8 5 Early Hominid Postcrania and Locomotor Adaptations Randall J. Thompkins The locomotor system of modern man is a complex and unique adaptation, and the evolutionary origin and history of bipedalism are of fundamental impor- tance in understanding the pattern of human evolu- tion. For this reason much attention has been focused upon the postcranial skeleton of the early hominids. The fossil evidence ot Australopithecus fortunately in- cludes numerous postcranial remains, and these have been analyzed to determine locomotor capabilities. Bipedal, erect posture inAustralopithecus has been dem- onstrated, but the full spectrum of locomotion in these hominids is not known. The use of extant models has placed undue emphasis on the striding gait and the question of its presence in Australopithecus. Methodological problems have hindered progress in reconstructive anatomy and biomechanical analysis. Since the locomotor system influences most activities of an organism, differences in locomotor capacities often have phylogenetic significance; if these differ- ences are reflected in bony morphology, they should be appropriate for taxonomic criteria. Discerning the range of locomotor behaviors of the aus- tralopithecines is a challenging and important task for human evolutionary studies if we are to assimilate the anatomy, ecology, behavior, and systematics of Au- stralopithecus into a synthetic, holistic view of the evolu- tion of the Hominidae. Analysis of Postcranial Material There are a number of paleontological problems central to the analysis of post-cranial remains; these include incomplete and damaged specimens plus the problem of variation. Several critical parameters: body stature and weight, sex, and age, are often, by the nature of the hominid fossil record, indeterminable. This frequent lack of data on gross body proportions and paleodemography of the sample hinders any at- tempt at biomechanical analysis of the postcranial material. Another frequent problem is that of associa- tion; do skeletal parts found in close proximity actually belong to one individual? The discovery in recent years of several partial or relatively complete skeletons of Plio-Pleistocene Hominidae will go far in illustrating the morphological pattern of Australopithecus and early Homo, in delineating any differences within these taxa, and particularly in yielding suitable material (of one individual) for the analysis ofjoints. Such analysis has been infrequently attempted yet is crucial to a compos- ite biomechanical study. The current lack of consensus about the morpholog- ical characteristics of early hominid1 postcrania is in considerable part a lack of agreement about approp- riate techniques and relevant parameters. Much of the earlier literature was on a purely descriptive level, and later statistical treatment gave ambiguous results. This is partly because of different workers measuring the same set of fossils and getting different results; also, access to the original specimens is not readily available to all workers, who therefore must rely upon pub- lished reports and casts. Many of the original speci- mens are not fully prepared for study and are in need of reconstruction. A general consensus on operationalism of method is needed with respect to: (a) the basic parameters involved, (b) how to obtain the necessary data from the fossils (what to measure on the bone and how to measure it), (c) how to standardize for effects of body size, and (d) the appropriate statistical methodologies. Isolated traits must undergo extensive comparative study, and then biomechanical recon- struction, to validate preliminary inferences. Ulti- mately one hopes to integrate fossil postcranial data into what Zihlman (1977) calls a behavioral- evolutionary framework which (a) relates the locomotor morphology and behavior to an environ- mental context, (b) correlates the locomotor behavior with the morphology, using living models, and (c) at- tempts a synthetic limb/joint analysis. These goals may be reached via a combination of techniques, including stress analysis (Preuschoft 1971), cineradiographic analysis of related extant forms (Jenkins 1972), biomechanical reconstructions, including bone load- ing patterns (Zihlman and Hunter 1972; Lovejoy et al. 1973; Lovejoy 1977), and joint reconstruction analysis (Zihlman 1977). Multivariate statistical approaches, while very prom- ising, have yielded some contradictory results with re- gard to australopithecine postcranial anatomy. Using multivariate techniques (specifically canonical analysis) on a number of postcranial fossils of Aus- tralopithecus, Oxnard (1975) has stressed the unique- ness of these fossil remains, their clear differences with fossils placed in Homo, and the elimination of Aus- tralopithecus as a direct ancestor of Homo. Other work- ers, using the same fossils and multivariate techniques (principal coordinate and canonical analysis), have concluded that early hominid lower limb (McHenry and Corruccini 1975a) and upper limb (Ciochon and Corruccini 1976) morphology is clearly allied with Hominidae, and approaches, but is not identical with, that of Homo. The problem of how to compensate for size effects in both uni- and multivariate analysis is critical. A frequent technique to eliminate size effects is to con- vert measurements to angles or indices, and create ratios between the structure being compared and some morphological feature used as a size standard. Ratio data often cannot be compared for lack of matching or suitably preserved body parts. Multivariate techniques have the potential to avoid latent size effects. Corruc- cini (1975) has criticized Oxnard and stressed the necessary distinction between size and shape. He notes that statistics computed directly from raw measure- ments are dominated by simple size differences. There has been a tendency to deal exclusively with isolated bones in such statistics, and to make conclusions solely from those data (e.g., Oxnard 1968a & b). The isolated bone is not the functional unit in limb mrorphology, and analysis should incorporate available joints (Ciochon and Corruccini 1976). Two important issues in statistical studies are what measurements are to be employed and how to inter- pret the results. When isolated bones instead of articu- lated joint complexes are utilized the measurements used severely limit their utility, and factors such as body weight are often ignored (Zihlman 1975). Great care must be taken not to "load" the multivariate pro- gram with redundant measurements. Lovejoy (1977) has emphasized that this problem of the interrelation- ship of metrically assessable characters is the greatest limitation to the interpretation of multivariate results. As Day (1973) has pointed out, multivariate studies have in recent years been at best a complement to anatomical description. With additional fossil material and progress in multivariate techniques, the full im- pact of advanced biometrics should be realized. An important consideration is the level of inference which is most appropriate for reconstructing postcra- nial anatomy and locomotor capacities in early hominids. Preliminary anatomical description should not be used as a final analysis. The observation of morphological differences is not in itself evidence of locomotor differences between two groups. The range of variation within a population must be accounted for; differences in morphology, if statistically signifi- cant, may have no obvious mechanical effect. As Lovejoy (1975) has emphasized, the important ques- tion is: does a given morphological feature have biomechanical importance? It is the biomechanical ef- fect of a morphological change that is important, and this level of inference has often been lacking. The range of variation, both metric and nonmetric, in postcranial morphology is considerable. Further- more, the correlation between environment, locomotor behavior, and morphology may be weak in extant groups (e.g., within Colobus species, Zihlman 1977). Analysis should begin with accurate description of the fossil anatomy, and functional morphology dis- cerned via extensive comparative work. Variation should be accounted for by use of appropriate statisti- cal techniques, and finally a total biomechanical pat- tern should be reconstructed. It is then possible to address the question of differences in locomotor be- haviors between a fossil taxon (Australopithecus) and the extant relative (Homo), as well as possible differences within the fossil taxon itself (i.e., between gracile and robust Australopithecus). Then paleoecological, phylogenetic, and taxonomic inferences may be de- rived, whether differences between the various groups under consideration exist or not, since either result would have significance in elucidating the pattern of evolution of early Hominidae. Early Hominid Fossil Postcrania Early hominid postcranial fossils are relatively rare when compared to more abundant dental, gnathic, and cranial parts. This is unfortunate since locomotor capabilities are so important in any organism's be- havioral repertoire, and especially so in the evolution of the Hominidae. Recently, however, the sample of postcrania has become fairly complete and most body parts are now represented, some by relatively com- plete and undistorted fossils, some by the remains of several partially complete skeletons (Table 1). A review of the fossil evidence will be offered for each major anatomical region, describing basic morphologies and delineating any biomechanical differences. Axial Skeleton Early hominid vertebrae are relatively rare in the fossil record, and the best specimens are fromjust two individuals, Sts 14 (Sterkfontein) and AL 288-1 (Afar). Both Day (1977) and Robinson (1972) stress the small size of the vertebral bodies. Small size is also evident in the two articulated vertebrae (in matrix) of KNM-ER 164 (East Lake Turkana, Day 1977). A relatively well developed lumbar curvature is evidenced in Sts 14 (Robinson 1972), which concurs with lower limb evi- dence of bipedal erect posture. The sacrum of Sts 14 is of typically hominid disposi- tion and shape, but in overall size significantly smaller than that of modern man. A smaller auricular surface, smaller lumbosacral articular surface, and smaller sagittal and transverse diameters of the vertebral bodies indicate a column not as well adapted for effi- cient stress and weight bearing, in an individual of light body build with a fairly large pelvis for its body size (Robinson 1972; Zihlman 1970). Body Proportions Reconstructions of height, weight, and body pro- portions of australopithecines usually assume that the limb proportions are equivalent to those of modern man, an assumption which is unwarranted since the evidence indicates relatively longer upper limbs in comparison with modern human populations (Genet-Varcin 1969; McHenry 1974; Olivier 1976). 8 6 This is demonstrated by material recovered from Sterkfontein, Kromdraai, and especially Hadar. The Afar AL 288-1 humerus is 83.9% the length of the femur (Johanson and Taieb 1976). This is also appar- ent if the KNM-ER 739 humerus and the Omo L 40-19 ulna are articulated (they form a fair match); the result is a relatively long and quite robust upper limb. Variability within one site (e.g., at Sterkfontein, the Sts 7 humerus and Sts 14 femur) is considerable, and marked sexual dimorphism has been suggested by McHenry (1974) to account for it. Estimates of stature and body weight for Aus- tralopithecus africanus (Sts 14) range from a low estimate of 107 cm height and 18-22 kg mass (Lovejoy and Heiple 1970) to a high estimate of 145 cm and 27.6 kg (McHenry 1974, 1976a). A narrower range has been discussed for robust australopithecines: 146 cm (Burns 1971) to 165 cm (Genet-Varcin 1969), and 43.2 kg (McHenry 1976a) to 68-91 kg (Robinson 1972). McHenry (1974, 1976a) has estimated the weight and stature of several East African specimens assigned to Homo sp. as 52.8 kg and 163 cm (mean). The low stature determination by Lovejoy and Heiple (for one A. africanus) is confirmed by preliminary descriptive reports on the Hadar 1974 partial skeleton; Johanson and Taieb (1976) reconstruct AL 288-1 as approxi- mately one meter tall. Appendicular Skeleton: Upper Limb Fossil evidence from early hominid forelimbs is not as abundant as that from hindlimb parts, but a consid- erable number of specimens have now been catalogued. Clavicula The fossil clavicle from Olduvai (Old. Hom. 48)2 was described by Napier (1965) as fundamentally like that of Homo sapiens. Oxnard felt that the clavicle, along with the Sts 7 glenoid fragment, indicated some sort of suspensory posture (1968a and b), based on the angle of clavicular torsion. The Olduvai clavicle is damaged by an internal fracture and is missing both sternal and acromial extremities; because of this Oxnard had to reconstruct both ends and assume no significant dis- tortion. Day (1976a), in a series of multivariate tests, was unable to discriminate Old. Hom. 48 from Homo