Functional Osteology of the Primate Carpus With SpecialReference to Strepsirhini

MARK W. HAMRICK*Department of Anthropology, Kent State University, Kent, Ohio 44242

KEY WORDS: wrist morphology; positional behavior;lemuriforms; postcranial evolution

ABSTRACT Preuschoft et al. ([1993] in H. Preuschoft and D. Chivers(eds.): Hands of Primates. New York: Springer-Verlag, pp. 245–256) used atheoretical biomechanical analysis to generate several predictions relatingsubordinal differences in primate hand proportions to differences in carpalmorphology. This study tests these predictions using quantitative analyses ofcarpal morphology between extant haplorhine and strepsirhine primates.Results show that living strepsirhines have a significantly larger hamatehamulus than do haplorhines, supporting Preuschoft et al.’s (1993) predic-tions. Extant strepsirhines also have a significantly shorter pisiform bodythan do haplorhines and arboreal nonprimate eutherians and a largerscaphoid tubercle than New and Old World monkeys. These results contrastmarkedly with those expected under Preuschoft et al.’s (1993) model. Further-more, strepsirhines and haplorhines do not differ significantly in the relativesize of their radiocarpal articulations. These morphometric observations donot match the predicted morphological patterns because the kinematicassumptions upon which the biomechanical models are based are incorrect.Living strepsirhines appear to be derived in having very deep radial and ulnarmargins of the carpal tunnel for well-developed extrinsic digital flexors.Moreover, tooth-combed prosimians differ from most haplorhines, earlyTertiary adapiforms, and arboreal nonprimate eutherians in having a rela-tively short pisiform body, which gives the flexor carpi ulnaris less power toflex the wrist from extended (5 dorsiflexed) positions. These structuralobservations suggest that powerful manual grasping and an emphasis onleaping and climbing, rather than palmigrade quadrupedal walking andrunning, are morphotypic for extant Strepsirhini. Am J Phys Anthropol104:105–116, 1997. r 1997 Wiley-Liss, Inc.

A general survey of hand morphologyacross the Order Primates reveals a tremen-dous degree of variation in relation to differ-ent postural, locomotor, and manipulatorybehaviors (e.g., Midlo, 1934; LeGros Clark,1959; Bishop, 1964; Etter, 1974; Jouffroy etal., 1991; Hamrick, 1996a,b; Hamrick andAlexander, 1996). Primate hand proportionsare in particular quite variable, with manystrepsirhines such as lorises having an elon-gate fourth digit, a condition known as ec-taxony (Jouffroy et al., 1991). In contrast,

the third digit is the longest ray of the handin many haplorhine primates such as tar-siers and humans, a condition referred to asmesaxony (Jouffroy et al., 1991).

Preuschoft et al. (1993) defined some ste-reotyped primate hand postures which they

Contract grant sponsor: Boise Fund of Oxford University;Contract grant sponsor: Leakey Foundation; Contract grantsponsor: National Science Fund; Contract number SBR-932037.

*Correspondence to: Mark W. Hamrick, Department of Anthro-pology, Box 5190, Kent State University, Kent, OH 44242.

Received 9 June 1996; accepted 18 June 1997.

AMERICAN JOURNAL OF PHYSICAL ANTHROPOLOGY 104:105–116 (1997)

r 1997 WILEY-LISS, INC.

believed characterized ‘‘the ectaxonic handof most strepsirhines and the mesaxonic[hand] of many, especially ground-dwelling,monkeys, apes, and humans’’ (Preuschoft etal., 1993, p. 245). Their analysis of strepsi-

rhine hand biomechanics was based uponthe following assumptions: 1) the grasp is apincer grasp between digit I and the post-axial rays and 2) the hand is ulnarly devi-ated on arboreal supports (Fig. 1). In con-

Fig. 1. a: Ectaxonic hand of strepsirhines and mesax-onic hand of haplorhines showing the more ulnarlydeviated hand of strepsirhines. The resultant muscleforce (Cu) from the digital flexors (M) has a largecomponent (Fu) directed towards the ulnar margin ofthe wrist in strepsirhines. In contrast, the resultantforce (Cr) from the flexors (M) has a large component(Fr) directed towards the radial margin of the wrist inhaplorhines. b: The ulnar borders of the carpal tunnel(e.g., the pisiform body and hamate hamulus) and the

ulnocarpal articular surfaces are therefore expected tobe relatively expanded in strepsirhines in order tostabilize the carpus during contraction of the digitalflexors. The radial border of the carpal tunnel (e.g., thescaphoid tubercle) and the radiocarpal articular sur-faces are expected to be relatively expanded in haplo-rhines in order to stabilize the carpus during contractionof the digital flexors. Modified and redrawn from Preus-choft et al. (1993).

