Comparative and Functional Morphology of Hominoid Fingers RANDALL L. SUSMAN Department of Anatomical Sciences, Health Sciences Center, State University of New York at Stony Brook, Stony Brook, New York 1 1 794 KEY WORDS Ape hands . Hominoid morphology . Knucklewalking . Ape locomotion . Morphometrics ABSTRACT Comparisons of hominoid metacarpals and phalanges reveal differences, many of which are closely linked to locomotor hand postures. The African apes display features of the metacarpals and phalanges which distinguish them from t h e other Hominoidea. These features are most evident in digits 111 and IV. The orangutan hand is demonstrably less well adapted to knuckle-walking and is distinctive in its adaptation to power and hook grasping of vertical and horizontal supports, respectively. Orangutan fingers possess a "double-locking'' mechanism (Napier, '601, and a slight ulnad shift in the axis of the hand which results in lengthened phalanges of ray IV. Hylobatid apes are more like orangutans in their finger morphology than any of the other Hominoidea, but exhibit unique features of their own. These include elongate phalanges of fingers 11-V. Human metacarpals 11-V form two sets composed of 11-111,and IV-V. The heads of both metacarpals I1 and 111are characterized by axial torsion. This reflects the enhanced manipulatory role of the third finger in humans. Human distal phalanges are unique in the development of pronounced apical tufts. Multivariate analysis of metacarpal 111and proximal I11 yields variables that array the extant apes along an arboreal-terrestrial axis, from hylobatid apes to male gorillas. The positions of taxa on this discriminant concur with observations on the locomotion of free-ranging apes. The comparative morphology of the hand has been a central topic in discussions of human phylogeny for over a century (Huxley, 1863; Gregory, '28; Osborn, '28; Wood-Jones, '42; Straus, '40, '49; Marzke, '71). More recently, the discovery and accounts of hominoid hand fossils have spurred interest in the functional and evolutionary morphology of hominoid hands (Napier, '59; Napier and Davis, '59; Zapfe, '60; Leakey, '60; Musgrave, '71, '73; Day and Scheuer, '73; Lewis, '73; Johanson, '76). A number of workers have studied the functional morphology of ape and human hands (Napier, '60, '62; Tuttle, '67, '69; Lewis, '69, '72a,b, '73, '771, Others, primarily interested in the wrist, have also considered aspects of the descriptive and evolutionary morphology AM. J. PHYS. ANTHROP. (1979) 50: 215-236. of hominoid hands (Schon and Ziemer, '73; Corruccini et al., '75; Etter, '74; Jenkins and Fleagle, '75; O'Connor, '76). Tuttle ('67) and Lewis ('69) have identified features of wrist morphology that distinguish the hominoids and that separate hominoid from cercopithecoid primates, but functional interpretations of wrist morphology are almost as numerous as the studies themselves (e.g., Lewis, '72a,b; Tuttle, '67, '74; Jenkins and Fleagle, '75; Schon and Ziemer, '73). Correspondingly, inferences on hand use in fossil primates have varied. Lewis ('72a,b) identified certain suspensory features in the wrist of Dryopithecus africanus while others have suggested cercopithecoid-like quadrupedalism (Preuschoft, '73a; Morbeck, '75; Corruccini et al.; '75; O'Connor, '76), cebid-like quadrupedalism 215 216 RANDALL L. SUSMAN (Schon and Ziemer, ’731, and knuckle-walking (Conroy and Fleagle, ’72). In spite of these differences of opinion, few workers have a t tempted t o extract quantitative, functional information from the morphology of the manual rays. The present study investigates the relationship of metacarpal and phalangeal morphology to locomotor hand use in the Hominoidea. The data from the manual rays complements earlier work on the primate wrist and provides a firmer basis for determining hand use in fossil Hominoidea. hancing mechanical properties of long bones (Currey, ’68). In some instances approximations to the ideal representation of these features had to be made. A series of measurements were made on each individual metacarpal and phalanx of rays 11-V. Figure 1 illustrates the points from which the variables are taken. Variables for metacarpals include length (c-h), radio-ulnar midshaft diameter (r-s), dorso-palmar midshaft diameter (b-el, dorso-palmar head diameter (d-f), radio-ulnar head diameter (n-o), biepicondylar diameter (p-q), dorso-palmar diameter of the medullary cavity, and set of the metacarpal head. l Other variables taken from these points were also tested. The phalanges were similarly characterized by a set of linear and computed variables. For the proximal phalanges these include: length ( a d , radioulnar midshaft diameter (h-i), dorso-palmar midshaft diameter (b-d),radio-ulnar diameter of the base (f-g),radio-ulnar diameter of the trochlea (j-k),“biepicondylar” diameter (1-m), and flexor sheath ridge height (across points h-i). Other variables including cortical thickness and longitudinal curvature were computed from these points. Middle and distal phalanges were described by a set of measurements taken from points illustrated in figure 1. The variation in hominoid fingers is thus described with multivariate procedures that employ sets or subsets of the above variables. Dissections were performed on hands of four Pan troglodytes, two Gorilla, two Pongo, two Hylobates lar, and two Homo sapiens. Other dissection data from a n extensive series of ape forelimbs was provided by Professor Russell H. Tuttle. MATERIALS AND METHODS This work is based on observations and measurements of a sample of 208 adult hominoid hand skeletons (table 1). In all but a few cases specimens are wildshot. The human sample is drawn from the Todd Collection housed a t the Cleveland Museum of Natural History. Articulated hands were used for determining identifiable patterns in individual ray segments and for determining ray segment formulae. Such skeletal material was then disarticulated and cleaned. After a method for recognizing individual phalanges was determined, loose finger bones were entered in the study. Measurements were taken on individual bones and from radiographs. The latter were made with an Oralix 415 portable X-ray unit. Bones were placed directly on the film plane and exposed with the tube head a t a distance of 74 cm. Lateral and frontal exposures were made on all specimens. Radiographs permit inspection of internal bony features and provide an accurate permanent record of specimens studied. Data attesting to the accuracy of the radiograph technique are presented in table 2. For statistical comparisons a set of variables were selected that represents morphological responses to patterns of stress in loading. Features such as curvature, cortex thickness, joint articular surface area, and linear dimensions have been discussed as means for en- Descriptive morphology of metacarpals II- V Gorilla The metacarpal of gorillas are comparatively short and stout, with heavily developed secI This was computed as the central angle of triangle abc using the B‘ + C’ A‘ = COS t) where flis the central angle. 2BC ~ fUl-IIIUla: TABLE 1 Specimens (adult, mostly wildshot) Pan troglodytes Pan paniscus M F M F M F M F M F M F Total 19 19 5 6 20 17 13 24 20 20 21 24 208 H. concolor, H. lar, H.hoolok, Gorilla Homo sapiens Pungu H.syndactylus, H. funerur, H. alhrminus, If. klossi, Ii. s p p ? Hylobates SPP. 217 FUNCTIONAL MORPHOLOGY OF HOMINOID FINGERS TABLL 2 Comparison o f measurements f r o m bones and radiographs .. v, - J , Mean difference lmmr so. Length proximal phalanx IV Diameter base proximal phalanx V Length metacarpal V Diameter, dorso-palmar metacarpal 111 Diameter, dorso-palmar distai phalanx I1 Ri-epicondyle diameter middle phalanx TI Length middle phalanx V Length distal phalanx I1 Diameter, radio-ulnar middle phalanx V 1 12 12 10 12 5 12 r y.yA' 0.999 0.998 0.998 0.996 0.959 0.995 0.998 0.994 0.998 0.358 0.125 0.700 -0.016 - 0.200 0.117 0.275 -0.740 0.067 12 5 ~ 12 r, correlation coefficient; y , . meamrements on art,ual hones; y l . mrasurements taken from radiographs b =&+9, 0. a C. 4 c i F f. Fig. 1 Measurement points for variables on metacarpals and phalanges. Linear and computed variables are taken from these points and from radiographs. ondary features in the adult. The predominant metacarpal length formula is 11-111-IV-V (82x1, but in 11 specimens metacarpals IV and V are equal in length. This equality is in marked contrast to the length pattern in the other Hominoidea. The metacarpal shafts are robust and essentially triangular in cross section. The metacarpal heads are large with well defined secondary features such as epicondyles, articular surface relief and capsular impressions. The head of metacarpal I1 displays a vari- 218 RANDALL L. SUSMAN A C F Fig. 2 