sapiens. Scapulae The only scapular fragment to be analyzed in detail is Sts 7, described by Broom et al. (1950). These au- thors noted an unusual combination of features seen inHomo and inPongo. Oxnard (1968a) saw evidence of a suspensory adaptation in the fossil; however, the scapular fragment was embedded in matrix and bro- ken at the glenoid fossa (also the present state of the fossil), the coracoid had not been adequately exposed, and the acromion was lost. Shoulder (Glenohumeral) Joint The fossil evidence of the glenohumeral joint in Australopithecus is limited to Sts 7. In a recent paper on this glenohumeral joint3, Ciochon and Corruccini (1976) note the rather large coracoid process and con- clude, after multivariate comparative analysis, that this glenohumeral joint possibly evidences not only a sus- pensory posture ancestry (synplesiomorphy) but also partial suspensory locomotor posture in these early hominids. This hypothesis will receive further testing when full description and analysis of the same joint in AL 288-1 is completed; at present it is tentatively sup- ported by some of the fossil arm, forearm, and hand bone evidence. Humeri Fossil humeri in a state of preservation which allows detailed study are few in number. Consequently much attention has been focused upon relatively few speci- mens (Sts 7, TM 1517, Kanapoi, KNM-ER 739). Un- fortunately comparable body parts for the proximal humerus are only of one taxon, Australopithecus af- ricanus. The Sterkfontein proximal humerus was de- scribed by Broom and Robinson (Broom et al. 1950), who noted several Pongo-like features. The fossil exhibits a well rounded head, a deep intertubercular groove, and prominent muscular markings (observ- able on Wenner-Gren cast). Day (1977) feels that the specimen evidences a powerful m. biceps brachii and well developed stabilizer (short scapular) muscles (also seen in KNM-ER 1473). Omo 119-73-2718 and AL 288-1 (left proximal humeri) have yet to be described. Omo 119-73-2718 is a relatively large bone with a hemispherical articular head, a prominent intertuber- cular groove, well developed greater and lesser tuber- cles, and a robust shaft; it is comparable in size to Sts 7 (observations on casts; Boaz et al. n.d.), and has been tentatively referred to Australopithecus africanus (How- ell and Coppens 1976). Distal humeri are better known and have been ex- tensively described. Broom described TM 1517 as pos- sessing a large and more irregularly spherical capitulum than modern man (Broom and Schepers 1946). Clark (1947) felt that the trochlea was not set as far anterior (to a frontal plane through the axis of the distal humerus) as in man or chimpanzee, and this might have limited flexion but allowed considerable hypertension. Straus (1948) emphasized numerous nonhominid characters and felt that the distal humerus was not useful for taxonomic discrimination (also Robinson 1972). McHenry and Corruccini (1975a) concluded that distal humeri are discrimina- tory, and placed TM 1517 as intermediate between pongids and modern humans. Day (1977) concluded that the specimen was less like Homo sapiens than the Kanapoi humerus. Kanapoi 1, larger in size than TM 1517 and designated Australopithecus cf. africanus (Pat- terson and Howells 1967), is fundamentally like that of 8 7 modern man (McHenry 1976b; McHenry and Cor- ruccini 1975a; McHenry et al. 1976); Day has suggested placing the fossil in Homo (1974a). The humerus from East Lake Turkana, KNM-ER 739, has often been the subject of speculation concern- ing the locomotor capabilities of robust aus- tralopithecines. The specimen is a very long4 and robust humerus with strong muscular ridging and a well rounded capitulum. Day (1977) describes the ridging as indicative of massive deltoid, brachioradialis, as well as common hand flexor and extensor muscles. Several workers have cited the u- nique morphology of the fossil (McHenry 1973a; McHenry and Corruccini 1975a; McHenry et al. 1976). Kay (1973) felt that the specimen indicated "facultative rather than habitual bipedalism," and McHenry's (1976b) sorting via locomotor categories placed KNM-ER 739 within the suspensory posture category. Forearm Fossil evidence of the early hominid forearm is scarce and poorly preserved. There are no complete radii. Available material is confined to proximal and shaft fragments. These illustrate a rounded, recessed head and a rounded neck (SK 18b, Omo 75s-70-'1317, KNM-ER 1500e, AL 288- 1), a marked m. pronator teres insertion (SK 18b, Old. Hom. 49, KNM-ER 803d), and a well delineated interosseous border (Old. Hom. 49, KNM-ER 1500k). The total observable pattern is simi- lar to Homo sapiens (Day 1977). Obviously, description of the distal radii from Hadar is anticipated. Only then, with knowledge of the morphology of the carpal articular surface and the anterior and posterior distal surfaces, can we draw firm conclusions about radial morphology. We are fortunate to have one virtually complete and several fragmentary ulnae. The complete right ulna from the Shungura Formation, in unit E-5, has been described and assigned toAustralopithecus boisei by Howell and Wood (1974). The Omo ulna (L 40-19) is quite long (315 mm), has a moderate dorso-ventral curvature to the shaft, and is somewhat flattened an- teroposteriorly. The shaft does not project proximally beyond the olecranon process (as in Papio). Both the margin of the coronoid process and the ulnar tuberos- ity are damaged, as well as details of the radial notch. The insertions of m. triceps brachii and especially m. anconeus and m. pronator teres are well delineated. The supinator crest is not well marked. The ulna exhibits no noticeable interosseous margin, the articular head is less crescentic than in modern human ulnae, and a prominent styloid process is not evident (Howell and Wood 1974; personal observations). Manus A number of hand bones of the early hominids are available; however, there is a distinct lack of compa- rable body parts and several carpals are absent from the fossil record (lunate, triquetrum, pisiform). The trapezoid and hamate bones, previously absent in the fossil sample, are present in the (undescribed) material from the Hadar locality AL 333. The TM 1526 capitate from Sterkfontein, small and well-preserved, was described by Broom and Schepers (1946) as morphologically intermediate between pon- gids and man. Clark (1947, 1967) concurred, specify- ing that the fossil differed from modern human capi- tates in the pronounced excavation of the lateral sur- face; this resulted in a "waisted" appearance which he felt to be the site of a strong interosseous (capitate- trapezoid) ligament. Robinson (1972) notes pongid features in the lack of a definite articular facet for metacarpal IV, a laterally-waisted profile, and rela- tively smaller articular surfaces for the scaphoid and the trapezoid. OJ. Lewis (1973) has described (from casts) the Sterkfontein and Olduvai (Old. Hom. 7r) capitates, in a major review of the comparative mor- phology of the wrist joint in the Hominoidea. He con- cludes that lateral (radial side) waisting of the capitate is not for attachment of a trapezoid-capitate interosse- ous ligament but is part of a unique midcarpal locking mechanism best developed in the African, knuckle- walking apes. Lewis attempts to demonstrate that the waisting on the capitate radial surface in Pan troglodytes is contacted by the scaphoid head in wrist extension (as the scaphoid is rotated, its head becomes locked into the constricted neck of the capitate). In Homo sapiens, this lateral waisting is not apparent, and an anterior trapezoid-capitate articulation (in front of the interos- seous ligament, a neomorph) has replaced the primi- tive posterior articulation. The trapezoid is now larger in Homo, the metacarpal II is now articulating more distally in the carpus, and the metacarpal III develops the prominent styloid process (near the posterolateral corner of the capitate). His observations included the notation of a distinct lateral waisting in TM 1526 and Old. Hom. 7r. He describes TM 1526 as exhibiting a laterally swollen head, a posteriorly located trapezoid facet, and a pongid-like articulation for the medial side of the base of metacarpal II. He concludes that both specimens are far more Pan-like than previously esti- mated. The right juvenile trapezium from Olduvai (Old. Hom. 7q) can be articulated with the Old. Hom. 7p scaphoid and the base of a second metacarpal (Old. Hom. 7o) to form a reasonable reconstruction of the radial side of a right hand. Further material from Olduvai includes two damaged proximal, four inter- mediate, and three terminal phalanges (Old. Hom. 7i-h, d-g, and a-c, respectively), perhaps all of the same hand (Day 1976a). The terminal pollical phalanx is broad and stout, and bears the impression for m.flexor pollicis longus. Napier's original analysis of the hand bones emphasized: (1) their differences from Homo sapiens in overall robustness, dorsal curvature of the 8 8 phalangeal shafts, and strength of the fibrotendinous markings, and (2) their similarity to modern man in the breadth and stoutness of the terminal phalanges, and the ellipsoidal contour of the metacarpo- phalangeal joint surfaces. Three metacarpals are known from Swartkrans; two specimens, SK 85 and SKW 14147, are generally comparable with Homo sapiens (Napier 1959; Day 1977; Day and Scheur 1973). SK 84 is a short, very robust metacarpal which is strongly curved in its longitudinal axis. It exhibits strong medial and lateral muscular markings, on the radial side form. opponens pollicis, and on the ulnar side for the first palmar interosseous muscle. The saddle- shaped proximal articular surface is laterally elon- gated (observations on cast; Napier 1959; Day 1977; Rightmire 1972). SK 85 has been assigned to Homo erectus and SK 84, with some pongid features, to Aus- tralopithecus robustus (Napier 1959, Robinson 1972). Napier, discussing Old. Hom. 7, offers what might be an accurate description of the australopithecine manus. "The overall picture presented by this as- semblage is of a short, powerful hand with strong, curved digits, surmounted by broad, flat nails and held in marked flexion. The thumb is strong and opposable, though possibly rather short." (1962:414). Differences between the species of Australopithecus cannot be determined with any accuracy. A sufficient sample with suitably preserved corresponding body parts is not available. Description of the hand bones from the Hadar hominid AL 288-1 and new, unpub- lished material from the AL 333 locality should com- plement our knowledge of gracile australopithecine hand morphology. The new Afar specimens are re- markable not only in their abundance but also for their superb state of preservation. If they prove to be consis- tent with the morphological pattern discussed above, then one may draw firm conclusions about the form of the early hominid hand (in at least one lineage). The problem will then be to delineate behavioral potentialities from that pattern. Lower Limb, and Bipedal Efficiency in Australopithecus A great deal of the earlier literature focused upon: a) whether the australopithecines were bipedal, and b) if they were bipedal, was the bipedal adaptation the same as seen in modern man. Naturally, attention was centered upon the lower limb, particularly pelvic and femoral, anatomy of the early hominids. The essentially hominid nature of the aus- tralopithecine pelvis is well known. The focus here will be on the delineation of differences in pelvic (and lower limb) morphology between Australopithecus and Homo (and between species of Australopithecus), and the elucidation of their biomechanical significance. It is suggested that much of the work claiming significant gait differences between Australopithecus and Homo, and between gracile and robust australopithecines, has been based upon inadequate reconstruction, and these conclusions are at least partially in error. Os Coxae The controversy in the study of early hominid in- nominates has been in the interpretation of those fea- tures in Australopithecus which are not identical with Homo. The recovery of a complete coxal bone from Sterkfontein, the basis for most of our knowledge of the pelvic anatomy of A. africanus, has been com- plemented by the discovery of a juvenile ilium and ischium from the Makapan Limeworks (MLD 7 and 8, Dart 1949) and, more recently, of a relatively complete left innominate from Hadar, AL 288-1 Uohanson and Taieb 1976). Fortunately this sample includes most of the major morphological landmarks of all three pelvic bones. The ilium of Australopithecus robustus (and A. boisei) is represented by SK 50 and SK 3155b; the ischium is represented only by SK 50, a poorly pre- served specimen. The ilium of Australopithecus, the best known ele- ment of its pelvis, is typically hominid in its low, wide, and laterally disposed shape. It shows both the convex anterior and concave posterior gluteal surfaces of the ala. The iliac crest is less arched in its mid-length (is more horizontal anteroposteriorly) than in modern man but does form a sigmoid curve. The auricular surface is relatively small, although the relationship of this to body size is problematic (Broom et al. 1950; Clark 1955; Day 1973; Lovejoy 1975; McHenry 1975c; Robinson 1972; Zihlman 1967). Distinctions from Homo are in the structure and disposition of the an- terior margin of the ilium: in more projecting anterior superior and inferior iliac spines, with a more re- curved margin between them ipsilaterally, in a longer acetabulospinous distance, and in a greater interspin- ous distance (between contralateral anterior superior spines, forming greater lateral iliac flare). This distinct morphology is witnessed in all specimens, and has been the subject of considerable speculation. The orig- inal descriptions of Sts 14 and SK 50 emphasized the more anterior position of the anterior superior spines, giving the pelvis a "beaked" appearance (Broom et al. 1950; Broom and Robinson 1952). The anterior in- ferior iliac spine is well developed, particularly in SK 50. The anterior margin in SK 50 forms a distinct notch between superior and inferior spines, and pro- jects forward in a pronounced fashion to reach the ventrally protuberant anterior superior iliac spine (the same pattern is visible, although less accentuated, on Sts 14). The anterior inferior spine bears the impres- sions of m. rectusfemoris and the iliofemoral ligament, and the impression of the reflected head of m. rectus femoris is visible on the body of the ilium (just posterior to the anterior inferior spine and just superior to the rim of the acetabulum) on the original specimens (Day 1973; Robinson 1972). 