The relative lengths and weights of metacarpals and metatarsals in baboons (Papio hamadryas)

Transcription

1 Available online at R Hormones and Behavior 43 (2003) The relative lengths and weights of metacarpals and metatarsals in baboons (Papio hamadryas) Dennis McFadden* and Mary S. Bracht Department of Psychology and Institute for Neuroscience, University of Texas, Austin, TX, USA Received 9 November 2001; revised 29 April 2002; accepted 10 July 2002 Abstract The lengths and weights of metacarpals and metatarsals were measured in two collections of baboon skeletons 33 animals, all born and raised in the wild, and 60 animals, nearly all born and raised in captivity. For both length and weight, ratios were constructed for all possible pairings of the five bones in each individual hand and foot. The pattern of results was not identical for the two collections, but substantial sex differences existed in both collections for many of the ratios calculated. Nearly all of the large sex differences were in the direction of the length or weight ratio being smaller for males than for females. For the length ratios showing the largest sex differences, those differences were larger for the right hand than for the left, and larger for the left foot than for the right, but this pattern was less evident for the weight ratios. Some length ratios for human fingers show sex differences of the sort seen in the metacarpals and metatarsals of these baboons, and the human differences in relative length exist early in life. The implication is that the marked sex difference in androgen exposure during prenatal development may effect the relative size of the individual bones of the extremities in male and female humans and baboons. The relative sizes of the bones of the hand and foot may provide helpful supplemental information on the relations between species, thus making the study of hand and foot bones in other primates and mammals worthwhile Elsevier Science (USA). All rights reserved. Keywords: Sex differences; Baboons; Metacarpals; Metatarsals; 2D:4D ratio In humans, the relative lengths of the fingers on the same hand are different in males and females (George, 1930; Phelps, 1952). Specifically, the relative length of the index and ring fingers (the 2D:4D ratio) is close to 1.0 in females, but is smaller than 1.0 in males. This sex difference exists in both hands, but there is some evidence that the 2D:4D ratio is more asymmetric in females than in males (Williams et al., 2000). The 2D:4D ratio has been shown to be relevant to such issues as homosexuality (Brown, et al., 2002; Mc- Fadden and Shubel, 2002a; Robinson and Manning, 2000; Williams et al., 2000), depression (Martin et al., 1999), breast cancer (Manning and Bundred, 2000), heart disease (Manning and Bundred, 2000), autism (Manning et al., 2001), congenital adrenal hyperplasia (Brown et al., 2001b), * Corresponding author. Department of Psychology, Seay Building A8000, University of Texas, 1 University Station, Austin, TX , USA. Fax: address: mcfadden@psy.utexas.edu (D. McFadden). attention-deficit hyperactivity disorder (McKay, 2001), and assertiveness in women (Wilson, 1983). The sex difference in the 2D:4D ratio exists at least as early as 2 years of age (Manning et al., 1998). The strong implication is that relative finger length is a marker for degree of androgen exposure during early development, probably during prenatal development. The fact that finger length is controlled by certain homeobox genes (Kondo et al., 1997) makes these facts of interest to scientists studying a number of different topics in the general areas of human development, hormonal effects, and sexual differentiation. Recent research on the 2D:4D ratio is summarized in Manning (2002). In an attempt to extend the study of relative size of the bones in the extremities to other primates, measurements were made on two collections of baboon (Papio hamadryas) skeletons. Although the results differed somewhat in the two collections, taken together the data revealed that both the metacarpals and the metatarsals of baboons exhibit sex differences similar to those exhibited by humans. Following X/03/$ see front matter 2003 Elsevier Science (USA). All rights reserved. doi: /s x(02)00048-x