Distal end view of metacarpal heads 11-V.A. Siamang (AMNH 106584 Lt. d). B. Modern Human (S.U.N.Y. 15 Lt. d). C. Bonobo (Tervuren 29042 Lt. 7). D. Chimpanzee (AMNH 89354 Lt. 7). E. Gorilla F. Orangutan (AMNH 61584 3 ) . (AMNH 1673384 6). able dorsal articular ridge (fig. 2E) but the radio-ulnar diameter of the articular surface does not equal that of the third or fourth fingers. The head is twisted axially so that the dorsal aspect is oriented toward the third ray. The epicondyles and excavations for the collateral ligaments are strongly represented. The head of metacarpal I1 is distinct from that of 111 and IV in having its maximum radioulnar diameters on the palmar surface in 27 of 30 cases, Metacarpal I11 is particularly well developed. The shaft is heavily constructed with pronounced impressions resulting from the 219 FUNCTIONAL MORPHOLOGY OF HOMINOID FINGERS origins of the second and third dorsal in- ridge. Asymmetry of the head when viewed terosseous muscles. A strong anterior crest dorsally is not marked in Pan troglodytes as in often separates the anterior extent of these Gorilla, indicating reduced conjunct rotation muscles and ends in a tubercle for the glenoid and radial deflection of the finger during hyligament of the metacarpnphalangeal joint. perextension of t h e metacarpophalangeal The dorsal articular ridge is marked and spans joint. heavily developed epicondyles. In 76%of speciThe fourth metacarpal in Pan troglodytes mens (n = 37), the maximum radio-ulnar di- recalls the shape of metacarpal I11 in features ameter of the head is on the dorsal surface. of the head and shaft (fig. 2D). It is always Metacarpal IV resembles 111in the morphol- shorter than I11 and less robust. The greatest ogy of its shaft and head. The base may have breadth of the articular surface is dorsal, and either one or two facets for contact with a well developed, continuous dorsal ridge is a metacarpal 111. The hamate surface extends consistent feature. anteriorly onto the surface for metacarpal V Metacarpal V is small and does not apand may continue distally to a variable proach the length of metacarpal IV to the exdegree. The anterior interosseous crest and tent observed in gorillas. The head is broader glenoid tubercle are conspicuous. The head is on the palmar than on the dorsal surface in broad and resembles that of 111. In 77% of 89% of specimens (n=37). The head lacks a cases (n=35), the dorsal articular surface is dorsal ridge and displays an axial twist which broader than the palmar surface. In profile, mirrors that of metacarpal 11. the heads of both metacarpals I11 and IV flatPan paniscus ten prior to the dorsal ridge. In dorsal view the The metacarpals of Pan paniscus differ distal articular surfaces of these metacarpals are extended on their ulnar aspect. This pro- from those of chimpanzees in a number of subvides for a slight axial rotation and radial tle ways. The present bonobo sample, although deflection upon hyperextension of the meta- small, indicates that females and males have carpophalangeal joints. Metacarpal V is proportionately longer in TABLE 3 gorillas than in other Hominoidea. In gorilla males i t attains 97%of the length of metacarLength proportion of metacarpal V in the apes and man (adults) pal IV (table 3). The shaft is broad and flat and often possesses a lateral curvature, conLength metacarpal V/rnetacarpal IV cave radially. The head is relatively narrow with an axial twist t h a t mirrors the condition Males Female8 of metacarpal 11. In 29 of 30 specimens the Gorilla gorilla palmar surface of the head is broader than the 22 20 !! dorsal surface. X Pan troglodytes Chimpanzee metacarpals are long and thin compared to those of gorillas, but they resemble the latter closely in features of the metacarpal head. The head of metacarpal I1 is asymmetrical, as in Gorilla, and lacks a dorsal articular ridge in 42%of cases. The epicondyles are well developed as are the excavations for the metacarpophalangeal joint collateral ligaments. Metacarpal I11 displays a marked dorsal ridge on the distal articular surface and this ridge spans well developed epicondyles (fig. 2D). The maximum radio-ulnar diameter of the third metacarpal head is always located dorsally (n=33). In profile the head recalls the pattern observed in gorillas, with less pronounced dorsal flattening prior to the dorsal S 97.33 2.97 95.18 2.34 17 91.18 2.63 19 91.21 2.20 5 88.95 3.85 89.19 0.67 13 92.31 2.10 24 91.92 2.30 12 88.08 1.83 15 88.20 1.74 20 93.00 2.29 20 92.35 1.93 Pan troglodytes n X S Panpaniscus 11 X 8 Pongo pygmaeus n - X S Hylobates andsymphalangus n - X 8 6 Homo sapiens sapiens !! x S 220 RANDALL L. SUSMAN similar mean metacarpal lengths (table 5 ) . If this is indeed the case, then the pattern is unique among the great apes. The metacarpal heads are long and narrow compared to those of the other apes (fig. 2C). The pattern of dorsal articular ridges on metacarpals 11-V differs in bonobos from the pattern observed in chimpanzees. Slight dorsal ridges are found in only 6 of 11 adult specimens. In two cases slight ridges occur on all four metacarpals. In two other instances dorsal ridges are found only on metacarpal 111, and in two cases faint ridges occur on 111and IV. All second metacarpals display a wider palmar articular surface. but only 7 of 11 fifth metacarpals retain this condition. Metacarpals 111 and IV display the expanded dorsal surface observed in chim panzees and gorillas. In 2 third and 2 fourth metacarpals the palmar articular surface is broader than the dorsal aspect. Pongo The metacarpals of orangutans differ from those of the knuckle-walking apes in a number of features of the shaft and head (fig. 2F). The morphology of the bases of metacarpals 11-IVis for the most part similar to that of the African apes (Lewis, '77: p. 173). Orangutan metacarpals are elongated with less pro nounced muscle and ligament impressions on the shaft and head. The shafts of metacarpals 11-IVare greater in antero-posterior diameter than in radio-ulnar breadth. The sides of the shafts diverge more abruptly than in the African apes, and they terminate in relatively reduced epicondyles. The cortices of the shafts are thicker in orangutans than in the other great apes or humans. The metacarpal heads of all four metacarpals (11-V) are normally broader on their palmar surface, and the most dorso-proximal extent of the articular surface of III and IV does not end in a dorsal ridge but rather it ends in bipartite extensions of the a r ticular surface. The head of metacarpal I1 is invariably wider on the palmar surface (n=31). The ulnar moiety of the distal articular surface commonly extends more dorsally than its radial counterpart, contributing to the axial torsion of the second metacarpal head. Metacarpal 111 is more symmetrical than 11, lacking torsion of the head. In 27 of 32 cases the maximum breadth is palmad in contradistinction to the African apes. In four cases the dorsal surface is broader and in one specimen the two surfaces are equal. Dorsal articular ridges are absent, and in dorsal view the epicondyles are not pronounced. Metcarpal IV is similar t o I11 in morphology of the shaft and head. In 27 of 30 cases the palmar aspect of the articular surface exceeds the dorsal aspect in breadth. Metacarpal V mirrors I1 in asymmetry of the head, while the shaft cross-section attains its greatest diameter in the radio-ulnar plane, similar to the other great apes and humans. Metacarpal V in orangutans may possess a marked lateral curvature, concave ulnad. Hylobatid apes The metacarpals of the hylobatid apes are long and thin with nondistinctive heads. Muscle markings for the intrinsic hand muscles are slight in gibbons, but may be moderately developed in siamangs. The metacarpals are lacking in curvature for the most part, although metacarpal V often displays a lateral (radio-ulnar)curve, concave ulnad. The cortex of the metacarpal shafts is relatively thin, especially when compared with that of Pongo and Pan (see below). In some individuals metacarpal I1 has the most robust shaft, while in others metacarpal 111 is more robust.' In both gibbons and siamangs the dorsal extent of the distal articular surface on metacarpals 11-V is more restricted than that of the great apes (fig. 2A). The palmar portions of the metacarpophalangeal joint capsules contain sesamoid bones in contrast to the other apes (Calhoun, '77). Homo sapiens The metacarpals of humans are characterized by well developed muscle markings, globular heads and a short, stout appearance. The third metacarpal in humans may exceed the second metacarpal in length (27%of cases) due to the development of a prominent basal styloid process in the latter bone. Metacarpals I1 and I11 resemble each other and differ from those of IV and