8 9 There has been considerable disagreement concern- ing the distribution and relative thickness of cortical bone in the anterior aspect of the ala. In modern man this area illustrates two thickenings of bone, radiating out from the acetabulum to the anterior iliac crest in a "V"-shaped fashion: the acetabulocristal and acetabulospinous pillars. The former (the iliac pillar) is a pronounced bony buttress which runs from the center of the body of the ilium to the iliac or cristal tubercle. The latter is a bony ridge from the acetabulum to the anterior superior iliac spine. Since in Homo the anterior superior spine is curved medially, the iliac pillar divides the gluteal surface of the ilium into anterior and posterior planes. The question arose as to whether Australopithecus possessed an iliac pillar and cristal tubercle. Although Mednick (1955) and Napier (1964) as- serted that no cristal tubercle was present, other work- ers have noted its presence, albeit in less pronounced form (Dart 1949; Howell 1955; Robinson 1972). The disagreement probably arose from focusing upon the cast of the Sts 14 right innominate, which is lacking the anterior iliac crest. The tubercle is visible on the left innominate from Sterkfontein (illustrated by Robin- son 1972:299). The iliac pillar can be seen in Aus- tralopithecus, though clearly less pronounced (Dart 1949; Day 1973; Genet-Varcin 1969; McHenry 1975c; Zihlman 1970). The pillar is also of a somewhat differ- ent form, being more anterior in position. This is probably an effect of lack of modern medial rotation of the anterior superior spines, thus placing both acetabulocristal and acetabulospinous pillars closer to one plane (Lovejoy 1975; Lovejoy et al. 1973). This results in effectively merging the two buttresses into one more subtle thickening directed to the area of the anterior iliac crest (Robinson 1972). The pronounced beaking in SK 50 is not seen in the SK 3155b innominate (Brain et al. 1974; McHenry 1975c), but it is clearly from ajuvenile individual. The relatively minor iliac differences between gracile and robust australopithecines are not demonstrably reflec- tive of any functional difference but may simply be correlative of body size differences (Robinson 1972). The pubic bones are not well known for Aus- tralopithecus. The pubic bone in Sts 14 has been dis- cussed as relatively long (Broom et al. 1950; Robinson 1972). Sts 14 is not accurately reconstructed at the symphysis pubica since the left pubic bone crosses the midline into the right side. Both Lovejoy (1975) and Day (1973) have expressed the opinion that a correc- tive reconstruction would reduce the apparent "pon- gid" length and place the length of the pubis within the modern human range. The pubic symphysis is rela- tively small (Zihlman 1967). All specimens with the relevant area preserved (just lateral to the iliopubic eminence) show a well developed iliopsoas groove (Day 1973; Zihlman 1967; clearly visible on Wenner- Gren casts of Sts 14, SK 50, and SK 3155b). A distinct pectineal line can be seen on Sts 65 and is apparent on Sts 14 (although here the superior pubic ramus is damaged, Day 1973). This damage is quite evident on the superior ramus where pecten pubis and the ob- turator crest would approximate each other in the vicinity of the pubic tubercle. The distorted inferior pubic ramus might account for the unique shape of the obturator foramen (noted by Broom et al. 1950 as distinctly larger in superior-inferior diameter than in medio-lateral). The morphology and dimensions of the ischium in Australopithecus have been a source of controversy; this disagreement has focused upon the length of the is- chium and the nature of the ischial tuberosity. Broom and Robinson described the Sts 14 and SK 50 ischia as more anthropoid-like than the ischia of modern man. They emphasized the greater distance between the inferior margin of the acetabulum and the superior tip of the ischial tuberosity (Broom et al. 1950; Broom and Robinson 1952). It was then suggested by several au- thors that the "long" ischium in Australopithecus rep- resented an incomplete bipedal adaptation, by retain- ing a more pongid-like, power oriented hip extensor mechanism (Clark 1955; Napier 1964; Straus 1962; Zihlman 1967, 1970). Cursory examination of casts of Sts 14 and SK 50 plus numerous modern human ischia appears to confirm the above observation, but the ap- parently "long" ischium of Sts 14 is in fact a visual artifact, resulting from an ischial tuberosity of some- what different morphology and surface area, a rela- tively wide sulcus tuberoglenoidalis, and a relatively small acetabulum. The distance between the acetabular margin and the ischial tuberosity (superior margin) reflects the combination of the above characters and has no biomechanical significance. What is significant is the distance from the center of the acetabulum to the impression on the ischial tuberosity of the common head of origin of the hamstring muscles (the mechani- cal lever arm of the hamstrings); this metric dimension must then be corrected for gross body size differences, usually via a ratio involving acetabular diameter. When the biomechanical ischial length is thus deter- mined, the Sts 14 ischium is very short, in the upper ranges of modern human populations (Howell 1955; Lovejoy 1973, 1975; Lovejoy et al. 1973; McHenry 1975b; Robinson 1972; Schultz 1969). Robinson (1972, 1977) maintains, however, that the ischium is long in SK 50. The length of the SK 50 ischium (biomechanical length, as defined above) is considerably longer than that of Sts 14 (60 and 43 mm, respectively, Lovejoy et al. 1973). The problem is how to correct this length for body size in SK 50. As noted, most workers have used maximum acetabular diame- ter as the correcting factor. In SK 50, this dimension is 50 mm (Lovejoy 1975). Thus the relative length of the Swartkrans ischium is expressed as: Maximum Acetabular Diameter Biomechanical Length of Ischium This computation is equal to Robinson's (1972) value 9 0 for Sts 14 (83%). (Robinson,1 977, cites an ischial length of 69 mm for SK 50, but states that the SK 50 ratio is 62%). Despite the striking visual impression of a long ischial shank in this fossil, SK 50 may well possess a functional ischial length equivalent to that in Sts 14 and well into the range of modern man. Given the condition of the acetabulum in SK 50 (it is clearly distorted), it is impossible to accurately determine is- chial length. As Lovejoy (1975) has illustrated, regard- less of the measurements and indices Robinson utilized, his data on modern humans (x = 68, range 56-8 1, n = 40) do not merit excluding his 62% figure for the "Paranthropus" innominate. If a small acetabulum is a populational character of Aus- tralopithecus, compared to modern Homo sapiens, then acetabular diameter may be an unsuitable parameter for size correction, especially given the state of preser- vation of SK 50. Lovejoy, Heiple, and Burstein (1973) suggest the use of maximum iliac height but this is not determinable on SK 50. Standardizing via the auricu- lar surface area of the ilium may also be inappropriate, since this is smaller in Australopithecus than in modern populations, and the auricular surface is absent in SK 50.5 Obviously, suitable parameters for size correc- tion applicable to Australopithecus are needed. The im- portance of ischial length is that wide ranging infer- ences about locomotor differences between A. africanus and A. robustus have been based upon this one morphological observation (i.e., Robinson 1972), an observation clearly not substantiated by the available sample. The form of the ischial tuberosity in Australopithecus is somewhat different than that of modern man, in- cluding differences in shape, disposition, and extent up the ischial shank. The narrow, flattened tuberosity in Sts 14, with a rather sharp lateral edge, was noted by Broom et al. (1950). The tuberosity does not extend as far towards the acetabulum (because it is interrupted by a large tuberoglenoidal sulcus). The distinct impres- sions made by the common origin of m. semitendinosus and m. biceps femoris (long head) on the lower part of the tuberosity, and by m. semimembranosus on the more superior part, are not visibly differentiated. Chopra (1962) reported that all primates with a flattened tuberosity and sharp lateral ischial border possess a m. gluteus maximus which passes laterally, not over, the tuberosity. Since Sts 14 possessed this type of tuberos- ity, he concluded that in Australopithecus africanus m. gluteus maximus was an abductor as in pongids, not an extensor (as in hominids). To the contrary, the aus- tralopithecine ilium indicates an origin of m. gluteus maximus posterior to the hipjoint and thus an action as a powerful extensor. Robinson (1972) also has re- ported that in the Sts 14 original specimen the tuberos- ity surface is absent and areas of cancellous bone are exposed. The SK 50 ischial tuberosity exhibits an un- usual flange-like eversion (Howell 1955; visible on cast); the tuberosity surface is damaged also (Robinson 1972). The known acetabula of australopithecines are abso- lutely and relatively smaller in diameter than Homo (Day 1973; Lovejoy 1976b; Napier 1964; Robinson 1972; Zihlman 1967), which correlates with smaller femoral head diameters and smaller body size. The SK 50 acetabulum is relatively small and would be smaller if corrected for distortion. The SK 3155b acetabulum is almost exactly equal in diameter to Sts 14 (McHenry 1975c). Napier (1964) states that the acetabulum of SK 50 is more laterally disposed (more pongid-like) than in Sts 14 but given the condition of the former fossil this is not justified. The only other robust aus- tralopithecine fossil with an intact acetabulum is SK 3155b; here, however, the essential landmarks neces- sary to accurately orient the pelvis (minimally, the an- terior superior iliac spine and the pubic symphysis) are absent. All known early hominid acetabula are rela- tively deep: Femora The hominid femur is a stout bone which clearly reflects a bipedal adaptation, and thus fossil hominid femora are very important in gait reconstruction. For- tunately australopithecine femora are fairly numer- ous, but complete and undistorted femora are rare (KNM-ER 999, 1472, 1481a; AL 288-1). Most fossil femora are proximal, distal, or diaphyseal fragments. The proximal femoral anatomy of Australopithecus is well known and a fairly distinct morphological pattern is obvious. All observers have noted the presence of a small femoral head (even when standardized via femoral shaft diameter, measured just inferior to the lesser trochanter, Day 1977; McHenry and Corruccini 1976; Napier 1964; Wood 1976). Our knowledge of head diameters for A. africanus is based upon the re- constructed head in Sts 14, which is severely damaged. We may note, however, the lack of significant differ- ences in acetabular diameter between gracile and robust early hominids and logically correlate the small acetabula with small femoral heads (Zihlman 1970). The small femoral head was not proportionately small for the length of the femur (Lovejoy 1975; Walker 1973). Two hominid femora from East Lake Turkana, KNM-ER 1472 and 148 la, have femoral heads consid- erably larger than Australopithecus and comparable to Homo (Leakey 1973a; McHenry and Corruccini 1976). Jenkins (1972) has stated that the femoral head coverage is different in A. robustus, and intermediate in form between the posterior elongation of the articular surface characteristic of Pan, and the anterior prolon- gation seen in Homo (correlating with the more shal- low, laterally disposed acetabulum of chimpanzee, and the deeper, more ventrally disposed acetabulum of man). In man, a small articular facet encroaching onto the anterior aspect of the neck is present, in the ab- sence of an anterior sub-capital sulcus, but with a fre- quency of only 10 per cent (Harty 1973). As Lovejoy (1975) has noted, the Jenkins (1972) reconstruction is 9 1 based upon the damaged and distorted SK 50 acetabulum and the SK 82 proximal femur, which on the basis of gross size differences are clearly not a suitable match. Australopithecus possessed long femoral necks (Day 1969; Lovejoy and Heiple 1972; McHenry and Cor- ruccini 1976; Napier 1964; Walker 1973; Zihlman 1970), relatively flattened anteroposteriorly (Day et al. 1976; Johanson and Coppens 1976).6 The neck is set upon the femur at a lower collo-diaphyseal angle (coxa vara) than in modern human populations (coxa valga) (Day 1977; Lovejoy 1976b; Walker 1973; Zihlman 1970). The values listed in Lovejoy (1977) and Lovejoy and Heiple (1970) range from 1150 (SK 83, Old. Hom. 20, KNM-ER 738 & 815) to 1200 for SK 82; the mean in one Amerindian sample was 128?