AMERICAN JOURNAL OF PHYSICAL ANTHROPOLOGY 84:249-260 (1991)
Mechanical Advantages of the Neandertal Thumb in Flexion: A Test of an Hypothesis ERIK TRINKAUS AND ISABELLE VILLEMEUR Department of Anthropology, University of New Mexico, Albuquerque, New Mexico 87131 (E.T.); U A . 376 du C.N.R.S., Laboratoire d’Anthropologie, Universite de Bordeaux I, 33405 Talence, France (E.T.); Laboratoire d’Anthropologie, Universite de Bordeaux I, 33405 Talence, France (I.V.)
KEY WORDS
Human paleontology, Hand morphology, Thumb
ABSTRACT It has been proposed that the pollical phalangeal length proportions of the Neandertals provided them with a greater mechanical advantage relative to recent humans for their pollical flexor muscles in power grips across the interphalangeal (IP)joint at the expense of the mechanical advantage of those pollical flexor muscles in precision grips at the finger tip. To test these related hypotheses, we compared the pollical load arm dimensions (phalanx lengths) to power arm dimensions (dorsopalmar articular heights) for the European and Near Eastern Neandertals and for European and Amerindian samples of recent humans. It was found, initially, that the proximal articular height of the pollical distal phalanx is a poor predictor of the power arm at the IP articulation, even though the proximal articular height of the pollical proximal phalanx was an adequate indicator of the power arm size at the metacarpophalangeal (MCP) joint. In addition, differences in distal pollical ulnar deviation at the IP joint appeared to make little difference in the mechanical advantage comparisons. More importantly, the relative shortness of Neandertal proximal pollical phalanges and the relative lengthening of their distal pollical phalanges was confirmed, and it was determined that, despite some minor differences in articular dimensions between Neandertals and recent humans, these pollical phalangeal length contrasts translated into significant differences in mechanical advantages for the flexor muscles across the MCP and IP articulations. Research on the evolution of human manipulative behavior has increasinglyfocused on the preserved hand remains of archaic fossil hominids, and especially on the relatively abundant and associated hand remains of European and Near Eastern late archaic humans (Neandertals sensu lato) (e.g., Musgrave, 1970, 1971, 1973; VlEek, 1975; Kimura, 1976; Stoner, 1981; Stoner and Trinkaus, 1981; Heim, 1982; Trinkaus, 1983;Riley and Trinkaus, 1989).In addition to documenting the exaggerated muscularity and general articular robusticity of the Neandertals relative to modern humans, as well as some articular contrasts between these two human groups, this research has noted the relatively consistent differencesin pollical phalangeal length roportions between Neandertals and bot early and recent modern humans.
K
@ 1991 WILEY-LISS. INC.
The contrast between the Neandertals and modern humans in pollical phalangeal length roportions consists of the Neandertals exflibiting relatively shorter proximal pollical phalanges and proportionately longer distal ones. The result is a summed phalangeal length relative t o the pollical metacarpal length similar to that of modern humans, despite a marked contrast in distal to proximal pollical phalangeal proportions (Fig. 1). It has been suggested (Stoner, 1981; Stoner and Trinkaus, 1981; Trinkaus, 1983, 1986, 1989a) that this contrast in pollical phalangeal length proportions altered the mechanical advantages of the pollical flexor muscles in the Neandertal hand relative to Received November 15,1989; accepted June 27,1990.
250
E. TRINKAUS AND I. VILLEMEUR
Fig. 1. Palmar (above) and radial (below)views of casts of the Shanidar 4 left pollical bones (above in each case)adjacent t o the left pollicalbones of a recent Euroamerican (UNM #OMI-67) with the same MC-1 articular length. Scale is in centimeters.
those of modern humans. More specifically, it was proposed that 1)it increased the mechanical advantages of the pollical flexors across the metacarpophalangeal (MCP) articulation in “power” grips utilizing the mid (or interphalangeal [IPI) region, and 2) it consequently decreased the mechanical advantage of the pollical flexor (M. flexor polli-
cis longus) across the IP articulation for “precision” grips utilizing the distal phalangeal pad. This interpretation was proposed on the basis of length comparisons between Neandertals, early modern humans, and recent human samples. However, it only took into consideration one of the relevant variables:
NEANDERTAL THUMB FLEXION
thumb segment lengths as assumed to be proportional to the load arms of the thumb flexor muscles. It did not take into consideration the relative lengths of the power arms involved, which would be proportional to the dorsopalmar (DP) diameters of the MCP and IP articulations. We have therefore collected and analyzed human paleontological and recent human data to test whether the differencesin Neandertal pollical phalangeal length pro ortions significantly altered the mechanica advantages of their thumbs in “power”and “precision” grips. More precisely, we tested the null hypotheses that 1) the relative shortness of Neandertal proximal pollical phalanges (relative to those of modern humans) did not alter the mechanical advantages of the flexor muscles operating across their MCP articulations with respect to grips utilizing the IP region, and 2) the relative lengthening of Neandertal distal pollical phalanges (relative to those of modern humans) did not alter the mechanical advantages of their Mm. flexor pollicis longus operating across their IP articulations with respect to grips employing the distal phalangeal palmar pad.
