 0 1 0 1 2 1 1 1 2 1 1 1 1 1 0 0 0 0 0 0 0 0 0 0 0 1 0 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 0 1 0 1 0 0 1 0 0 0 1 1 1 0 1 1 0 0 1 1 1 1 1 1 1 1 0 0 0 0 1 1 0 0 1 1 0 0 1 1 0 0 1 1 1 1 0 0 0 0 0 0 0 0 1 1 1 2 1 1 1 2 1 1 1 2 1 2 1 1 1 1 0 0 0 2 0 0 1 1 1 1 1 1 1 0 1 1 1 0 1 1 1 0 1 1 0 0 0 0 0 0 1 1 1 0 1 1 1 0 1 1 1 0 1 1 1 1 1 0 1 0 0 0 0 0 1 2 1 1 1 2 1 0 1 2 1 0 0 2 1 1 1 2 1 1 0 1 0 0 1 1 1 1 1 1 1 1 1 1 1 0 0 0 1 1 0 0 0 0 0 0 0 0 1 1 1 0 1 2 1 0 1 2 0 0 1 2 0 1 1 1 0 0 0 1 0 0 1 2 0 0 1 2 0 0 1 2 0 0 1 2 0 1 1 1 0 0 0 0 0 0 1 2 1 1 1 2 0 1 1 2 0 1 1 2 0 2 1 2 0 0 0 1 0 0 Appendix 2 1 0 1 0 1 2 1 0 1 2 1 1 1 2 1 1 1 2 0 0 0 1 0 0 1 1 0 1 1 1 0 1 1 1 0 1 1 1 0 0 1 0 0 0 0 1 0 0 Character number 1 1 2 1 1 1 0 1 1 1 0 1 1 1 0 1 1 1 0 0 0 0 0 0 1 0 1 2 1 1 0 2 1 1 0 0 1 1 2 2 1 1 0 2 1 0 0 0 1 2 1 0 1 2 1 0 1 2 1 1 1 2 1 1 1 0 1 0 1 1 0 0 1 1 1 2 1 1 1 2 1 1 1 2 1 1 1 1 1 0 0 0 0 0 0 0 1 2 1 0 1 2 1 0 1 2 1 0 1 2 1 0 1 2 1 0 0 1 0 0 1 3 2 0 1 2 0 1 1 2 0 1 1 2 0 1 0 1 0 1 0 0 0 0 0 2 1 0 0 2 1 0 0 2 1 0 0 2 1 1 0 1 0 0 1 0 0 0 1 2 0 0 1 2 0 0 1 2 0 0 1 2 0 1 0 1 0 1 0 0 0 0 1 0 0 1 1 1 0 1 1 1 0 0 1 1 1 1 1 1 0 1 0 0 0 0 1 1 1 1 1 1 0 1 1 1 0 1 1 1 0 1 1 1 0 0 0 0 0 0 1 2 1 1 1 2 0 0 1 2 0 0 1 2 0 1 1 1 0 0 0 0 0 0 1 1 1 0 1 1 0 0 1 1 0 0 1 1 1 0 1 1 0 0 0 0 0 — 1 2 1 1 1 2 1 1 1 2 1 0 1 2 1 1 1 1 0 1 0 0 0 0 1 2 1 1 1 2 1 1 1 2 1 1 1 2 1 1 1 1 0 0 0 0 0 — 1 2 1 1 1 2 1 1 1 2 1 1 1 2 1 0 1 1 0 0 0 1 0 0 1 2 1 1 1 2 1 1 1 2 1 1 1 2 1 0 1 1 0 0 0 0 0 0 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 0 1 1 0 0 0 0 0 0 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 0 1 0 0 0 0 0 0 — 1 1 1 2 1 1 1 2 1 1 1 2 1 1 1 1 1 1 0 0 0 0 0 0 0 1 1 1 0 1 1 1 0 1 1 1 0 1 1 1 0 1 0 1 1 0 0 0 1 0 1 2 1 1 1 2 1 1 1 1 1 1 1 1 1 1 0 0 0 0 0 0 Character by taxon matrix for the Hominoidea. 1234567891011121314151617181920212223242526272829303132333435363738394041424344454647484950 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 0 1 1 1 0 0 0 0 0 Australopithecus Stratigraphic character state: ‘‘j&k’’ Pan Stratigraphic character state: ‘‘k’’ Gorilla Stratigraphic character state: ‘‘k’’ Pongo Stratigraphic character state: ‘‘k’’ Hylobates Stratigraphic character state: ‘‘k’’ Proconsul Stratigraphic character state: ‘‘a&b&c&d’’ 646 JOHN A. FINARELLI AND WILLIAM C. CLYDE 0 0 0 0 0 2 1 0 1 1 0 — — — — — — — — — — — — — 0 0 0 0 0 0 0 0 1 0 1 — — — — — — — — — — — — — 0 0 0 0 0 0 0 0 1 1 0 — — — — — — — — — — — — — 0 1 1 0 1 1 1 1 1 — — — — — — — — — — — — — — — 0 0 0 0 1 0 1 — — — — — — — — — — — — — — — — — 0 1 0 0 0 0 1 1 1 2 — — — — — — — — — — — — — — 1 0 0 0 0 1 1 0 0 0 1 1 1 — — — — — — — — — — — 0 0 0 1 0 1 1 — — — — — — — — — — — — — — — — — 0 0 0 0 0 0 1 1 0 1 1 — — — — — — — — — — — — — 0 0 0 1 0 0 1 1 — — — — — — — — — — — — — — — — 0 0 0 0 1 1 — — — — — — — — — — — — — — — — — — 0 0 0 0 0 0 1 — — — — — — — — — — — — — — — — — 0 0 1 1 0 0 1 0 0 1 — — — — — — — — — — — — — — 0 0 0 0 1 1 — — — — — — — — — — — — — — — — — — 0 0 0 0 0 0 1 — — — — — — — — — — — — — — — — — 0 0 1 0 1 0 1 1 — — — — — — — — — — — — — — — — 0 1 0 0 1 0 1 1 0 — — — — — — — — — — — — — — — 0 0 0 0 0 0 0 0 0 0 1 0 0 — — — — — — — — — — — 2 1 0 0 2 0 0 1 1 1 1 1 1 1 1 0 — — — — — — — — 0 0 0 0 1 0 0 0 1 0 — — — — — — — — — — — — — — 0 0 0 0 0 1 0 0 0 1 0 — — — — — — — — — — — — — 1 1 0 1 1 1 1 1 1 1 1 0 — — — — — — — — — — — — 0 0 0 0 0 0 0 0 0 1 1 0 — — — — — — — — — — — — 0 0 0 0 0 0 1 0 0 1 0 — — — — — — — — — — — — — 0 0 0 0 1 0 0 0 0 0 1 0 — — — — — — — — — — — — 1 0 0 0 1 0 0 0 0 1 0 0 0 1 1 0 — — — — — — — — 0 1 0 1 0 1 0 0 1 0 0 0 1 0 — — — — — — — — — — 0 1 0 1 0 1 1 1 0 1 0 1 — — — — — — — — — — — — Character number 0 0 0 0 0 0 0 0 0 0 0 — — — — — — — — — — — — — 0 1 0 1 1 0 1 1 1 1 — — — — — — — — — — — — — — 0 1 0 0 0 0 0 1 1 1 0 — — — — — — — — — — — — — 1 1 0 1 1 0 0 0 1 1 0 — — — — — — — — — — — — — 1 0 0 0 0 0 1 1 0 0 1 1 0 — — — — — — — — — — — 0 1 1 0 1 0 0 1 0 1 1 0 — — — — — — — — — — — — 0 0 0 0 1 0 1 0 — — — — — — — — — — — — — — — — 0 0 0 0 0 0 0 0 1 1 0 0 — — — — — — — — — — — — 2 0 2 0 0 0 1 0 — — — — — — — — — — — — — — — — 0 0 0 0 0 0 0 1 0 0 — — — — — — — — — — — — — — 0 0 1 1 0 1 1 1 0 0 — — — — — — — — — — — — — — 0 1 1 1 1 0 0 1 1 0 — — — — — — — — — — — — — — 