(1 S.D. = 3.6+). The morphology of the greater trochanter is notice- ably different in Australopithecus. This traction apophysis appears smaller, and exhibits less lateral flare and vertical (proximal) extension than in modern humans (Day 1969; McHenry and Corruccini 1976; Zihlman 1970). Impressions for the insertion of mm. gluteus minimus et medius, obturator internus and gemellus superioret inferior are visible (Day 1969; Day et al. 1976; Johanson and Coppens 1976). The morphology of this region in KNM-ER 1472 and 1481a is again more like that of Homo (greater lateral flare, Lovejoy 1977; also KNM-ER 999, Day and Leakey 1974). The quadrate tubercle (on the intertrochanteric crest, at the distal end of the greater trochanter) is apparent in KNM-ER 1503 (Day et al. 1976), in the proximal femora from Hadar (Johanson and Coppens 1976), and can be seen on a cast of SK 82 (this area is damaged in SK 97). Although some australopithecine femora illustrate a posteriorly positioned lesser trochanter (Day 1969; Napier 1964), there is considerable variation in both australopithecine (Walker 1973) and modern popula- tions with regard to this trait. The position of the lesser trochanter does not lie significantly outside the range of Homo sapiens (Lovejoy 1975). Day (1973) states that the vertical groove for the iliopsoas conjoint tendon is visible as it approaches the lesser trochanter, in SK 82 and 97. Although Napier (1964) suggested intrageneric dif- ferences in the depth of the trochanteric fossa, all Australopithecus femora possess deep fossae (Day 1969; Day et al. 1976; Johansori and Taieb 1976; Lovejoy 1975a), with a clearly delineated groove for the tendon of m. obturator externus (Day 1973; Lovejoy 1975; Walker 1973). Controversy has arisen over the presence and sig- nificance of the intertrochanteric line and the femoral tubercle in Australopithecus. Day (1969) and Napier (1964) had stated that these were absent. Many speci- mens, however, do exhibit a weak intertrochanteric line and femoral tubercle (Day 1977; Day et al. 1976; Johanson and Coppens 1976; Robinson 1972; Walker 1973). The weak line resembles more a synovial reflec- tion than a ligamentous attachment, and is known to be an age-related feature with considerable variability (Lovejoy 1975). The line and tubercle are not apparent on casts of SK 82 and 97, but Robinson (1972) states that they are visible on the originals, with the femoral tubercle being more incorporated into the greater trochanter than in modern man. Cresting on the proximal posterior aspect of the shaft conforms to a modern pattern, with a well de- veloped crista hypotrochanterica and pectineal line (Day 1969). More distally, the spiral line and the lateral lip of linea aspera are seen on femoral shaft fragments from Hadar (Johanson and Coppens 1976). The femoral shaft can be quite stout in certain specimens. The distal femoral anatomy of Australopithecus is ex- tremely modern. This includes the relatively deep (non-pongid-like) patellar fossa with an elevated lat- eral lip; a narrow but deep intercondyloid fossa, and the relatively elongated and flattened lateral condyles (Broom and Schepers 1946; Clark 1947; Lovejoy et al. 1973). The proportions of the condylar articular sur- faces are modern (Walker 1973), contrary to Kern and Straus (1949) and Preuschoft (1971). The posterior 1/2 of the groove for the m. popliteus tendon is visible in TM 1513 near the lateral margin of the lateral condyle (Clark 1947). The attachment of the posterior cruciate ligament is observable on the medial condyle in the Hadar specimen (Johanson and Coppens 1976), as well as a deep popliteal groove postero-inferior to the lateral epicondyle. Day (1973) found grooving for m. popliteus and asymetrical grooves for the menisci (on the condylar surfaces) in Sts 34 and TM 1513, evidence for the presence of a knee un-locking and locking mechanism (respectively) similar to that of Homo. The bicondylar (femoro-condylar) angle of TM 1513 was determined by Clark (1947) as ca. 70, simulat- ing the midline of the shaft via the highest point of the intercondyloid fossa. Kern and Straus (1949) felt that an angle of 70 was neither unique nor discriminatory. Heiple and Lovejoy reconstructed the bicondylar angle in TM 1513 and Sts 34 (as the angle of the central axis of the femur with the vertical to the bicondylar plane) and found the previous estimates too low, since the centroid axis of the femur did not necessarily in- clude the highest point in the intercondyloid fossa. Their estimates were 140 for TM 1513 and 150 for Sts 34, figures higher than those for modern man (x = 11 0) and substantially beyond those for Pan (60; Heiple and Lovejoy 1971; Lovejoy and Heiple 1970). The very high femoro-condylar angles have been confirmed by additional material and by other workers (Day 1973; Johanson and Taieb 1976; Walker 1973). The lengths of australopithecine femora apparently varied considerably (as did their robusticity). Broom et al. (1950) reconstructed the Sts 34 femur as 310 mm long. Lovejoy and Heiple (1970), utilizing the new bicondylar angle determination, reconstructed a length of 276 mm for Sts 14, and, using a composite Sts 9 2 14/SK 82, 310 mm for A. robustus. Walker (1973) re- constructs a length of 250 mm for Sts 14, and 360 mm for a composite (Old. Hom. 20/KNM-ER 993) femur of A. boisei. The crushed femur of AL 288-1 is ca. 280 mm long (Johanson and Taieb 1976). Leg: Tibiae and Fibulae The tibia (shaft and distal end) from Olduvai, Old. Hom. 35, was described by Davis (1964). He concluded that the horizontal, slightly anteriorly disposed distal articular surface indicated an ankle region of modern morphology. The proximal shaft morphology, accord- ing to Davis, indicated significant differences from Homo. He described the area of origin for m. tibialis posterior as more extensive than that of m. flexor dig- itorum longus (on the tibia), a pongid condition. He suggested that m. popliteus pulled in a vertical direction and m. soleus had a less extensive tibial attachment. Day (1977) and Lovejoy (1975) independently concluded that Davis had misidentified the soleal line. Appar- ently the "strong marked crest" discussed by Davis (1964) is the soleal line, the upper, more anteriorly located portion. The soleal line is more vertically dis- posed than in some modern tibiae, but this trait is variable (e.g., Lovejoy 1977). There is thus no need to suggest anything but an oblique line of action of m. popliteus (Lovejoy 1975). The area for m. popliteus is damaged and surface bone is missing (visible on the Wenner-Gren cast). A prominent soleal line is evident in the AL 129- lb proximal tibia from the Afar triangle (Johanson and Coppens 1976). A well preserved tibial fragment from East Lake Turkana, KNM-ER 741, illustrates a basically mnodern morphology. The Old. Hom.35 tibia was originally part of Old. Hom. 6, a "Homo habilis" Bed I paratype (Leakey, Tobias, and Napier 1964); KNM-ER 741 was assigned to Aus- tralopithecus, although Day (1977) has suggested that it has more affinity with Homo (along with KNM-ER 1471, 1476b, and 1481a). Unfortunately, matching tibial parts are not available for comparison in taxonomic assignment, and all such attempts are strictly suggestive. The Old. Hom. 35 figula is essentially modern in morphology (Davis 1964). The right distal fibula and tibia of AL 288- 1, undescribed, should hopefully con- firm our interpretations of this area for Au- stralopithecus. Pes Several relatively complete early hominid tali are known. The first recovered, TM 1517 from Krom- draai, is a small specimen with considerable-damage to the inferior surface. It exhibits a relatively narrow articular surface for the malleoli, and a broader, more horizontal (less angulated), more medially extensive neck with a relatively large articular surface for the navicular. The horizontal neck angle (angle between the axis of the neck and that of the trochlea) is wide (estimated as 32?by Day and Wood 1968; Wood 1974c). The medial lip of the trochlea is elevated as in Homo (observation on cast; Clark 1947; Gregory 1949; Preuschoft 1971; Wood 1974c). The Olduvai talus (Old. Hom. 8) also exhibits a relatively high horizontal neck angle (280, incorrectly cited as 32.50 by Robinson 1972), as does KNM-ER 1476a (Day 1976a; Day et al. 1976; Day and Wood 1968; Wood 1974c), although the Old. Hom. 8 talus has a higher torsion angle of the neck (40?) than TM 1517 (340) (Day and Wood 1968). Multivariate analysis has consistently segregated Old. Hom. 8 and TM 1517 from the more modern talus from East Lake Turkana, KNM-ER 813a (Day and Wood 1968; Wood 1974a and b). The KNM-ER 813a talus, and probably KNM-ER 1464 (per Day 1976a and b, and contra Leakey 1973b), would be placed in Homo on the basis of their strikingly modern morphol- ogy: a narrow neck/body angle, a high torsion angle of the head, and a less extensive navicular articular facet (Leakey 1972; Wood 1974c).7 There is evidence of two morphological patterns, one which is assigned to Homo (KNM-ER 813) and another to Australopithecus (TM 1517, and Old. Hom. 8, originally a paratype of "Homo habilis" Leakey, Tobias, and Napier 1964, and suggested as more appropriately placed in Aus- tralopithecus, Day 1976a and b; Wood 1974c). The only published calcaneum is that of Old. Hom. 8, fragmentary but with a horizontal, very modern sustentaculum tali (Day and Napier 1964). The Old. Hom. 8 medial cuneiform, according to LewNTis (1 972a), has a primitive (Pan gorilla-like) distal articular sur- face, being markedly convex superiorly and concave inferiorly. The Olduvai metatarsals show the thick proximal cross-section, with large proximal joint sur- faces, characteristic of a hominid pedal loading pat- tern (Preuschoft 1971). The index of robusticity is hominid (Archibald et al. 1972, contra Day and Napier 1964). The general proportions of the pes, with a relatively long tarsus, average-length metatarsus, and short phalanges (compared to non-human primates), is characteristic of hominid entaxonic weight-bearing (Lessertisseur and Jouffroy 1975). The terminal hallucial phalanx from Olduvai, Old. Hom. 10, is typically hominid in being extremely broad and stout with axial torsion and valgus deviation, and has been discussed as the toe bone of a "fully moderni bipedal strider" of the genus Homo (Day and Napier 1966; Day 1967, 1974b) although Wood (1974c) has discussed placing it into Australopithecus. Biomechanics of Early Hominid Locomotion Forelimb The forelimb remains of Australopithecus illustrate a distinct morphological pattern. The lack of compara- ble body parts between gracile and robust forms is critical and any conclusions must be qualified (as to whether a given part is available for study in both taxa). We can conclude that all australopithecines pos- sessed relatively longer arms than modern man, and 9 3 that the upper limb morphology has some distinctly pongid-like features. These two observations are obvi- ously related and the crux of the issue is whether the second reflects (1) retention of adaptive complexes to suspensory/supportive posture and locomotor be- haviors in the common African ape ancestor (synplesiomorphy), or (2) the actual use of such mor- phological adaptations in habitual forelimb suspensory/supportive behavior. The evidence for the latter suggestion with regard to the pectoral girdle is the glenoid fragment from Sterkfontein (Oxnard 1968a), particularly dealt with in the morphometric analysis of this glenohumeral joint by Ciochon and Corruccini (1976). The morphology of