106 M.W. HAMRICK

trast, their analysis of haplorhine handbiomechanics assumed that 1) the digits areplaced in line with the long axis of thesupport and 2) the hand is not significantlyulnarly deviated (Fig. 1).

Preuschoft et al.’s (1993) model predictsthat the strepsirhine carpus should experi-ence an ulnarly directed resultant muscleforce produced by contraction of the longdigital flexors during manual grasping (Fig.1). Strepsirhines should therefore have rela-tively large ulnocarpal articular surfacesand ulnar borders of the carpal tunnel inorder to reduce stress on the bones andjoints produced by this resultant force. Incontrast, haplorhine carpals should experi-ence a radially directed resultant muscleforce produced by the long flexors (Fig. 1).Thus, haplorhines are expected to possessexpanded radial articular elements and ra-dial borders of the carpal tunnel.

The purpose of this study is to test themorphological predictions germane to thesebiomechanical models in order to elucidatesome similarities and contrasts in carpalmorphology between strepsirhine and haplo-rhine primates. These morphologies are thenanalyzed from a functional perspective inorder to reconstruct the pattern of carpalform and function morphotypic for the pri-mate Suborder Strepsirhini. Finally, thesedata on carpal form and function are used asa basis for inferring the locomotor and pos-tural behaviors present in the last commonancestor of tooth-combed primates.

MATERIALS AND METHODSSample

Subordinal differences in carpal morphol-ogy were examined in an extant skeletalsample of 34 haplorhine and strepsirhineprimate species of which 33 are differentgenera (Table 1). Over 350 individuals wereincluded in this sample, composed of approxi-mately 90 haplorhines and over 250 strepsi-rhines. The haplorhines range in body sizefrom approximately 130 g in the case ofTarsius up to 20 kg in Papio anubis, whereasthe strepsirhines range in size from approxi-mately 215 g in Galago senegalensis up to 10kg in Indri (Fleagle, 1988). Several pre-served specimens of each suborder were alsodissected. This sample is designed to include

both haplorhines and strepsirhines coveringa similar range of body masses. The large-bodied hominoids were excluded from analy-sis primarily because there are no livingstrepsirhines of similar size that could beincluded for comparison.

The overwhelming majority of these haplo-rhines are mesaxonic—that is, the third rayis over 3% longer than the fourth ray. Thestrepsirhines are primarily ectaxonic, wherethe fourth ray is over 3% longer than thethird ray. Paraxonic species are defined asthose in which the third and fourth rays aresimilar in length to within 3% of one an-other. These distinctions (Table 1) are basedon the hand proportion data set publishedby Jouffroy et al. (1991). Hand skeletons ofone early Tertiary primate (Notharctus tene-brosus, AMNH 127167) (Hamrick and Alex-ander, 1996) and three nonprimate euthe-rian mammals, Procyon (Carnivora), Sciurus(Rodentia), and Tupaia (Scandentia), werealso included to expand the comparativedata set. These taxa are intended to repre-sent outgroups useful for assessing the phy-logenetic valence of some of the traits dis-cussed in this paper.

Osteometrics and statistical analysis

Linear dimensions of the carpal tunnelborders, such as dorsopalmar lengths of the

TABLE 1. Extant strepsirhine and haplorhine skeletalspecimens included for analysis1

Haplorhines Strepsirhines

Tarsius spp.2 (7) Cheirogaleus major3 (8)Saguinus oedipus2 (6) Lemur fulvus4 (20)Saimiri sciureus2 (10) Lemur catta4 (19)Callithrix jacchus2 (6) Varecia variegata3 (19)Callimico goeldii3 (5) Hapalemur griseus4 (21)Aotus trivirgatus3 (5) Lepilemur mustelinus4 (24)Ateles paniscus3 (3) Indri indri4 (22)Alouatta palliata2 (4) Propithecus verreauxi4 (22)Pithecia pithecia2 (3) Avahi laniger4 (13)Miopithecus talapoin2 (5) Daubentonia madagascaren-

sis4 (8)Macaca spp.2 (4) Otolemur crassicaudatus4

(19)Papio anubis3 (6) Galago senegalensis4 (7)Colobus guereza3 (3) Euoticus elegantulus4 (11)Presbytis frontatus2 (3) Loris tardigradus4 (8)Procolobus verus2 (3) Nycticebus coucang4 (20)Hylobates spp.2 (6) Arctocebus calabarensis4 (4)

Perodicticus potto4 (17)1 The number of individuals included for each species is shown inparentheses.2 Mesaxonic.3 Paraxonic.4 Ectaxonic.