P
A NOTE ON NEANDERTAL HAND FUNCTION
In the following discussion, it should be kept in mind that, despite comments to the contrary (e.g., Boule, 1911-13; Sarasin, 1931; Bonch-Osmolovskij, 1941; Musgrave, 1971),there is no aspect of Neandertal hand morphology which precludes them from having had a range of manipulative postures and movements indistinguishable from those of early modern and recent humans. Their overall hand-to-arm and thumb-tomiddle finger length proportions are indistinguishable from modern human ranges of variation (Musgrave, 1970; Trinkaus, 1983), all of their evident muscular insertions are similar (even if relatively hypertrophied) to those of modern humans (Musgrave, 1971; Kimura, 1976; Trinkaus, 1983; contra Week, 1978))and all of their hand articulations, to the extent currently understood, would have allowed the ranges of movement common among modern humans (Stoner and Trinkaus, 1981; Trinkaus, 1983; Riley and Trinkaus, 1989). There are contrasts between their carpometacarpal (CMC)1and 5 articulations and those of recent humans, with the Neandertals exhibiting more condyloid surfaces and recent humans usually presenting sellar shaped surfaces; however, both configurations permit three degrees of freedom
251
and an possible limitations of movement at those MC articulations are likely to have been compensated at adjacent intercarpal and MCP articulations (Landsmeer, 1955; Kauer, 1974; Stark et al., 1977; Riemann and Ebner, 1980; Dubousset, 1981; Kapandji, 198l), given the general similarities between the intercarpal and MCP articular surfaces of Neandertals and modern humans. Furthermore, the contrasts in their capitate-to-MC-2 articular facets and relatively small MC-3 styloid processes (Riley and Trinkaus, 1989) imply only differences in habitual joint reaction force trajectories, not in ranges of movement. As a result, it is reasonable to apply models based on modern human hand functional morphology for the interpretation of the relative pollical phalangeal length proportions of these archaic hominids.
F
MATERIALS AND METHODS
This analysis is based on metric comparisons between the available European and Near Eastern late archaic human pollical metacarpals and phalanges and those of recent humans. All of the fossil individuals either belong within the Neandertal morphological group or (such as the Krapina remains, Shanidar 4 and 6 and possibly Tabun C1) represent populations immediately ancestral to their regional Neandertal populations. The sample contains all of the currently known sufficiently complete pollical phalanges and, when available, associated pollical metacarpals (MC-1s). The sample, with the measurements each swecimen Drovides (see measurement definitions and-abbreviations below and-in table 11, includes La Ferrassie 1 (MCL, MCH, PPL, PPB, PPH, DPL, DPB, PPA, DPA), La Ferrassie 2 (MCL, MCH, PPL, PPB, PPH, DPL. PPA) (Heim. 1982). Kebara 2 (MCL. MCH, PPL, PPB,’ PPH,”DPL, DPB,‘ PPA; DPA) (Arensburg et al., 1985; Vandermeersch, in press), Kiik-Koba 1 (MCL, MCH, DPL, DPB, DPA) (Bonch-Osmolovskij, 1941), Krapina 202 and 253.3 (PPL, PPB, PPH each) and KraDina 203.1 to 203.4 (DPL. DPB, DPA each) (RadoGie et al., 19881, Re: gourdou 1 (MCL, MCH, PPL, PPB, PPH, DPL, DPB, PPA, DPA) (Duday and Villemeur, in prep.), Shanidar 3 (DPL, DPB, DPA), Shanidar 4 (MCL, MCH, PPL, PPB, PPH, DPL, DPB, PPA, DPA),and Shanidar 5 and 6 (PPL, PPB, PPH, DPL, DPB, PPA, DPA each) (Trinkaus, 1983), Spy 25H (PPL, PPB, PPH) (Fraipont and Lohest, 18871,and
252
E. TRINKAUS AND I. VILLEMEUR
Tabun C1 (MCL, MCH, PPL, PPB, PPH, PPA) (McCown and Keith, 1939). It does not include the Neandertal MC-1s which are not associated with pollical phalanges (e.g., Amud 1and La Chapelle-aux-Saints 1). The recent human samples consist of preand proto-historic Amerindians from Pecos Pueblo (collections of the Harvard Peabody Museum [Stoner, 19811) (desi ated “Pecos”), prehistoric Amerindians rom central Rio Grande pueblos (collections of the Maxwell Museum, University of New Mexico) (designated “Puebloan”), medieval Yugoslavs from Mistihalj (collections of the Harvard Peabody Museum [Stoner, 19811) (designated ‘Yugoslav”), modern British (Musgrave, 1970)(designated “British),and modern Euroamericans (collections of the Maxwell Museum, University of New Mexico) (designated “Euroamerican”). All of the paleontological measurements, except those on Kiik-Koba 1, were taken by us on the original fossil remains. The KiikKoba 1 measurements were determined on casts of the hand remains in the Natural History Museum (London). The recent human samples consist of two Amerindian and three European (or European derived) samples. Only the Puebloan and Euroamerican samples provide all of the measurements employed, but the other samples provide significant proportions of the measurements and were included to broaden the modern human comparative framework. In addition to comparisons between the individual recent human samples and the Neandertals, comparisons were made between the total Amerindian and European samples and the Neandertals, given similarities within the modern regional samples and minor contrasts between them. The measurements employed here were defined so as to best characterized the skeletal morphology of interest (see below). Most of them differ in definition from the maximum dimensions usually provided for metacarpals and phalanges, and therefore they should not be considered directly comparable to the measurements in many of the original, primary descriptions of the fossils listed above with the sample specimens.
