Marshall University Marshall Digital Scholar

Biological Sciences Faculty Research Biological Sciences

Spring 3-2007 The volutE ion of Zinjanthropus boisei Paul J. Constantino Biological Sciences, [email protected]

Bernard A. Wood George Washington University

Follow this and additional works at: http://mds.marshall.edu/bio_sciences_faculty Part of the Biological and Physical Anthropology Commons, Biology Commons, Paleobiology Commons, and the Commons

Recommended Citation Constantino, Paul J. and Wood, Bernard A., "The vE olution of Zinjanthropus boisei" (2007). Biological Sciences Faculty Research. 38. http://mds.marshall.edu/bio_sciences_faculty/38

This Article is brought to you for free and open access by the Biological Sciences at Marshall Digital Scholar. It has been accepted for inclusion in Biological Sciences Faculty Research by an authorized administrator of Marshall Digital Scholar. For more information, please contact [email protected], [email protected]. Evolutionary Anthropology 16:49–62 (2007)

ARTICLES

The Evolution of Zinjanthropus boisei

PAUL CONSTANTINO AND BERNARD WOOD

Many people assume that OH 5, the specimen of boisei, HISTORY OF DISCOVERY collected in 1959, was the first evidence of that taxon to be found, but OH 3, The first evidence of a megadont recovered in 1955, predated the discovery of OH 5 by four . Thus, Para- hominin (that is, a hominin with very nthropus boisei recently celebrated the equivalent of its fiftieth birthday. This large postcanine tooth crowns relative review marks that milestone by examining the way our understanding of this to its estimated body size) from East taxon has changed during its fifty, or so, history. was found at in 1955 (Fig. 1).6,7 Labeled OH 3, for Several hominin taxa have been hominin taxon that began life as Zin- ‘‘Olduvai Hominid 3,’’ it was a speci- established longer or have a larger fos- janthropus boisei, is unusual for several men consisting of two teeth, a decidu- sil record, but Paranthropus boisei, the reasons. First, much of its mor- ous canine and a large deciduous phology is apparently derived1–3 and molar crown (Fig. 2). The huge size of the teeth made them unique among Paul Constantino is a doctoral candidate highly distinctive (Table 1). Second, in hominid paleobiology at the George most of its hypodigm comes from sites East African hominins known at the Washington University in Washington, with good stratigraphic and chrono- time. The of OH 3 therefore DC. His interest in Paranthropus led to remained uncertain until the 1959 re- his dissertation study on how fracture-re- logical control and can be dated with sistant foods influence masticatory mor- relative precision.4,5 Third, the covery of a well-preserved subadult phology and function in and record of Paranthropus boisei is cranium with similar teeth (OH 5) bears. In addition to his laboratory- almost exclusively cranial, consisting that was assigned to Zinjanthropus based morphology work, Paul studies 8 the dietary ecology of in south- mostly of jaws and teeth. This means boisei (Fig. 3). ern Africa and giant pandas in central No to match the OH 5 China. Hominid Paleobiology Doctoral there are reasonably sized, relatively Program, Department of Anthropology, well-dated samples for some morpho- cranium have been found at Olduvai The George Washington University, 2110 logical regions like the and Gorge, but the recovery in 1964 of a G Street NW, Washington, DC 20052. well-preserved, robust-bodied, adult Phone: 202-994-7894, Fax: 202-994- the mandibular dentition. Researchers 6097, E-mail: [email protected] can thus trace the evolution of metri- mandible with megadont postcanine cal and nonmetrical variables across tooth crowns (Peninj 1) from a site on Bernard Wood is University Professor of Origins at the George Washing- hundreds of thousands of years. the western shore of ’s Lake ton University. His main research inter- Finally, all these factors, together with Natron seemed to provide evidence of ests have been the alpha taxonomy and the fifty, or so, years that have elapsed the type of mandible that would be systematics of the hominin fossil record, 9 and he has always had a particular inter- since its discovery, mean that it is pos- compatible with OH 5. Three years est in Paranthropus boisei. His current sible to use Paranthropus boisei as an later, another hominin mandible with preoccupation is how to make detailed postcanine megadontia was recovered data about the hominin fossil record example of how our understanding of more widely available for the purposes a hominin taxon changes over time. from the Omo Shungura Formation 10 of education and research. For example, to what extent has the in southern Ethiopia (Omo 18.18). Center for the Advanced Study of Human Paleobiology, Department of Anthropol- increase in the size of the sample Since 1967, many additional mandi- ogy, The George Washington University, changed our perception of the taxon? bles and isolated teeth similar to these 2110 G Street NW, Washington, DC Have important sample parameters megadont have been collected 20052, USA, and Human Origins Pro- 11–14 gram, National Museum of Natural His- changed significantly as the hypodigm from these same sediments. tory, Smithsonian Institution, Washington, has increased in size? To what extent Beginning in 1968, a series of P. boi- DC 20560, USA. E-mail: bernardawood@ gmail.com can we disentangle the influence of an sei fossils was uncovered near what enlarged fossil record from the bene- was then known as East Rudolf (now fits that have accrued from advances called ) in northern . Key words: Paranthropus; ; Oldu- in analytical methods? Thus, this This series included the recovery in vai Gorge; Koobi Fora; cranial morphology review not only summarizes what we 1969 of a well-preserved but edentu- know of the evolutionary history of lous adult cranium (KNM-ER 406) VC 2007 Wiley-Liss, Inc. Paranthropus boisei, but also traces and the recovery in 1970 of an adult DOI 10.1002/evan.20130 Published online in Wiley InterScience the evolution of our understanding of hemicranium preserving the majority (www.interscience.wiley.com). its evolutionary history. of the vault, the right side of the face, 50 Constantino and Wood ARTICLES

Koobi Fora in the late 1960s and early 1970s resulted in the rate of discovery of P. boisei fossils reaching its peak in the second decade (1966–1975) of the taxon’s history (Fig. 6). After that time, cranial remains belonging to P. boisei continued to be recovered from Koobi Fora and began to be found at West Turkana.1,4,16–18 Two notewor- thy P. boisei-like specimens from West Turkana, a mandible (KNM-WT 16005) and a cranium (KNM-WT 17000), were recovered in sediments dating to ca. 2.5 Ma.16,19 Along with Omo 18.18 and others, these speci- mens have been suggested as possible representatives of a second megadont taxon in East Africa, Paranthropus aethiopicus, which was geologically older than P. boisei. They therefore have particular relevance to the debate about the origin and subse- quent evolution of megadont archaic hominins. Additional evidence of P. boisei has come from sites elsewhere in East Africa, most notably in 1993, when a well-preserved skull of P. boi- sei (KGA 10-525) was recovered in Ethiopia at Konso (formerly called Konso Gardula).20 This skull was not only the geologically youngest known specimen of P. boisei, at 1.4 Ma, but also increased the geographical range of the taxon. Moreover, this was the first time that cranial and mandibular evidence from the same individual had been found in proximity, affirm- ing the presumed associations between the ‘‘robust’’ crania and large jaws made by previous researchers. Also, in 1999, a P. boisei-like maxilla was recovered from Malema, Malawi, thus substantially expanding the Figure 1. Map of P. boisei sites in East Africa highlighting the change in its known geo- southern extent of P. boisei’s range by graphic range over the past five decades. The only new site discovered between 1975 more than 1,000 km.21 and 1985 was West Turkana; no new sites were discovered between 1985 and 1995. The last decade has seen a significant increase in the range of P. boisei due mainly to the discovery of Malema in Malawi, but also Konso in Ethiopia. Adapted from Delson and coworkers.99 FOSSIL EVIDENCE Cranial More than 111 craniodental speci- and part of the right side of the cranial mation at Chesowanja, Kenya also mens are now attributed to P. boisei base (KNM-ER 732) (Fig. 4). Several provided evidence of P. boisei, most stricto; an additional 56 are mandibular specimens have also been notably in the form of a cranium assigned to P. aethiopicus. Only the found at Koobi Fora and elsewhere, discovered in 1970 that preserved most complete and well-preserved of so that numerically it is the best- the right side of the face and the ante- these are reviewed here. sampled region of the P. boisei skele- rior part of the cranial base (KNM- Similarities between the type speci- ton (Fig. 5).1 About the same time that CH 1).15 men (OH 5, Fig. 3) and the large, many of the Koobi Fora fossils were The intensive prospecting in the crested cranium from Koobi Fora being discovered, the Chemoigut For- Omo Shungura Formation and at (KNM-ER 406, Fig. 4) were noted at ARTICLES The Evolution of Zinjanthropus boisei 51