107STREPSIRHINE CARPAL MORPHOLOGY

scaphoid tubercle (STL) and hamate hamu-lus (HAL), radioulnar diameters of the radio-carpal (RCW) and ulnocarpal (UCW) articu-lar surfaces, and dorsopalmar height of thepisiform body (PISH), were taken with dialcalipers (Fig. 1). Breadth of the radiocarpalarticular surface (RCW) was measured asradioulnar diameter of the radiocarpal ar-ticular facet on the distal radius, whereasbreadth of the ulnocarpal articular surface(UCW) was measured as radioulnar diam-eter of the ulnocarpal articular facet on thedistal ulna. Previous workers (Swartz, 1989;Godfrey et al., 1991) have suggested thatjoint surface areas are preferable over lineardiameters as estimates of joint size. In orderto test whether or not radioulnar diameterof the radiocarpal articular surface is areliable predictor of radiocarpal articulararea, I regressed logradiocarpal articular area(dependent variable) against logradioulnardiameter of the radiocarpal articular surface(independent variable). Radiocarpal areawas calculated from dorsopalmar and radi-oulnar arc lengths taken from epoxy-resincasts with a Reflex microscope (see Ham-rick, 1996a,c) using the formula for the areaof an ellipse (p*1/2dorsopalmar arc length*1/2radioulnar arc length). Results indicate that,within a sample of 266 individuals from over 20haplorhine and strepsirhine species, radiocar-pal area is significantly correlated with radioul-nar diameter of the radiocarpal facet (r 5 .99,P , .001, slope 5 1.88, y intercept 5 2.56,S.E.E. 5 .13), indicating that radiocarpal diam-eter is a reliable predictor of radiocarpal articu-lar area.

Subordinal differences in primate carpaljoint morphology were examined by perform-ing several between-group statistical testson pooled-sex mean species values. The bodysize estimate used as the independent vari-able in these bivariate analyses is the meanof humeral midshaft diameter and mediolat-eral diameter of the distal calcaneal articu-lar facet. A regression between publishedmean species body weights and speciesmeans for the skeletal estimate used heredemonstrate that this composite measure isboth significantly correlated (r 5 .99) andvirtually isometric (slope 5 .34) with bodysize. Ordinary least-squares (OLS) analysis

of covariance (ANCOVA) was used to test forsubordinal differences in carpal joint dimen-sions relative to this body size estimate.Tsutakawa and Hewett’s (1977) quick testwas used as a second, nonparametric bivari-ate approach. The reduced major axis regres-sion model (RMA) (Clarke, 1980) was chosenfor this nonparametric test because neitherthe X nor Y variable was measured withouterror (Harvey and Pagel, 1991). Correla-tions between carpal dimensions and esti-mated body size are quite high in mostcases, and results using each regressiontechnique are very similar (Table 2). Subor-dinal differences in carpal joint shape werealso explored using a multivariate discrimi-nant function analysis performed on log-transformed variables.Amultivariate analy-sis of variance (MANOVA) was the preferredtechnique for this analysis since the taxo-nomic groups were defined a priori (Neff andMarcus, 1980).

RESULTS

Bivariate plots of radiocarpal and ulnocar-pal diameters against estimated body sizeshow considerable overlap between the twosuborders in relative size of the carpal jointarticular surfaces (Fig. 2). Strepsirhines tend

TABLE 2. Summary of regression analyses forhaplorhine and strepsirhine carpal dimensions against

estimated body size1

Measurement Y intercept Slope r

HaplorhinesRCW .24 .98 .99

(.23) (.99)UCW 2.19 .71 .97

(2.22) (.73)PISH .41 .83 .98

(.37) (.85)STL 2.73 1.04 .95

(2.83) (1.09)HAL 21.24 1.05 .89

(21.49) (1.18)Strepsirhines

RCW .15 1.04 .97(.11) (1.07)

UCW 2.63 .99 .90(2.81) (1.10)

PISH 2.35 1.11 .93(2.49) (1.19)

STL 2.44 1.02 .93(2.56) (1.09)

HAL 2.53 .83 .82(2.81) (1.0)

1 Ordinary least-squares values are shown above reduced major-axis values, which are in parentheses. Measurement abbre-viations are explained in the text.

108 M.W. HAMRICK

to have a higher slope for their ulnocarpaldimensions relative to body size (Table 2)and in fact differ significantly (P , .05) fromhaplorhines in this respect (Table 3). Visualinspection of the bivariate scatter, however,reveals that there is considerable overlapbetween the two suborders within intermedi-ate body size ranges (Fig. 2a). The smallerlorises, which are the most ectaxonic of allthe primates (Jouffroy et al., 1991), haverelatively reduced ulnocarpal diameters,whereas the paraxonic Varecia has a rela-tively expanded ulnocarpal joint (Fig. 2a).Thus, if there is any correlation betweendegree of ectaxony and relative size of theulnocarpal joint, it would appear to be anegative correlation. Bivariate analyses(Table 3) and visual inspection of bivariateplots (Fig. 2b) show no differences between

strepsirhine and haplorhine primates in rela-tive breadth of the radiocarpal articularsurface.