0 0 0 0 0 1 0 1 1 1 — — — — — — — — — — — — — — 1 0 0 1 1 1 2 1 1 — — — — — — — — — — — — — — — 0 1 0 1 0 1 1 1 1 — — — — — — — — — — — — — — — 0 0 0 1 0 1 0 1 1 1 1 — — — — — — — — — — — — — 0 0 0 1 1 1 1 1 1 1 — — — — — — — — — — — — — — 0 1 0 1 0 1 1 1 1 1 — — — — — — — — — — — — — — 0 0 1 0 1 1 1 1 . Continued. — — — — — — — — — — — — — — — — 0 0 0 0 1 0 1 1 1 1 — — — — — — — — — — — — — — 0 1 0 0 1 1 1 — — — — — — — — — — — — — — — — — 1234567891011121314151617181920212223242526272829303132333435363738394041424344454647484950 1 0 0 1 1 0 0 1 1 — — — — — — — — — — — — — — — Appendix 2 Turkanapithecus Stratigraphic character state: ‘‘b’’ Afropithecus Stratigraphic character state: ‘‘b’’ Kenyapithecus Stratigraphic character state: ‘‘d’’ Equatorius Stratigraphic character state: ‘‘c’’ Griphopithecus Stratigraphic character state: ‘‘c&d’’ Dryopithecus Stratigraphic character state: ‘‘f&g’’ DATA CONGRUENCE IN HOMINOID PHYLOGENY 647 1 1 2 1 0 0 1 0 2 2 0 1 1 0 2 — — — — — — — — — 1 0 3 0 0 1 0 0 1 0 1 0 1 0 2 — — — — — — — — — 1 0 2 1 0 1 0 1 0 0 0 0 0 3 1 — — — — — — — — — 1 2 2 1 1 1 1 1 0 1 2 — — — — — — — — — — — — — 1 0 2 1 0 0 1 0 1 0 0 1 — — — — — — — — — — — — 0 1 1 1 1 1 2 0 1 1 2 0 2 — — — — — — — — — — — 0 0 1 1 1 1 0 0 1 1 0 0 1 1 — — — — — — — — — — 0 1 0 0 1 0 1 1 2 1 1 1 0 0 — — — — — — — — — — 0 1 1 1 0 0 1 1 1 0 1 1 — — — — — — — — — — — — 1 2 1 – 1 0 0 – 1 0 0 0 — — — — — — — — — — — — 1 0 0 1 0 1 1 0 0 — — — — — — — — — — — — — — — 0 1 1 0 0 0 0 1 0 1 — — — — — — — — — — — — — — 0 1 0 1 0 0 1 0 — — — — — — — — — — — — — — — — 2 1 1 1 0 0 0 1 — — — — — — — — — — — — — — — — 0 0 1 1 1 0 0 1 0 — — — — — — — — — — — — — — — 1 1 1 1 1 0 1 1 — — — — — — — — — — — — — — — — 1 0 1 1 0 0 0 1 — — — — — — — — — — — — — — — — 51–100; the third, 101–150 and the fourth, 151–200. The number in a particular 0 0 1 1 0 0 1 0 0 0 0 0 0 — — — — — — — — — — — ond, 2 1 1 2 1 0 1 0 1 1 1 0 1 0 — — — — — — — — — — 1 1 1 1 1 0 1 1 1 0 0 1 1 — — — — — — — — — — — 1 1 1 1 1 0 1 1 1 1 0 0 1 1 — — — — — — — — — — 2 1 1 2 0 0 1 1 1 0 0 0 1 1 — — — — — — — — — — 0 1 1 0 1 0 1 0 1 0 0 0 1 0 — — — — — — — — — — 2 0 0 1 2 0 0 0 0 1 0 0 0 0 — — — — — — — — — — 1 0 1 1 2 0 0 0 0 0 0 1 0 0 0 — — — — — — — — — 1 1 0 2 1 2 1 1 1 1 1 0 0 0 0 — — — — — — — — — 0 0 2 1 0 0 0 0 0 1 0 0 2 — — — — — — — — — — — 0 0 1 1 0 0 0 1 1 1 0 1 1 — — — — — — — — — — — Character number 0 0 1 1 0 1 0 0 1 1 0 0 1 1 — — — — — — — — — — 1 2 1 1 2 2 2 1 0 0 2 1 — — — — — — — — — — — — denotes missing data. 2 1 0 1 2 0 0 1 0 0 0 2 0 1 1 — — — — — — — — — 0 1 1 1 1 1 1 1 1 0 0 0 1 — — — — — — — — — — — 1 1 1 0 1 2 1 1 1 0 1 0 1 1 1 — — — — — — — — — 1 1 0 0 1 0 0 0 0 2 0 1 — — — — — — — — — — — — 0 1 1 0 2 0 0 1 0 1 0 0 1 1 — — — — — — — — — — 0 0 1 1 0 0 0 0 0 0 1 0 — — — — — — — — — — — — 0 1 0 1 1 0 1 0 0 0 0 0 — — — — — — — — — — — — 0 0 1 1 1 0 0 1 1 0 1 0 — — — — — — — — — — — — 0 1 1 1 2 0 0 1 1 0 1 1 — — — — — — — — — — — — 0 1 1 1 1 0 1 0 1 1 1 — — — — — — — — — — — — — 0 1 0 1 2 1 1 1 1 1 0 1 0 — — — — — — — — — — — 1 0 1 2 0 1 1 1 2 1 0 0 — — — — — — — — — — — — arranged in four rows of fifty characters each (sum 200 characters). The first row contains 1–50; the sec 0 1 0 1 2 1 1 1 1 2 0 — — — — — — — — — — — — — 1 1 0 1 2 1 1 1 1 2 1 — — — — — — — — — — — — — able are 1 1 0 1 0 0 1 1 1 — — — — — — — — — — — — — — — 1 1 1 1 1 1 0 — — — — — — — — — — — — — — — — — 0 1 1 1 0 0 1 1 1 . Continued. — — — — — — — — — — — — — — — 0 1 0 1 0 1 1 1 — — — — — — — — — — — — — — — — 1 1 1 0 1 — — — — — — — — — — — — — — — — — — — * Characters in this t 1234567891011121314151617181920212223242526272829303132333435363738394041424344454647484950 1 0 1 1 1 1 0 — — — — — — — — — — — — — — — — — Appendix 2 Sivapithecus Stratigraphic character state: ‘‘e&f&g&h&i’’ Oreopithecus Stratigraphic character state: ‘‘i’’ Lufengpithecus Stratigraphic character state: ‘‘i’’ Ouranopithecus Stratigraphic character state: ‘‘g&h’’ Morotopithecus Stratigraphic character state: ‘‘a’’ Ankarapithecus Stratigraphic character state: ‘‘f’’ cell position gives the coded value of the character for the particular taxon. A dash 648 JOHN A. FINARELLI AND WILLIAM C. CLYDE