V in being long and robust. All four metacarpals exhibit thin cortices. Metacarpal I1 in humans is normally the most robust in contradistinction to the great apes. The base is well developed and variations in carpometacarpal and inter-metacarpal articulations have been noted by Singh ('59). The metacarpal shaft is marked by the impressions for the first two dorsal and first palmar interosseous muscles. The metacarpal Shaft robusticit" is defined as: radioulnar ameter a t m i d s h a f t k n g t h x 100. X antervpvsterivr di- FUNCTIONAL MORPHOLOGY OF HOMINOID FINGERS head is markedly twisted and bears no trace of a dorsal articular ridge. The flexion set of the head is slight compared to the great apes. The third metacarpal in humans is distinctive in its asymmetric head (fig. 2B). It re.,qembles the second metacarpal in that the head i s axially twisted. During flexion both the second and third fingers rotate and deviate radially. Thus in humans, not only does the index finger undergo rotation and radial deviation during flexion (Landsmeer, '55; Lewis, ' 7 7 ) , but the medius also endorotates and radially deviates during metacarpophalangeal joint flexion. This is a clear accommodation to pollex-indicus-mediusopposition during precision grip. The third metacarpal head in humans, like that of metacarpal I l l in Pongo and the hylobatid apes, maintains its greatest radio-ulnar diameter on the palmar surface. The length and robusticity of metacarpal IV is substantially less than that of I1 or 111. In the human hand metacarpals IV and V form a functional unit t h a t provides ancillary support in the precision grip and plays a more prominent supportive role in the power grip. This functional pairing is unlike the 11, 111-IV, V grouping of the fingers in the great apes, or the 11, 111, IV, V sequence in the hylobatid apes. Metacarpal IV is more symmetrical in humans with less torsion of the head than either 11, 111, or V. The fifth metacarpal is diminutive. The base lacks t h e anterior lipping of the hamate surface seen in the great apes. In some specimens there is moderate torsion of the metacarpal head while in others torsion is only slight. Summary of metacarpal differences Differences in the metacarpals of rays 11-V can be summarized as follows: (1) The metacarpals of the African apes possess more strongly developed heads than other hominoids with a variable occurrence of a dorsal articular ridge. Within African apes the expression of the dorsal ridge ranges from slight in the bonobo to marked in the gorillas; it is more pronounced on metacarpals I11 and IV than on I1 or V. ( 2 ) Widening of the dorsal aspect of the distal articular surface is common in metacarpals I11 and IV in the African apes, but this expansion is normally lacking in metacarpals I1 and V. Expansion of the dorsal articular surface is absent in other Hominoidea. (3) The strongly developed secondary fea- 22 1 tures observed in the metacarpals of the African apes are less well developed in Pongo and lacking for the most part in humans, and the hylobatid apes. (4)The metacarpals of Pongo possess thicker cortices than those of the other Hominoidea while the shaft cross sectional areas assume an ovoid shape with the long axis in the dorso-palmar plane. (5) Human metacarpals differ in two unique and pronounced ways from the other Hominoidea: (a) the head of metacarpal I11 is twisted relative to the long axis of the shaft, and (b) the pairs 11-111and IV-V form two distinct size groups which reflect their different roles in precision and power grip. Descriptive morphology ofphalanges 11-V Gorilla The proximal phalanges of gorillas are stout, heavily constructed, and notable in their lack of longitudinal curvature (fig. 3). All articulated gorilla hands adhere to a I I I > I V > I I > V length formula for the proximal phalanges. On the dorsal aspect of the base of proximal phalanx I1 there is an extension of the palmar radial tubercle which receives the insertion of the pennate portion of the first dorsal interosseous muscle. This tubercle imparts a characteristic asymmetry to the base (fig. 3). Basal tubercles for the collateral ligaments are also marked. In gorilla males the flexor sheath ridges are extremely well developed. The trochlea is flattened on its distal end and raised well above the anterior surface of the body. Proximal phalanx I11 is notable in its symmetry in dorsal view. Proximal phalanx IV is distinguished from I11 by its reduced length and basal asymmetry induced by the insertion of the fourth dorsal interosseous muscle. The flexor sheath ridges are considerably more pronounced on phalanges I11 and IV than on I1 or V. The anterior portion of the phalanx base projects well above the surface of the body increasing the moment of the long flexor tendons as they course over the metacarpophalangeal joint. The fifth proximal phalanx is markedly asymmetrical and considerably smaller than the others. The base protrudes on its ulnar side where the abductor and flexor digiti minimi muscles insert. The radial side of the body bows out a t the flexor sheath ridge. The middle phalanges of Gorilla are very robust with marked antero-posterior flattening. The bases are extremely well developed 222 RANDALL Id. SUSMAN 1cm m Fig. 3 Proximal phalanges 11-Vat. to rt.) of gorilla (adult d! TABLE 4 Antero-posterror diameter of distal phalanx baselradio-ulnar diameter of distal phalanx base in the Hominoidea Hominoid No. I I1 111 IV V Chimpanzees Gorilla Humans Orangutans Hylobatids 6 6 6 67 58 58 80 71 76 69 79 70 60 60 60 6 68 73 6 57 74 80 71 72 70 80 69 62 73 71 and the bodies lack longitudinal curvature. The rugose surfaces for the insertions of flexor digitorum superficialis are marked, particularly in phalanges I11 and IV. Immediately laterally are well defined flexor sheath ridges. All but two specimens of gorilla display a TI1 > IV > I1 >V length formula for the middle phalanges. The trochleae of middle phalanges 11, IV, and V often incline toward the third ray. The bases of IT-V are broad and deep with well developed protuberances for the col- lateral ligaments of the first interphalangeal joint. The trochlear circumference, in lateral view, is not as extensive as that of the proximal phalanges. The distal phalanges do not conform to a regular size sequence in Gorilla. However, the base of distal phalanx I is distinctive in its “flatness” and kidney-shaped appearance. This obtains in all of the apes, and t o a lesser extent, humans (table 4).As a result, the pollical distal phalanx may be distinguished from that of the other rays. FUNCTIONAL MORPHOLOGY OF HOMINOID FINGERS 223 1cm Fig. 4 Proximal phalanges 11-V flt. to rt.1 of chimpanzee (adult d). The apical tufts of distal phalanges 11-Vare well developed, second only to humans. The apical tuft of the pollical distal phalanx, however, is slight compared t o the other four. Pan troglodytes The proximal phalanges of chimpanzees have a constant length formula of III>IV> I I > V . As with the proximal phalanges of gorillas, subtle asymmetries of bones, centered around the axial third ray, permit identification of individual phalanges. The development of secondary features such as flexor sheath ridges, basal protuberances and longitudinal curvature are variable in adult chimpanzees. Flexor sheath ridges range from marked to very slight in adults of both sexes (fig. 4). Lateral radiographs reveal a proportionate- ly thicker cortex in the proximal phalanges than in the metacarpals of chimpanzees and other great apes. The proximal phalanges develop the greatest amount of cortex in the distal portion of the body. Asymmetry of the trochlear surface is not consistent, but often the radial moiety of proximal phalanx I1 and the ulnar portion of V are larger than their counterparts. The length formula for chimpanzee middle phalanges is like that for the proximal phalanges; however, the lack of consistent asymmetry in individual bones makes their identification difficult in the absence of a complete set. The distal phalanges are similar t o those of gorillas. The expression of the apical tuft and overall robusticity are less than in gorillas, but the distal phalanges are relatively longer 224 RANDALL L. SUSMAN in chimpanzees. Distal phalanges I1 through V adhere t o the III>IV>II>V pattern in all cases (n = 17), and the pollical distal phalanx can be distinguished on the basis of its basal index (table 4). Pan paniscus The overall shape of the proximal phalanges is similar in Pan paniscus and Pan troglodytes. However, the proximal phalanges of bonobos differ in a number of respects from those of chimpanzees. The flexor sheath ridges are faint and may be absent in adult bonobos. Moreover, the palmar surface of the body, in some cases, is raised above the margins where