The two suborders do, however, differ inmorphology of the proximal carpal tunnelborders—that is, in relative size of the scaph-oid tubercle and pisiform body. Tree shrewsand monkeys tend to have a relatively shortscaphoid tubercle, whereas the tubercle ismore elongate in strepsirhines (Fig. 3a).Notharctus and Tarsius, like living strepsi-rhines, have a relatively elongate scaphoidtubercle, suggesting that this is likely aprimitive primate trait (Fig. 4a). Bivariatetests show that strepsirhines differ signifi-cantly (P , .001) from haplorhines in rela-tive scaphoid tubercle length (Table 3), withstrepsirhines having a relatively longerscaphoid tubercle than haplorhines across

Fig. 2. Bivariate plots of (a) ulnocarpal and (b) radiocarpal breadth against estimated body size instrepsirhines (open squares) and haplorhines (closed squares). Fossil and nonprimate taxa (shaded) arelabeled.

TABLE 3. ANCOVA tests (haplorhines vs. strepsirhines) for the carpal dimensions included for analysis1

Measurement RMA quick test OLS slope test OLS y intercept test

RCW ns ns nsUCW ** *

(strepsirhines . Y) (strepsirhines . slope)PISH ** **

(haplorhines . Y) (strepsirhines . slope)STL *** ns ***

(strepsirhines . Y) (strepsirhines . Y)HAL ** ns **

(strepsirhines . Y) (strepsirhines . Y)1 Measurement abbreviations are explained in the text. ns, P . .05.* P , .05.** P , .01.*** P , .001.

109STREPSIRHINE CARPAL MORPHOLOGY

common body size ranges (Fig. 4a). Thispattern is the exact opposite of that expectedunder the biomechanical model discussedearlier (Fig. 1), where the predominantlymesaxonic haplorhines were expected to havea relatively larger scaphoid tubercle thanthe primarily ectaxonic strepsirhines.

Bivariate plots also show that a morpho-cline exists between the two suborders inwhich strepsirhines have a relatively shortpisiform body and haplorhines tend to havea somewhat taller pisiform (Fig. 3b). Forexample, monkeys and nonprimate eutheri-ans such as Tupaia have a tall pisiform body,whereas that of Lemur is somewhat reduced(Fig. 3b). Bivariate plots (Fig. 4b) under-score this point and show that the twosuborders differ significantly from one an-other in relative height of the pisiform body(P , .01) (Tables 2, 3). A morphocline isapparent within the sample where mon-keys, tarsiers, Notharctus, tree shrews,squirrels, and racoons all tend to share arelatively tall pisiform body, whereas strep-sirhines appear derived in having a rela-tively reduced pisiform body (Fig. 4b). Thispattern is again precisely the opposite ofthat expected under the biomechanical model

described earlier (Fig. 1), where the moreectaxonic strepsirhines were expected tohave a tall pisiform body.

Finally, strepsirhine primates have a well-developed hamate hamulus compared to hap-lorhines and climbing nonprimate eutheri-ans. The hamate hamulus, like the scaphoidtubercle, is relatively short in haplorhinescompared to strepsirhines and is altogetherabsent in many tree shrews and squirrels(Fig. 3c). Bivariate plots again show that amorphocline exists between the two subor-ders in which monkeys, tarsiers, and Nothar-ctus all have a relatively short hamate hamu-lus, whereas strepsirhines tend to have awell-developed hamulus (Fig. 4c). Bivariatetests show that strepsirhines differ signifi-cantly from haplorhines in relative length ofthe hamate hamulus (P , .01) (Table 3),with strepsirhines having a relatively longerhamulus than haplorhines across commonbody size ranges (Fig. 4c).

The aforementioned osteometric resultsare summarized using a multivariate dis-criminant function analysis. The discrimi-nant function classified 97% of the casescorrectly and is highly significant (P , .001,Wilkes-Lambda). The discriminant axisshows a strong negative correlation withheight of the pisiform body (Table 4), wherethe strepsirhines tend to have high scores onthis axis and a relatively small pisiformbody in contrast to the haplorhines, whichhave lower scores and a relatively largerpisiform body (Fig. 5). The only case whichwas classified incorrectly is Daubentonia,which groups with haplorhines in the dis-criminant analysis. Daubentonia has a tall,rod-like pisiform which contacts the radius,a condition unknown in other eutherianmammals (Nayak, 1933). The strong loadingof pisiform height on the first discriminantaxis explains why Daubentonia is classifiedwith the haplorhines rather than the strepsi-rhines, since Daubentonia, like haplorhines,has a relatively expanded pisiform body(Nayak, 1933).