It was assumed that the primary “power” and “precision”grip positions would involve loads applied near the IP joint and the distal apical tuft respectively. It is fully recognized that many common grips use intermediate load application points and/or utilize much of the palmar surface of the thumb during gripping, and that the load application point on the distal pollical segment is not always at the distal tip of the distal phalanx (Napier, 1956; Landsmeer, 1962; Cooney and Chao, 1977; Marzke and Shackley, 1986). This assumption should not, however, distort the results greatly. As a result, the interarticular lengths (mid proximal articulation to mid distal articulation or apical tuft [Musgrave, 1970; Trinkaus, 19831) of the proximal (PP-1)and distal (DP-1) pollical phalanges (PPL and DPL, respectively, Fig. 2) were employed to characterize these load arms. The interarticular length of the MC-1 (MCL) was employed to provide a reference length for the pollical phalanges. The pollical flexor muscles of concern at the MCP articulation consist of M. flexor pollicis brevis with contributions from M. adductor pollicis and M. abductor pollicis brevis plus the tendon of M. flexor pollicis longus. The intrinsic flexors operating across the MCP joint pass close to the palmar surface of the MC-1 head, even though they insert in part through almar sesamoid bones. The tendon for M. lexor pollicis longus similarly passes close to the palmar MC-1 head, as it spans the MCP joint. It is here assumed, given the absence in the fossil record of pollical sesamoid bones, that the additional distance from the MC-1 palmar head to the mean line of action of the flexors would have been similarly proportional to the head size in all of the samples em loyed. With respect to the IP articulation, t e tendon for M. flexor pollicis longus passes within the sulcus on the head of the PP-1 to insert into the palmar surface adjacent to the base of the DP-1. As a result, it is assumed that the power arms for these pollical flexor muscles will be proportional to half of the DP diameters of the MC-1 and PP-1 heads (MCH and PPH, respectively). PPH was taken as the maximum diameter from the palmar radial and METHODS ulnar margins of the trochlea to the most The linear and angular measurements em- dorsal margin of the head; MCH was taken ployed were chosen to characterize, within from the dorsal articular surface to the palthe limits of the paleontological remains mar median crest, so as to avoid variation available, the relevant load arms and power introduced by differences in the size of the arms for flexion of the human thumb. radial-palmar process (a process which is
P
P
R
253
NEANDERTAL THUMB FLEXION
absent from most Neandertals and some recent humans) (Fig. 2). Given the relative rarity of associated thumb bones in the fossil record (complete MC-1, PP-1,and DP-1 are preserved for only five Neandertals), it is not always possible to compare the length of a pollical phalanx to the head diameter of the next proximal element. It was therefore assessed whether it would be appropriate to em loy the maximum articular height of the p alangeal base (PPB and DPB respectively for the PP-1 and DP-1) as a substitute for the adjacent MCH or PPH. Although both articular heights are generally correlated with size, there is no necessary reason for the actual measures to be highly correlated, given the compounding effects of articular surface area, radius of curvature and joint excursion on their dimensions. Least-squares regressions between MCH and PPB provided r values of 0.680 for the entire modern sample and 0.700, 0.838, 0.584, and 0.735 for the Pecos, Yugoslav, Euroamerican, and Puebloan samples, respectively. This provides confidence in using PPH as a substitute for MCH for those individuals who lack a MC-1. Least squares regression for PPH to DPB provides a slightly lower r value (0.581) for the entire modern human sample, but the r values for individual samples varied from 0.282 (Pecos), to 0.347 (Yugoslav), to 0.427 (Euroamerican), to 0.657 (Puebloan) across the modern human samples, giving less confidence in the use of the DPB as a substitute for PPH in the comparisons. The lower r values between the IP articular heights is probably due t o a combination of a lower degree of correlation in their functionally significant DP dimensions, combined with greater difficulty in repeating the DPB measurement. The normal irregularities in the palmar margin of the DP-1 proximal articulation, combined with the presence near the middle of the palmar margin of a non-articular area of variable size which projects dorsally into the surface, introduce variation of questionable functional significance into the DPB measurement. In addition, since the Neandertal sample is too small to determine the correlation between these values within that sample, a ratio of head DP diameter to adjacent proximal articular height was computed. Across the MCP articulation, most of the Neandertal to modern human sample comparisons showed insignificant differences across the