known (at least gracile) Aus- tralopithecus humeri is consistent with that of an animal with fully developed rotatory capabilities at the glenoid fossa, with the head of the humerus secured via strong rotator-cuff muscles and, presumably, a strong superior glenohumeral ligament (since the en- tire weight of the upper limb at rest is countered by the superior glenohumeral ligaments, Basmajian 1972). There is evidence of a powerful m. biceps brachii, long head. The humeral shaft (robust forms) indicates de- velopment of powerful abduction (also seen in short scapulai- muscles). The humeri for all aus- tralopithecines indicate strong common forearm and hand flexors. If the distal humerus can be used to discriminate knucklewalkers (possessing a steep lateral wall of the olecranon fossa, McHenry 1976b; Day 1976b), then there is no evidence of knuckle-walking in the australopithecine elbow join. Conclusions about the forearm are virtually limited to the Omo L40-19 ulna, which exhibits traits resem- bling pongids, hominids, and numerous features which are totally unique. The ulna (known only for robust forms) suggests some sort of supportive, post- ural, and/or active locomotor use of the forelimb (Howell and Wood 1974; Day 1977). When the mor- phology of the ulna is considered in conjunction with that of the humerus, there is evidence in at least the robust lineage of the use of the upper limb in a non- Homo sapiens fashion. The australopithecine manus is also illustrative of a myriad of features, some of which are hominid-like, somne pongid-(particularly, Pan-)like, and some totally unique. Again the lack of comparable body parts be- tween the early hominid lineages obstructs full analysis, yet one can conclude that at least some carpal bones, in all forms, were quite primitive (i.e., Pan-like). The robustness of the digits and particularly the well developed lateral ridging on the palmar phalangeal surfaces combines with evidence in the shoulder and in the distal humerus to indicate an animal capable of very strong flexion, pronation, and supination in the superior member. McHenry (1937b) has emphasized that body size is critical in interpreting the morphology of the KNM-ER 739 humerus, and in a sense this applies to the entire forelimb anatomy. Certainly the interpreta- tion of the robustness of the KNM-ER 739 humerus depends upon whether Australopithecus boisei is recon- structed at 100 or 200 pounds. The total morphologi- cal pattern of the upper limb, albeit known from an insufficient sample, is reasonably consistent and is suggestive of use of the forelimb in behaviors not ordinarily adopted by modern man. The evidence is found in the morphology, size, robusticity, and pro- portions of the various skeletal elements, and is appar- ent in all forms of Australopithecus. The lack of suitable parameters for discussing locomotor behaviors in early hominids is a serious problem. The use of broad categories such as "brachia- tion" in this context is inadequate, since the forelimb anatomy of the extant pongids varies considerably (particularly in the manus, Lewis 1973; Tuttle 1975), as does the manner in which these apes utilize the forelimb in locomotor, postural, and feeding be- havior. The unique forelimb anatomy of Aus- tralopithecus may be the result of a different complex of postural-locomotor adaptations and behaviors, with several underlying broad similarities to extant pongids but in toto different from them all. Our sample of extant hominoids is very limited when compared to the total range of adaptive complexes which must have occurred in hominoid evolution. The Neogene faunas give us an indication of this diversity, but unfortu- nately the hominoid sample is more often dental and gnathic remains rather than postcranial. The post- crania preserved do not show strong affinities to mod- ern ape morphology. The anatomy and morphomet- ric affinities of the Miocene ape postcrania are in many ways strikingly primitive (i.e., cercopithecoid-like; Corruccini et al. 1976; Napier and Davis 1959; Morbeck 1972; O'Connor 1974) and should not be equated with those of modern apes (e.g., Lewis 1973; Conroy and Fleagle 1972; Zwell and Conroy 1973). Tuttle (1967) has shown that the human hand does not reflect the same kind of adaptations to knuckle- walking seen in the extant African apes, but this does not preclude the possibility of some form of suppor- tive role of the forelimb (terrestrial or arboreal), espe- cially if the adaptation were more pronounced in the soft tissues and would not be fossilized. Lovejoy (1977) has suggested that if one can delineate a complete bipedal adaptation (striding gait) in Australopithecus, then the use of other locomotor behaviors is a priori excluded. This viewpoint is based upon the modern human gait, seen as a very restrictive locomotor adap- tation; the existence of a different type of bipedal adaptation, in an animal of different body propor- tions, means that bipedal behavior in Australopithecus did not necessarily preclude the resumption of other postures and locomotor behaviors. The type of biomechanical analysis seen in studies on the human hindlimb are not yet available for the 9 4 hominoid forelimb, although considerable progress has been made in recent years. A full analysis of the australopithecine forelimb must be done in the con- text of a comparative study of pongid forelimb anatomy, with emphasis on how locomotor behaviors and soft tissue adaptations are reflected in the bony skeleton. An emphasis on the varied and complex use of the limbs in postural, locomotor, and feeding be- haviors in the Hominoidea is essential in reconstruct- ing such possible behaviors in early hominids. I sus- pect such analysis would find the role of the forelimb in australopithecine behavior to be unique among hominoids. Hindlimb The functional morphology and biomechanics of modern human bipedalism are well known, as are the major evolutionary changes involved. The focus with Australopithecus is to what degree are these changes in anatomical adaptive complexes developed. A major, perhaps pivotal, adaptation to bipedalism, of a broad, short, laterally displaced iliac blade, and re-orientation of the gluteal muscles, is witnessed in Australopithecus. Thus australopithecines had achieved the placement of m. gluteus maxcimus to a position behind the hip joint, enabling the muscle to be a powerful thigh extensor (Clark 1955; Washburn 1950). Although Robinson et al. (1972) suggest that there has been no major change in the function of m. gluteus maximus from hominoid to hominid, the hypertrophy of the muscle, particularly in its cross-sectional area, directly relates to its power- ful role in hip extension (Ishida 1968; Lovejoy 1975, 1977; Reynolds 1931). M. Gluteus maximus in apes actu- ally consists of two divisions: (a) a more superior por- tion, the sacroiliac part (in Pongo, m. gluteus maximus proprius), innervated by the inferior gluteal nerve and acting as a powerful abductor and lateral rotator, and (b) a more inferior portion, the ischiofemoral part innervated by the flexores femoris nerve and posses- sing some extensor capability. The evolution to the hominid condition entailed the loss of the is- chiofemoral portion (Sigmon 1975; Tuttle et al. 1975). M. gluteus maximus became most important in full extension of the thigh, especially in situations demand- ing considerable power, as well as aiding in the maintenance of stability at the knee joint (via its partial insertion into the iliotibial tract) and in checking the forward momentum of the limb in swing phase (Robinson et al. 1972; Basmajian 1972; Washburn 1950). One of the most significant aspects of the modern human bipedal adaptation is the system of lateral pel- vic balance control (Reynolds 1931). The reorientation of the iliac blades, positioning the anterior gluteal sur- face and margin in a lateral-ventral position, aligns mm. gluteus medius et minimus for abduction of the hip joint (losing their extensor function in pongids). This abduction acts to limit the displacement of the center of gravity in a horizontal plane. When one leg is inI stance phase the abductors cantilever the pelvis from the femur and limit contralateral pelvic lowering of the limb in (unsupported) swing phase. Action on ver- tical displacement is accompanied by limitation of dis- placement in a sagittal plane, via internally (medially) rotating the thigh beginning at heel-strike, and exter- nally (laterally) rotating from toe-off. Absence of this abductor-rotator control of limb progression results in a clinical manifestation (Trindelenberg gait) which in- volves a smaller cadence, the feet placed further apart, and considerably more side-to-side "rolling." The greater lateral iliac flare and greater distance between anterior superior iliac spines have been con- sidered indicative of less efficient lateral balancing in Australopithecus (Zihlman 1967), via positioning the in- ternal rotators more laterally (therefore in a less effi- cient porsition). Furthermore, the supposed lack of an iliac tubercle and pillar (Napier 1964; Mednick 1955) would indicate less developed gluteus medius and minimus musculature. Zihlman and Hunter (1972) have attempted a biomechanical reconstruction of the hip joint, focusing upon internal rotation and abduc- tion in MLD 7 (adolescent left ilium). They conclude that Australopithecus developed considerably more torque in internal rotation than modern man, but had to exert more force to rotate the trunk for less distance at a lower speed. The medial movement of the anterior superior spine from Australopithecus to Homo allowed more direct pull and shortened the power arm of the internal rotators, facilitating more efficient internal rotation. The authors concluded that the relatively more massive upper limbs aided in internal rotation (also Genet-Varcin 1969). Lovejoy and his associates have studied the abductor complex in Australopithecus and have reached very dif- ferent conclusions. They feel that the beaked anterior superior spine is the result of compensation for great- er iliac flare. The beaking is simply the adjustment necessary to place the anterior superior spine into the correct ventral plane for alignment of the inguinal ligament and m. sartorius (Lovejoy 1976, 1977; Lovejoy et al. 1973). The combined effect of a more lateral anterior iliac region, a longer acetabulospinous dis- tance, a longer femoral neck, a smaller collo- diaphyseal angle, and a very. high femoro-condylar angle is to give the abductors a long lever arm (right angle distance between abductor line of action and the center of the acetabulum) for control of pelvic tilt, reducing the muscle force necessary for equilibrium and the total force on the hip joint (Lovejoy 1973, 1975, 1977). Thus the long femoral neck, greater iliac flare, and more protuberant anterior superior spine form an adaptive complex for minimizing pelvic sag- ging, hip joint pressure, and energy expenditure in bipedal striding (Lovejoy 1976; Lovejoy et al. 1973). Lovejoy concludes that the longer femoral neck and more protuberant anterior superior iliac spine would 9 5 allow effective internal rotation like that of modern man (1973). He further correlates this complex in australopithecines with the distinctly less flared great- ter trochanter. Efficient abductor lever arm length is accomplished by long femoral neck, lateral iliac flare, and protuberant spine, and therefore a flared greater trochanter is not necessary. The small acetabulum and femoral neck diameter would simply reflect small body size and relatively low femoral head pressure (Lovejoy et al. 1973). This entire pattern of pelvic and femoral morphol- ogy may be related to one important dimension: in- teracetabular diameter. The australopithecine pelvis is characterized by a relatively large false-pelvis diameter and relatively small true-pelvis diameter. The major change fromAustralopithecus toHomo, according to this paradigm, would entail maximizing interacetabular (pelvis minor) dimensions for the passage of full-term fetuses with relatively enlarged cranial capacities (Lovejoy 1973; Lovejoy et al. 1973). In this regard, Leutenegger (1972, 1974) has shown that Aus- tralopithecus had relatively small fetal cranial diameters compared to pelvic outlet breadth. Changes in Homo, accompanying rapid encephalization in the lower and middle Pleistocene (McHenry 1975a), would include increasing anteroposterior pelvic outlet diameter via lengthening the superior pubic ramus and increasing retroflexion of the inferior end of the sacrum; in- creased coronal interacetabular distance would be ac- complished via an increase in sacral breadth, which would place the alae of the ilium in a more sagittal orientation (reducing lateral iliac flare and beaking). The shortening of the abductor lever arm (less iliac flare, shorter femoral necks, and a greater collo- diaphyseal angle) would result in less efficient abduc- tion and greater femoral head pressure. Greater femoral head pressure