DISCUSSION

The above morphometric analyses revealthat patterns of strepsirhine and haplorhinecarpal morphology do not fit those expected

Fig. 3. Drawings of the (a) scaphoid, (b) pisiform,and (c) hamate in Tupaia (left), Saimiri (middle), andLemur (right) showing the relatively elongate scaphoidtubercle (shaded), reduced pisiform body, and elongatehamate hamulus (shaded) of strepsirhines. Scale bars 51 mm.

110 M.W. HAMRICK

Fig. 4. Bivariate plots of (a) scaphoid tubercle length,(b) pisiform body height, and (c) hamate hamuluslength against estimated body size in strepsirhines(open squares) and haplorhines (closed squares). Fossiland nonprimate taxa (shaded) are labeled. Dotted linesrepresent ordinary least-squares regression lines forhaplorhines and strepsirhines.

Fig. 5. Univariate plot of the first discriminant axisof a discriminant analysis performed on log-transformedlinear carpal dimensions of strepsirhine (open squares)and haplorhine (closed squares) primates. The only caseclassified incorrectly is Daubentonia.

TABLE 4. Canonical loadings and coefficients betweencarpal measurements and the first discriminant

function axis of a MANOVA performed onlog-transformed haplorhine and strepsirhine

carpal dimensions1

MeasurementCanonical

loadingCanonicalcoefficient

PISH 2.39 23.82RCW 2.20 .66UCW 2.18 1.41STL 2.07 1.84HAL 2.01 2.32

1 Measurement abbreviations are explained in the text.

111STREPSIRHINE CARPAL MORPHOLOGY

under Preuschoft et al.’s (1993) model. Arecent study of primate hand postures(Schmitt and Lemelin, 1995) suggest thatthis is so because the kinematic assump-tions upon which the biomechanical modelsare based are incorrect. Specifically, al-though ectaxonic primates do indeed useabducted hand postures on arboreal sup-ports, the degree of abduction is often thesame or less than that observed in arborealmesaxonic primates (Schmitt and Lemelin,1995; Lemelin and Schmitt, submitted). Thehand postures and related muscle resultantforces (Fig. 1) attributed to haplorhines andstrepsirhines by Preuschoft et al. (1993)therefore do not accurately reflect those ofmany arboreal mesaxonic and ectaxonic pri-mates. Although both haplorhines and strep-sirhines use ulnarly deviated hand postures,strepsirhines produce this movement pri-marily at the midcarpal joint, whereas hap-lorhines produce most of this movement atthe metacarpophalangeal joints (Lemelinand Schmitt, submitted). The relatively largehamate and associated centrale-hamate con-tact of living strepsirhines (Beard et al.,1988) are then likely to be functionallyrelated to frequent ulnar deviation at themidcarpal joint.

The underlying functional significance ofthe subordinal differences in carpal morphol-ogy presented here can be further under-stood by examining the soft tissue anatomyof the primate hand. The transverse carpalligament, which forms the carpal tunnel,attaches primarily to the scaphoid tubercleand hamate hamulus, giving the ligamentan oblique orientation (Fig. 6). Scaphoidtubercle length and hamate hamulus lengthare, not surprisingly, highly correlated(r 5 .89, P , .001) with one another in thissample. These bony structures are elongatein living strepsirhines, which deepens thecarpal tunnel for well-developed tendons ofthe extrinsic digital flexors. The elongatescaphoid tubercle also serves as a windlassmechanism, which helps the pollical tendonof m. flexor digitorum profundus pull thethumb towards the palm in primates with adivergent, grasping pollex (Fig. 6). Relativepisiform size appears to have little influenceupon carpal tunnel depth but is an impor-tant feature related to quadrupedal walk-

ing, running, and bounding. The elongatepisiform body of haplorhines increases themoment arm of m. flexor carpi ulnaris (FCU),a muscle which flexes the wrist from ini-tially extended (5 dorsiflexed) positions dur-ing the propulsive phase of quadrupedallocomotion (Whitehead, 1993), by placing itsinsertion further away from the joint center(Fig. 6) (Sarmiento, 1988; Hamrick, 1996b).