TABLE 1. Some Characteristic and Distinctive Features of Paranthropus boisei Most researchers now agree with this sensu stricto assessment, interpreting these crania as smaller-bodied, presumably female Cranium 1,26,27 – Orthognathic facial profile representatives of P. boisei, thus – Broad mid-face with anteriorly positioned and laterally flaring zygomatic providing evidence of substantial cra- nial within this – High degree of postorbital constriction taxon (Fig. 4). Another cranium re- – Overlap of parietotemporal sutures with marked striae parietales covered from Koobi Fora (KNM-WT – Polymorphic pattern of sagittal and nuchal cresting in presumed males 17400) has lost some of the diagnostic – Anteriorly positioned, heart-shaped foramen magnum morphology of the face, but what Mandible remains of its osseous and dental – Deep and wide corpus with a rounded base morphology leaves little doubt that – Substantial superior and inferior transverse tori at the symphysis this specimen represents another – Tall and wide ramus Dentition example of the small, presumably – Very small anterior teeth relative to the posterior teeth female morph of P. boisei. – Very large posterior teeth relative to skull size – Maxillary premolars having large crowns with a relatively large paracone; 2 or 3 roots – Maxillary molars, M1 < M2 ¼ M3; large crowns with high incidence of a Carabelli’s ; 3 roots Postcranial – Mandibular premolars, P4 crown base area is ca. 140% that of P3; talonid The only sure way for researchers to area is positively allometric relative to total crown area; 2 plate-like roots know what the postcranial hypodigm < < > – Mandibular molars-M1 M2 M3; expanded talonid; usually 6 cusps; of a hominin taxon looks like is to find usually at least one C6; unlikely to have a C7; 2 plate-like roots diagnostic craniodental remains to- – ‘‘Hyper-thick’’ enamel with relatively little enamel decussation on postcanine crowns gether with elements of the postcra- nial skeleton. Unfortunately, there is no compelling evidence of an associ- ated skeleton for P. boisei (but see Grausz and coworkers28). Thus, the the time of the latter’s announce- cluded that the differences between problem researchers have faced at ment.22 Most researchers have con- these two large, presumed male crania Olduvai Gorge, Koobi Fora, and else- are best interpreted as evidence of where is how to tell which of the unas- intraspecific and not interspecific var- sociated hominin postcranial fossils iation.1 The morphology of two other should be assigned to P. boisei and partial crania from Koobi Fora (KNM-ER 13750 and 23000) and of the skull from Konso is also suffi- ciently similar to that of OH5 and KNM-ER 406 to suggest that all five crania belong to P. boisei. Some aspects of their morphology, such as the topography of the ectocranial crests and the face, vary within the presumed male morph of P. boisei, while other regions, such as the cra- nial base, are relatively invari- ant.17,23,24 A calvarium (KNM-ER 407) and a partial hemicranium (KNM-ER 732) differ in both size and shape from the presumed male P. boisei crania. In the case of KNM-ER 407, the contrast in morphology was considered to be so great that its initial taxonomic assess- Figure 3. Early photo of OH 5 in lateral view ment concluded it was ‘‘either a grac- shortly after its reconstruction. The mandi- ile of Australopithecus or else a ble was created based on the known mor- Figure 2. The first evidence of Paranthropus phology of the cranium. Its size and shape 2 very early representative of boisei. The deciduous right M of OH 3 in ... 22 were later affirmed by the discovery of the occlusal (A), alveolar (B), mesial (C), distal .’’ In contrast it was suggested Peninj mandible at Lake Natron, which (D), buccal (E), and lingual (F) views. From that KNM-ER 406 and 732 ‘‘represent closely resembles the one created for OH 5. 25 Tobias.100 the two sexes of the same species.’’ Photo by Bob Campbell. 52 Constantino and Wood ARTICLES

Figure 4. Right anterosuperior view of the presumed male cranium KNM-ER 406 (A) and anterior view of the presumed female cranium KNM-ER 732 (B).

Figure 5. Occlusal views of KNM-ER 729 (A) and KNM-ER 992 (B) casts. KNM-ER 729 has been attributed to P. boisei. Easily visible are its huge molar teeth, molarized premo- lars, and thick corpus. KNM-ER 992 is the type specimen of . Although its postcanine teeth are larger than those of modern , they are clearly much smaller than those of P. boisei. which should be assigned to the con- 82 and 97). These features prompted 1517).31,32 Wood31 argued that the temporaneous species . Day29 to remark that ‘‘it would seem logic that had led to the interpretation The only unassociated hominin reasonable to allocate the new femo- of Olduvai foot fossils (OH 8 and 10) postcranial fossil recovered at Olduvai ral fragment to Australopithecus cf. as modern human-like was flawed and Gorge to be explicitly assigned to P. boisei’’ (Note that Australopithecus suggested that they have as much boisei is a proximal femoral fragment boisei is the same taxon as Paranthro- claim to be attributed to P. boisei as (OH 20, Fig. 7).29 A detailed analysis pus boisei. The reason for the contin- they have to H. habilis. of OH 20 suggested that it shares a ued disagreement about whether to At approximately the same time suite of features with two proximal attribute these fossils to Australopithe- that these reassessments of the Oldu- femora attributed to P. robustus (SK cus or Paranthropus will be discussed later). At that time, no one knew what an H. habilis femur looked like, so it seemed logical to link a P. robustus- like proximal femur with P. boisei, as well as to link what were then inter- preted as the more modern human- like Olduvai Bed I and II postcranial specimens (for example, OH 8 and 10) with H. habilis. The discovery of a modern human- like ankle (the talus KNM-ER Figure 6. The number of P. boisei speci- 813) at Koobi Fora30 weakened the mens discovered each decade for the case for assuming that the OH 8 foot past 50 years. The large number of speci- mens discovered between 1965 and 1975 belonged to H. habilis, since the OH 8 was mostly due to the beginning of field talus more closely resembles a talus Figure 7. The proximal left femur OH 20 in work at Koobi Fora and the Omo. assigned to P. robustus (TM anterior (A) and posterior (B) views. ARTICLES The Evolution of Zinjanthropus boisei 53 vai hominin postcranial fossils were SYSTEMATICS include the pre-2.3 Ma fossils, among taking place, Richard Leakey and his them Omo 18.18, L338y-6, and KNM- Upon the discovery of OH 5, Louis team had begun to recover hominin WT 17000, then P. aethiopicus Leakey8 drew attention to twenty dif- postcranial remains from Koobi Fora. becomes a junior synonym of P. boisei ferences between this cranium and Most were femoral specimens; provi- sensu lato. The point at issue is crania already attributed to Australo- sional comparisons suggested that whether the differences between pithecus and Paranthropus, which at they could be divided into those that KNM-WT 17000 and the < 2.3 Ma cra- the time were only known from south- were more like the femora of modern nia belonging to the P. boisei sensu ern Africa. The following year Robin- humans and those that were ‘‘not stricto hypodigm (for example, OH 5, son34 went through Leakey’s claimed unlike other femoral fragments that KNM-ER 406) justify a specific dis- ‘‘major differences’’ and, to the for- have been collected elsewhere and tinction for the West Turkana cra- mer’s satisfaction, refuted the vast assigned to Australopithecus.’’25 How- nium and the early megadont jaws majority. Robinson suggested that the ever, the attributions of the more and teeth attributed to P. aethiopi- name Zinjanthropus be aban- primitive specimens to Australopithe- cus.10 Two studies have looked at this doned and proposed that OH 5 be cus and, by inference, to Australopi- problem in detail. Suwa35 concluded included within the existing genus thecus cf. boisei, depended on the that there are differences in mandibu- Paranthropus. untested assumption that the femora lar premolar cusp morphology of H. habilis were recognizably more between the pre-2.3 Ma and the post- modern human-like than were the How Many Species? 2.3 Ma megadont fossil evidence, with femora of P. boisei. The subsequent the later material having larger and There are currently three main discovery of an H. habilis partial skel- more elaborate talonids. Wood, debates about the taxonomy of P. boi- eton (OH 62) challenged this assump- Wood, and Konigsberg23 found that sei. The first focuses on whether tion. The partial skeleton is poorly several features of the mandible and there are sufficient differences preserved and its interpretation re- the mandibular dentition (other than mains controversial. Nevertheless, its the premolar morphology referred to femoral morphology is evidently simi- by Suwa35) also change at ca. 2.3 Ma. lar enough to that of femora from ... These authors supported the interpre- Olduvai and Koobi Fora that had we are badly in need tation that the early stage of the Para- been assigned to Australopithecus and of an associated nthropus lineage in East Africa should thence to P. boisei33 to make these lat- be recognized as a different species. ter attributions suspect. Conse- skeleton that includes The main differences between the two quently, there is no current way of tell- cranial elements East African Paranthropus species are ing to which taxon or taxa the diagnostic of P. boisei. the greater facial prognathism, larger ‘‘Australopithecus-like’’ hominin post- incisors, flatter cranial base, smaller cranial evidence from Olduvai and mandibular corpus, and shorter post- Koobi Fora belongs. For the time canine tooth row of P. aethiopicus.23 being, it would be prudent to regard Tobias36 cogently made the case for this fossil evidence as gen. between P. boisei and P. aethiopicus to distinguishing between P. boisei and et sp. indet. justify retaining P. aethiopicus as a P. robustus and, until recently, the Attempts to identify the postcranial separate species. The second asks enlargement of the two hypodigms skeleton of P. boisei at sites other than whether recent discoveries have had not materially altered that assess- Olduvai and Koobi Fora have been ad blurred the distinction between P. boi- ment. However, the discovery of a hoc, and none has been particularly sei and the megadont taxon from megadont hominin skull (KGA 10- convincing. So where does that leave southern Africa, Paranthropus robus- 525) at Konso prompted Suwa and us? In short, we are badly in need of tus. The third debate concerns coworkers20 to suggest that some an associated skeleton that includes whether the differences in size and aspects of its morphology are com- cranial elements diagnostic of P. boi- shape subsumed within P. boisei are mon to both P. boisei and P. robustus. sei. Finding archaic-looking femora is consistent with a sexually dimorphic This, together with the recovery of not enough because at least one other early hominin taxon or whether they additional cranial material from synchronic East African hominin, H. are an indication that even P. boisei Koobi Fora (for example, KNM-ER habilis, exhibits similar morphology. sensu stricto subsumes more than one 23000)17 and the publication of the It may be that some of the hominin taxon. detailed analysis of a cranium from postcranial specimens assigned to If the hypodigm of P. boisei is re- the Omo (Omo 323–896)14 prompted H. habilis actually belong to P. boisei stricted to the post-2.3 Ma fossil re- at least one commentator to suggest but, for the time being, we have no cord, then the name P. boisei sensu that Paranthropus taxonomy should way of telling. We suggest it is better stricto should be retained for the main be reassessed.37 In addition, the re- to accept that for various reasons no hypodigm and P. aethiopicus should covery of a large, ‘‘P. boisei size’’ molar postcranial remains can be confi- be used as the species name for the from Gondolin38 suggests that it dently assigned to P. boisei than to earlier fossil evidence. However, if the would be worthwhile to consider continue with the present confusion. P. boisei hypodigm is judged to whether the recent discoveries in 54 Constantino and Wood ARTICLES