Appendix 3 al. 1990; Hill et al. 1985, 2002). Using the most encompassing dates for Proconsul then places the FAE at Meswa Bridge, and Chronostratigraphy of the Hominoidea the LAE in the Ngorora Formation. However, although the tax- Morotopithecus onomic assignment of the Kisingiri and Tinderet material is Until recently, the age of the Moroto localities was thought to rather secure, the attribution of the hominoid fossils from the be significantly younger than either of the two genera with Ngorora Formation and Meswa Bridge to Proconsul is less cer- which the hominoid material was originally allied in the liter- tain (see Andrews et al. 1981; Senut et al. 2000; Hill et al. 2002). ature: Proconsul and Afropithecus. Bishop et al. (1969) reported Thus, both the FAE and LAE of the OTU Proconsul, as used here, a K/Ar age for the beds at Moroto of 14.3 Ϯ 0.3 Ma. Pickford et may change dramatically with the recovery of more complete al. (1986) had revised this date on the basis of biochronological material. The effect of these changes on the results of this anal- correlations to older than 17.5 Ma, and recent revision of the ysis is minimal, however, and the topology of the stratocladistic radiometric dates using the more precise 40Ar/39Ar method hypothesis is not affected. See text for discussion. places the age of the Moroto I and II localities at older than 20.61 An additional locality exists that is purported to extend the Ϯ 0.05 Ma (Gebo et al. 1997). range of Proconsul into the latest Oligocene. Material from the Eragaleit Beds of the Lothidok Formation was also referred to Proconsul Proconsul in beds dated to 24.3–27.5 Ma (Brochetto et al. 1992). Dating of Proconsul fossil material is generally well con- This material has been reassigned to the genus Kamoyapithecus, strained as a result of the geological setting in East Africa ca. a late Oligocene catarrhine (Leakey et al. 1995b). Although the 20 Ma. The beginnings of the East African Rift resulted in the authors did note morphological affinities of this taxon to the ge- formation of large volcanoes that produced tuffs intercalated nus Proconsul and tentatively placed this taxon within the Hom- with the fossiliferous sedimentary strata, providing radiometric inoidea, they pointed out that the level of preservation for this dates bracketing hominoid-bearing units. taxon is as yet insufficient for use in cladistic analysis (Leakey The fossil record of Proconsul in Kenya is known from the Tin- et al. 1995b), and consequently Kamoyapithecus is not considered deret and Kisingiri. Both have been reconstructed as sedimen- in the present study. tary basins on the flanks of large shield volcanoes (Pickford and Andrews 1981; Retallack 1991; Retallack et al. 1995). The Tin- Afropithecus deret includes multiple fossil localities, including Koru, Leget, Songhor, and Chamtwara. The stratigraphic sequence of inter- Afropithecus is known from the region around Lake Turkana est includes the Koru Formation, overlain by the Leget Forma- from the early Miocene. The genus was described from material tion, and no fewer than 23 fossil sites bearing hominoid material recovered at the Kalodirr locality in Kenya (Leakey and Leakey have been reported in these formations (Pickford 1984). This se- 1986a) consisting of the single species Afropithecus turkanensis. quence is bracketed by unconformities, and K/Ar dating places Material recovered from the Warata Formation at Buluk (Leakey these beds between 19.5 and 19.6 Ma (Pickford and Andrews and Walker 1985) and Locherangan, Kenya (Anyonge 1991), is 1981). Above these beds lie the Kapurtay Agglomerates, includ- also referred to this genus. Fossils from Morourot Hill (An- ing Songhor and Chamtwara. The Chamtwara Member has been drews 1978) originally referred to Proconsul major are also now dated to ca. 19.7 Ma (K/Ar), and Songhor correlates stratigraph- included in Afropithecus (Leakey and Walker 1997). ically to just above the Leget Formation (Pickford and Andrews Afropithecus fossils have been recovered from 21 separate sites 1981). Although a slight discrepancy appears to exist in the throughout the Kalodirr Member of the Lothidok Formation. dates, it is within the margin of error and supports a 19.7–19.5 Originally dated to between 16 and 18 Ma on the basis of faunal Ma date for