the fibrous flexor sheaths insert. This condition recalls that often observed in orangutan proximal phalanges. The trochlear sulcus is deeper in bonobos than in chimpanzees. Pongo The phalanges of Pongo differ markedly from those of the African apes both in the comparison of individual bones and as a set. This morphological divergence is observed in proximal, middle, and distal phalanges. The length formula of the proximal phalanges varies in Pnngo and reveals a pattern of sexual dimorphism. In 58% (n=36) of orangutans the length pattern of proximal phalanges is III>IV>II>V, while 28% display a IV > I11> I1 >V formula. Fourteen percent possess a I11= IV > I1 >V pattern. Fifty percent of articulated male orangutan hands display the IV > I11 > I1 >V pattern while only 5%of females (1 specimen) show this pattern of proximal phalangeal length. The proximal phalanges in both sexes are long, thin, and ex- Fig. 5 Proximal phalanges 11-V (It. to rt.) of orangutan (adult d). FUNCTIONAL MORPHOLOGY OF HOMINOID FINGERS tremely curved. The bones widen distally and achieve their greatest radio-ulnar diameter a t the flexor sheath ridges. Curvature of the proximal phalanges contributes in large part to the exaggerated flexion set of the orangutan hand (fig. 5). The flexor sheath ridges are located opposite the point of maximum arc, increasing the sheath's ability to prevent bowstringing of the long flexor tendons and perhaps also providing a bilateral buttress for the body of the bone. The radio-ulnar concavity of the anterior surface of the body is lacking in Pongo for the most part. The anterior surface is, on the contrary, often raised above the height of the flexor sheath ridges. The trochlear sulcus is deep. Asymmetry of the second and fifth proximal phalanges of orangutans is similar to the other Pongidae. The third and fourth phalanges are variable in their relative degree of asymmetry. Often, when proximal phalanx IV exceeds 111 in length, the former bone is also more symmetrical. In this case the third phalanx takes on a pattern normally seen in I1 wherein the tubercle for the insertion of the pennate portion of the second dorsal interosseous muscle is enlarged and the ipsilatera1 flexor sheath ridge may extend more distad. In these specimens, however, the overall robusticity of proximal phalanx I11 is still greater than that of IV. The third phalanx maintains a greater proximal articular surface, greater radio-ulnar diameter of the body, and a broader trochlea. Thus, the hypertrophy of proximal phalanx IV (mainly in male orangutans) is primarily a linear development. The middle phalanges of Pongo reflect the size relationships of the proximal phalanges. The most notable feature of the middle phalanges is the relatively strong longitudinal curvature. Only 32%of orangutans possess the basic 111>IV > I1 > V length pattern in the middle phalanges. In 40% phalanges I11 and IV are equal in length, and in 28% the pattern is IV > I11 > I1 > V. Thus the middle phalanges deviate more frequently from the I11 > IV> I I > V sequence than the proximal phalanges. The middle phalanges constrict immediately beyond the base, but expand to their maximum breadth a t the midpoint of the insertions of m. flexor digitorum superficialis. These areas are well defined and bordered by moderately developed flexor sheath ridges. There is no striking asymmetry of the middle 225 phalanges and, unlike t h e proximal phalanges, the overall robusticity of IV may exceed that of 111. Often the distal articular surfaces of I1 and V incline toward the midline of the hand. The conical tip of orangutan fingers is reflected in the reduced apical tufts of the distal phalanges. Often the pollical tuft is no wider (radio-ulnad) than the body itself (29% of cases). Malformations including lateral splaying and resorption of the apical tufts, or fusion of the pollical distal phalanx are frequently observed in orangutans. Fifty-four percent of specimens deviate from the 111> IV> I1 > V length pattern. Although distal phalanges are highly variable the basal indices (table 4) subscribe t o the pattern observed in other apes. Hylobatid apes The individual proximal phalanges of the lesser apes are similar in shape, and it is sometimes difficult to identify individual bones. However, subtle variation in basal morphology normally permits identification of I1 and TV. These two bones are the most similar in size. The bodies of the phalanges are concave medio-laterally on the anterior surface and the flexor sheath ridges are well defined. Longitudinal curvature is pronounced and the flexor sheath ridges lie opposite the point of maximum arc as in Pongo (or slightly distal). Lateral radiographs reveal a constriction of the anterior posterior cortical moieties a t about the point of maximum arc. Overall, however, the cortex is relatively thin. The sulcus which divides the trochlea is deep, particularly in siamangs. The distal articular surface is narrow (radio-ulnad) but it retains a large circumference that extends well onto the anterior aspect. Asymmetry of the individual middle phalanges is not demonstrable, but the length formula of I11 > IV > I1 > V is constant among the hylobatid species sampled (n =45). The middle phalanges, together with the proximal, are responsible for the curved character of hylobatid fingers. The intrinsic longitudinal curvature of the middle phalanges is marked in both gibbons and siamangs. Well developed insertions for the m. flexor digitorum superficialis are bounded by extensive flexor sheath ridges. The bodies achieve their ' H klossr (4). H. eoncolor (41, H Zar (3). H hooiok (31, H ,spp. (221, H. syndoctylus (91, H.spp. includes the ahove plus H. funerus, H. albirnanus, but excludes H. klossi. 226 RANDALL L. SUSMAN greatest breadth a t the flexor insertions. The bases are characterized by well developed biconcavities for the trochlae of the proximal phalanges and a deep antero-posterior diameter. The distal phalanges possess very distinctive apical tufts compared to the other apes especially on distal phalanx I. Homo sapiens langes. These segments are distinctive in the shape of the base and in the development of the apical tufts. Distal phalanges 11-V have flat bases and bodies. with large, spade-like, apical tufts. The pollical distal phalanx is recognized by its enlarged area of insertion for the tendon of flexor pollicis longus. Summary of phalangeal differences The proximal phalanges of the African apes The proximal phalanges of modern humans are shorter in relationship to the metacarpals differ from the other hominoid primates in than those of orangutans or hylobatid apes. lacking well defined secondary features such The proximal phalanges of gorillas are short as curvatures, concavities, and muscle mark- and stout while those of orangutans and lesser ings. The bones have no longitudinal cur- apes are long, thin, and markedly curved. The vature and bear only faint flexor sheath proximal phalanges of chimpanzees, and esperidges. The trochlae are less circumferential cially bonobos, represent an intermediate than in the apes. The short, straight appear- morphology between gorillas and orangutans ance and reduced extension of the trochleae with respect to length, curvature, and robusaccount for the reduced flexion set of human ticity. In the African apes the I11 >IV > I1 > V fingers. length pattern is constant, while 50%of male In 36 of 40 individuals the proximal pha- orangutans and 5%of female orangutans devilangeal length pattern is III>IV>II>V. In ate from this pattern. The middle phalanges follow a similar trend three cases phalanx I1 exceeds IV, while in one individual phalanges I1 and IV are equal in to that of the proximal phalanges from gorillas through the orangutan and hylobatid length. The second proximal phalanx is identified apes. The size sequence of middle phalanges in by its relatively robust body and enlarged African apes is I11 > IV > I1 >V, while that of radial tubercle, the dorsal portion of which Pongo deviates from this pattern in 68% of accommodates the prominent first dorsal in- cases. Human distal phalanges are distinguished terosseous muscle. This feature characterizes living (Landsmeer, ' 5 5 ) and fossil (Musgrave, from those of other Hominoidea by their well '71) Hominidae. Although length alone may developed apical tufts and flat bases. Gorillas not distinguish proximal phalanges I1 and IV, have the most well developed finger tips the stouter base, asymmetry, and linear body among the apes, while orangutans often lack identify the second phalanx. The third prox- apical tufts altogether. imal phalanx is notable among the four for its Multivariate analysis combined length and its stout base and basal tubercles. Proximal phalanx IV is long, thin, Multiple discriminant