Figure 8. Adaptations of two proposed phylogenies for Plio- hominin evolution with the ‘‘robust’’ australopiths highlighted in bold. The phylogeny by Skelton and McHenry48 shows the ‘‘robust’’ australopiths as a polyphyletic group descended from two different ancestors, with many of the morphological features that are shared between Au. (P.) aethiopicus and the Au. (P.)boisei/Au. P. robustus clade having evolved through parallel or convergent evolution. The phylogeny by Strait, Grine, and Moniz46 shows the ‘‘robust’’ austral- opiths as monophyletic (all evolving from the same recent common ancestor). The length of the lines is arbitrary and not meant to be reflective of time. In the phylogeny by Strait, Grine, and Moniz,46 ‘‘Au. afarensis’’ is in quotes because the authors designated a new taxon name for members of this group. ‘‘Au. afarensis’’ is retained here for clarity of comparison with the Skelton and McHenry phylog- eny. Also for ease of comparison, additional taxa such as platyops and Au. garhi are not included.

southern Africa at this site and at Dri- Fig. 2E), with both small and large nize the hypothesis of ‘‘robust’’ mono- molen39 may have closed the morpho- mandibles in the sample retaining phyly? First, the average confidence logical gap between P. robustus and P. their characteristically robust corpus interval for hominin cladistic analyses boisei. However, recent studies and rounded base.1,23 of ca. 0.65 means that approximately addressing these issues have found no 35% of the characters used in the support for taxonomic revision based Paranthropus analyses must have been independ- on the current evidence. Wood and Is Monophyletic? ently acquired; that is, they are homo- 24 Lieberman concluded that ‘‘the Regardless of whether one or two plasies. If these homoplasies are con- Konso specimens fit within the popu- Paranthropus species are recognized centrated in one anatomical region lation parameters of P. boisei pre- in East Africa, dental metrical evi- such as the skull, and if that region dicted by the ‘pre-Konso’ hypodigm. dence still indicates that P. boisei is has a major influence on the shape of 40 When Constantino and Wood com- distinct from P. robustus.40 But did the cladograms, then an analysis of pared the regional hypodigms of Par- the East and southern African mega- the preserved morphology may not anthropus before and after addition of dont taxa evolve from a recent com- result in an accurate reconstruction of the new material from Drimolen and mon ancestor exclusive to themselves evolutionary relationships. Second, Gondolin, they found that the number and thus form a monophyletic group, many of the characters that link Para- of significant metrical differences be- or did the various regional Paranthro- nthropus taxa are related to the masti- tween the postcanine dentition from pus taxa evolve independently (Fig. 8)? catory system. For example, when 47 East and southern Africa increased In 1988 Wood44 reviewed fifteen Wood and Chamberlain organized rather than decreased. hominin cladistic studies, all of which characters according to anatomical As for whether the P. boisei hypo- concluded that the two regional var- region, they found that support for a digm samples more than one hominin iants of Paranthropus are sister taxa. Paranthropus clade relied heavily on taxon, it has been suggested that the An independent review published in characters from the face, palate, and degree of size variation in the mandib- the same year also found that one mandible. These regions all reflect ular hypodigm of P. boisei is excep- of the few reliable parts of the homi- masticatory adaptations and are thus 41,42 tional. However, much of this ‘‘ex- nin cladogram is the Paranthropus likely to be functionally integrated. cessive’’ variation can be explained by clade.45 The major cladistic study by Therefore, the characters derived post mortem cracks that fill with ma- Strait, Grine, and Moniz46 also found from those regions are potentially trix and artificially inflate the size of that the most parsimonious clado- ‘‘non-independent’’ and if so, should the mandibular corpus of larger indi- grams support Paranthropus mono- not be coded as individual characters viduals, while erosion of surface bone phyly. As part of their comprehensive in a cladistic analysis. Skelton and 48 has reduced the size of the corpus of morphological analysis of the cranial McHenry reached a comparable some of the smaller individuals in the remains of Au. afarensis, Kimbel, Rak, conclusion. Evidence from other 1,43 hypodigm. Apart from these ex- and Johanson3 also found consistent groups of (see, for example, 49 50,51 trinsic causes of differences in overall support for a ‘‘robust’’ australopith clade. Maglio and Vrba ) also suggests size, the size and shape of the mandib- Given the near unanimity of the that the masticatory system might be ular corpus of P. boisei is remarkably conclusions of these studies, what rea- the equivalent of a ‘‘homoplasy 23 stable through geological time (see sons are there to continue to scruti- ghetto.’’ It should be noted, however, ARTICLES The Evolution of Zinjanthropus boisei 55 that even when Strait, Grine, and What about the evolutionary rela- that the currently accepted temporal Moniz46 excluded masticatory charac- tionships between Paranthropus and range of ca. 1.2 Ma is likely to be close ters from one of their cladistic analy- nonrobust taxa? A recent attempt to to the true temporal range. ses they still found strong support for use morphological evidence to reeval- a Paranthropus clade. uate the phylogenetic relationships of Paleohabitat A third argument for questioning Au. afarensis has revealed several sim- Paranthropus monophyly is that a ilarities in cranial morphology There have been various interpreta- 52 study that used Patterson’s similar- between that taxon and P. boisei sensu tions of the habitat preferences of Par- ity and ontogenetic criteria to test the lato,3 but failed to support a direct anthropus in East Africa. Shipman hypothesis of Paranthropus mono- phyletic link between these hominins. 55 44 and Harris concluded that P. boisei phyly with respect to mandibular Interestingly, another hominin with sensu lato probably preferred closed molar cusp morphology and mandib- postcanine megadontia, Australopi- and wet habitats, while Reed56 sug- ular premolar root form produced thecus garhi, has been recovered from gested that these hominins lived in mixed results in terms of supporting ca. 2.5 Ma sediments in the Middle more open environments such as eda- or falsifying the hypothesis. In a fur- Awash of Ethiopia.54 While it does not phic grasslands. Some P. boisei sites ther examination of Paranthropus seem to be a member of the Para- not included in these analyses, such postcanine cusp morphology, Suwa, nthropus clade based on other aspects as Peninj and Chesowanja, do not Wood, and White53 noted that ‘‘the of its craniofacial morphology, it is have detailed paleoenvironmental individual cups involved in the talonid data available for the relevant strati- expansion are not always the same,’’ too early to know how it is related to other hominins. Finally, most graphic levels, but habitats have been with the hypoconid and the entoconid reconstructed at the more recently contributing more to talonid expan- researchers accept that the derived discovered sites of Konso and Mal- sion in P. boisei and P. robustus, res- morphology shared between Para- ema. The nine Konso specimens are pectively. This suggests that the de- nthropus and Homo, such as cranial associated with a ‘‘predominantly dry rived dental morphology in the two base flexion, either evolved independ- grassland fauna’’20 close to a paleo- Paranthropus taxa may not have the ently in the two lineages or was inher- lake,57 with no P. boisei fossils coming same developmental basis. ited from an as-yet undiscovered