R
samples; only the Puebloan to Neandertal comparison indicated a difference at the P < .05 level (P = .037).However, across the IP articulation only the Pecos to Neandertal comparison showed general similarity, with the other comparisons having P values from .068 to .0001, in each case with DPB being relatively smaller in the recent human sample. Therefore, both com arisons were made, using both sides of eac articulation to estimate the power arms so as to maximize sample size. However, the limitations of using DP-1 base dimensions need to be kept in mind, given its low correlations with PPH and smaller relative size among most of the recent human samples. In addition, it has been noted (e.g., Musgrave, 1971; Trinkaus, 1983)that the DP-1s of the Neandertals exhibit a consistent ulnar deviation relative to their IP articulation. Even though their degree of ulnar deviation has been shown to be within modern human ranges of variation (Stoner, 1981; Table 11, their values tend to be higher on the average than those of recent humans. It was therefore decided to correct, to the extent possible, for possible contrasts in the ulnar deviation of the distal phalanx relative to the proximal one. The ulnar deviation of DP-1 was measured as the angle between the radioulnar plane of the IP articular facet and the perpendicular to the diaphyseal axis; a positive angle indicates an ulnar deviation of the apical tuft. Similarly, the ulnar or radial
R
,
Fig. 2. Radial view of a modern human thumb showing the primary measurements employed. PPL, proximal pollical phalanx (PP-1) articular length; DPL, distal pollical phalanx (DP-1) articular length; MCH, pollical metacarpal (MC-1) distal articular dorsopalmar height (not including the radial palmar rocess); PPB,, PP-1 proximal articular height; PPH, P1-! distal maximum articular height; DPH, DP-1 proximal articular height. One-half of each head height (MCH and PPH) is taken to be proportional to the power arm for the pollical flexors; the phalangeal lengths are taken to be proportional to the load arms during grips near the interphalangeal joint and at the fingertip.
254
E. TRINKAUS AND I. VILLEMEUR
deviation of the PP-1 head was measured as the angle between the plane tangent to the distal trochleae of the head and the perpendicular to the PP-1 diaphyseal axis; a positive angle indicates an ulnar deviation of the head, and a negative one indicates the more common radial deviation of the PP-1 head. Consequently, comparisons between DPL (as a load arm) and IP DP diameters (as reflecting power arms) were made using the full DPL, DPL corrected by the cosine of DP-1 ulnar deviation angle, and DPL corrected by the cosine of the summed DP-1 plus distal PP-1 ulnar deviation angles. The last, when available, should provide the most accurate reflection of the M. flexor pollicis longus mechanical advantage across the IP articulation. Consequently, two forms of comparisons were made. First, the relative lengths of pollical phalanges were compared, using ratios, to assess the proportions of thumb segment lengths between the Neandertals and recent humans. Second, the lengths of the pollical phalanges (direct and corrected for ulnar deviation) were compared to articular diameter approximations (or reflections) of the flexor muscle moment arms around the MCP and IP articulations, producing “mechanical advantage” ratios for the samples. Even though the fossil sample is too small to determine its distributions, comparisons were made using two-tailed t-tests assuming unequal variances and the probability levels indicated in the tables reflect these comparisons. In cases where right and left sides were preserved for an individual, the values were averaged to provide a mean value by individual; the N values therefore represent individuals and not bones. The isolated Krapina pollical phalanges are assumed to derive from separate individuals. Since only Neandertal to modern human comparisons were made, the probability levels reflect pairwise comparisons without a multiple sample Bonferroni correction; given that many of the relevant comparisons provide P values
Mechanical Advantages of the Neandertal Thumb in Flexion: A Test of an Hypothesis ERIK TRINKAUS AND ISABELLE VILLEMEUR Department of Anthropology, University of New Mexico, Albuquerque, New Mexico 87131 (E.T.); U A . 376 du C.N.R.S., Laboratoire d’Anthropologie, Universite de Bordeaux I, 33405 Talence, France (E.T.); Laboratoire d’Anthropologie, Universite de Bordeaux I, 33405 Talence, France (I.V.)