the Tinderet. associations, two tuffs bracketing the Kalodirr Member have The Kisingiri localities consist of fossil sites on Rusinga and now provided K/Ar ages for the Kalodirr Member. The Kalodirr Mfangano Islands in Lake Victoria. At least eight fossil sites tuff, found immediately below the Kalodirr Member has been have been reported on Rusinga Island, including one site that dated to ca. 17.5 Ma and the capping Naserte tuff has been dated has yielded nine individuals of the species P. nyanzae (Walker to ca. 16.8 Ma (Brochetto et al. 1992). Additionally, the beds and Teaford 1988; Walker et al. 1993). Proconsul is known from from Morourot were assigned an age of ca. 17.5 Ma on the basis the uppermost portion of the Kinhera Formation, the entire Ru- of their position between the same radiometric ages (Brochetto singa Agglomerate, and the earliest portion of the Hiwegi For- et al. 1992). mation, which form a conformable sequence of strata (Retallack McDougall and Walthers (1985) reported an age of between et al. 1995). Drake et al. (1988) reported K/Ar dates for tuffs in 17.5 and 17.2 Ma for the hominoid-bearing beds of Buluk. These the Rusinga Agglomerates and the Hiwegi Formations of 17.9 Ϯ 0.1 and 17.8 Ϯ 0.2 Ma respectively. Three sites on Mfangano Is- beds span ten meters of a 120-meter section bracketed by K/Ar land have yielded hominoid fossils (Ward et al. 1993) and cor- dates of 18 and 17.2 Ma, with the fossil-bearing beds just below relate to the Hiwegi Formation on Rusinga Island (Drake et al. the capping basalt. The fossil beds at Locherangan have been 1988). Proconsul material has been recovered from the Napak dated to between 17.5 and 17 Ma (Anyonge 1991). Formation in Uganda (Allbrook and Bishop 1963). This locality The early Miocene hominoid species Heliopithecus leakeyi,rep- was dated to between 18 and 19 Ma (Bishop et al. 1969). resented by a single maxillary fragment and isolated teeth, is Outside of the well-established localities in the Tinderet and grouped by most workers with Afropithecus and Turkanapithecus Kisingiri, Andrews et al. (1981) reported fossils that they as- in the tribe ‘‘Afropithicini’’ (Andrews 1992; Andrews et al. signed to Proconsul from the latest Oligocene/earliest Miocene 1996). The similarity of Heliopithecus to Afropithecus argues for Muhoroni Agglomerates at Meswa Bridge, Kenya. These deposits inclusion of Heliopithecus within the Afropithecus OTU for this its have been dated to 23.5 Ma (Pickford and Andrews 1981; Tas- study. Fossil remains of Heliopithecus have been recovered from sy and Pickford 1983). In addition, several sites in the Ngorora Ad Dabtiyah, Saudi Arabia (Andrews and Martin 1987b). The Formation of the Tugen Hills have produced fragmentary fossil age of Ad Dabtiyah is often cited as approximately 17 Ma (see remains of primates that have been tentatively assigned to the Andrews and Martin 1987b). Analysis of the vertebrate fauna at genus Proconsul. These beds in the Ngorora Formation have the site (Gentry 1987a,b) more conservatively places Ad Dabti- been dated through a variety of techniques (K/Ar, 40Ar/39Ar, yah between Rusinga (just younger than 18 Ma; see Proconsul) paleobotanical correlation) to approximately 12.5 Ma (Deino et and Maboko (younger than 16 Ma; see Equatorius). DATA CONGRUENCE IN HOMINOID PHYLOGENY 649

Turkanapithecus community (Andrews and Alpagut 1990; Bestland 1990). Pas¸a- lar is well studied, and faunal analyses have conventionally The genus Turkanapithecus is known only from the Kalodirr placed the hominoid-bearing unit in the lower subzone of MN locality in Kenya (Leakey and Leakey 1986b) and contains the 6 on the basis of the appearance of rhinocerotids and probos- species Turkanapithecus kalakolensis. Recovered Turkanapithecus cideans and analysis of cricetine rodents (Bernor and Tobien material consists of 20 specimens including a nearly complete 1990). However, Pelaéz-Campomanes and Daams (2002) placed cranium and mandible and several well-persevered postcranial Pas¸alar in later MN6 on the basis of a more complete analysis elements. Dating the Turkanapithecus material follows the dating of the rodent assemblage. The type locality for Griphopithecus,at for the Kalodirr locality, but