from the more mesic (moderately The question of Paranthropus mo- recent common ancestor of the Para- wet) localities. The faunal assemblage nophyly is therefore unresolved. nthropus and Homo clades. The dis- found with the P. boisei maxillary Future research will have to deter- tinctness of other aspects of their fragment21 at Malema is sparse and mine whether the shared skull mor- morphology leaves little doubt that H. highly biased, but it is dominated by phology of East and southern African habilis and P. boisei are at least sepa- open-habitat mammals including Paranthropus is due to common an- rate species, probably belonging to dif- alcelaphines, antilopines, Hipparion, cestry or convergent evolution. If one ferent genera. Thus, the existing fossil and the pig Notochoerus scotti, sug- is sanguine that hard-tissue morphol- evidence suggests that the genus Para- gesting that P. boisei was preserved in ogy is capable of recovering phyloge- nthropus went extinct without contrib- a relatively open environment at that netic relationships established on the uting to the evolution of later hominins. site as well. The results of earlier basis of independent genetic evidence, 58 work also indicate the proximity of a then Paranthropus monophyly must large paleolake at Malema. Thus, at be the hypothesis of choice. But if one CONTEXT both Konso and Malema, the evidence is more skeptical about its ability to Geographical and supports the assertion that P. boisei do so, then what others interpret as Temporal Range favored open habitats near permanent overwhelming evidence for Paranthro- water, as advocated by Reed.56 pus monophyly looks less compelling. Paranthropus boisei sensu lato (P. This has taxonomic implications since boisei sensu stricto plus P. aethiopicus) BEHAVIOR the genus name Paranthropus is only is currently known from eight sites in appropriate if these taxa form a mo- East Africa and spans a geological age Given the lack of postcranial bones nophyletic group. Otherwise, they range of ca. 2.6–1.4 Ma (Figs. 1 and 9). that can be confidently attributed to P. cannot be included in their own sepa- With the exception of the Omo, no boisei, there is not much that can be rate genus and should be included as major East African sites are known to said of this taxon’s dexterity, posture, members of the genus Australopithe- have fossiliferous deposits dated to or locomotion except that data on the cus. Because the majority of phyloge- between ca. 1.4 and 1.0 Ma or between relative position of the foramen mag- netic analyses currently support Para- 3.0 and 2.5 Ma. Therefore, it could be num67,68 indicate that the habitual nthropus monophyly, we suggest that argued that there is a gap of half a mil- posture of P. boisei was similar to that until the evidence demonstrates oth- lion years on each side of the temporal of modern humans. This section will erwise, the genus name Paranthropus range of P. boisei sensu lato. We do therefore focus on the evidence we should be used to recognize the strong not know how far into those gaps Par- have regarding diet and the function possibility that megadont taxa in both anthropus fossils extend, but the appa- of the masticatory system, followed by regions form an adaptively distinctive rent absence of Paranthropus fossils a brief discussion of attempts to infer and monophyletic group. in the older Omo sediments suggests social structure from morphology. 56 Constantino and Wood ARTICLES

Figure 9. Stratigraphic location of the P. boisei hypodigm. Shaded areas indicate levels where P. boisei fossils have been found. The spe- cific positions of key specimens mentioned in the text are shown. The oldest known specimens of P. aethiopicus are from the Omo and are dated to ca. 2.6 Ma, while the oldest fossils of P. boisei sensu stricto are from Malema and are approximately 2.3 Ma, based on fau- nal correlations with the Omo. The specimens from Konso are the youngest known for P. boisei, at ca. 1.4 M. Diet ering the predominance of cranial, for diet to provide information gnathic, and dental evidence in its regarding the divergence between There has not been much success hypodigm, the likely role diet played Paranthropus and early Homo. We in determining the diet of P. boisei. in the evolution of its derived masti- will review what has been done to This is somewhat surprising consid- catory morphology, and the potential understand P. boisei’s diet in terms ARTICLES The Evolution of Zinjanthropus boisei 57

roots, and rhizomes made up a signifi- 72,73 BOX 1: Was Paranthropus boisei a maker and cant proportion of P. boisei’s diet. user of stone tools? All post-4 Ma hominins have rela- tively thick enamel but, as with post- When the OH 5 cranium was discov- boisei to the sidelines of cultural canine crown area, P. boisei was even ered at FLK I, it was assumed to be evolution. However, a few have more derived along this morphocline the manufacturer of the stone tools entertained the possibility that than were other hominins; its enamel found on the ‘‘living floor.’’8 How- members of more than one hominin has been described as ‘‘hyper-thick’’74 ever, the career of P. boisei as the lineage may have had the ability to (p. 33). The functional role of thick maker of the stone tools manufacture simple stone tools.60– enamel is still unclear, but it has been ended just five years later, when 64 The truth is that there is no direct suggested that it is part of a strategy was forced to consider evidence linking any of these early to resist wear and/or withstand high the implications of the subsequent hominin species with stone tools. occlusal loads caused by abrasive or discovery of Homo habilis fossils in No hand bones can be confidently hard foods. Gantt and Rafter75 sug- association with the Oldowan cul- assigned to either P. aethiopicus or gested that the hyper-thick enamel of ture at three other localities at Oldu- P. boisei, so we do not know whether P. boisei was linked with ‘‘increased vai. He concluded that ‘‘while it is these taxa were dexterous enough to crushing and grinding and adaptation possible that Zinjanthropus and make or use stone tools. Hand bones to savanna habitat.’’ While few would Homo habilis both made stone tools, tentatively attributed to P. robustus disagree with this statement, there is it is probable that the latter was the in southern Africa reportedly indi- still no consensus as to what it was that more advanced tool maker and that cate that these hominins were capa- P. boisei was ‘‘crushing and grinding.’’ the Zinjanthropus skull represents an ble of a human-like precision grip.65 intruder (or a victim) on a Homo Both stone and bone tools have been Masticatory biomechanics habilis living site.’’59 This conclusion found in loose association with P. The cranial attachments for the mass- was reached not because any evi- 61 robustus at and Sterk- eter muscles are more anteriorly posi- dence suggested that P. boisei could 66 fontein. However, even if these an- tioned in P. boisei than they are in not be the toolmaker, but because H. atomical and archeological associa- other early hominins76 and, to judge habilis, with its larger brain and tions prove to be sound, it is unclear from the size of the ectocranial crests, apparently modern human-like hand what the implications would be for zygomatic bones, and other muscle bones, seemed a more likely tool- attachment areas, P. boisei probably maker than did P. boisei. P. boisei. Thus far, there is no firm had larger masticatory muscles as Since 1964, virtually all research- evidence that P. boisei made and well (Fig. 10).77 However, it is unclear ers have come to agree that H. used stone tools, but there is also no whether these larger masticatory habilis made stone tools. Most of firm evidence that members of this muscles would have resulted in higher these researchers have relegated P. taxon were incapable of doing so.