KEY WORDS
Human paleontology, Hand morphology, Thumb
ABSTRACT It has been proposed that the pollical phalangeal length proportions of the Neandertals provided them with a greater mechanical advantage relative to recent humans for their pollical flexor muscles in power grips across the interphalangeal (IP)joint at the expense of the mechanical advantage of those pollical flexor muscles in precision grips at the finger tip. To test these related hypotheses, we compared the pollical load arm dimensions (phalanx lengths) to power arm dimensions (dorsopalmar articular heights) for the European and Near Eastern Neandertals and for European and Amerindian samples of recent humans. It was found, initially, that the proximal articular height of the pollical distal phalanx is a poor predictor of the power arm at the IP articulation, even though the proximal articular height of the pollical proximal phalanx was an adequate indicator of the power arm size at the metacarpophalangeal (MCP) joint. In addition, differences in distal pollical ulnar deviation at the IP joint appeared to make little difference in the mechanical advantage comparisons. More importantly, the relative shortness of Neandertal proximal pollical phalanges and the relative lengthening of their distal pollical phalanges was confirmed, and it was determined that, despite some minor differences in articular dimensions between Neandertals and recent humans, these pollical phalangeal length contrasts translated into significant differences in mechanical advantages for the flexor muscles across the MCP and IP articulations. Research on the evolution of human manipulative behavior has increasinglyfocused on the preserved hand remains of archaic fossil hominids, and especially on the relatively abundant and associated hand remains of European and Near Eastern late archaic humans (Neandertals sensu lato) (e.g., Musgrave, 1970, 1971, 1973; VlEek, 1975; Kimura, 1976; Stoner, 1981; Stoner and Trinkaus, 1981; Heim, 1982; Trinkaus, 1983;Riley and Trinkaus, 1989).In addition to documenting the exaggerated muscularity and general articular robusticity of the Neandertals relative to modern humans, as well as some articular contrasts between these two human groups, this research has noted the relatively consistent differencesin pollical phalangeal length roportions between Neandertals and bot early and recent modern humans.
K
@ 1991 WILEY-LISS. INC.
The contrast between the Neandertals and modern humans in pollical phalangeal length roportions consists of the Neandertals exflibiting relatively shorter proximal pollical phalanges and proportionately longer distal ones. The result is a summed phalangeal length relative t o the pollical metacarpal length similar to that of modern humans, despite a marked contrast in distal to proximal pollical phalangeal proportions (Fig. 1). It has been suggested (Stoner, 1981; Stoner and Trinkaus, 1981; Trinkaus, 1983, 1986, 1989a) that this contrast in pollical phalangeal length proportions altered the mechanical advantages of the pollical flexor muscles in the Neandertal hand relative to Received November 15,1989; accepted June 27,1990.
250
E. TRINKAUS AND I. VILLEMEUR
Fig. 1. Palmar (above) and radial (below)views of casts of the Shanidar 4 left pollical bones (above in each case)adjacent t o the left pollicalbones of a recent Euroamerican (UNM #OMI-67) with the same MC-1 articular length. Scale is in centimeters.
those of modern humans. More specifically, it was proposed that 1)it increased the mechanical advantages of the pollical flexors across the metacarpophalangeal (MCP) articulation in “power” grips utilizing the mid (or interphalangeal [IPI) region, and 2) it consequently decreased the mechanical advantage of the pollical flexor (M. flexor polli-
cis longus) across the IP articulation for “precision” grips utilizing the distal phalangeal pad. This interpretation was proposed on the basis of length comparisons between Neandertals, early modern humans, and recent human samples. However, it only took into consideration one of the relevant variables:
NEANDERTAL THUMB FLEXION
thumb segment lengths as assumed to be proportional to the load arms of the thumb flexor muscles. It did not take into consideration the relative lengths of the power arms involved, which would be proportional to the dorsopalmar (DP) diameters of the MCP and IP articulations. We have therefore collected and analyzed human paleontological and recent human data to test whether the differencesin Neandertal pollical phalangeal length pro ortions significantly altered the mechanica advantages of their thumbs in “power”and “precision” grips. More precisely, we tested the null hypotheses that 1) the relative shortness of Neandertal proximal pollical phalanges (relative to those of modern humans) did not alter the mechanical advantages of the flexor muscles operating across their MCP articulations with respect to grips utilizing the IP region, and 2) the relative lengthening of Neandertal distal pollical phalanges (relative to those of modern humans) did not alter the mechanical advantages of their Mm. flexor pollicis longus operating across their IP articulations with respect to grips employing the distal phalangeal palmar pad.