analysis is one of the and does not display the basal asymmetry of approaches employed to isolate and describe 11; however, the flexor sheath ridges are functionally related morphological differusually more distinct on IV than on 11. The ences in the fingers of the Hominoidea. This fifth proximal phalanx is small and weakly de- statistical technique is well described in the veloped. The insertions of the hypothenar morphometric literature (Howells, '73; Oxmuscle give the base a robust appearance rela- nard, '73; Creel and Preuschoft, '76). It distinguishes known groups from each other on the tive to the body. Human middle phalanges have broad bases basis of multivariate variables constituted and well defined insertions for m. flexor dig- from weighted linear combinations of original itorum superficialis. The bones lack the other variables. The weights are calculated so as to secondary features seen in the middle pha- maximize the variance between groups relalanges of adult apes. The I11 > IV> I1 > V pat- tive to that within groups (for a detailed distern obtains in 36 of 40 cases. In four in- cussion see Tatsuoka, '71). The hominoid groups in this study are morstances phalanges I11 and IV are equal in phologically quite distinct. Discriminant analength. The most striking feature of human fingers lysis can be expected to find combined variais the distinctive character of the distal pha- bles along which the positions of the groups 227 FUNCTIONAL MORPHOLOGY O F HOMINOID FINGERS differ significantly. It is not so certain that these combined variables will be readily interpretable with respect to their functional significance. This is particularly true if the original variables are not carefully chosen to reflect functional differences. Even so, there is no guarantee that some unknown, or unrecognized, functional adaptation is not reflected in the weighting of variables. Seemingly important aspects of morphology may represent the evolutionary history of particular taxa and this may or may not reflect functional affinities. Natural selection is not constrained t o effect the same solution to a particular problem each time it operates. Presumably a discriminant reflecting morphological differences among taxa caused by differing locomotor adaptations will array the centroids for the taxa in a way similar to their functional ranking. This ranking may reflect a continuum (rather than discrete categories) as in the case of locomotor differences in the apes and in humans. Moreover, weightings given to individual variables on the discriminant axis should lend themselves t o “meaningful (biomechanical) interpretation. The analysis of metacarpal I11 and proximal phalanx I11 was conducted on the linear and calculated variables listed in tables 5 and 6. These elements of ray I11 were chosen because they represent the principal weight bearing finger, and earlier studies and this work demonstrates the intimate relationship of locomotor hand posture and the metacarpophalangeal complex. Prior to the discriminant analysis the variables were adjusted for ‘‘size” according to a procedure designed by N. Creel. Metacarpal length and proximal phalanx I.000 LOO0 2.000 LOO0 *Hylobates $ Hylobates I -4.000 ** I -3.009 -2.000 1.on0 -1.000 Q+** R P * P.ta 3400 i 4.000 PP.? *pmeo* P.t-4 =+* 2.000 *pone0 d -1.000 -2.000 -3.000 - -I 0 0 0 Fig. 6 Plot of hominoid group centroids on discriminant axes I (horizontal) and I1 from analysis of metacarpal and proximal phalanx 111. Hs., modern humans; Gg., gorilla; P.P., bonobo; P.t., chimpanzee; Hylobates, Hylobates spp.; and Pongo. Scale, std. dev. units. 228 RANDALL L. SUSMAN length for each animal were summed to form a combined length variable. The combined variable and the other original variables were then logarithmized. A pooled within groups covariance matrix was calculated for the transformed original variables and combined length variable. From this a series of regression equations were calculated with each of the original measurements being used in turn as the dependent variable and the combined length measurement as the independent variable. These equations were used to calculate estimated values for each animal and each variable from the length variables. The estimated values were then subtracted from the measured values yielding residuals. The estimated value of each original variable group mean on the length variable grand mean was computed from the resultant regression equation. The residual of each individual in the group was then added to the estimated value for the group mean to yield a “size corrected” variable in which all variance attributable to variation in the combined length variable has been removed. In other words, the combined length variable is used as a covariate for each of the measured variables. A plot of the groups on the first discriminant reveals an axis from the hylobatid apes to gorillas (fig. ti). This axis represents an arboreal-terrestrial trend with orangutans closest of the pongid apes to the lesser apes. The position of man is, of course, difficult to interpret in a locomotor context. Gorillas and, to a lesser extent, humans and chimpanzees are distinguished by the dorso-palmar diameter of the metacarpal shaft, dorso-palmar diameter of the metacarpal head, a wide trochlea on the proximal phalanx, and a stout phalangeal base. Hylobatid apes are distinctive primarily in features of the proximal phalanx. These include flexor sheath ridge development on the proximal phalanges, the dorso-palmar diameter of the phalangeal body, the “biepicondylar” diameter, and the enlarged medullary cavity. Axis two separates the apes from humans. The basal development of the proximal phalanx, wide medullary cavit y of the metacarpal, and dorso-palmar diameter metacarpal shaft are prominent variables on axis two. Features used in the discriminant analysis can be viewed in the context provided by the plot of group centroids (fig. 6) and the relevant behavioral data on horninoid locomotion and posture. Sexes are combined and groups are thus arranged in table 8. Row one expresses the proportion of the metacarpal and proximal phalanx, and reveals the lengthened proximal phalanges of the Asian apes. Humans have a high value for this index, but this reflects the relatively shortened metacarpals TABLE 5 Variable statistics-meta~arpalIII Mean Length Radio-ulnar midshaft diameter Dorso-palmar midshaft diameter Dorso-palmar head diameter Radio-ulnar breadth head Biepicondylar diameter Medullary cavity diameter Dorsal ridge height Set metacarpal head (in radians) Standard deviation Length Radio-ulnar midshaft diameter Dorso-palmar midshaft diameter Dorso-palmar head diameter Radio-ulnar breadth head Biepicondylar diameter Medullary cavity diameter Dorsal ridge height Set metacarpal head @ cf @ Bf Gm c f 03 of 18 16 5 6 I8 17 11 23 90.35 87.00 8.41 8.19 9.50 8.64 15.78 14.71 13.99 13.07 16.86 15.70 3.34 3.34 3.39 2.47 2.65 2.65 6.02 0.89 1.05 1.11 1.35 1.58 1.14 0.91 0.05 6.64 0.67 0.78 0.81 0.71 1.00 0.79 0.89 0.04 & € 15 Hylf H z 19 20 83.27 82.57 97.98 82.51 109.77 93.18 57.45 58.89 7.01 7.22 12.27 9.67 8.36 6.80 3.93 3.96 8.03 7.72 8.64 14.30 9.89 7.76 4.06 4.19 13.80 13.26 20.65 16.07 17.45 14.19 6.84 6.98 11.39 11.42 20.40 15.09 14.87 12.03 6.46 6.47 13.70 13.83 23.26 17.90 18.17 15.00 7.06 7.05 3.41 3.48 7.18 5.17 3.28 2.63 1.85 1.88 3.12 3.05 5.02 3.89 1.72 1.48 0.89 0.92 2.63 2.66 2.56 2.59 2.71 2.73 2.79 2.79 1.01 1.05 1.04 1.05 1.10 1.04 1.09 1.16 1.01 1.01 1.09 1.06 1.04 1.08 1.06 1.23 1.15 1.01 6.81 1.00 1.47 1.35 1.40 2.24 1.20 0.92 0.04 4.50 0.93 1.03 1.11 0.89 1.39 1.08 0.97 0.04 C, chimpanzee; B, honoho; G, gorilla; 0. orangutan; Hyl, hylnhatid apes, Hn,modern humans 8.01 0.92 1.47 2.01 1.50 1.99 1.39 1.04 0.04 5.55 0.71 0.83 0.99 0.84 1.18 0.76 0.87 0.05 6.49 0.61 0.63 0.86 0.96 0.83 0.37 0.34 0.03 4.42 0.47 0.44 0.82 0.71 0.68 0.32 0.31 0.02 HBf 20 69.82 65.71 8.81 7.87 10.08 8.85 14.28 13.38 13.54 12.39 14.74 13.48 4.87 4.37 0.58 0.56 2.59 2.64 5.33 0.69 0.81 0.90 0.91 1.33 0.95 0.38 0.03 3.49 0.40 0.68 1.00 0.83 0.89 0.74 0.35 0.04 229 FUNCTIONAL MORPHOLOGY OF HOMINOID FINGERS of humans rather than the elongation of the proximal phalanges. The cortex thickness reveals a linear trend in the Pongidae from Gorilla to Pongo with Homo between Pan troglodytes and Gorilla. The low value for hylobatid apes violates this trend. This may reflect a scaling problem related to decreased body proportions. The flexion set of the metacarpal head is pronounced in gorillas and most reduced in the hylobatid apes. Lastly, the proportionate depth of the metacarpal head and the length of the proximal phalanx also displays a trend from the knuckle-walkers, through the orangutan and hylobatid apes. This latter index is of critical importance for