of dental morphology, masticatory The very large, bunodont postca- biomechanics, dental microwear, nine tooth crowns of P. boisei may and chemical analyses of bones and have been an adaptation to disperse teeth. high occlusal loads or simply to increase the surface area over which food could be processed at any one time. Lucas, Corlett, and Luke,70 Dental morphology showed that the ratio of the area of M1 to M3 (both upper and lower) in The small incisors of P. boisei appear to primates was inversely related to the Figure 10. KNM-ER 406 (A) and KNM-ER indicate a diet that did not require a sig- amount of leaves consumed and sug- 3733 (B) in superior view. The large and lat- nificant amount of incisal preparation erally flaring zygomatic arches of KNM-ER gested that the low M1:M3 ratio of 406 can be seen clearly. Coupled with the such as one consisting of leaves or ber- Paranthropus indicates that they were 69 shorter anteroposterior length of the cra- ries. Therefore, if P. boisei was eating ‘‘probably consuming small mouth- nium, this gives P. boisei a much more cir- fruit, then the fruit either lacked thick fuls of leaves and seeds.’’ However, the cular cranial circumference when viewed husks or fleshy pulp or was small low shearing crests71 and rounded from this angle. Also apparent in this pho- enough to need little preparation before cusps of these teeth suggest that a diet tograph are the high degree of postorbital constriction and sagittal crest develop- mastication with the postcanine teeth. high in fibrous leaves or grasses was ment in P. boisei, both related to its smaller However, the possibility that foods were unlikely. Nevertheless, it is still possi- relative . One can also see evi- being prepared outside of the mouth ble that seeds and plant underground dence of the convergence in facial must be considered (see Box 1). storage organs such as tubers, bulbs, orthognathy between P. boisei and Homo. 58 Constantino and Wood ARTICLES occlusal forces. As Demes and Creel78 that P. boisei had lower Sr:Ca ratios collectively termed its life history. have pointed out (and as Walker79 than did early Homo. This is broadly Knowledge of the pattern and timing pointed out earlier for P. robustus), P. consistent with later analyses of P. of an organism’s life history can reveal boisei would have been able to gener- robustus from Swartkrans86 and sug- information about its strategies for ate exceptionally high bite forces rela- gests that Paranthropus may have survival and reproduction. Unfortu- tive to those of other hominins, but been consuming more meat than did nately, little is known about the life the occlusal forces at the molars the earliest members of our own ge- history of P. boisei. Here we discuss would not have been exceptional if nus. However, the results of the Boaz why neither life-history variables they were distributed over the entire and Hampel study need to be inter- (ones directly related to life history) occlusal surface. Only if P. boisei was preted with caution due to the small nor life-history-related variables (ones feeding on limited numbers of small samples used for early Homo. indirectly related to life history) are objects at any one time would increased currently able to tell us much about force on any single object result. life history in this taxon. Since, under certain circumstances, Social Structure P. boisei could generate significantly Plavcan87 has been at the forefront higher bite forces than could other of efforts to predict the social struc- Life-History Variables hominins, it is not surprising that they ture of extinct taxa through compari- also appear to have had the ability to sons of sexual dimorphism in body Age at weaning is an indicator of 80 relative offspring dependence and withstand high bite forces. Hylander size and canine crown size. Unfortu- maternal investment. Because lacta- has suggested that the thick mandibu- nately, we currently lack reliable ways tion suppresses ovulation in great lar symphysis and the deep, thick of determining the sex of most early mandibular corpus of P. boisei, which apes and humans, it is also a determi- hominins but, as mentioned earlier, is approximately three times larger nant of interbirth interval. Among liv- the pattern of craniodental size than would be expected in a general- ing primates, weaning appears to dimorphism apparent in P. boisei sug- 89 ized hominoid of the same body coincide with M1 eruption, but gests the presence of distinctive male mass,81 allowed these creatures to within the hominin clade some evi- and female morphs. These morphs resist high stresses caused by masti- dence suggests that weaning may pre- are somewhat unusual, however, in 90 cating very hard or tough food items. date M1 emergence. This is certainly that they both possess relatively small the case for modern human groups canines, despite their evident cranio- who wean infants at about the age of dental and inferred body size differen- 2.5 years, whereas the first permanent Dental microwear ces. Therefore, even if we assume that mandibular molar does not erupt Thus far, no study of P. boisei dental the larger specimens are males and until children are approximately 6 microwear has been reported in the the smaller specimens are females, years old. Determination of age at literature. This is due, at least in part, there are no modern higher weaning in fossil hominins has thus to the fact that most P. boisei fossils analogues that show a comparable been based on an assessment of the are found on the surface, and surface pattern of within-species variation. degree and timing of the deciduous specimens typically exhibit a substan- Inferences could be made about dental attrition associated with die- tial proportion of nondiet-related the social structure of P. boisei by tary supplementation. Aiello, Mont- microwear due to erosion, weather- atomizing its morphology and then gomery, and Dean91 showed that ing, and trampling.82 Although the comparing the predictions from each specimens of P. boisei and P. robustus analysis of dental microwear has pro- of the components (canine crown judged to be between 2.5 and 3.5 years ven to be a valuable technique for height, canine crown buccolingual of age exhibit high levels of deciduous understanding diet in the megadont width, body size, and so on), but dental attrition compared to specimens hominins from southern Africa,83,84 it Plavcan87 suggested that the relation- of Au. afarensis. However, these authors is unclear what these results say, if ships among these variables in living acknowledged that this difference could anything, about the diet of P. boisei. higher primates are ‘‘not strong be related to either an earlier age at enough to make detailed inferences weaning or to differences in diet. about mating systems or behavior on the basis of dimorphism alone.’’ In a Chemical analysis 88 later paper, Plavcan cautions Life-History-Related Variables The chemical analysis of bones and against assuming that any living teeth is another method that has sig- hominoid is a suitable analogue for Body size nificantly increased our understand- early hominins with respect to pre- There are strong correlations between ing of hominin diet but, as with dental dicting social structure. body size and life-history variables microwear analysis, it has not yet such as gestation length, weaning age, been used in an in-depth study of the age at first reproduction, interbirth LIFE HISTORY diet of P. boisei. One of the only chem- interval, and maximum life span ical studies that included P. boisei was The stages through which an indi- across subfamilies of primates.92 by Boaz and Hampel,85 who found vidual passes during its lifetime are Researchers who have used cranial ARTICLES The Evolution of Zinjanthropus boisei 59

Figure 11. The effect of additional specimens on three P. boisei dentognathic variables including (A) P4 buccolingual breadth, (B) M3 buccolingual breadth, and (C) mandibular corpus area (h w) at M1. Variables were chosen based on the number of available fossil specimens. For the three variables selected, the median values stabilize after approximately 7 to 15 specimens have been assessed. The dark horizontal line across each box is the median value; the box itself represents the interquartile range (middle 50%); the lines (whiskers) show the complete range (excluding outliers). [Color figure can be viewed in the online issue, which is available at www. interscience.wiley.com.]