Haplorhines and adapiforms (includingAdapis [Hamrick, 1996b]) share with nonpri-mate eutherians 1) a relatively tall pisiformbody, which increases the moment arm form. flexor carpi ulnaris (FCU), and 2) arelatively short hamate hamulus, indicativeof a carpal tunnel that is relatively shallowcompared to that of modern strepsirhines(Figs. 3, 4). Functionally, these bony fea-tures are related to 1) a flexor carpi ulnarismuscle that is afforded greater mechanicaladvantage to flex the wrist from extendedpositions compared to strepsirhines and 2)extrinsic digital flexors that are not so welldeveloped as those of extant strepsirhines.Behaviorally, such features are associatedwith primarily quadrupedal locomotor be-haviors. The presence of these features inadapiforms, tarsiers, monkeys, squirrels, ra-coons, and tree shrews suggests that theseare primitive mammalian features whichwere retained by the last common ancestorof Primates.

The taxonomic distribution of carpal char-acters referred to here indicates that anelongate scaphoid tubercle is shared by tar-siers, adapiforms, and strepsirhines but nottree shrews. Godinot and Beard (1991) con-sidered a long scaphoid tubercle, large pisi-form, and small hamulus to be primitive forprimates, although they did not specifywhich of these features were derived foreuprimates and which were not. Evidence

Fig. 6. Palmar views of the superficial hand muscula-ture in Otolemur and Papio showing the obliquelydirected proximal border of the transverse carpal liga-ment (L) which can be seen attaching primarily to thehamate hamulus (H) and scaphoid tubercle (St). Theflexor carpi ulnaris muscle (Fc) inserts upon the palmarsurface of the pisiform (P, Star). Note also that thepollical tendon of m. flexor digitorum profundus (I) isdirected at a right angle relative to the long axis of theforearm in Otolemur, which has a very divergent pollex.The long flexor tendon to the second digit is also shown(II). Scale bars 5 1 cm.

112 M.W. HAMRICK

Fig. 6.

113STREPSIRHINE CARPAL MORPHOLOGY

presented in this paper suggests that anelongate scaphoid tubercle is a derived fea-ture of euprimates. A long scaphoid tubercledeepens the radial margin of the carpaltunnel and also acts as a windlass mecha-nism for the pollical branch of flexor digi-torum profundus, which permits powerfulthumb adduction from abducted positions.Altner (1971) also observed that primatesdiffer from tree shrews in having a thumbcomplex that is relatively independent fromthe rest of the hand, and he too related thisform-function complex to more effectivemanual grasping in primates. Behaviorally,structural adaptations for powerful pollicalgrasping are, like those adaptations for hal-lucial grasping, most likely related to clasp-ing small-diameter arboreal substrates ofirregular orientation (Cartmill, 1972, 1974).

The tendency for extant strepsirhines topossess both a relatively short pisiform bodyand an elongate hamate hamulus is notobserved in the majority of primates (homi-noids excepted) and climbing eutherians.Although there are certainly cases in whichsome strepsirhine species overlap with hap-lorhines in the relative expression of thesetraits and vice versa (e.g., Fig. 4b,c), there isnevertheless a clear indication that modernstrepsirhines as a whole differ from otherprimate and eutherian groups in the develop-ment of these postcranial features. Hence,the relatively reduced pisiform and ex-panded hamulus of extant strepsirhines arelikely derived features for primates andeutherians. The short pisiform body givesthe flexor carpi ulnaris less power to flex thewrist from extended positions (Jouffroy, 1991;Hamrick, 1996b), indicative of a decreasedcommitment to quadrupedal locomotion andthus more frequent leaping and/or climbingbehaviors (Dagosto, 1990; Hamrick, 1996b).The relatively short pisiform body of strepsi-rhine vertical clingers and even shorter pisi-form body of strepsirhine slow climbers(Hamrick, 1996b) suggest further that pisi-form reduction is correlated with a de-creased commitment to palmigrade, quadru-pedal locomotion. Pisiform reduction isespecially obvious in humans, where thehand has been liberated entirely from itsrole as a locomotor organ (Jouffroy, 1991).The elongate hamulus increases the depth

of the carpal tunnel for well-developed ex-trinsic digital flexors, which are importantfor powerful manual grasping. As Napier(1961, p. 117) noted, ‘‘[T]he carpal arch isdeepest in those forms in which finger flex-ion takes a dominant role in locomotion, andshallowest in those in which locomotion is ofthe more generalized quadrupedal type.’’