would also result from in- creased body weight, thus selecting for maximum diameter of the femoral articular head and the acetabulum. Given some absolute limit on maximum body width at the hip joint, relative femoral neck length would be decreased as birth canal diameter increased. The medial approximation of the anterior superior spines would maintain efficient internal rota- tion (by maintaining the length of the internal rotator level arm), and develop the characteristic acetabulo- cristal and acetabulospinous pillars of Homo sapiens. The distinct characteristics of Australopithecus in the ilium (confirmed in multivariate sorting by McHenry and Corruccini 1 975b) and the femur can be related to a bipedal adaptation in a form of small body size, small interacetabular diameters, and full-term fetuses with small cranial diameters. Lovejoy, and Wood (1973), emphasize that a different pattern need not imply less efficient bipedalism. I feel that the tendency to equate anything morphologically not identical with Homo sa- piens to some sort of less efficient bipedalism is the result of (a) looking at the striding gait as an adaptive pinnacle, and as the sole activity selected for in lower limb evolution, and (b) the lack of animals with which to compare. Certainly striding gait is not the only bipedal locomotor activity of modern man, and such a gait was the result of the interaction of selective pres- sures and adaptive compromises for bipedal locomo- tion and the obstetrics of large-brained newborn. The human gait is unique, so we have little knowledge of the possible adaptive pathways involved in reaching that pattern. Presumably the postcrania of Aus- tralopithecus evidence one such adaptive complex dif- ferent from that of modern man. The existence of two such morphologies need not preclude their involve- ment in an ancestral-descendent lineage. The two morphologies can be used to sort Australopithecus from Homo, at least for the ilium and proximal femur; this has been confirmed by the discovery of two well pre- served innominate fragments from eastern Africa. Both Old. Hom. 28 (Olduvai Gorge) and KNM-ER 3228 (East Lake Turkana) exhibit a large acetabulum, a distinct acetabulocristal pillar of modern form and position, and an essentially modern ischial tuberosity. The Olduvai specimen was recovered from deposits of Bed IV and has been assigned to Homo erectus (Day 1971). KNM-ER 3228 was recovered from the Lower Member, Koobi Fora Formation, and has been as- signed to Homo sp. (Leakey 1976). The available australopithecine ischia do not imply any significant difference from modern humans in the biomechanics of hip extension (i.e., longer ischia and more power-oriented hamstrings; Robinson 1972; Sigmon 1975). Australopithecus (all forms) had already evolved a shortened ischium which developed the hamstrings into purely propulsive muscles, extending the thigh with greater speed and less power, through a longer arc, than in pongids (Robinson et al. 1972). The femur lengthened as a propulsive strut, and the inser- tion of m. gluteus maximus moved superiorly on the femoral shaft, specializing for power extension. Pre- sumably this indicates (along with the protuberant an- terior superior spine for m. sartorius insertion and anterior inferior spine for m. rectus femoris, straight head insertion) greater development of the quad- riceps than the hamstring muscle group and the re- quirements of the principal stresses of the hindlimb met by one-joint muscles (both characterize modern man, Haxton 1947). The evidence for the morphology of the hip and knee joints in Australopithecus indicates modern form. There is a relatively broad superior aspect of the lu- nate facies of the acetabulum and modern femoral head coverage. The visible markings for the Y-shaped ligament of Bigelow (iliofemoral) indicate the pres- ence of this mechanism for prevention of hyperex- tension at the hip joint, via ligamentous, not muscular, action (Joseph 1960). Habitual full extension of the lower limb reduced the stress upon the long bones (Kummer 1968), eliminating much tonic muscular ac- 9 6 tivity reqjuired for idealized erect posture and ap- proximating the "best fit" position of the femoral head in the acetabulum (Hall 1965). Such habitual exten- sion of the lower limb is evidenced in the well defined iliopsoas groove and its vertical fiber markings near the lesser trochanter. The size of the distal femoral condyles, the high valgus angle to the femur, and details of the intercon- dyloid fossa indicate a knee joint fundamentally mod- ern in design. The relatively narrow spatium interos- seum, the lack of fibular robusticity, and the lack of developed tibial curvature indicate decreased bending stresses at the knee compared to pongids. The funda- mentally modern femoral condyles, with elliptical pro- longation of femoral-tibial articulation, maximize the surface area of vertical weight transmission. Maximum cross-sectional area of weight transmission (perpendicular to the line of gravity) over the support- ing strut, with dorsal/caudal migration of the center of gravity, optimizes stability of the lower limb. The sta- bility of the knee joint, dependent upon its supportive soft tissues (Hall 1965), is apparent in Australopithecus from impressions of the intra-articular cruciate liga- ments (Day et al. 1976; Johanson and Coppens 1976). The morphology of the ankle joint is not well under- stood, but based upon the horizontal articular surface of the distal tibia and the talar trochlea, one may pre- sume that the stability of the talo-crural joint was main- tained (as in man) by tonic contraction of m. triceps surae (particularly m. soleus, Joseph 1960; Tuttle 1975). The australopithecine foot appears to illustrate the type of bow-string loading adaptation (longitudinal and transverse arching) seen in modern humans (Kummer 1975). There are definite limitations to our reconstructions of lower limi and joint biomechanics in early hominids. The analyses of Lovejoy are (quite sound methodologically, but are based upon work on casts of Sts 14 (the proximal femur of which is entirely recon- structed, Day 1973) and the distal femora f'rom Sterkfontein, which are considerably larger than Sts 14 (Walker 1973). I doubt that these reconstructions are fundamentally in error, but until confirmation is made via the analysis of the AL 288-1 lower limb (complete body parts of one associated individual), they must remain suggestive. The implications of Lovejoy's work is that the anterior iliac region and the proximal femur of modern man should show consid- erable sexual dimorphism (demonstrated for the pel- vis by Coleman 1969). Males, without the necessity of maximizing the birth canal diameter, should retain more australopithecine-like features (more lateral iliac flare, more pronounced anterior superior spine, nar- rower sacrum, shorter pubis, and longer femoral neck and/or smaller neck-shaft angle). He sees evidence of this dimorphism in a sample of Amerindian pelves (Lovejoy 1975), but a thorough review of the literature on modern human sexual dimorphism in these areas of the pelvis and femur ought to be attempted to substantiate this. In this regard, Walker (1973) has noted sexual dimorphism in the collo-diaphyseal ang- le, with females having angles up to 3? greater than males. There is substantial evidence for two separate mor- phologies in the early hominid talus; these can be used to discriminate A ustralopithecus and Homo, yet interpre- tation of such differences in a functional paradigm is difficult. There is considerable complexity in the cor- relation of talar neck angles to locomotor habits (Bar- nett 1955), and we know that a high horizontal neck angle does not necessarily imply a habitually abducted hallux (i.e., Old. Hom. 8). Lovejoy (1977) has em- phasized talar variability in modern man, and the role of the three joints which lie between the talus and hallux in modifying the influence of the talar horizon- tal neck angle upon hallucial position. The Old. Hom. 8/TM 1517 talar morphology does not necessarily iInply lack of a striding gait (i.e., Day and Napier 1964; Day and Wood 1968). A logical assumption is that the australopithecine foot was used in a wider variety of locomotor contexts, on a more varied set of substrates, than modern man. It is difficult, however, to dem- onstrate this directly, with the bony morphology, or indirectly, with inferences about the long tendons and ligaments of the foot. The foot is used in modern man to adjust to different terrain and loading patterns; this compensates for the simple and relatively rigid design of the knee joint (Elftman 1945). If the aus- tralopithecine foot is indicative of greater flexibility, then this function of the ankle joint would be critical since a modern knee morphology had already been accomplished. Data correlating the morphology (and variation) of the talus and other tarsals with locomotor behaviors in a wide variety of primates could aid in this matter. More foot fossils have been recovered, yet to be described, froin Hadar. These include the talus from AL 288-1, and several fragmentary calcanea, navicula, and tali, a cuneiform, metatarsals, and phalanges from locality AL 333. Limitations and Applications: Postcranial Data The methodological problems discussed above im- pose limitations on the analysis of early hominid post- cranial fossils. One of the most difficult problems is in the unique nature of the fossil morphology, and the lack of living analogies. Thus, it is difficult to concep- tualize the locomotor behaviors of these early hominids and correlate it with morphology. We can conclude that, as yet, there is no consistent and demonstrable difference in the postcranial skele- ton of gracile and robust australopithecines other than size. This is a conclusion reached by a number of workers familiar with the original material (Day 1977; Walker 1973; Wood 1974c). The central thesis of Robinson (1972), that there are sufficient morpholog- ical (and inferred locomotor) differences between the 9 7 two lineages to warrant generic distinction (correlated with his interpretation of the dental and cranial evi- dence), is not supported by the presently available material. Far too much of this thesis depends upon the interpretation of the SK 50 ischium, discussed above, which cannot be substantiated. The non-modern morphology of the "Paranthropus" talus and carpals is also reflected in gracile australopithecine foot and hand morphology and thus do not support his generic distinctions. What is needed to determine possible intra-generic differences is a partial or relatively complete as- sociated skeleton (of one individual) in direct associa- tion with cranial and dental parts clearly assignable to a robust early hominid. Then the morphology of robust forms could be determined with a high level of confi- dence and 'possible specific differences delineated. Without such evidence there is the danger of circular- ity of reasoning, assigning postcranial fossils to a given species on the basis of assumed differences in mor- phology and behavior. It is interesting to note that all relatively complete associated skeletons recovered be- long to the gracile australopithecine/early Homo lineage. If this situation persists in areas of good to excellent (complete) preservation (i.e., Koobi Fora, Hadar), then perhaps some taphonomic process was operating which has some paleoecological signifi- cance. The postcranial morphology of Australopithecus can be used as taxonomic criteria at the generic level. There is a distinct set of characters, well seen in most of the forelimb, ilium, femur, and foot, which is not modern and is clearly sortable from that of Homo, even in the lowermost Pleistocene. The diagnosis of the genus Australopithecus can thus be extended to include elements of the postcranial morphology. The consis- tency and distinctiveness of this morphology supports the generic separation from Homo and confirms the notion of an australopithecine adaptive grade. Locomotion in Australopithecus Locomotor adaptations in Australopithecus were a mosaic of pongid, modern Homo, and totally unique morphologies and behaviors. These early hominids had rather primitive limb proportions and apparently used both forelimbs and hindlimbs in a greater variety of behaviors than modern man. The upper limb evi- dence suggests use in postural and supportive ac- tivities. Whether this can be correlated with paleoecological data (e.g., Vrba's 1975 reconstruction of the environment at the Transvaal sites, as suggested in Ciochon and Corruccini 1976) remains to be seen and should prove quite illuminating. The lower limb evidence indicates that Aus- tralopithecus was an erect biped with a fully developed striding gait and stabilization adaptations at the pelvis. If Zihlman and Hunter (1972) are correct, in that the capacity for internal rotation at the hip joint was achieved in a different fashion in Austrclopithecus, then the longer arms of these hominids