variables as proxies for body mass93,94 derived from brain volume, and brain differences between P. boisei and have concluded that the estimated volume can be derived from endocra- closely related taxa. body mass of P. boisei is similar to nial volume if allowance is made for that of archaic hominin taxa in the ge- the space occupied by the endocranial nus Australopithecus. Thus, differen- vasculature and the intracranial extra- ces in body mass cannot be used to cerebral cerebrospinal fluid. The Dental ontogeny account for any differences in predic- mean of the ten endocranial volumes In some respects, such as incisor tions about the life history of P. boisei in the P. boisei sample is 481 cm3, with crown formation and eruption and other archaic hominins. a range of 400–545 cm3. This range sequence, dental development in P. substantially overlaps with that of boisei resembles that of modern extant nonhominin higher primates humans, while in others, such as the Brain size and that of archaic hominins, includ- rate of root formation, it resembles Brain mass has also been shown to be ing H. habilis sensu stricto.95 Like that of . In still other highly correlated with many life-his- body size, therefore, brain size is not respects, dental development in P. boi- tory variables.92 Brain mass can be very useful in elucidating life-history sei is unique (for example, enamel for- 60 Constantino and Wood ARTICLES mation in the postcanine teeth). The taxon should encourage field re- same general grade of archaic homi- premolar and molar crowns of P. boi- searchers and funding agencies to nins as do taxa presently included in sei take the same time, or less, to form recognize that even relatively modest- the genera Australopithecus and Keny- than do those of modern humans and sized samples can provide useful in- anthropus, albeit as the most derived chimps, despite the fact that P. boisei formation about the original popula- member of that grade grouping. Two crowns are approximately twice the tion from which the fossil sample was judgments about the phylogenetic size of human and chimp crowns. drawn. Larger samples are still impor- relationships of P. boisei have fared This is due to a combination of more tant, however. As this review has high- less well. First, the enlargement of the enamel secretion per day by amelo- lighted, larger samples of P. boisei fos- hypodigm has effectively falsified blasts and a faster rate of ameloblast sils have increased our understanding Louis Leakey’s claim that OH 5 activation.96 We need more informa- about morphological character com- ‘‘differs from both Australopithecus tion before we can determine whether binations, the geographic and tempo- and Paranthropus much more than these two genera differ from each these differences are due to selection ral range of the taxon, and variation in nonmetrical traits such as the other.’’8 Because phylogenetic analy- operating on life history, on diet, or patterns of ectocranial cresting. ses seldom, if ever, put H. habilis on a combination of the two. sensu stricto and P. boisei in the same clade, Phillip Tobias’36 suggestion that it is ‘‘unlikely’’ that P. boisei and H. TRACKING MORPHOLOGICAL The premolar and molar VARIABLES OVER TIME habilis ‘‘were genetically isolated’’ (p. crowns of P. boisei take 244) can be refuted. Most contributions to the paleoan- This review highlights how much thropological literature either focus the same time, or less, to remains to be learned about most on the latest tranche of fossil discov- form than do those of aspects of P. boisei’s paleobiology, eries or look in depth at the fossil evi- including its diet, life history, tool use, dence from a single site or from a par- modern humans and temporal and geographical range, phy- ticular region or body part. Relatively chimps, despite the fact logenetic relationships, and even what few studies focus attention on the that P. boisei crowns are it looked like from the neck down. So accumulated hypodigm of a single how can we improve our understand- 3,97 early hominin taxon. approximately twice the ing of the paleobiology of P. boisei? Mandibular and dental remains are size of human and The first priority is to find at least one, particularly well represented in the and preferably several, taxonomically hypodigm of P. boisei and for some chimp crowns. This is unambiguous associated skeletons. > parameters the sample sizes (N 20) due to a combination of This would help researchers work out are respectable, at least by paleonto- which limbs go with which heads at logical standards. For some variables, more enamel secretion Olduvai and Koobi Fora. The second therefore, it is possible to investigate per day by ameloblasts priority is to extract more information how much the sample parameters of about the functional morphology of P. P. boisei have changed as the sample and a faster rate of boisei’s masticatory system, and thus sizes have increased. We track the pa- about its diet, from the fossil record. rameters of three such variables (Fig. ameloblast activation. Third, we need to clarify the evolution- 11). These variables have been chosen ary relationships between P. boisei and at random to the extent that they hap- the other ‘‘robust’’ taxa. Fourth on the pen to be the ones represented by the list would be a way of teasing out what largest number of specimens. The pa- rameters of all three variables are CONCLUSIONS AND PROSPECTS aspects of dental growth and develop- ment are diet-related and what aspects remarkably consistent, and other vari- Half a century of collection and carry a signal about life history. Finally, ables with smaller hypodigms show analysis has vindicated some of the comparable patterns. The two largest and perhaps the trickiest of all, is the initial assessments of the taxonomy of task of finding ways of establishing samples are for M3 buccolingual P. boisei by Leakey8 and Tobias.36 For breadth and mandibular corpus area 8 whether, and to what extent, P. boisei example, two judgments of Leakey at M . There is very little change in the was a cultural . Did its cultural 1 appear to have been supported by the sample parameters once the hypo- repertoire go beyond that of a modern new evidence. First, only Robinson34 digm reaches N ¼ ca. 10. These data ? If so, how did its culture contradict the notion that paleoan- has seriously doubted the wisdom of differ from that of the other less mega- thropological samples are generally establishing a new taxon for the P. dont archaic hominins? too small to support claims of mor- boisei hypodigm. Second, although The rate of discovery is on the phological distinctiveness. Richard support for recognizing P. boisei decline for P. boisei fossils (Fig. 6), Smith has provided a chastening dem- and its ilk at the level of a separate probably due mainly to changes in onstration that this is the case for very ‘‘sub-family Australopithecinae’’8 has research priorities within paleoanthro- small samples,98 but our post-hoc waned, few would demur from the pology. Nevertheless, there is much to investigation of one early hominin judgment that P. boisei belongs to the learn about this intriguing taxon, and ARTICLES The Evolution of Zinjanthropus boisei 61 we need both additional fossils and clusions from the Omo research expedition’s 34 Robinson JT. 1960. The affinities of the new new analytical methods and strategies studies, ed. Konigsson LK. Oxford, Pergamon. Olduvai . 186:456– 13 Coppens Y, Sakka M. 1983. Un nouveau crane 458. in order to move forward. d’australopitheque. Evolutive Morphogenese du 35 Suwa G. 1988. Evolution of the ‘‘robust’’ Cranie et Anthropogenese. Paris, CNRS. p 185– in the Omo succession: evi- dence from mandibular premolar morphology. ACKNOWLEDGMENTS 194. 