P
A NOTE ON NEANDERTAL HAND FUNCTION
In the following discussion, it should be kept in mind that, despite comments to the contrary (e.g., Boule, 1911-13; Sarasin, 1931; Bonch-Osmolovskij, 1941; Musgrave, 1971),there is no aspect of Neandertal hand morphology which precludes them from having had a range of manipulative postures and movements indistinguishable from those of early modern and recent humans. Their overall hand-to-arm and thumb-tomiddle finger length proportions are indistinguishable from modern human ranges of variation (Musgrave, 1970; Trinkaus, 1983), all of their evident muscular insertions are similar (even if relatively hypertrophied) to those of modern humans (Musgrave, 1971; Kimura, 1976; Trinkaus, 1983; contra Week, 1978))and all of their hand articulations, to the extent currently understood, would have allowed the ranges of movement common among modern humans (Stoner and Trinkaus, 1981; Trinkaus, 1983; Riley and Trinkaus, 1989). There are contrasts between their carpometacarpal (CMC)1and 5 articulations and those of recent humans, with the Neandertals exhibiting more condyloid surfaces and recent humans usually presenting sellar shaped surfaces; however, both configurations permit three degrees of freedom
251
and an possible limitations of movement at those MC articulations are likely to have been compensated at adjacent intercarpal and MCP articulations (Landsmeer, 1955; Kauer, 1974; Stark et al., 1977; Riemann and Ebner, 1980; Dubousset, 1981; Kapandji, 198l), given the general similarities between the intercarpal and MCP articular surfaces of Neandertals and modern humans. Furthermore, the contrasts in their capitate-to-MC-2 articular facets and relatively small MC-3 styloid processes (Riley and Trinkaus, 1989) imply only differences in habitual joint reaction force trajectories, not in ranges of movement. As a result, it is reasonable to apply models based on modern human hand functional morphology for the interpretation of the relative pollical phalangeal length proportions of these archaic hominids.
F
MATERIALS AND METHODS
This analysis is based on metric comparisons between the available European and Near Eastern late archaic human pollical metacarpals and phalanges and those of recent humans. All of the fossil individuals either belong within the Neandertal morphological group or (such as the Krapina remains, Shanidar 4 and 6 and possibly Tabun C1) represent populations immediately ancestral to their regional Neandertal populations. The sample contains all of the currently known sufficiently complete pollical phalanges and, when available, associated pollical metacarpals (MC-1s). The sample, with the measurements each swecimen Drovides (see measurement definitions and-abbreviations below and-in table 11, includes La Ferrassie 1 (MCL, MCH, PPL, PPB, PPH, DPL, DPB, PPA, DPA), La Ferrassie 2 (MCL, MCH, PPL, PPB, PPH, DPL. PPA) (Heim. 1982). Kebara 2 (MCL. MCH, PPL, PPB,’ PPH,”DPL, DPB,‘ PPA; DPA) (Arensburg et al., 1985; Vandermeersch, in press), Kiik-Koba 1 (MCL, MCH, DPL, DPB, DPA) (Bonch-Osmolovskij, 1941), Krapina 202 and 253.3 (PPL, PPB, PPH each) and KraDina 203.1 to 203.4 (DPL. DPB, DPA each) (RadoGie et al., 19881, Re: gourdou 1 (MCL, MCH, PPL, PPB, PPH, DPL, DPB, PPA, DPA) (Duday and Villemeur, in prep.), Shanidar 3 (DPL, DPB, DPA), Shanidar 4 (MCL, MCH, PPL, PPB, PPH, DPL, DPB, PPA, DPA),and Shanidar 5 and 6 (PPL, PPB, PPH, DPL, DPB, PPA, DPA each) (Trinkaus, 1983), Spy 25H (PPL, PPB, PPH) (Fraipont and Lohest, 18871,and
252
E. TRINKAUS AND I. VILLEMEUR
Tabun C1 (MCL, MCH, PPL, PPB, PPH, PPA) (McCown and Keith, 1939). It does not include the Neandertal MC-1s which are not associated with pollical phalanges (e.g., Amud 1and La Chapelle-aux-Saints 1). The recent human samples consist of preand proto-historic Amerindians from Pecos Pueblo (collections of the Harvard Peabody Museum [Stoner, 19811) (desi ated “Pecos”), prehistoric Amerindians rom central Rio Grande pueblos (collections of the Maxwell Museum, University of New Mexico) (designated “Puebloan”), medieval Yugoslavs from Mistihalj (collections of the Harvard Peabody Museum [Stoner, 19811) (designated ‘Yugoslav”), modern British (Musgrave, 1970)(designated “British),and modern Euroamericans (collections of the Maxwell Museum, University of New Mexico) (designated “Euroamerican”). All of the paleontological measurements, except those on Kiik-Koba 1, were taken by us on the original fossil remains. The KiikKoba 1 measurements were determined on casts of the hand remains in the Natural History Museum (London). The recent human samples consist of two Amerindian and three European (or European derived) samples. Only the Puebloan and Euroamerican samples provide all of the measurements employed, but the other samples provide significant proportions of the measurements and were included to broaden the modern human comparative framework. In addition to comparisons between the individual recent human samples and the Neandertals, comparisons were made between the total Amerindian and European samples and the Neandertals, given similarities within the modern regional samples and minor contrasts between them. The measurements employed here were defined so as to best characterized the skeletal morphology of interest (see below). Most of them differ in definition from the maximum dimensions usually provided for metacarpals and phalanges, and therefore they should not be considered directly comparable to the measurements in many of the original, primary descriptions of the fossils listed above with the sample specimens.