it was noted by Brochetto et al. Ç andır, had also been faunally correlated to MN 6 (Steininger (1992) that the single Turkanapithecus site ‘‘directly underlies’’ et al. 1996). Comparison of the faunas at these localities suggests the Naserte tuff (see Afropithecus above). that Pas¸alar is slightly older than Ç andır (Bernor and Tobien Equatorius 1990). However, several biostratigraphic and magnetostratigraphic correlations place both the Ç andır and Pas¸alar localities The genus Equatorius was only recently described by Ward et within MN 5, ca. 16.5 Ma (Steininger 1999; Heizmann and Be- al. (1999). By their definition all Kenyapithecus material not as- gun 2001). The European hominoid locality, Klein Haddersdorf, signed to K. wickeri (known only from Fort Ternan, Kenya) is re- also correlates to MN 6, and is believed to be younger than Pas¸- ferred to the species Equatorius africanus. Equatorius is known alar (Begun 1992c; Heizmann and Begun 2001). from several contemporary sites in the Maboko Formation, from Magnetostratigraphic, lithostratigraphic, and biostratigraph- Kipsaramon in the Tugen Hills, and from Nachola in the Sam- ic correlations to before the Langhian transgression, between buru Hills (Ward et al. 1999). The Maboko Formation sites (Ka- 16.5 and 17 Ma indicate that Engelswies is the oldest evidence loma, Ombo, Nyakach, Maboko, Majima) are estimated through for hominoids outside of Africa (Heizmann and Begun 2001). stratigraphic correlations to be between 15 and 16 Ma, on the However, Heizmann and Begun (2001) noted that hominoid ma- basis of their relative positions to radiometrically dated pho- terial from Engelswies is tentatively assigned to Griphopithecus, nolitic units (Andrews 1981; Pickford 1985). The hominoid- and possibly represents a distinct genus. With Ç andır and Pas¸- bearing beds of the Muryur Formation of the Tugen Hills are alar recalibrated to MN 5, the coding of the stratigraphic char- slightly older than 15 Ma (Ward et al. 1999); they lie under the acter for Griphopithecus is identical with or without the inclusion Tiim Phonolite series dated (K/Ar) to between 15 and 13 Ma of the Engelswies material in the OTU. Therefore, the FAE of (Hill et al. 1985). Recent dating (40Ar/39Ar) of the Muryur bed Griphopithecus is older than the western Kenyan Equatorius lo- has produced an age for the BPRP#122 locality of between 15.58 calities, and is penecontemporaneous with the LAE of Afropi- Ma and 15.36 Ma (Behrensmeyer et al. 2002). Specimens origi- thecus. nally attributed to Kenyapithecus africanus from the Aka Aiteputh and Nachola Formations in the Samburu Hills (Nakatsukasa et Sivapithecus al. 1998) were transferred to Equatorius by Ward et al. (1999). These beds have been estimated to ca. 14.5 Ma (Nakatsukasa et The hominoid genus Sivapithecus is known from more than al. 1998). However, Ishida et al. (1999) have placed these fossils 100 fossil localities in the Siwalik molasse of India, Nepal, and in a separate taxon, Nacholopithecus kerioi, and the morphological Pakistan. Although the dentition and mandible are well known distinctness of this taxon from Equatorius was recognized by and the face is represented by a fairly complete specimen (GSP Kelley et al. (2002). Awaiting a detailed analysis of this new tax- 15000 [Raza et al. 1983]), the postcranial anatomy of Sivapithecus on, we did not include Nacholopithecus in this study. is rather poorly known. Several partial long bones and assorted bones of the hand and foot are known; however, no remains of Kenyapithecus the trunk have been recovered (Ward 1997b). The stratigraphy With the separation of Equatorius africanus, Kenyapithecus is of Sivapithecus is particularly well resolved. The commencement here restricted to K. wickeri, known only from the ‘‘B’’ fossil of movement along the Main Central and Main Boundary beds at Fort Ternan (Pickford 1985). The Baraget phonolitic lava Thrust systems of the Himalaya Mountains is responsible for a underlying the fossil-bearing unit has been dated by using K/ continuous flux of sediment into the Siwalik foreland for the last Ar to ca. 15 Ma (Shipman et al. 1981). K/Ar results on micas in 18 million years, producing strata locally in excess of 4 km thick