14 Alemseged ZY, Coppens Y, Geraads D. 2002. In: Grine FE, editor. Evolutionary history of the ‘‘robust’’ australopithecines. New York: Aldine PC is supported by an NSF IGERT Hominid cranium from Omo: description and taxonomy of Omo-323-1976-896. Am J Phys de Gruyter. p 199–222. grant. BW is grateful to Richard Lea- Anthropol 117:103–112. 36 Tobias PV. 1967. Olduvai Gorge, vol. 2. The key for giving him the opportunity to 15 Gowlett JAJ, Harris JWK, Walton D, Wood cranium and maxillary dentition of Australopi- get in at the ‘‘ground floor’’ in terms of BA. 1981. Early archaeological sites, hominid thecus (Zinjanthropus) boisei. Cambridge: Cam- remains and traces of fire from Chesowanja, bridge University Press. the fossil evidence for P. boisei. That Kenya. Nature 294:125–129. 37 Delson E. 1997. One skull does not a species opportunity is still greatly appreci- 16 Leakey REF, Walker A. 1988. New Australo- make. Nature 389:445–446. ated. BW thanks the Henry Luce pithecus boisei specimens from East and West Lake 38 Menter CG, Kuykendall KL, Keyser AW, Turkana, Kenya. Am J Phys Anthropol 76:1–24. Conroy GC. 1999. First record of hominid teeth Foundation, the George Washington from the Plio-Pleistocene site of Gondolin, 17 Brown B, Walker A, Ward CV, Leakey RE. South Africa. J Hum Evol 37:299–307. University, and the National Science 1993. New Australopithecus boisei calvaria from Foundation for their generous sup- East , Kenya. Am J Phys Anthro- 39 Keyser AW, Menter CG, Moggi-Cecchi J, port. We thank Chris Campisano for pol 91:137–159. Pickering TR, Berger LR. 2000. Drimolen: a new hominid-bearing site in Gauteng, South 18 Prat S, Brugal J-P, Roche H, Texier P-J. providing the geological ages of P. boi- Africa. S Afr J Sci 96:193–197. 2003. Nouvelles de´couvertes de dents d’homi- sei specimens and for creating Figure nide´s dans le membre Kaitio de la formation de 40 Constantino P, Wood B. 2004. Paranthropus 9, as well as Elizabeth Harmon, Fer- Nachukui (1, 65-1, 9 Ma), Ouest du lac Turkana paleobiology. Miscela´ nea en homenaje a Emi- (Kenya). CR Palevol 2:685–693. liano Aguirre. Madrid: Museo Arqueolo´ cico Re- nando Ramirez-Rozzi, Gen Suwa, Fred gional. p 136–151. 19 Walker AC, Leakey RE, Harris JM, Brown 41 Grine, Peter Ungar, and Alan Walker FH. 1986. 2.5 Myr Australopithecus boisei from Dean MC. 1988. Growth of teeth and devel- for information and advice. Finally, we west of Lake Turkana, Kenya. Nature 322:517–522. opment of dentition in Paranthropus. In: Grine 20 FE, editor. Evolutionary History of the ‘‘Ro- thank John Fleagle for his persistence. Suwa G, et al. 1997. The first skull of Aus- bust’’ Australopithecines. New York, Aldine de tralopithecus boisei. Nature 389:489–492. Gruyter. p 43–53. 21 REFERENCES Kullmer O, Sandrock O, Abel R, Schrenk F, 42 Groves CP. 1989. A theory of human and Bromage TG, Juwayeyi YM. 1999. The first Par- primate evolution. Oxford: Clarendon Press. anthropus from the Malawi Rift. J Hum Evol 1 Wood BA. 1991. Koobi Fora research project, 43 37:121–127. Silverman N, Richmond B, Wood B. 2001. vol. 4. Hominid cranial remains. Koobi Fora Testing the taxonomic integrity of Paranthropus 22 Research Project, series ed. Leakey RE. Oxford: Leakey REF. 1970. New hominid remains boisei sensu stricto. Am J Phys Anthropol Clarendon Press. and early artefacts from North Kenya. Nature 115:167–178. 226:223–224. 2 Strait DS, Grine FE. 2004. Inferring hominoid 44 Wood BA. 1988. Are ‘‘robust’’ australopithe- 23 and early hominid phylogeny using cranioden- Wood BA, Wood CW, Konigsberg LW. 1994. cines a monophyletic group? In: Grine FE, edi- tal characters: the role of fossil taxa. J Hum Paranthropus boisei: an example of evolutionary tor. Evolutionary history of the ‘‘robust’’ Aus- Evol 47:399–452. stasis? Am J Phys Anthropol 95:117–136. tralopithecines. New York: Aldine de Gruyter. 3 Kimbel WH, Rak Y, Johanson DC. 2004. The 24 Wood B, Lieberman DE. 2001. Craniodental p 269–284. skull of Australopithecus afarensis. Oxford: variation in Paranthropus boisei: a developmen- 45 Corruccini RS. 1994. How certain are homi- Oxford University Press. tal and functional perspective. Am J Phys noid phylogenies? The role of confidence intervals Anthropol 116:13–25. 4 Feibel CS, Brown FH, McDougall I. 1989. in cladistics. In: Corruccini RS, Ciochon RL, edi- Stratigraphic context of fossil hominids from 25 Leakey REF. 1971. Further evidence of tors. Integrative paths to the past: paleoanthropo- the Omo group deposits: Northern Turkana Lower Pleistocene hominids from East Rudolf, logical advances in honor of F. Clark Howell. Basin, Kenya and Ethiopia. Am J Phys Anthro- North Kenya. Nature 231:241–245. Englewood Cliffs: Prentice Hall. p 167–183. pol 78:595–622. 26 Robinson JT. 1972. The bearing of East 46 Strait DS, Grine FE, Moniz MA. 1997. A 5 McDougall I, Brown FH. 2006. Precise Rudolf fossils on early hominid systematics. reappraisal of early hominid phylogeny. J Hum 40Ar/39Ar geochronology for the upper Koobi Nature 240:239–240. Evol 32:17–82. Fora Formation, , northern 27 Wood BA, Li Y, Willoughby C. 1991. Intra- 47 Wood BA, Chamberlain AT. 1986. Australo- Kenya. J Geol Soc London 163:205–220. specific variation and sexual dimorphism in pithecus: grade or clade? In: Wood B, Martin L, 6 Leakey LSB. 1958. A giant child among the cranial and dental variables among higher pri- Andrews P, editors. Major topics in primate giant of Olduvai? A huge fossil milk mates and their bearing on the hominid fossil and . Cambridge: Cambridge molar which suggests that Chellean Man in record. J Anat 174:185–205. University Press. p 220–248. Tanganyika may have been gigantic. The Illus- 28 Grausz HM, Leakey RE, Walker AC, Ward 48 Skelton RR, McHenry HM. 1992. Evolution- trated London News 232:1104–1105. CV. 1988. Associated cranial and postcranial ary relationships among early hominids. J Hum 7 Leakey LSB. 1958. Recent discoveries at Old- bones of Australopithecus boisei. In: Grine FE, Evol 23:309–349. uvai Gorge, Tanganyika. Nature 181:1099–1103. editor. Evolutionary history of the ‘‘robust’’ Aus- 49 Maglio VJ. 1975. Origin and evolution of the 8 Leakey LSB. 1959. A new fossil skull from tralopithecines. New York: Aldine de Gruyter. Elephantidae. Trans Am Philos Soc 63:1–149. Olduvai. Nature 184:491–493. p 127–132. 50 Vrba ES. 1979. Phylogenetic analysis and 29 9 Leakey LSB, Leakey MD. 1964. Recent discov- Day MH. 1969. Femoral fragment of a ro- classification of fossil and recent Alcelaphini eries of fossil hominids in Tanganyika, at Oldu- bust australopithecine from the Olduvai Gorge, (Mammalia: Bovidae). Biol J Linnaean Soc vai and near Lake Natron. Nature 202:5–7. Tanzania. Nature 221:230–233. 11:207–228. 30 51 10 Arambourg C, Coppens Y. 1968. Decouverte Wood BA. 1974. Evidence on the locomotor Vrba ES. 1984. Evolutionary pattern and d’un australopithecien nouveau dans les Gise- pattern of Homo from Early Pleistocene of process in the sister-group Alcelaphini-Aepycer- ments de L’Omo (Ethiopie). S Afr J Sci 64:58–59. Kenya. Nature 251:135–136. otini (Mammalia: Bovidae). In: Eldredge N, 31 11 Coppens Y. 1978. Evolution of the hominids Wood BA. 1974. Olduvai Bed I post-cranial Stanley SM, editors. Living fossils. New York: and of their environment during the Plio-Pleis- fossils: a reassessment. J Hum Evol 3:373–378. Springer-Verlag. p 62–79. tocene in the lower Omo Valley, Ethiopia. In: 32 Gebo DL, Schwartz GT. 2006. Foot bones 52 Patterson C. 1982. Morphological characters Bishop WW, editor. Geological background to from Omo: implications for hominin evolution. and homology. In: Joysey KA, Friday AE, edi- fossil man. Edinburgh: Scottish Academic Am J Phys Anthropol 129:499–511. tors. Problems of phylogenetic reconstruction. Press. p 499–506. 33 Johanson DC, et al. 1987. New partial skele- London: Academic Press. p 21–74. 12 Coppens Y. 1980. The differences between ton of Homo habilis from Olduvai gorge, Tanza- 53 Suwa G, Wood BA, White TD. 1994. Further Australopithecus and Homo; preliminary con- nia. Nature 327:205–209. analysis of mandibular molar crown and cusp 62 Constantino and Wood ARTICLES