The morphoclines observed in osteometricanalyses lend additional support to thesefunctional interpretations. Hylobates, whichhas very powerful grasping hands and well-developed digital flexors (Tuttle, 1969, 1972),has the highest score of all the haplorhinesin the discrimant function analysis (Fig. 5)and is therefore the most strepsirhine-likehaplorhine in terms of its carpal tunnelmorphology. Daubentonia, Varecia, and Chei-rogaleus, which use horizontal supports fre-quently (Gebo, 1987; Oxnard et al., 1990;Dagosto, 1994), have the lowest scores of allthe strepsirhines in the discriminant func-tion analysis (Fig. 5) and are therefore themost haplorhine-like strepsirhines in termsof their carpal tunnel morphology. The pow-erful grasping hand of extant strepsirhinesis well suited for the more frequent use ofsmall-diameter supports and the use of thesesupports for clinging and climbing posi-tional behaviors. This may also suggest abias towards the use of vertical supports bythe last common ancestor of tooth-combedprosimians (Gebo, 1986; Dagosto, 1988).

CONCLUSIONS

Extant strepsirhine and haplorhine pri-mates are quite similar to one another in therelative size of their radiocarpal and ulnocar-pal articular surfaces. Extant strepsirhines(5 lemuriforms) do, however, tend to differfrom haplorhines, early Tertiary adapi-forms, and nonprimate eutherians in havinga reduced pisiform body and therefore areduced lever arm for the flexor carpi ulna-ris muscle. Lemuriforms also differ fromthese taxa in having a relatively well-developed hamate hamulus, whereas lemu-riforms share with tarsiers and early Ter-tiary adapiforms a well-developed scaphoidtubercle. The elongate hamate hamulus andscaphoid tubercle of lemuriforms deepen thecarpal tunnel for relatively large digitalflexor tendons. The tendency for tooth-

114 M.W. HAMRICK

combed strepsirhines to possess both a re-duced pisiform and elongate hamulus ap-pears to be derived relative to other primatesand nonprimate arborealists. Results pre-sented in this study show that lemuriformprimates are distinguished from the major-ity of primates (including adapiforms) andnonprimate eutherians by several bony post-cranial features that can be identified infossils, in addition to their tooth comb. More-over, results from this study indicate thatthe origin of tooth-combed primates wasaccompanied by a significant change in handmorphology, function, and substrate use.

ACKNOWLEDGMENTS

I am grateful to the following individualsand institutions for access to skeletal andcadaver specimens of extant and extinctprimates in their care: Dr. G. Musser andMr. Wolfgang Fuchs, American Museum ofNatural History; Dr. R. Thorington, Jr., andMs. Linda Gordon, United States NationalMuseum; Ms. M. Rutzmoser, Museum ofComparative Zoology; Dr. B. Latimer andMr. L. Jellema, Cleveland Museum of Natu-ral History; Ms. Paula Jenkins, British Mu-seum of Natural History; Dr. Michel Tranier,Musee National d’Histoire Naturelle; Dr. C.Smeenk, Rijksmuseum van Natuurlijke His-torie; Dr. Renate Angermann, Museum furNaturkunde der Humboldt-Universitat; andMr. John Alexander, American Museumof Natural History, for allowing me tostudy newly discovered hand specimens ofNotharctus. Drs. M. Dagosto, M. Ravosa, B.Shea, and W. Jungers provided importantinsights and suggestions throughout thecourse of this study, and three anonymousreviewers provided helpful comments whichimproved the quality of the manuscript.

LITERATURE CITED

Altner G (1971) Histologische und vergleichend-anato-mische untersuchungen zur ontogenie und phylogeniedes handskeletts von Tupaia glis und Microcebusmurinus. Folia Primatol. 14:1–106.

Beard KC, Dagosto M, Gebo D, and Godinot M (1988)Interrelationships among primate higher taxa. Na-ture 331:712–714.

Bishop A (1964) Use of the hand in lower primates. In J.Buettner-Janusch (ed.): Evolution and Genetic Biol-ogy of Primates, Vol. 2. New York: Academic Press, pp.133–225.

Cartmill M (1972) Arboreal adaptations and the originof the order Primates. In RH Tuttle (ed.): Functionaland Evolutionary Biology of Primates. Chicago: Aldine-Atherton, pp. 97–122.

Cartmill M (1974) Pads and claws in arboreal locomo-tion. In FA Jenkins Jr. (ed.): Primate Locomotion. NewYork: Academic Press, pp. 133–225.

Clarke MRB (1980) The reduced major axis of a bivari-ate sample. Biometrika 67:441–446.

Dagosto M (1988) Implications of postcranial evidencefor the origin of euprimates. J. Hum. Evol. 17:35–56.

Dagosto M (1990) Models for the origin of the anthropoidpostcranium. J. Hum. Evol. 19:121–139.

Dagosto M (1994) Testing positional behavior of Mala-gasy lemurs: A randomization approach. Am. J. Phys.Anthropol. 94:189–202.