Fort Ternan paleosols give an age of 14 Ma with a hydrothermal with few hiatuses (N. M. Johnson et al. 1982; Opdyke et al. 1982; overprint in the paleosols of 13.5 Ma, implying that these units Sorkhabi and Macfarlene 1999). Because much of the Siwalik se- had been deposited, pedogenically altered, and buried prior to quence has been calibrated to the GPTS (G. D. Johnson et al. this date (Shipman et al. 1981). The Fort Ternan beds correlate 1982), Sivapithecus can be accurately placed into a worldwide stratigraphically higher than Equatorius fossil localities (Wardet chronology. Kappelman et al. (1991) placed the FAE of Sivapi- al. 1999). thecus at the Y750 locality near Chinji, Pakistan, which has been paleomagnetically dated to Chron 5Ar.1 of the GPTS (12.68– Griphopithecus 12.71 Ma [Cande and Kent 1995]). Speculation over the presence of Sivapithecus older than 13 Ma (the GSP 15000 face [Raza et al. Griphopithecus alpani is known from the type locality at Ç andır 1983]) was discounted with a stratigraphic reassessment plac- and from Pas¸alar (both in Turkey) (Andrews et al. 1996). Cur- ing this locality stratigraphically higher than Y750 (Kappelman rently a single species is recognized at Pas¸alar, although differ- et al. 1991). The LAE of Sivapithecus is likely recorded at the GSI- ent tooth morphologies suggest two species are probably pre- D-185 locality near Haritalyangar, India (Patnaik and Cameron sent (Alpagut et al. 1990). Fossils from the European localities 1997). The paleomagnetic results of Johnson et al. (1983) esti- of Engelswies (cf. Griphopithecus) and Klein Haddersdorf, Aus- mate this site to be ca. 7.4 Ma in age, although some disagree- tria, and Neudorf Sandberg (Griphopithecus cf. G. darwini) are in- ment exists over the correlations at Haritalyangar (Kelley and cluded known (Andrews et al. 1996; Kordos 2000). It is not clear Pilbeam 1986). if Griphopithecus material from these localities represents the same taxon as the unnamed Pas¸alar species, and there is some Ankarapithecus debate as to the assignment of the Engelswies material to this genus (see below). Hominoid fossils from the locality of Yassıo¨ ren, Turkey, were It appears that the hominoid-bearing unit at Pas¸alar is the re- originally known from two gnathic fragments with teeth (Ozan- sult of a single flood event, possibly preserving an ecological soy 1965) and a palate with the right zygomatic arch (Andrews 650 JOHN A. FINARELLI AND WILLIAM C. CLYDE

and Tekkaya 1980). Both were found at Locality 8A in the Sinap (de Bonis and Koufos 1993). Therefore, Ouranopithecus is limited Formation and assigned to the taxon Sivapithecus meteai (An- in this study to the three sites for which verifiable stratigraphic drews and Tekkaya 1980). Restoration and reevaluation of the data are available, and it is treated here as ranging from latest palate led to the separation of this material from Sivapithecus MN 10 through MN 11. (Begun and Gu¨leç 1995). In 1996, a face and unassociated mandible were recovered from Locality 12 at Yassıo¨ren (Alpagut et Lufengpithecus al. 1996). Although separated by over a kilometer, localities 8A and 12 correlate stratigraphically to within a few meters of one Lufengpithecus is a late Miocene hominoid from the Shihuiba another and are considered contemporaneous. Paleomagnetic colliery site 9 km north of Lufeng, China. Although Lufengpi- surveys of the Sinap Formation at Yassıo¨ren place the hominoid- thecus lufengensis is the only species formally named for this ge- bearing beds in Chron C5n.1n of the GPTS (Alpagut et al. 1996). nus, some debate exists as to whether larger and smaller spec- Recent paleomagnetic surveys of the Valle`s Penede`s, Spain, cor- imens recovered from this locality represent males and females relate Ankarapithecus as contemporary with the MN 9 Zone for in a highly dimorphic ape (such as Pongo), or if they represent western Europe (Krijgsman et al. 1996; Agustı´ et al. 1998, 2001), two congeneric species (see Kelly and Plavcan 1998). Faunal indicating that Ankarapithecus was a contemporary of both Si- analysis of rhyzomyid rodents at the site are originally pub- vapithecus and Dryopithecus (see below). lished in Chinese (Wu et al. 1986) and are generally quoted as ca. 8 Ma in the English literature (Kelly and Plavcan 1998; Dryopithecus Schwartz 1997), although they may indicate an age as young as 7 Ma (see Badgely et al. 1988). Pilbeam et al. (1996) however, Currently four species are recognized in the genus Dryopithe- correlated Lufeng biostratigraphically to the Dhok Pathan For- cus. Dryopithecus fontani is known from St. Gaudens, France (Be- mation in the Siwaliks, roughly 9 Ma. An age of approximately gun 1994), and from St. Stephan, Austria (Andrews et al. 1996). 