areas in and Early Pleistocene homi- 70 Lucas PW, Corlett RT, Luke DA. 1986. Post- robustus and Homo sp. in Swartkrans. J Hum nids. Am J Phys Anthropol 93:407–426. canine tooth size and diet in anthropoid prima- Evol 28:277–285. tes. Z Morphol Anthropol 76:253–276. 54 Asfaw B, White T, Lovejoy O, Latimer B, 87 Plavcan JM. 2000. Inferring social behavior Simpson S, Suwa G. 1999. Australopithecus 71 Kay RF. 1975. The functional adaptations of from animal dimorphism in the fossil record. J garhi: a new species of early hominid from primate molar teeth. Am J Phys Anthropol Hum Evol 39:327–344. Ethiopia. Science 284:629–635. 43:195–216. 88 55 72 Plavcan JM. 2003. Scaling relationships Shipman P, Harris JM. 1988. Habitat prefer- Hatley T, Kappelman J. 1980. Bears, pigs, between craniofacial sexual dimorphism and ence and paleoecology of Australopithecus boi- and Plio-Pleistocene hominids: a case for the body mass dimorphism in primates: implica- sei in eastern Africa. In: Grine FE, editor. Evo- exploitation of belowground food resources. tions for the fossil record. Am J Phys Anthropol lutionary history of the ‘‘robust’’ Australopithe- Hum Ecol 8:371–387. 120:38–60. cines. New York: Aldine de Gruyter. p 343–381. 73 Laden G, Wrangham R. 2005. The rise of the 89 56 Reed KE. 1997. Early hominid evolution hominids as an adaptive shift in fallback foods: Smith BH. 1991. Age of weaning approxi- and ecological change through the African Plio- plant underground storage organs (USOs) and mate age of emergence of the first permanent Pleistocene. J Hum Evol 32:289–322. australopith origins. J Hum Evol 49:482–498. molar in non-human primates. Am J Phys An- thropol (suppl):163–164. 57 Nagaoka S, Katoh S, Wolde Gabriel G, Sato 74 Grine FE, Martin LB. 1988. Enamel thick- H, NaKaya H, Beyene Y, Suwa G. 2005. Lith- ness and development in Australopithecus and 90 Dean MC. 2000. Progress in understanding ostratigraphy and sedimentary environments Paranthropus. In: Grine FE, editor. Evolution- hominoid dental development. J Anat 197:77–101. of the hominid-bearing Pliocene-Pleistocene ary history of the ‘‘robust’’ australopithecines. 91 Aiello LC, Montgomery C, Dean C. 1991. Konso Formation in the southern Main Ethio- New York: Aldine de Gruyter. p 3–42. The natural history of deciduous tooth attrition pian Rift, Ethiopia. Palaeogeogr Palaeocl 216: 75 Gantt DG, Rafter JA. 1998. Evolutionary and in hominoids. J Hum Evol 21:397–412. 333–357. functional significance of hominoid tooth 92 Harvey PH, Clutton-Brock TH. 1985. Life 58 enamel. Connect Tissue Res 39:195–206. Schrenk F, Bromage TG, Gorthner A, San- history variation in primates. Evolution 39:559– drock O. 1995. Paleoecology of the Malawi Rift: 76 Rak Y. 1988. On variation in the masticatory 581. vertebrate and invertebrate faunal contexts of system of Australopithecus boisei. In: Grine FE, 93 the Chiwondo Beds, northern Malawi. J Hum editor. Evolutionary history of the ‘‘robust’’ Aus- Aiello L, Wood B. 1994. Cranial variables as Evol 28:59–70. tralopithecines. New York: Aldine de Gruyter. predictors of hominine body mass. Am J Phys 59 Leakey LSB, Tobias PV, Napier JR. 1964. A p 193–198. Anthropol 95:409–426. new species of the genus Homo from Olduvai 77 Tobias PV. 1963. Cranial capacity of Zinjan- 94 Kappelman J. 1996. The evolution of body Gorge. Nature 202:7–9. thropus and other australopithecines. Nature mass and relative brain size in fossil hominids. 60 Isaac GL. 1984. The of human 197:743–746. J Hum Evol 30:243–276. origins: studies of the Lower Pleistocene in East 78 Demes B, Creel N. 1988. Bite force, diet, and 95 de Sousa A, Wood B. 2007. The hominin Africa 1971–1981. Advances in World Archaeol cranial morphology of fossil hominids. J Hum fossil record and the emergence of the 3:1–87. Evol 17:657–670. modern human central nervous system. In: 61 Brain CK, Churcher CS, Clark JD, Grine FE, 79 Walker AC. 1981. Diet and teeth: dietary Kaas JH, editor. The evolution of primate Shipman P, Susman RL, Turner A, Watson V. hypotheses and human evolution. Philos Trans nervous systems. Oxford: Elsevier. p 291– 1988. New evidence of early hominids, their 336. culture and environment from the Swartkrans R Soc Lond B 292:57–64. 96 cave, South Africa. S Afr J Sci 84:828–835. 80 Hylander WL. 1988. Implications of in vivo Beynon A, Wood B. 1987. Patterns and rates of enamel growth in the molar teeth of early 62 Clark JD. 1990. The earliest cultural eviden- experiments for interpreting the functional sig- hominids. Nature 326:493–496. ces of hominids in southern and south central nificance of ‘‘robust’’ australopithecine jaws. In: Africa. In: Sperber G, editor. New York: Wiley- Grine FE, editor. Evolutionary history of the 97 Weidenreich F. 1935. The Sinanthropus pop- Liss. p 1–15. ‘‘robust’’ Australopithecines. New York: Aldine ulation of Choukoutien (locality) with a prelimi- de Gruyter. p 55–83. 63 Susman RL. 1991. Who made the Oldowan nary report on new discoveries. Bull Geo Soc tools? Fossil evidence for tool behavior in Plio-Pleis- 81 Wood BA, Aiello LC. 1998. Taxonomic and func- China 14:427–466. tocene hominids. J Anthropol Res 47:129–151. tional implications of mandibular scaling in early 98 Smith RJ. 2005. Species recognition in pale- hominins. Am J Phys Anthropol 105: 523–538. 64 Wood BA. 1997. The oldest whodunnit in oanthropology: implications of small sample the world. Nature 385:292–293. 82 Walker A, personal communication. 2006. sizes. In: Lieberman DE, Smith RJ, Kelley J, 65 Susman RL. 1988. Hand of Paranthropus 83 Grine FE. 1981. Trophic differences between editors. Interpreting the past: essays on human, robustus from Member 1, Swartkrans: fossil evi- gracile and robust Australopithecus: a scanning primate, and evolution in honor of dence for tool behaviour. Science 240:781–784. electron microscope analysis of occlusal events. David Pilbeam. Boston: Brill Academic Publish- ers. p 207–219. 66 Kuman K, Clarke RJ. 2000. Stratigraphy, ar- S Afr J Sci 77:203–230. tifact industries and hominid associations for 84 Scott RS, Ungar PS, Bergstrom TS, Brown 99 Delson E, Tattersall I, Van Couvering JA, , Member 5. J Hum Evol 38:827–847. CA, Grine FE, Teaford MF, Walker A. 2005. Brooks AS, editors. 2000. Encyclopedia of human evolution and , 2nd ed. New 67 Dean MC, Wood BA. 1981. Metrical analysis Dental microwear texture analysis shows York: Garland Publishing. of the basicranium of extant hominoids and Aus- within-species diet variability in fossil homi- tralopithecus. Am J Phys Anthropol 54:63–71. nins. Nature 436:693–695. 100 Tobias PV. 1991. Olduvai Gorge, vol. 4. The 68 Dean MC, Wood BA. 1982. Basicranial anat- 85 Boaz NT, Hampel J. 1978. Strontium con- , endocasts and teeth of Homo habilis. omy of Plio-Pleistocene hominids from East and tent of fossil tooth enamel and diet of early Cambridge: Cambridge University Press. South Africa. Am J Phys Anthropol 59:157–174. hominids. J Paleontol 52:928–933. 69 Hylander WL. 1975. Incisor size and diet in 86 Sillen A, Hall G, Armstrong R. 1995. Stron- anthropoids with special reference to Cercopi- tium calcium ratios (Sr/Ca) and strontium iso- thecidae. Science 189:1095–1098. topic ratios (87Sr/86Sr) of Australopithecus VC 2007 Wiley-Liss, Inc.