It was assumed that the primary “power” and “precision”grip positions would involve loads applied near the IP joint and the distal apical tuft respectively. It is fully recognized that many common grips use intermediate load application points and/or utilize much of the palmar surface of the thumb during gripping, and that the load application point on the distal pollical segment is not always at the distal tip of the distal phalanx (Napier, 1956; Landsmeer, 1962; Cooney and Chao, 1977; Marzke and Shackley, 1986). This assumption should not, however, distort the results greatly. As a result, the interarticular lengths (mid proximal articulation to mid distal articulation or apical tuft [Musgrave, 1970; Trinkaus, 19831) of the proximal (PP-1)and distal (DP-1) pollical phalanges (PPL and DPL, respectively, Fig. 2) were employed to characterize these load arms. The interarticular length of the MC-1 (MCL) was employed to provide a reference length for the pollical phalanges. The pollical flexor muscles of concern at the MCP articulation consist of M. flexor pollicis brevis with contributions from M. adductor pollicis and M. abductor pollicis brevis plus the tendon of M. flexor pollicis longus. The intrinsic flexors operating across the MCP joint pass close to the palmar surface of the MC-1 head, even though they insert in part through almar sesamoid bones. The tendon for M. lexor pollicis longus similarly passes close to the palmar MC-1 head, as it spans the MCP joint. It is here assumed, given the absence in the fossil record of pollical sesamoid bones, that the additional distance from the MC-1 palmar head to the mean line of action of the flexors would have been similarly proportional to the head size in all of the samples em loyed. With respect to the IP articulation, t e tendon for M. flexor pollicis longus passes within the sulcus on the head of the PP-1 to insert into the palmar surface adjacent to the base of the DP-1. As a result, it is assumed that the power arms for these pollical flexor muscles will be proportional to half of the DP diameters of the MC-1 and PP-1 heads (MCH and PPH, respectively). PPH was taken as the maximum diameter from the palmar radial and METHODS ulnar margins of the trochlea to the most The linear and angular measurements em- dorsal margin of the head; MCH was taken ployed were chosen to characterize, within from the dorsal articular surface to the palthe limits of the paleontological remains mar median crest, so as to avoid variation available, the relevant load arms and power introduced by differences in the size of the arms for flexion of the human thumb. radial-palmar process (a process which is
P
P
R
253
NEANDERTAL THUMB FLEXION
absent from most Neandertals and some recent humans) (Fig. 2). Given the relative rarity of associated thumb bones in the fossil record (complete MC-1, PP-1,and DP-1 are preserved for only five Neandertals), it is not always possible to compare the length of a pollical phalanx to the head diameter of the next proximal element. It was therefore assessed whether it would be appropriate to em loy the maximum articular height of the p alangeal base (PPB and DPB respectively for the PP-1 and DP-1) as a substitute for the adjacent MCH or PPH. Although both articular heights are generally correlated with size, there is no necessary reason for the actual measures to be highly correlated, given the compounding effects of articular surface area, radius of curvature and joint excursion on their dimensions. Least-squares regressions between MCH and PPB provided r values of 0.680 for the entire modern sample and 0.700, 0.838, 0.584, and 0.735 for the Pecos, Yugoslav, Euroamerican, and Puebloan samples, respectively. This provides confidence in using PPH as a substitute for MCH for those individuals who lack a MC-1. Least squares regression for PPH to DPB provides a slightly lower r value (0.581) for the entire modern human sample, but the r values for individual samples varied from 0.282 (Pecos), to 0.347 (Yugoslav), to 0.427 (Euroamerican), to 0.657 (Puebloan) across the modern human samples, giving less confidence in the use of the DPB as a substitute for PPH in the comparisons. The lower r values between the IP articular heights is probably due t o a combination