maintaining joint stability in knucklewalkers (fig. 7). Functional interpretations o f observed differences in the metacarpophalangeal joints The observed differences in the size and shape of hominoid ray segments are inter- TABLE 6 Variable statistics-proximal phalanx 111 Mean Length Radio-ulnar midshaft diameter Dorso-palmar midshaft diameter Radio-ulnar diameter base Radio-ulnar diameter trochlea Biepicondylar diameter Flexor ridge height Longitudinal curvature (in radians) Standard deuiation Length Radio-uln a r midshaft diameter Dorso-palmar midshaft diameter Radio-ulnar diameter base Radio-ulnar diameter trochlea Biepicondylar diameter Flexor ridge height Longitudinal curvature @ 17 cf B m Bf Cfl c f Om of Hylf Hsm Ef 16 5 6 1R 15 10 23 13 15 20 20 60.84 12.71 7.54 16.74 12.45 12.50 1.08 3.81 58.73 11.53 7.05 15.52 11.96 12.06 1.03 3.84 51.04 9.82 6.27 3.64 0.50 0.09 1.16 2.54 52.74 14.99 7.08 18.34 12.90 13.90 3.38 3.84 78.83 12.63 8.62 18.08 13.25 13.95 0.79 3.64 69.10 10.50 6.75 14.93 11.20 12.26 0.83 3.68 43.77 5.93 3.50 7.59 6.15 6.65 1.13 3.83 44.42 6.07 3.52 7.72 5.98 6.58 3.34 1.99 0.86 1.21 0.74 1.22 0.94 0.07 4.94 0.87 0.54 1.01 0.85 1.27 0.83 0.07 5.01 2.21 1.09 1.69 0.82 1.32 0.98 0.15 3.93 0.94 0.73 1.03 0.56 1.32 0.70 0.10 5.41 0.81 0.57 1.14 0.76 0.77 0.67 0.06 3.84 0.89 0.35 0.72 0.82 0.98 0.36 0.08 1.02 1.02 1.10 1.04 1.07 1.10 1.27 1.01 51.77 64.89 9.84 21.08 7.83 6.27 13.11 23.56 10.00 16.88 9.78 19.14 1.40 4.88 3.87 2.53 1.02 1.06 1.08 1.05 1.09 1.04 1.30 1.01 4.02 1.90 0.77 1.26 1.02 2.67 1.07 1.01 3.18 1.20 0.67 1.53 1.11 1.45 1.02 1.01 47.77 10.88 7.36 16.86 12.14 11.35 0.09 1.10 3.92 3.83 45.14 9.31 6.47 15.01 10.97 10.23 0.24 3.99 4.23 0.88 0.62 1.10 0.88 0.80 0.27 0.09 2.33 0.61 0.40 0.78 0.46 0.51 0.36 0.04 C. chimpanzee; B, bonabo, G , gorilla; 0, orangutan; Hyl, hylobatid apw: Hs, modem humans. TABLE 7 Standardized dicriminant function coefficients from combined metacarpal-proximal phalanx III analysis Metacarpal III Metacarpal length Midshaft diameter (r-u) Midshaft diameter (d-p) Head diameter (d-p) Head diameter (r-u) Biepicondylar diameter Medullary cavity (d-p) Dorsal ridge height Set of metacarpal head Proximalphalanx 111 Midshaft diameter (r-u) Midshaft diameter (d-p) Base diameter (r-u) Trochlear diameter (r-uj Biepicondylar diameter Flexor sheath ridge height Curvature Relative percentage of trace Function 1 Function 2 Function 3 -0.14568 - 0.17168 - 0.32612 - 0.24684 0.06767 -0,00998 0.10538 0.02479 - 0.00978 -0.11582 -0.03444 0.41738 - 0.47640 -0,02361 - 0.59205 0.30149 - 0.35861 0.03804 - 0.05832 - 0.33390 0.17629 0.04319 1.11748 - 0.09529 -0.27868 -0.03689 0.29549 - 0.45279 -0.11805 0.11449 - 0.19440 -0.20097 0.10438 0.11091 -0.08199 68.10 - 14.92 0.08678 0.12927 - 0.64757 0.52773 -0.33771 - 0.24783 0.30293 0.48359 0.92843 - 0.01545 - 0.02542 -0.24824 -0.31837 11.55 230 RANDALL L. SUSMAN pretable in t h e divergent contexts of knucklewalking and suspensory hand postures. Figure 7 illustrates t h e third rays of Pun troglodytes and Pongo in knuckle-walking and hypothetical knuckle-walking postures, respectively. I t h a s been demonstrated above t h a t i n the African apes features associated with knuckle-walking are more pronounced in rays 111 and IV. Several additional features of metacarpal and phalangeal morphology, some of which are underscored by t h e discriminant function coefficient values, are noteworthy: (1) In t h e African apes the proximal phalanges are short relative to t h e length of t h e metacarpal, and moreover, in relation to t h e dorso-palmar diameter of t h e metacarpal head (table 8, row 2 ) . (2) In chimpanzees and gorillas there is a PAN FgC = FppA f well developed fibrocartilaginous p a l m a r plate (=glenoid plate) and ligament on t h e ventral surface of t h e joint capsule (fig. 8 ) . This structure is absent for the most part in Pongo. (3) The epicondyles on t h e metacarpal heads are pronounced in t h e African apes and are most often connected by a dorsal articular ridge on metacarpals 111 and IV. Adult Pun paniscus, may altogether lack this feature on metacarpals 11-V. (4) Although variable, t h e widest radio-uln a r dimension on the heads of metacarpals 111 and IV is usually dorsal in the African apes and palmar in Pongo and humans. (5) The cortex in the metacarpal shafts is thicker in the orangutan and bonobo than in either chimpanzees or gorillas. Thus, this fea- PONGQ FmmB+ F j O + F c O Fig. 7 Ray I11 of Pun and Pongo in knuckle-walking and hypothetical knuckle.walking postures, respectively. F, and F,, counteract the torque imposed by the ground reaction force, F,. Moments of F, and F,, are determined by depth of the metacarpal head and the thickness of the palmar plate. Moment of Fg is determined by length of the proximal phalanx. FUNCTIONAL MORPHOLOGY OF HOMINOID FINGERS ture also displays a pattern of intergroup variation that follows a suspensory to knucklewalking continuum (table 8). These morphological differences are elements of a system that in part functions to reduce compressive stress in the metacarpophalangeal joints of the knuckle-walking apes, and helps to balance the torque that tends t o hyperextend the proximal phalanges (fig. 7). The metacarpal heads are deeper and more flexed in the African apes. This increases the moment of the long digital flexors and the palmar plate (Fmm and Fr respectively) for resisting the torque of t e ground reaction force (FgI. The hypertrophy of the palmar plate in the knuckle-walking apes also increases the moments (A and B) of the long flexors. Concomitantly, the short proximal phalanx (especially pronounced in gorillas), reduces the moment of the ground reaction force (FJ. The result of these differences between knuckle-walkers and non-knuckle-walkers is 231 that the potential influence of the long flexors and the palmar plate to counter force F g , and to provide an active (or passive) component of resistance (Fmm and Fpp), is increased in the African apes. The metacarpophalangeal joints in Pongo, with long proximal phalanges, shallow metacarpal heads and reduced palmar plates, are less well suited to balance the forces imposed in a knuckle-walking mode. The “power armiload arm index” given in table 8 further illustrates this point. It should be noted, however, that this index is an underestimation of the difference between Pongo and the African apes because i t does not include the thickness of palmar plate (fig. 8). In the African apes the predominant occurrence of expanded dorsal articular surfaces on metacarpals 111and IV reduces the joint compressive stress in normal loading. The dorsal articular ridge is coincident with the epiphyseal line and does not reach its full expression until well into adulthood. The enlarged dorsal ridge and expanded epicondyles serve to Fig. 8 Metacarpophalangeal joint capsule of ray III in gorilla. Proximodistal incision of the capsule exposes the palmar surface of the metacarpal head. Note t h e extreme development of the fibrocartilagenous palmar (=glenoid) plate. 232 RANDALL L. SUSMAN TABLE 8 Proportions and indices of ray 111 in the Hominoidea Gorilla n=15 Relative length PPhIII (pphllIimetaIlI X 100) Power arm: load arm index Metacarpophalangeal joint I11 ' Cortex thickness index Meta. I11 "Set" of the metacarpal head (in") Dorso-palmar midshafti Metacarpal length x 100 Biepicondylar diameter] length X 100 Pan troglodytes n-16 Pan panisrus n=ll Pongo Iiylohates spp. n=15 11-20 77 Humans c-15 65 67 62 73 69 16.0 13.3 13.2 10.2 7.6 15.8 48.9 65.3 69.2 70.5 54.7 54.4 147.6 152.1 151.6 155.9 160.1 150.2 12.16 9.36 8.57 7.47 6.78 12.31 22.80 18.36 16.60 16.34 12.14 20.82 ' This index is determined by one-half the dorso-palmar diameter of t h e metacarpal headflengh Cortex IS determined from lateral radiogmpha. The index 18 of the proximal phalanx. computed as cortexldorswpalmar diameter a t midshaft x 100 buttress the joint (Preuschoft, '73a) while the flattened dorsal extent of the surface provides a mechanism to tauten the collateral ligaments and inhibit axial rotation of the proximal phalanx while it is hyperextended. An additional function of the protuberant epicondyles may be to help prevent bowstringing of the interosseous and lumbrical tendons in the hyperextended position of the joint. Suspensory grasp and finger morphology Attenuated ray segments and lack of strong secondary features of the metacarpophalangea1 joints characterize the manual rays of the Asian apes. Muscle force t o counter the external force on the fingers is increasingly necessary as the metacarpophalangeal and interphalangeal joints are fle ed (Preuschoft, '73b). Lengthening of the ray segments increases the compass of the hand while markedly longitudinal curvatures are a remodelling response to strong bending moments