Etter HUF (1974) Morphologisch- und metrisch-ver-gleichende untersuchung am Handskelet rezenterprimaten. Gegenbaurs Morphol. Jahrb. 120:1–21.

Fleagle JG (1988) Primate Adaptation and Evolution.New York: Academic Press.

Gebo DL (1986) Anthropoid origins—the foot evidence.J. Hum. Evol. 15:421–430.

Gebo DL (1987) Locomotor diversity in prosimian pri-mates. Am. J. Primatol. 13:271–281.

Godfrey L, Sutherland M, Boy D, and Gomberg N (1991)Scaling of limb joint surface areas in anthropoidprimates and other mammals. J. Zool. 223:603–625.

Godinot M, and Beard KC (1991) Fossil primate hands:A review and an evolutionary inquiry emphasizingearly forms. Hum. Evol. 6:307–354.

Hamrick MW (1996a) Functional morphology the lemu-riform wrist joints and the relationship between wristmorphology and positional behavior in arboreal pri-mates. Am. J. Phys. Anthropol. 99:319–344.

Hamrick MW (1996b) Locomotor adaptations reflectedin the wrist joints of early Tertiary primates (Adapifor-mes). Am. J. Phys. Anthropol. 100:585–604.

Hamrick MW (1996c) Articular size and curvature asdeterminants of carpal joint mobility and stability instrepsirhine primates. J. Morphol. 230:113–127.

Hamrick MW, and Alexander JP (1996) The hand skel-eton of Notharctus tenebrosus (Primates, Adapidae)and its significance for the origin of the primate hand.Am. Mus. Nat. Hist. Novitates 3182:1–20.

Harvey PH, and Pagel MD (1991) The ComparativeMethod in Evolutionary Biology. Oxford: Oxford Uni-versity Press.

Jouffroy FK (1991) La’main sans talon’ du primatebipede. In Y. Coppens and B. Senut (eds.). Origine(s)de la Bipedie Chez Les Hominides. Paris: CNRS, pp.21–35.

Jouffroy FK, Godinot M, and Nakano Y (1991) Biometri-cal characteristics of primate hands. Hum. Evol.6:269–306.

LeGros Clark WE (1959) The Antecedents of Man.Edinburgh: Edinburgh University Press.

Lemelin P, and Schmitt D (submitted) The relationshipbetween hand morphology and quadrupedalism inprimates. Am. J. Phys. Anthropol.

Midlo C (1934) Form of hand and foot in primates. Am. J.Phys. Anthropol. 19:337–389.

Napier JR (1961) Prehensility and opposability in thehands of primates. Symp. Zool. Soc. Lond. 5:115–132.

Nayak UV (1933) The articulations of the carpus inChiromys madagascarensis with reference to certainother lemurs. J. Anat. 68:109–115.

Neff NA, and Marcus LF (1980) A Survey of MultivariateMethods for Systematics. New York: American Mu-seum of Natural History.

Oxnard CE, Crompton RH, and Lieberman SS (1990)Animal Lifestyles and Anatomies: The Case of the

115STREPSIRHINE CARPAL MORPHOLOGY

Prosimian Primates. Seattle: University of Washing-ton Press.

Preuschoft H, Godinot M, Beard C, Nieschalk U, andJouffroy FK (1993) Biomechanical considerations toexplain important morphological characters of pri-mate hands. In H Preuschoft and D Chivers (eds.):Hands of Primates. New York: Springer-Verlag, pp.245–256.

Sarmiento EE (1988) Anatomy of the hominoid wristjoint: Its evolutionary and functional implications.Int. J. Primatol. 9:281–345.

Schmitt D, and Lemelin P (1995) Hand position duringarboreal and terrestrial quadrupedalism in primates.Am. J. Phys. Anthropol. 20 (Suppl.):190–191 (ab-stract).

Swartz SM (1989) The functional morphology of weight-bearing: Limb joint surface allometry in anthropoidprimates. J. Zool. 218:441–460.

Tsutakawa RK, and Hewett JE (1977) Quick test forcomparing two populations with bivariate data. Bio-metrics 33:215–219.

Tuttle RH (1969) Quantitative and functional studies onthe hands of Anthropoidea. J. Morphol. 128:309–364.

Tuttle RH (1972) Functional and evolutionary biology ofhylobatid hands and feet. In DM Rumbaugh (ed):Gibbon and Siamang, Vol. 1. Basel: Karger, pp. 136–206.

Whitehead P (1993) Aspects of the anthropoid wrist andhand. In D Gebo (ed.): Postcranial Adaptation inNonhuman Primates. DeKalb: Northern Illinois Uni-versity Press, pp. 96–120.

116 M.W. HAMRICK