of a lower degree of correlation in their functionally significant DP dimensions, combined with greater difficulty in repeating the DPB measurement. The normal irregularities in the palmar margin of the DP-1 proximal articulation, combined with the presence near the middle of the palmar margin of a non-articular area of variable size which projects dorsally into the surface, introduce variation of questionable functional significance into the DPB measurement. In addition, since the Neandertal sample is too small to determine the correlation between these values within that sample, a ratio of head DP diameter to adjacent proximal articular height was computed. Across the MCP articulation, most of the Neandertal to modern human sample comparisons showed insignificant differences across the
R
samples; only the Puebloan to Neandertal comparison indicated a difference at the P < .05 level (P = .037).However, across the IP articulation only the Pecos to Neandertal comparison showed general similarity, with the other comparisons having P values from .068 to .0001, in each case with DPB being relatively smaller in the recent human sample. Therefore, both com arisons were made, using both sides of eac articulation to estimate the power arms so as to maximize sample size. However, the limitations of using DP-1 base dimensions need to be kept in mind, given its low correlations with PPH and smaller relative size among most of the recent human samples. In addition, it has been noted (e.g., Musgrave, 1971; Trinkaus, 1983)that the DP-1s of the Neandertals exhibit a consistent ulnar deviation relative to their IP articulation. Even though their degree of ulnar deviation has been shown to be within modern human ranges of variation (Stoner, 1981; Table 11, their values tend to be higher on the average than those of recent humans. It was therefore decided to correct, to the extent possible, for possible contrasts in the ulnar deviation of the distal phalanx relative to the proximal one. The ulnar deviation of DP-1 was measured as the angle between the radioulnar plane of the IP articular facet and the perpendicular to the diaphyseal axis; a positive angle indicates an ulnar deviation of the apical tuft. Similarly, the ulnar or radial
R
,
Fig. 2. Radial view of a modern human thumb showing the primary measurements employed. PPL, proximal pollical phalanx (PP-1) articular length; DPL, distal pollical phalanx (DP-1) articular length; MCH, pollical metacarpal (MC-1) distal articular dorsopalmar height (not including the radial palmar rocess); PPB,, PP-1 proximal articular height; PPH, P1-! distal maximum articular height; DPH, DP-1 proximal articular height. One-half of each head height (MCH and PPH) is taken to be proportional to the power arm for the pollical flexors; the phalangeal lengths are taken to be proportional to the load arms during grips near the interphalangeal joint and at the fingertip.
254
E. TRINKAUS AND I. VILLEMEUR
deviation of the PP-1 head was measured as the angle between the plane tangent to the distal trochleae of the head and the perpendicular to the PP-1 diaphyseal axis; a positive angle indicates an ulnar deviation of the head, and a negative one indicates the more common radial deviation of the PP-1 head. Consequently, comparisons between DPL (as a load arm) and IP DP diameters (as reflecting power arms) were made using the full DPL, DPL corrected by the cosine of DP-1 ulnar deviation angle, and DPL corrected by the cosine of the summed DP-1 plus distal PP-1 ulnar deviation angles. The last, when available, should provide the most accurate reflection of the M. flexor pollicis longus mechanical advantage across the IP articulation. Consequently, two forms of comparisons were made. First, the relative lengths of pollical phalanges were compared, using ratios, to assess the proportions of thumb segment lengths between the Neandertals and recent humans. Second, the lengths of the pollical phalanges (direct and corrected for ulnar deviation) were compared to articular diameter approximations (or reflections) of the flexor muscle moment arms around the MCP and IP articulations, producing “mechanical advantage” ratios for the samples. Even though the fossil sample is too small to determine its distributions, comparisons were made using two-tailed t-tests assuming unequal variances and the probability levels indicated in the tables reflect these comparisons. In cases where right and left sides were preserved for an individual, the values were averaged to provide a mean value by individual; the N values therefore represent individuals and not bones. The isolated Krapina pollical phalanges are assumed to derive from separate individuals. Since only Neandertal to modern human comparisons were made, the probability levels reflect pairwise comparisons without a multiple sample Bonferroni correction; given that many of the relevant comparisons provide P values