imposed by the lengthening of the fingers. In orangutans and the hylobatid apes this results in pronounced phalangeal curvatures. The metacarpals are sometimes also subjected to strong bending moments, especially during combined flexion of the three finger joints (Preuschoft, '73b). The metacarpals in Pongo resist increased bending stresses by increased longitudinal curvature and increased cortical thickness (and perhaps increased density as well). These features are individually variable, especially with regard to shaft curvature, but together they form a pattern distinct from chimpanzees and gorillas. An additional feature of the orangutan hand (especially in males) is the increased length and concomitant high degree of variation in the fourth ray. The pattern of asymmetry suggests that the axis of the hand has shifted ulnad. The increased emphasis on the ulnar side of the hand seems to be related more to grasping vertical supports (as in climbing) rather than horizontal superstrata. I t is interesting to consider this possibility, especially in light of recent suggestions that climbing per se may be the basic locomotor adaptation in the hominoid radiation (e.g., Fleagle, '76; Stern, '76). In animals that climb vertical supports it is the ulnar side of the hand that provides friction against the downward force of the individual's weight. The ulnar rays provide force against the branch or vine, while the radial fingers, I1 and I11 (and occasionally I) provide ancillary support against the upward displacement of the fourth and fifth fingers. The latter component would not be available if the main muscular force opposing the branch or trunk was exerted by the radial side of the hand. DISCUSSION Fleagle ('76) and Tuttle ('77) have summarized the locomotor activities of the extant apes. Gorillas a r e primarily terrestrial knuckle-walkers, but occasionally climb trees. Chimpanzees are also terrestrial knucklewalkers but frequently climb and knucklewalk in trees. Orangutans are versatile climbers when in the canopy and only adult males are frequently found on the ground. When on FUNCTIONAL MORPHOLOGY OF HOMINOID FINGERS the ground orangutans do not knuckle-walk but instead place the proximal phalanges or palms on the ground. The lesser apes climb, brachiate and, to a lesser extent, walk bipedally in the trees. They are rarely observed on the ground. The morphology of the manual rays in living apes is intimately related to positions of the hand in locomotor and postural modes. These morphological differences, however, are subtle and continuous, and grouping extant apes on the basis of metacarpals and phalanges (11-V) is best accomplished hy the use of multivariate variables. Morphological patterns of one ray cannot be generalized to the others as both the pattern of loading and the manipulatory role of each ray differ within a single hand. The human hand emphasizes the pollex along with rays TI and IIT. The increased robustness of metacarpal I1 and the torsion of the third metacarpal head help identify the human pattern. Torsion of the head of the index metacarpal provides for pronation (endorotation) of the proximal phalanx during flexion (Lewis, ’77: p. 169); however, this action also occurs a t the third metacarpophalangeal joint. Rays IV and V comprise a functional unit that contrasts with the index and medius (Kaplan, ’ 6 5 : p. 6). The bones of rays 11-Vare notable in their lack of secondary features and the reduced proportions of the metacarpus. The hand of the African apes emphasizes rays TI1 and IV. Features associated with knuckle-walking hand postures are observed in the metacarpals and proximal phalanges. However, rays I1 and V commonly lack such features as a bilaterally expanded dorsal articular surface of the metacarpal head, a prominent dorsal articular ridge, and large epicondyles. In fact, metacarpals of these rays frequently resemble those of Pongo, and to a lesser extent humans, in these isolated features. In Gorilla metacarpal V achieves a length approaching t h a t of IV. In all three African apes (gorilla, chimpanzee, and bonobo), the proximal and middle phalanges are individually recognizable on the basis of relative length and pattern of asymmetry. Of the distal phalanges only that of the pollex is distinctive. The fingers of Pongo are notable in their length and curvature. The latter characteristic is primarily a result of proximal and middle phalangeal curvature. Suspensory features are especially marked in the proximal 233 phalanges. In Pongo the phalanges of ray IV display increased length relative to the other rays. The pattern of asymmetry also shifts (especially in males) whereby proximal phalanx IV appears to be more symmetrical than 111. This increased emphasis on the fourth ray and the occurrence of the “double locking” mechanism in orangutan fingers (Napier, ’60) reflects the increased importance of vertical climbing in Pongo and an enhanced capacity for grasping thin, vertical supports. Orangutan metacarpophalangeal joints are demonstrably less well adapted for knuckle-walking hand postures owing to the reduced dorso-palmar dimension of the metacarpal head, absence of a developed palmar plate, and a proportionately long proximal phalanx. Differences in finger morphology between Pongo and the hylobatid apes reflect distinctive differences in locomotor patterns and effects of disparate body size. The characteristic features of the pollex in the lesser apes are most distinctive (Lorenz, ’71; Van Horn, ’72; Tuttle, ’69, ’72). Because of the reduced body size of the lesser apes, they do not fit into the arboreal-terrestrial, or suspensory knucklewalking continuum, when the criterion is a single character (e.g., thickness of the metacarpal shaft cortex). Notable in hylobatid apes is the increased emphasis on the index finger. The second metacarpal of hylobatid apes is the most robust of the five. A metacarpal length formula of IT > TI1 > TV> V is found in all gi bbons and siamangs sampled (n = 22, and n = 9, respectively). It is noted above (table 8 ) that gibbons have the greatest proximal phalanx IIIimetacarpal I11 index among the Homoinoidea, and Midlo (’34) also noted that the lesser apes possess the highest phalangealipalm length ratio (ray 111) among the Hominoidea. The bones of rays 11-V are notable in their lack of secondary features (save in some specimens of Symphalangusl. The flexion set of the fingers is due to strongly curved proximal and middle phalanges. Pan paniscus It is interesting to note the position of Pan paniscus among the other apes in the discriminant analysis of metacarpal I11 and proximal phalanx 111. The relative position of bonobos between chimpanzees and orangutans suggests a more arboreal finger morphology in bonobos than in chimpanzees. Little has been reported on the behavior of bonobos (Pan paniscus) either in the wild or in captivity. 234 RANDALL L. SUSMAN Fig. 9 Pan troglodytes versus (left) and Pan paniscus (right) in similar knuckle-walking postures. Note the angle of the trunk to t h e horizontal (both animals are captive, young adult females). Recent reports, however, of pilot studies and work in progress indicate differences between bonobos and chimpanzees in locomotion and posture. Early reports (Schouteden, '31; Yerkes and Learned, '25) and recent brief reports of free-ranging bonobos (MacKinnon, '76; Badrian and Badrian, '77) suggest that they are more arboreal than are chimpanzees. My observations of captive animals a t the Antwerp Zoo, San Diego Zoo, and Yerkes Regional Primate Research Center, support this impression. Captive bonobos adopt knuckle-walking hand postures similar to chimpanzees and gorillas. The employment of knuckle-walking obtains, as in chimpanzees, both on the ground and on firm arboreal supports. The hands assume the same variety of angular placements and finger supports reported for other African apes (Tuttle? '67, '69. '72, '74, '77). Bonobos, however, appear to assume a more pronograde attitude than either chimpanzees or gorillas in normal quadrupedal stance and gait (fig. 9). This difference may reflect the lower intermembral index in bonobos than in chimpanzees (101 ' and 107. respectively). Badrian and Badrian ('77) report that in the wild bonobos are more arboreal than chimpanzees. Small groups of one t o four animals were observed t o nest high in the canopy (Badrian and Badrian, '77; MacKinnon, '76). Nests are similar to those of Pan troglodytes but bonobo nest heights exceed those of chimpanzees and even orangutans (MacKinnon, '76). Mean nest heights of 82 and 78 feet were reported by MacKinnon and the Badrians, respectively. Like chimpanzees, bonobos normally travel from one feeding source to another along terrestrial routes, but bonobos were observed to trek more than one kilometer via arboreal routes (Badrian and Badrian, '77: p. 466). In a number of important features bonobos fall between chimpanzees and orangutans while retaining their overall relationship with Pan troglodytes. One such feature is cortical thickness index of the metacarpal shaft (table 8). As noted earlier, this is probably related to large bending moments induced by strong power grasp. Bonobos reveal a value for the cortical thickness index very close to that of Pongo (wildshot specimens) and distinct from that of Pan troglodytes. Other features of the metacarpal head are less pronounced in bonobos than in chimpanzees and recall the unspecialized condition seen in Pongo. These include the lack of well defined dorsal articular ridges and absence of broad flat dorsal surfaces of the metacarpal heads. The phalanges also appear to be somewhat more "suspensory adapted." The curvature of the proximal phaThis difference is significant a t the 0.05 level (n= 13, d.f. = 11). 