8–9 Ma correlates Lufengpithecus to near the European MN Zone In eastern and northern Spain, the taxa D. laietanus and D. cru- 11/12 transition (Harrison et al. 2002; Steininger et al. 1996), safonti have been recovered from seven localities in the Valle`s and thus with the lower (V1) portion of the Oreopithecus range Penede`s and from Seu d’Urgell (‘‘El Firal’’). Dryopithecus brancoi (see below). Additional material from the Xiaohe Formation in is known from several sites in Eastern Europe, including Sal- the Yaunmou Basin correlates biostratigraphically to just older mendigen, Germany (type) and Rudabanya, Hungary (Begun than the Shihuiba locality near Lufeng (Harrison et al. 2002). and Kordos 1993). Reports of hominoid material have been reported from the Yan- The sites in the Valle`s Penede`s have been correlated to Zones gyi locality, although these fossils have yet to be completely de- MN 8 (San Quirze, Can Vila) through MN 10 (La Tarumba) (Be- scribed. Dates for this locality have been reported at roughly 4 gun et al. 1990). Recent paleomagnetic surveys in the Valle`s Pe- Ma on the basis of proboscideans, although these dates are, as nede`s have restricted the correlations of these sections to MN 9 yet, not reliably established (Harrison et al. 2002). As such, the (Chron C5r.1r: 11.1 Ma) to middle MN 10 (ca. 9.2 Ma), with the discussion of Lufengpithecus here will incorporate the Shihuib site of Can Llobateres straddling the MN9/10 transition at ca. and Yaunmou localities for which there is formally described 9.7 Ma (Agustı´ et al. 2001). This is in agreement with the MN 9 material and the stratigraphy is reasonably well constrained. faunal age given to the Seu d’Urgell, which has been correlated to Can Llobateres (Begun 1992a). Salmendigen is correlated to the earliest part of MN 10 (Begun 1994), roughly equivalent to Oreopithecus La Tarumba, whereas Rudabanya, has been correlated to MN 9 The hominoid genus Oreopithecus is represented by a single on the basis of the mollusk and vertebrate faunas (Begun 1992c). species, O. bambolii. Oreopithecus has been recovered from five St. Stephan is correlated to MN 8 (Andrews et al. 1996). The ma- localities in the Maremma Valley in southwestern Tuscany (Az- terial from St. Gaudens presents the most troublesome data for zaroli et al. 1986), a lignite bed in Serrazzano, Italy, and one lo- Dryopithecus. The locality is not well sampled, and the age is cor- cality in Sardinia (Harrison and Rook 1997). Early work divided respondingly not well known. Begun (1992c) places the site at the Maremma localities into three faunal horizons: V1, V2 and latest MN 7 or, more probably, early MN 8. V3 (Delson and Szalay 1985). Oreopithecus material has been recovered from the lower V1 and V2 faunas. Hominoid material Ouranopithecus also has been recovered from the V1/V2 intermediate ‘‘Car- Ouranopithecus (ϭ Graecopithecus [see de Bonis and Koufos dium Horizon’’ at Baccinello (Harrison and Rook 1997). Bacci- 1993 for summary of the debate]) is known from three fossil lo- nello is the only locality where Oreopithecus has been docu- calities in Greece. Two of these localities, Ravin de la Pluie and mented spanning both horizons. The Serrazzano and Sardinia Xirochori, lie in the Nea Messimbria Formation. Multiple max- localities correlate to the V2 horizon of Baccinello (Rook et al. illary and mandibular remains have been recovered from Ravin 1996). de la Pluie, and Xirochori has yielded a distorted face (de Bonis A recent 40Ar/39Ar date of a felsic tuff in the V2 horizon at and Koufos 1993). The third locality lies within the Nikiti For- Baccinello gives an age of 7.55 Ϯ 0.03 Ma (Rook et al. 2000). The mation and has produced a mandible