might have been very useful in assisting internal rotation. Striding bipedalism was achieved via a different pattern; it is entirely possible that the outward manifestation of the striding gait in Australopithecus was unlike that of Homo sapiens. This pattern may have involved a difference in cadence as well as different use of the forelimb in walking. Arm swinging is important in man's gait and its nature and vigor varies considerably between indi- viduals, and also varies with the nature and inclination of the terrain, cadence, and speed and type of move- ment. Given the difference in body proportions and size in Australopithecus, a unique type of locomotor behavior is expected for these hominids. What emerges from the study of australopithecine postcranial morphology is a distinctive pattern: a u- nique combination of body size, limb proportions, and limb morphology. The earlier emphasis in the litera- ture, on fitting this postcranial morphology into either modern human or extant pongid models, has limited progress in understanding the data. Given the nature of the evolutionary process and the sampling prob- lems of paleontological research, the discovery of major adaptive grades in any lineage is not surprising. There is every reason to expect the early hominids to have undergone some sort of adaptive radiation and to have evolved different adaptive complexes. The aus- tralopithecine data evidence just such evolutionary di- versity in adaptation. Delineating and understanding these adaptations, and the total way of life of the early hominids, constitutes a major goal of paleoanthropol- ogy and demands an understanding of their locomotor capabilities. The postcranial skeleton of Australopithecus has been instrumental in'reconstruct- ing early hominid locomotion, and continuing study should fully delineate the pattern of evolution which led to the modern human gait. TABLE I Early Hominid Postcranial Fossils8 Catalog code: TM Transvaal Museum Old. Hom. Olduvai Sts Sterkfontein Omo. L and F Omo SE Sterkfontein Member 5 KNM-ER East Lake Turkana SK and SKW Swartkrans AL Hadar (Afar) MLD Makapansgat AXIAL SKELETON Vertebrae cervical: SK 854 (C2=axis) KNM-ER 164 (C7) thoracic: Sts 14 & 73 SK 3981a (T12) KNM-ER 164 (TI) AL 288-1 (6 thoracics) lumbar: Sts 14 (6 lumbars) SK 853, 3981b (L5) L 105-7 9 8 9 9 Al, 288-1 (one) sacirl: Sts 14 AL 288-1 f'tagmenits: Sts 65 (cenitrumii) SE 2398 AL 288-1 Costae Old. Hom. 50 (shaf't) KNM-ER 164b-c AL 288-1 (numerous fragments) Joints Sts 14 (intervertebral and sacroiliac) Al, 288-1 (intervertebral and sacroiliac) APPENDICULAR SKELETON Superior Member Clavicula Old. Hom. 48 (formerly Old. Hom. 8) missing both ends MLD 20 "acromial extremity" NOT HOMINID (equid phalanx? per Day) Scapulae Sts 7 glenoid fragment KNM-ER 1500o left (L) glenoid fragment Al, 288-1 frag. right (R) Humeri TM 1517 distal R end (Kromdraai) Sts 7 proximal R SK 50 distal half SK 860 crushed distal end MLD 14 frag. R distal NOT HOMINID? (per Day) MLD 39 frag. R distal Kanapoi Hominid I (KP 271) distal L Old. Hom. 13 distal half of shaft L 338Y-9 juvenile shaft Omo 119-73-2718 prox. L KNM-ER: 739 R, missing just proximal end 740 distal 1L 1473 prox. R 15001 distal supracondylar portion 1504 distal R 1591 prox. R 3735 distal R Al, 225-4 distal AL 288-1 prox. R and distal L Radii Sts 68 prox. NOT HOMINID (baboon? per Day) SK 18b prox. fragment MLD: 15 frag. R 16 prox. R 32 L prox. frag. Old. Hom. 49 R shaft (previously known as FLK NN 134) Omo 75s-70-1317 prox. KNM-ER: 803d L shaft frag. 803p frag. 1500e R. prox. 1500k R frag. AL 288-1 prox. and distal R, and distal L Ulnae TM 1517 prox. R (Kromdraai) Old. Hom. 36 L 40-19 complete R Omo 141-72-23 prox. L KNM-ER: 803c L shaft 1500f R prox. 1500i prox. R shaft Al 288-1 prox. and distal R and I, Manus TM 151 7h L metacarpal II, prox. hand phalanx (Kromdraai) - metacarpal is probably a R metatarsal and not necessarily hominid (Day and Scheuer 1973; Robinson 1972) TM 1526 R capitate (Sterkfontein) SK 84 1J metacarpal I (the canine tooth which has been listed as SK 84 is correctly catalogued as SK 845) SK 85 1, metacarpal IV SKW 14147 metacarpal V Old. Hom. 7: a terminal R pollical phalanx, juvenile b terminal phalanx, juvenile c terminal phalanx, juvenile d intermediate R juvenile phalanx, II e intermediate R juvenile phalanx, IV f intermediate Rjuvenile phalanx, III g intermediate Rjuvenile phalanx, V h prox. phalanx, juvenile? i prox. phalanx, juvenile? o base, metacarpal II, R, juvenile? p Rjuvenile scaphoid q R juvenile trapezium r L juvenile capitate (all other formerly included in Old. Hom. 7 now removed as NOT HOMINID, per Day 1976a) Old. Hom. 8 Old. Hom. 18 phalanx Omo 18-70-1848 L intermediate phalanx, III Omo 28-73-4570 intermediate phalanx, III KNM-ER 164b-c two phalanges KNM-ER 803t base of metacarpal IV AL 288-1 L capitate, two phalanges Joints a. Glenohumeral: Sts 7 KNM-ER 1500 AL 288-1 b. Elbow: TM 1517 (Kromdraai) AL 288-1 c. Hand: partial - AL 288-1 and new material from Al, 333 Inferior Member - (0s) Coxae TM 1605 small iliac frag., juvenile? (Kromdraai) Sts 14 Sts 65 R ilio-pubic frag. SK: 50 damaged R ilium/ischium 1590d 3155b R iliac fragment, adolescent MLD: 7 L ilium 8 R ischium (MLD 7/8 one individual, adolescent?) 21 frag. ilium 25 adolescent L ilium 35 infant ilium frag.? AL 288-1 L (Old. Hom. 28 L, minus pubis Homo erectu.s) (KNM-ER 3228 R cf. Homo erectus) Femora TM 1513 distal L (Sterkfontein) TM 3601 distal R (locality?) Sts: 14 damaged 1, 34 distal R 65 prox. SK: 82 prox. R rag. 83 prox. frag. 97 prox. R 1590d MLD 17 head only Old. Hom. 20 prox. shaft and neck, L Old. Hom. 28 I, shaft (Homo erectus) 1, 754-8 KNM-ER: 736 L shaft 737 L, shaft 738 prox. L 803a ?813b (see under tibiae) 815 prox. L 993 distal R 999 1, 1463 R shaft 1465 prox. L 1472 R 1475a prox. R 1475b f'rag. shaft 1481a L 1500b distal L, 1500d L, neck 1503 prox. R 1505 prox. L 1592 distal R 1807 R shaft 1809 R shaft 1822 frag. 3728 L, prox. end and shaft 3730 distal frag. 15)OOg R distal AL 288-1 R distal Pes TM 1517 R talus and distal phalanx (Kromdraai) Old. Hom. 8 7 tarsals, 5 metatarsals, 2 phalanges Old. Hom. 10 distal hallucial phalanx Old. Hom. 43 metatarsal L 576-18 distal R hallucial phalanx F 511-6 hominid metatarsal III KNM-ER: 803e L talus 803f base, L metatarsal V 803j I, metatarsal III 803k intermediate phalanx, III 8031 interrmediate phalanx, II 803m terminal phalanax, III or IV 803q frag. prox. phalanx of hallux 813a R talus 997 L, metatarsal III 1464 R talus 1476a L talus 1500m base, R metatarsal III 1815 R talus AL 288-1 R talus Joints a. Hip joint: Sts 14 AL 288-1 b. Knee joint: KNM-ER 1481a/b KNM-ER 1500a/b AL 129-la/lb AL 288-1 AL: 128-1 L prox. frag. 129-la (or 129-1) R distal 129-Ic (or 129-3) R prox. frag. 211- 1 R prox. f'rag. 211 -1 R shaft 288-1 I. Tibiae Old. Hom. 34 part of shaft Old. Hom. 35 L shaft and distal end (formerly OH 6) KNM-ER: 741 prox. L 803b I shaft 803g frag. prox. end 813b frag.? 1471 prox. R 1476b & c prox. L 148tb prox. L 1481c distal 1J 1500a prox. L 150(c distal L 1500h R distal 1500j R shaft frag. 15OOr crushed frag. 1810 prox. L 2594a prox. L, 2594b shaft frag. 2596 distal frag. AL 129-lb (or 129-2) R prox. AL 288-1 prox. and distal R Fibnlae Old. Hom. 35 1, shaft and distal end (formerly Old. Hom. 6) KNM-ER: 803n L frag. 803o R shaft frag. 1481 distal L PARTIAL TO SOMEWHAT COMPLETE ASSOCIATED SKELETONS (one individual in direct association) TM 1517 (Kromdraai) Sts 14 KNM-ER: 803 (cf. Homo erectus?) 1481a-d (lower leg partial skeleton) 1500 (Australopithecus aff. africanus) 1808 Al, 288-1 AL 333: at least several individuals represented by numerous cranial and postcranial fragments (undescribed and unpublished) Notes 1"Early hominid" here includes all species ofAustralopithecUs, and early members of the genusHomo (Homo sp. indet., Homo habilis). "Australopithecine" here refers to just the species of Australopithecus (A. africanus, A. robustus, A. boisei). 2There is only one clavicle from Olduvai, formerly catalogued as Old. Hom. 8, a paratype of Homo habilis (Leakey, Tobias, and Napier 1964); it is now catalogued as Old. Hom. 48. 3The Sts 7 scapula was recently prepared by E. Vrba, aiding Ciochon and Corruccini in assessing some key anatomical landmarks (especially the coracoid process). 4Estimated total length - 328 mm (McHenry et al. 1976), compared to 290-3 10 mm for Sts 7 (Broom et al. 1950), and 235 mm for AL 288-1 (Johnson and Taieb 1976). 5I have made a measurement (on casts) of minimum width of the ilium (anteroposterior), from the anterior-most point of the greater sciatic notch to the most posterior point in the interspinous notch of the anterior iliac margin. I then formed a ratio of minimum iliac width over ischial length, and determined for: 1 0 0 1 ( 1 Sts 14- 45 mm X 100= 105% 43 SK 50- 61 mm X 100= 102% 60 Correlating for size differenices in this manner places both Sts 14 and SK 50 in approximately the same range. 'The australopithecine femoral neck is relatively long, whether one measures length via (a) the distance from the intertrochanteric crest to the subcapital sulcus, standardized via femoral shaft diameter just distal to the greater trochanter (Wood 1976), or (b) the length of a line, perpen- dicular to the centroidal axis of the shaft, from the most lateral point of the greater trochanter to its tangential inter- sectioni with the most cephalward part of the femoral head (biomechanical femoral neck length, per Lovejoy 1975). 7Using Lovejoy's (1975) sample of 128 modern Homo sapiens tali (x= 19?, 1 S.D. =3.40), the horizontal neck angle of Old. Hom. 8 lies just within and TM 1517 just beyond three standard deviations from the modern mean. 8For cataloguing by site, a selected bibliography includes Day (1976a & b), Howell and Coppens (1974), Johanson and Taieb (1976), Leakey and Isaac (1976), M. Leakey (1971), Oakley and Campbell (1967), and Robinson (1972). REFERENCES CITED Archibald, J. David, C. Owen Lovejoy, and Kingsbury G. Heiple 1972 Implications of Relative Robusticity in the Olduvai Metatarsus. Am. J. Phys. Anthrop. 37:93-96. Barnett, C.H. 1955 Some Factors Influencing Angula- tion of the Neck of the Mammalian Talus. J. Anat. 89:225-230. Basmajian, John V. 1972 Biomechanics of Human Posture and Locomotion: Perspectives from Elec- tromyography. In The Functional and Evolu- tionary Biology of Primates. Russell Tuttle, ed. Chicago: Aldine-Atherton, Pp. 292-304. Boaz, N.T., Y. Coppens, and F.C. Howell n.d. in prep. New Fossil Hominids from Omo. Brain, C.K., E.S. Vrba, andJ.T. Robinson 1974 A New Hominid Innominate Bone from Swartkrans. Ann., Transvaal Mus. 29:55-63. Broom, Robert, and J.T. Robinson 1952 Swartkrans Ape-Man Paranthropus crassidens. Transvaal Mus. Mem. 6. Broom, Robert, J.T. Robinson, and G.W.H. Schepers 1950 The Sterkfontein Ape-Man Plesianthropus. Transvaal Mus. Mem. 4. Broom, Robert, and G.W.H. Schepers 1946 The South African Fossil Ape-Men, The Aus- tralopithecinae. Transvaal Mus. Mem. 2. Burns, Peter E. 1971 New Determination of Aus- tralopithecine Height. Nature 232:350. Chopra, S.R.K. 1962 The Innominate Bone of the Australopithecinae and the Problems of Erect Post- ure. Bibliotheca Primatol. 1:93-102. Ciochon, Russell L., and Robert S. Corruccini 1976 Shoulder Joint of Sterkfontein Australopithecus. S. Afr. J. Sci. 72:80-82. Clark, W.E. Le Gros 1947 Observations on the Anatomy of the Fossil Australopithecinae. J. Anat. 81:300-322. 1955 The Os Innominatum of the Recent Ponginae with Special Reference to that of the Aus- tralopithecinae. Am. J. Phys. Anthrop. 13:19-27. 1967 Man-Apes or Ape-Men? New York: Holt, Rinehart, and Winston. Coleman, William H. 1969 Sex Differences in the Growth of the Human Bony Pelvis. Am. J. Phys. Anthrop. 31: 125-152. Conroy, Glenn C., and John G. Fleagle 1972 Locomotor Behaviour in Living and Fossil Pongids. Nature 237:103-104. Corruccini, Robert S. 1975 Morphometric Assessment of Australopithecine Postcranial Affinities. Sys- tematic Zool. 24:226-233. Corruccini, Robert S., Russell L. Ciochon, and Henry M. McHenry 1976 The Postcranium of Miocene Hominoids: Were Dryopithecines Merely "Dental Apes"? Primates 17:205-223. 1 02 Dart, Raymond A. 1949 Innominate Fragments of Australopithecus prometheus. Am. J. Phys. An- throp. 7:301-333. Davis, P.R. 1964 Hominid Fossils from Bed I, Olduvai Gorge, Tanganyika - A Tibia and Fibula. Nature 291:967-968. Day, Michael H. 1967 Olduvai Hominid 10:A Mul- tivariate Analysis. Nature 215:323-324. 1969 Femoral Fragment of a Robust Aus- tralopithecine from Olduvai Gorge, Tanzania. Na- ture 221:230-232. 1971 Postcranial Remains of Homo erectus from Bed IV, Olduvai Gorge, Tanzania. Nature 232:383-387. 1973 Locomotor Features of the Lower Limb in Hominids. Symp., Zool. Soc. London 33:29-51. 1974a The Early Hominids from East Rudolf, North Kenya. J. Anat. 117:651. 1974b The Interpolation of Isolated Fossil Foot Bones into a Discriminant Function Analysis- A Re- ply. Am. J. Phys. Anthrop. 41:233-236. 