'This difference between honohos and chimpanzees IS significant at 0.05 level tn=24, d.f.=22). FUNCTIONAL MORPHOLOGY OF HOMINOID FINGERS langes and the attenuation of the middle and distal phalanges are suggestive of a more suspensory hand. Future field studies are necessary to confirm the position of bonobos in the hominoid locomotor spectrum. ACKNOWLEDGMENTS I thank R. H. Tuttle for his advice and help throughout the course of this work. Benjamin Beck provided numerous suggestions and behavioral insights during the early stages of this project. J. T. Stern provided helpful criticisms of biomechanical points, and N. Creel offered much statistical advice and many hours of enlightening discussion. J. G. Fleagle and W. L. Jungers also read and commented on the manuscript. Funds for this study were provided by the Henry Hinds Fund of the Committee on Evolutionary Biology, University of Chicago. I am grateful to L. Throckmorton and the Committee for their support throughout my graduate studies. Ms. Lucille Betti prepared the figures and Ms. Eileen Petite typed the manuscript. Special thanks are due to R. Thorington (National Museum of Natural History), S. Anderson (American Museum of Natural History), P. Helwig (Cleveland Museum of Natural History), and T. Van Den Audenaerde (Musee Royal de Lafrique Centrale, Tervuren) for access to specimens in their care. LITERATURE CITED Badrian, A,, and N. Badrian 1977 Pygmy chimpanzees. Oryx, 23: 463-468. Calhoun, T. P. 1977 Sesamoid structures in primate hands. Yearbook of Phys. Anthrop.. Vol. 20: 525-536. Conroy, G. C., and J. G. Fleagle 1972 Locomotor behavior in living and fossil pongids. Nature, 237: 103-104. Corruccini. R. S., R. L. Ciochon and H. M. McHenry 1975 Osteometric shape relationships in t h e wrist joint of some anthropoids. Folia Primat., 24: 250-274. Creel, N., and H. Preuschoft 1976 Cranial morphology ot the lesser apes. In: Gibbon and Siamang. D. M. Rumbaugh, ed. S . Karger, Basel, Vol. 4, pp. 219-283. Currey, J. D. 1968 The adaptation of bones to stress. J. Theoret. Biol., 20: 91-106. Day, M. H., and J. L. Scheuer 1973 SKW 14147: A new hominoid metacarpal from Swartkrans. J. Hum. Evol., 2: 429-438. Etter, H. F. 1974 Morphulogisch und metrisch-vergleichende Untersuchung am Handskelet rezenter Primaten, 111. Morph. Jahrb., 120: 299-322. Fleagle, J. G. 1976 Locomotion and posture of the Malayan Siamang and implications for hominoid evolution. Folia Primat., 26: 245-269. Gregory, W. K. 1928 Were the ancestors of man primi67: 129-150. tive brachiators? Proc. Am. Phil. SOC., Howells, W. W. 1973 Cranial Variation in Man. Papers Peabody Mus. Arch. and Ethol. No. 67. Harvard University Press. 235 Huxley, T. H. 1893 Man’s Place in Nature. D. Appleton and Co., New York. Jenkins, F A,, and J. G. Fleagle 1975 Knuckle walkingand t h e functional anatomy of the wrist in living apes. In: Primate Functional Morphology and Evolution. R. H. Tuttle, ed. World Anthropology, Mouton Publishers, The Hague, pp. 213-228. Johanson, D. C. 1976 Ethiopia yields first “family” of Early Man. National Geographic, 150: 791-811. Kaplan, E. B. 1965 Functional and Surgical Anatomy of t h e Hand. Second ed. Lippincott, Philadelphia. Landsmeer, J. M. F. 1955 Anatomical and functional investigations of t h e articulation of the human fingers. Acta Anatomica Supplementum 24-2,25: 1-69. Leakey, L. S. B. 1960 Recent discoveries a t Olduvai Gorge. Nature, 188: 1050-1052. Lewis, 0. J. 1969 The Hominoid Wrist Joint. Am. J. Phys. Anthrop., 30: 251-269. 1972a Osteological features characterizing the wrists of monkeys and apes, with a reconsideration of this region in Dryopithecus Proconsul) ufricanus. Am. J. Phys. Anthrop., 36: 45-58. 1972b Evolution of the hominoid wrist. In: The Functional and Evolutionary Biology of Primates. R. H. Tuttle, ed. Aldine-Atherton, Chicago. pp, 207-222. 1973 The hominoid 0s capitatum, with special reference to the fossil bones from Sterkfontein and Olduvai Gorge. J . Hum. Evol., 2: 1-11. 1977 Joint remodelling and the evolution of the human hand. J. Anat., 123: 157-201. Lorenz, R. 1971 The functional interpretation of t h e thumb of t h e hylobatidae. In: Proceedings of the 34th lntl. Cong. Primat. Biegert and Leutenegger, eds. S. Karger Basel, pp. 130-136. MacKinnon, J. 1976 Mountain gorillas and bonobos. Oryx, 13: 372-382. Marzke, M. W. 1971 Origin of the human hand. Am. J. Phys. Anthrop., 34: 61-84. Midlo, C. 1934 Form of t h e hand and foot in primates. Am. J. Phys. Anthrop., 19: 337-389. Morbeck, M. E. 1975 Dryopithecus africanus forelimb. J. Human Evol., 4: 39-46. Musgrave, J. H. 1971 How dextrous was Neanderthal man? Nature, 233: 538-541. 1973 The phalanges of Neanderthal and Upper Paleolithic hands. In: Human Evolution. Symp. Soc. Study Hum. Biol. M. H. Day, ed. Taylor and Francis, Ltd. London, Vol. XI, pp. 59-85. Napier, J. R. 1959 Fossil metacarpals from Swartkrans. Fossil Mamm. Africa, No. 17. Brit. Mus. Nat. Hist. 1960 Studies of the hands of living primates. London, 134: 647-656. Proc. Zool. SOC. 1962 Fossil hand bones from Olduvai Gorge. Nature, 196: 409-411. Napier, J. R., and P. R. Davis 1959 The forelimb skeleton and associated remains of Proconsul africunus. Fossil Mamm. Africa, No. 16. Birt. Mus. Nat. Hist. OConnor, B. L. 1976 Dryopithecus (Proconsul/ ufricanus: Quadruped or non-quadruped? J. Hum. Evol., 5: 279-283. Osborn, H. F. 1928 Present status of the problem of human ancestry, Proc. Am. Phil. SOC.,67: 151-156. Oxnard, C. E. 1973 Form and pattern in human evolution. University of Chicago Press, Chicago. Preuschoft, H. 1973 Body posture and locomotion in some East African Miocene Dryopithecinae. In: Human Evolution. Symp. SOC. Study Hum. Biol., M. H. Day, ed. Taylor and Francis, Ltd., London, Vol. XI, pp. 13-46. 236 RANDALL L. SUSMAN 1973h Functional anatomy of the upper extremity. In: The Chimpanzee. H. Bourne, ed. Karger, Basel, Vol. 6, pp. 34-120. Schon. M. A., and A. Ziemer 1973 Wrist mechanism and locomotor behavior ofDryoithecus Proconsul) ufricanus. Folia Primat., 20: 1-11, Schouteden, H. 1931 Le chimpanze de la rive gauche du Congo. Bull. du Cercle Zool. Congolais, 7: 114.119. Singh, 1. 1959 Variations in the metacarpal bones. J. Anat. London, 93: 262-267. Stern, J. T. 1976 Before bipedality. Yearbook of Phys. Anthrop., 19: 59-68. Strauas, W. L. 1940 The posture of the great ape hand in locomotion and its phylogenetic implications. Am. J. Phys. Anthrop., 27: 199-207. 1949 The riddle of man’s ancestry. Qtr. Rev. Biol., 24: 200-223. Tatsuoka, M. M. 1971 Multivariate Analysis. Techniques for Educational and Psychological Research. John Wiley & Sons, New York. Tuttle, H. H. 1967 Knuckle-walking and the evolution of hominoid hands. Am. J. Phys. Anthrop., 26: 171-206. 1969 Quantitative and functional studies on t h e hands of the anthropoidea. J. Morph., 128: 309-364. 1972 The functional evolutionary biology of hylohatid hands and feet. In: Gihhon and Siamang. D. M. Rumhaugh, ed. S. Karger, Basel. Vol. 1, pp. 136-206. 1974 Darwin’s apes, dental apes, and the descent of man: normal science in anthropology. Curr. Anthrop., 15: 389-398. 1977 Naturalistic positional behavior of apes and models of Hominoid Evolution, 1929-1976. In: Progress In Ape Research. G. H. Bourne, ed. Academic Press, New York, pp. 277-296. Van Horn, R. N. 1972 Structural adaptations to climbing in the gibbon hand. Am. Anthrop., 74: 326-334. Wood-Jones, F. 1942 Principles of Anatomy as Seen in t h e Hand. Second ed. Williams and Wilkins Co., Baltimore, Maryland. Yerkes, R. M., and B. W. Learned 1925 Chimpanzee Intelligence and i t s Vocal Expression. Williams and Wilkins, Co., Baltimore, Maryland. Zapfe, H. 7960 Die Primatenfunde aus der miozanen Spaltenfullung von Neudorf an der March (Devinska Nova Ves), Tschechoslowakei Schwciz. Paleont. Abhand., 78: 1-285.