and a female maxilla V1 horizon is known from 150 m below this date, although sed- (Koufos 1995). imentation rates are thought to be high at Baccinello, so the ab- Faunal analysis of Ravin de la Pluie and Xirochori indicate solute time represented by the section is not thought to be ex- that this site is in MN Zone 10 (latest Vallesian Land Mammal tensive. Faunal analysis, specifically the presence of Huerzeler- Age [de Bonis and Koufos 1993]). A paleomagnetic survey con- mys vireti (ϭ ‘‘Valerimys aff. vireti’’ of Huerzeler and Engesser firms this, correlating the overlying formation to Chron 8 and 1976), correlates the V1 and V2 horizons to the early Turolian. MN Zone 11 (Kondopulou et al. 1992). Faunal analysis of the The Cardium Horizon at Baccinello has produced an MN 12 and locality at Nikiti gives an age of MN 10 to MN 11 (Koufos 1995). MN 13 intermediate fauna. No paleomagnetic information has It should be added that an additional specimen from Pyrgos, been published for the Oreopithecus localities. Revised dating of Greece, unearthed during the construction of a swimming pool, the Neogene Mammal Zones (Steininger et al. 1996; Agustıét purports to extend the range of Ouranopithecus to beyond MN al. 2001) agrees with the radiometric age obtained for Baccinello Zone 11 on the basis of the presence of Hipparion. The associa- (Rook et al. 2000). Therefore, a correlation of the V1 horizon tion of the Hipparion fossils, however, has been called into ques- with latest MN 12 and the V2 horizon with earliest MN 13 is tion and the locality was destroyed during local construction used here. DATA CONGRUENCE IN HOMINOID PHYLOGENY 651

Australopithecus early human evolution. Orrorin is known from the Lukeino For- mation of the Tugen Hills and has been dated to ca. 6 Ma (Senut Australopithecus is used here as an OTU to represent hominins et al. 2001). If Orrorin is the direct ancestor to the modern hu- (ϭ human and all direct ancestors [Tattersall et al. 1988]). The man lineage as proposed by Senut et al. (2001), this date would oldest known fossils of this genus are from Kanapoi, Kenya, at- not affect the stratigraphic character. However, if Australopithe- tributed to the species A. anamensis (Leakey et al. 1995a). Ar/Ar cus represents an evolutionary ‘‘side branch,’’ then the use of the ages on bracketing tephra place Kanapoi between 4.17 and 4.12 OTU Australopithecus as a proxy for the human lineage (as in this Ma (Leakey et al. 1999). From this point, the hominin lineage study and Begun et al. 1997) would have to be reexamined care- has a fairly continuous fossil record through the Recent. An old- fully. However, Aiello and Collard (2001) urged caution in ac- er fossil hominin from Aramis, Ethiopia, originally assigned to Orrorin ‘‘Australopithecus’’ ramidus (White et al. 1994) but later renamed cepting the hypothesis that is the direct ancestor to Ardipithecus ramidus (White et al. 1995) has not yet been fully modern Homo, and the use of Australopithecus as a representative described in the literature, although potential evidence of bi- for the hominin lineage is retained here. Sahelanthropus was re- pedality has been noted, indicating a link with later hominins. cently described form the Toros-Menalla locality TM 266, in the The Aramis VP Locality 6 directly overlies the Ga`ala Vitiric Tuff western Djurab Desert of northern Chad. The fossil-bearing an- Complex (White et al. 1994), which has been dated by 40Ar/39Ar thracotheriid unit has been tentatively dated to between 6 and to 4.39 Ma (WoldeGabriel et al. 1994). The FAE of the OTU Aus- 7 Ma on the basis of biochronological correlations with Lukeino tralopithecus can be deﬁned using either A. ramidus or A. ana- and Lothagam (Brunet et al. 2002). As with Orrorin, this age mensis without creating overlap with Oreopithecus. would have no effect on the coding of the stratigraphic character Recent discoveries of potential hominins Orrorin tugenensis if Sahelanthropus represents an early member of the hominin (Senut et al. 2001) and Sahelanthropus tchadensis (Brunet et al. clade. However, a cladistic analysis of the Sahelanthropus mate- 2002) will likely have profound effects on our understanding of rial is, as yet, not possible.