1976a Hominid Postcranial Material from Bed I, Olduvai Gorge. In Human Origins: Louis Leakey and the East African Evidence. Glynn LI. Isaac and Elizabeth R. McCown, eds. Menlo Park, Ca.: W.A. Benjamin. Pp. 363-374. 1976b Hominid Postcranial Remains from the East Rudolf Succession. In Earliest Man and Environ- ments in the Lake Rudolf Basin. Yves Coppens, et al., eds. Chicago: Univ. Chicago Press. Pp. 507-521 1977 Functional Interpretation of the Morphology of Postcranial Remains of Early African Hominids. In African Hominidae of the Plio-Pleistocene. C. Jolly, ed. London: Duckworth. Day, Michael H., and Richard E.F. Leakey 1974 New Evidence of the Genus Homo from East Rudolf, Kenya (III). Am. J. Phys. Anthrop. 41:367-380. Day, Michael H., Richard E.F. Leakey, A.C. Walker, and B.A. Wood 1976 New Hominids from East Turkana, Kenya. Am. J. Phys. Anthrop. 45:369- 435. Day, Michael H., and J.R. Napier 1964 Fossil Foot Bones. Nature 201:969-970. 1966 A Hominid Toe Bone from Bed I, Olduvai Gorge, Tanzania. Nature 211: 929-930. Day, Michael H., and J.L. Scheur 1973 SKW 14147: A New Hominid Metacarpal from Swartkrans. J. Human Evol. 2:49-438. Day, Michael H., and B.A. Wood 1968 Functional Affinities of the Olduvai Hominid 8 Talus. Man 3:440-455. Elftman, Herbert 1945 The Orientation of the Joints of the Lower Extremity. Bull., Hospital for Joint Diseases 6:139-143. Genet-Varcin, Emilienne 1969 Structure et com- portement des Australopitheques d'aprds certains Gregory, William K. 1949 The Bearing of the Aus- tralopithecinae upon the Problem of Man's Place in Nature. Am. J. Phys. Anthrop. 7:485-512. Hall, Michael C. 1965 The Locomotor System: Func- tional Anatomy. Springfield, Ill.: Charles C. Thomas. Harty, Michael 1973 The Anatomy of the HipJoint. In Surgery of the Hip Joint. Raymond G. Tronzo, ed. Phila.: Lea and Febiger. Pp. 45-78. Haxton, H.A. 1947 Muscles of the Pelvic Limb (A Study of the Differences between Bipeds and Quadrupeds). Anat. Rec. 98:337-346. Heiple, Kingsbury G., and C. Owen Lovejoy 1971 The Distal Femoral Anatomy of Australopithecus. Am. J. Phys. Anthrop. 35:75-84. Howell, F. Clark 1955 The Pelvic Remains of Aus- tralopithecus Am. J. Phys. Anthrop. 13:396. Howell, F. Clark, and Yves Coppens 1974 Inventory of Remains of Hominidae from Pliocene/Pleistocene Formations of the Lower Omo Basin, Ethiopia (1967-1972). Am. J. Phys. Anthrop. 40:1-16. 1976 An Overview of Hominidae from the Omo Succession, Ethiopia. In Earliest Man and En- vironmnts in the Lake Rudolf Basin. Y. Coppens, et al., eds. Chicago: Univ. Chicago Press. Pp. 522-532. Howell, F. Clark, and B.A. Wood 1974 Early Hominid Ulna from the Omo Basin, Ethiopia. Nature 249:174-176. Ishida, Hidemi 1968 Anatomical Variations of Limb Muscles in Primates from the Viewpoint of Locomotion. Proc., VIII Inter. Congr. Anthrop. Ethnol. Sciences 1:334-335. Jenkins, Farish A., Jr. 1972 Chimpanzee Bipedalism: Cineradiographic Analysis and Implications for the Evolution of Gait. Science 178:877-879. Johanson, D. Carl, and Yves Coppens 1976 A Prelimi- nary Anatomical Diagnosis of the First Plio/ Pleistocene Hominid Discoveries in the Central Afar, Ethiopia. Am. J. Phys. Anthrop. 45:217-233. Johanson, D. Carl, Yves Coppens, and M. Taieb 1976 Pliocene Hominid Remains from Hadar, Central Afar, Ethiopia. Preprint, IXe Congres, Union In- ter. des Sciences Prehistoriques et Protohistoriques. Colloque VI- Les Plus Anciens Hominides. Johanson, D. Carl, and M. Taieb 1976 Plio-Pleistocene Hominid Discoveries in Hadar, Ethiopia. Nature 260:293-297. Joseph, J. 1960 Man's Posture- Electromyographic Studies. Springfield, Ill.: Charles C. Thomas. Kay, Richard F. 1973 Humerus of Robust Aus- tralopithecus. Science 182:396. Kern, Howard M.,Jr., and William L. Straus,Jr. 1949 The Femur of Plesianthropus transvaalensis. Am.J. Phys. Anthrop. 7:53-77. os post-craniens. Annales de Paleontologie 55:139-148. 1 03 Kummer, Benno K.F. 1968 General Problems in Biomechanics of the Upright Posture and Gait. Proc., VIII Inter. Congr. Anthrop. Ethnol Sciences 1:316-322. 1975 Functional Adaptations to Posture in the Pelvis of Man and Other Primates. In Primate Functional Morphology and Evolution. Russell Tuttle, ed. The Hague: Mouton, Pp. 281-290 L,eakey, L.S.B., P.V. Tobias, and J.R. Napier 1964 A New Species of Genus Homo from Olduvai Gorge. Nature 202:7-9. Leakey, M.D. 1971 Olduvai Gorge, Volume III. Cam- bridge: Univ. Cambridge Press. Leakey, Richard E.F. 1972 Further Evidence of Lower Pleistocene Hominids from East Rudolf, North Kenya, 1971. Nature 237:264-269. 1973a Evidence for an Advanced Plio-Pleistocene Hominid from East Rudolf, Kenya. Nature 242:447-450. 1973b Further Evidence of Lower Pleistocene Hominids from East Rudolf, North Kenya, 1972. Nature 242:170-173. 1976 New Hominid Fossils from the Koobi Fora Formation in Northern Kenya. Nature 261:574- 576. Leakey, Richard E.F., and Glynn LI. Isaac 1976 East Rudolf: An Introduction to the Abundance of New Evidence. In Human Origins: Louis Leakey and the East African Evidence. Glynn LI. Isaac and Elizabeth R. McCown, eds. Menlo Park, Ca.: W.A. Benjamin. Pp. 307-332. Lessertisseur, J., and F.K. Jouffroy 1975 Comparative Osteometry of the Foot of Man and Facultatively Bipedal Primates. In Primate Functional Morphol- ogy and Evolution. Russell Tuttle, ed. The Hague: Mouton. Pp. 327-340. Leutenegger, Walter 1972 Newborn Size and Pelvic Dimensions of Australopithecus. Nature 240:568- 569. 1974 Functional Aspects of Pelvic Morphology in Simian Primates. J. Hum. Evol. 3:207-222. Lewis, O.J. 1972a The Evolution of the Hallucial Tar- sometatarsal joint in the Anthropoidea. Am. J. Phys. Anthrop. 37:13-34. 1972b Evolution of the Hominoid Wrist. In The Functional and Evolutionary Biology of Primates. Russell Tuttle, ed. Chicago: Aldine. Pp. 207-222. 1973 The Hominoid Os Capitatum, with Special Reference to the Fossil Bones from Sterkfontein and Olduvai Gorge. J. Human Evol. 2:1-11. Lovejoy, C. Owen 1973 The Gait of Aus- tralopithecines. Yrbk. Phys. Anthrop. 17:147-161. 1975 Biomechanical Perspectives on the Lower Limb of Early Hominids. In Primate Functional Morphology and Evolution. Russell Tuttle, ed. The Hague: Mouton. Pp. 291-326. 1976 The Locomotor Skeleton of Basal Pleistocene Hominids. Preprint, IXe Congres, Union Inter. des Sciences Pre'historiques et Protohistoriques. Col- loque VI- Les Plus Anciens Hominide's. 1977 A Biomechanical Review of the Locomotor Diversity of Early Hominids. In African Hominidae of the Plio-Pleistocene. C. Jolly, ed. London: Duckworth. Lovejoy, C. Owen, and Kingsbury G. Heiple 1970 A Reconstruction of the Femur of Australopithecus af- ricanus Am. J. Phys. Anthrop. 32: 33-40. 1972 Proximal Femoral Anatomy of Aus- tralopithecus. Nature 235:175-176. Lovejoy, C. Owen, Kingsbury G. Heiple, and Albert H. Burnstein 1973 The Gait of Australopithecus. Am. J. Phys. Anthrop. 38: 757-779. McHenry, Henry M. 1973a Early Hominid Humerus from East Rudolf, Kenya. Science 180: 739-741. 1973b Comment on Kay, "Humerus of Robust Aus- tralopithecus." Science 182: 396. 1974 How Large were the Australopithecines? Am. J. Phys. Anthrop. 40: 329-340. 1975a Fossils and the Mosaic Nature of Human Evolution. Science 190: 425-431. 1975b The Ischium and Hip Ectensor Mechanism in Human Evolution. Am. J. Phys. Anthrop. 43:39-46. 1975c A New Pelvic Fragment from Swartkrans and the Relationship between the Robust and Gracile Australopithecines. Am. J. Phys. Anthrop. 43:245-261. 1976a Early Hominid Body Weight and Encephali- zation. Am. J. Phys. Anthrop. 45:77-83. 1976b Multivariate Analysis of Early Hominid Humeri. In The Measures of Man. Eugene Giles and Jonathan S. Friedlaender, eds. Cambridge, Ma.: Peabdy Museum Press (Harvard University). Pp. 338-371. McHenry, Henry M., and Robert S. Corruccini 1975a Distal Humerus in Hominoid Evolution. Folia primatol. 23:227-244. 1975b Multivariate Analysis of Early Hominid Pel- vic Bones. Am. J. Phys. Anthrop. 43:263-270. 1976 Fossil Hominid Femora and the Evolution of Walking. Nature 259:657-658. McHenry, Henry M., and Robert S. Corruccini, and F. Clark Howell 1976 Analysis of an Early Hominid Ulna from the Omo Basin, Ethiopia. Am. J. Phys. Anthrop. 44:295-304. Mednick, Lois W. 1955 The Evolution of the Human Ilium. Am.J. Phys. Anthrop. 13:203-216. Morbeck, Mary Ellen 1972 A Re-Examination of the Forelimb of the Miocene Hominoidea. Ph.D. Dis- sertation, U. California, Berkeley. Napier,J.R. 1959 Fossil Metacarpals from Swartkrans. Fossil Mammals of Africa 17. 1962 Fossil Hand Bones from Olduvai Gorge. Na- ture 196:409-411. 1 04 1964 The Evolution of Bipedal Walking. Archives de Biologie (Li'ege) 75 (suppl.):673-708. 1965 Comment on Tobias, "New Discoveries in Tanganyika: Their Bearing on Hominid Evolu- tion." Curr. Anthrop. 6(4):391-411. Napier,J.R., and P.R. Davis 1959 The Forelimb Skele- ton and Associated Remains of Proconsul africanus. Fossil Mammals of Africa 16:1-70. O'Connor, Brian Lee 1974 The Functional and Evolutionary Biology of the Cercopithecoid Wrist and Inferior RadioulnarJoints, Ph.D. Dissertation, U. California, Berkeley. Olivier, Georges 1976 La Stature des Aus- tralopitheques. Preprint, IXe Congr'es, Union In- ter. des Sciences Prehistoriques et Protohistoriques. Colloque VI- Les Plus Anciens Hominides. Oxnard, Charles E. 1968a A Note on the Fragmentary Sterkfontein Scapula. Am. J. Phys. Anthrop. 29:429-432. 1975 The Place of the Australopithecines in Human Evolution: Grounds for Doubt? Nature 258:389- 395. Patterson, Brian, and W.W. Howells 1967 Hominid Humeral Fragment from the Early Pleistocene of Northwestern Kenya. Science 156:64. Preuschoft, H. 1971 Body Posture and Mode of Locomotion in Early Pleistocene Hominids. Folia Primatol. 14:209-240. Reynolds, Edward 1931 The Evolution of the Human Pelvis in Relation to the Mechanics of the Erect Posture. Papers, Peabody Mus. Am. Archaeol. Ethnol. 11:255-333. Rightmire, G.P. 1972 Multivariate Analysis of an Early Hominid Metacarpal from Swartkrans. Science 176:159-161. Robinson, J.T. 1972 Early Hominid Posture and Locomotion. Chicago: Univ. of Chicago Press. 1977 Evidence for Locomotor Difference Between Gracile and Robust Early Hominids from South Africa. In African Hominidae of the Plio- Pleistocene. C. Jolly, ed. London: Duckworth. Robinson, J.T., L. Freedman, and B.A. Sigmon 1972 Some Aspects of Pongid and Hominid Bipedality. J. Human Evol. 1:361-369. Schultz, Adolph 1969 Observations on the Acetabulum of Primates. Folia Primatol. 11:181- 199. Sigmon, B.A. 1975 Functions and Evolution of Hominid Hip and Thigh Musculature. In Primate Functional Morphology and Evolution. Russell Tuttle, ed. The Hague: Mouton. Pp. 235-252. Straus, William L., Jr. 1948 The Humerus of Paran- thropus robustus. Am. J. Phys. Anthrop. 6(3):285- 312. 1962 Fossil Evidence of the Evolution of the Erect, Bipedal Posture. Clinical Orthopaedics 25:9-19. Tuttle, Russell 1967 Knuckle-walking and the Evolu- tion of Hominoid Hands. Am. J. Phys. Anthrop. 26:171-206. 1975 Parallelism, Brachiation, and Hominoid Phylogeny. In Phylogeny of the Primates. W. Pat- rick Luckett and Frederick S. Szalay, eds. New York: Plenum. Pp. 447-480. Tuttle, Russell,John V. Basmajian, and Hidemi Ishida 1975 Electromygraphy of the Gluteus Maximus Muscle in Gorilla and the Evolution of Hominoid Bipedalism. In Primate Functional Morphology and Evolution. Russell Tuttle, ed. The Hague: Mouton. Pp. 253-269. Vrba, E.S. 1975 Some Evidence of Chronology and Palaeoecology of Sterkfontein, Swartkrans and Kromdraai from the Fossil Bovidae. Nature 254:301-304. Walker, Alan 1975 New Australopithecus Femora from East Rudolf. J. Human Evol. 2:545-555. Washburn, S.L. 1950 The Analysis of Primate Evolu- tion with Particular Reference to the Origin of Man. Cold Spring Harbor Symp. Quant. Bio. 15:67-78. Wood, B.A. 1973 Comment in Day, "Locomotor Fea- tures of Lower Limb in Hominids." Symp., Zool. Soc. London 33:5 1. 1974a A Homo Talus from East Rudolf. J. Anat. 117:203-204. 1974b Evidence on the Locomotor pattern of Homo from early Pleistocene of Kenya. Nature 251:135- 136. 1974c Olduvai Bed I Post-cranial Fossils: A Reas- sessment. J. Human Evol. 3:373-378. 1976 Remains Attributable to Homo in the East Rudolf Succession. In Earliest Man and Environ- ments in the Lake Rudolf Basin. Coppens, et al., eds. Chicago: Univ. of Chicago. Pp. 490-506. Zihlman, Adrienne L. 1967 Human Locomotion: A Reappraisal of the Functional and Anatomical Evi- dence. Ph.D. Dissertation, U. of California, Ber- keley. 1970 The Question of Locomotor Differences in Australopithecus. Proc., III Inter. Congr. Primatol- ogy 1:54-66. 1975 Review: J.T. Stern and C.E. Oxnard, Primate Locomotion: Some Links with Evolution and Mor- phology. Am. J. Phys. Anthrop. 43:314-316. 1977 Interpretations of Early Hominid Locomo- tion. In African Hominidae of the Plio-Pleistocene. C. Jolly, ed. London: Duckworth. Zihlman, Adrienne, and W.S. Hunter 1972 A Biomechanical Interpretation of the Pelvis of Aus- tralopithecus. Folia Primatol. 18:1-19. Zwell, Mike, and Glenn C. Conroy 1973 Multivariate Analysis of the Dryopithecus africanus Forelimb. Na- ture 244:373-375.