2 348 D. McFadden, M.S. Bracht / Hormones and Behavior 43 (2003) the collection of these data, Tague (2002) reported a sex difference in the relative lengths of one pair of metacarpals in baboons. In one other previous attempt to study relative digit length in nonhumans, a sex difference was found in the 2D:4D ratio in the rear paws of mice, but only on the right side (Brown et al., 2001a). Methods Materials The first set of skeletons measured was collected by C.A. Bramblett over a period of 2 years beginning in 1963 in the vicinity of the Darajani Primate Research Station near Kibwezi, Kenya, East Africa (see Bramblett, 1967). At the time of collection, these animals were characterized as Papio cynocephalus (Bramblett, 1967), but under current nomenclature (Jolly, 2002), they would be characterized as Papio hamadryas anubis. All had the stocky build currently used as a characteristic of the Anubis subspecies. Their coloration was an olive-yellow mix. Given the small geographic region over which these skeletons were collected, it seems appropriate to presume that genetically they were reasonably homogeneous. Measurements were made on 19 male and 18 female skeletons, but not all skeletons were complete. Three partial specimens and one specimen whose sex was uncertain were discarded entirely, leaving 33 included specimens. Additional partial exclusions are described in the procedure section below. The second set of skeletons was also collected by Dr. Bramblett. These were animals used for research on atherosclerosis and other topics by scientists at the Southwest Foundation for Biomedical Research, San Antonio, Texas (e.g., Glassman et al., 1984). The majority of these animals were born and bred in captivity and were 7.5 years of age at the time of death. Their mothers were wild-born and olive in color; their fathers were wild-born and yellow in color. The sites of origin of these parent animals were largely unknown. Accordingly, it is highly likely that this second collection of skeletons was far more diverse genetically than was the first collection. Most of the mothers were probably Papio anubis, while the yellow fathers were likely to have been Papio cynocephalus. The ages of these wild-born mothers and fathers were unknown. The bones of three of the mothers and 11 of the fathers were measured along with the bones of the 46 hybrid offspring. Because the data from the wild parents and hybrid offspring were not noticeably different, they were pooled for this report. Initially, measurements were made on 62 skeletons, but two partial specimens were discarded entirely, leaving the 60 included specimens. Additional partial exclusions are described in the procedure section below. For both collections, the preparation of the skeletons involved warming all the bones in water and leaving them in the resulting fluid for differing periods of time for the different skeletons. After cleaning, all the bones of each individual animal were stored together in separate containers and kept in the Department of Anthropology at the University of Texas, Austin. For both sets of skeletons, sex was determined by a combination of field notes and judgments about the size of the bones. When there was any uncertainty about sex, that specimen was omitted from the data analyses. Procedure Because the individual fingers and toes could not be reconstructed with certainty from the collection of phalanges in each container, measurements were made on the metacarpals and metatarsals (together called the metapodials), which were unambiguously identifiable. Bone length was measured by using an osteometric board having orthogonal grid lines with 1.0-mm spacing. The distal (rounded) end of the bone was placed against the top edge of the osteometric board, and the shaft of the bone was aligned with the grid lines running parallel with it. A length of wood was placed against the proximal end of the bone, and the length was read from the grid lines to the nearest 0.5 mm. Two measurements were made on each bone by two experimenters working together, and the average was used as the estimate of length. The interrater correlations were all greater than Bone weight was measured using a scale (Ohaus model SC2020) that was accurate to hundredths of a gram. The bone weights measured surely were smaller than they would have been immediately after death. All of these bones had been in storage for at least 15 years and up to nearly 40 years, so presumably the drying process was long since complete, and presumably it had been about equal for all corresponding bones. Occasionally bones were encountered that had obviously undergone some loss of material. Those bones were not weighed, but their lengths were recorded as long as no flaking was obvious at either end of the bone. The primary interest here was not the actual lengths and weights of the bones; the substantial sexual dimorphism of baboons is well established (e.g., Glassman et al., 1984). Rather, the interest was a possible dimorphism in the relative sizes of bones of the hand and of the foot. Accordingly, ratios (for length and for weight) were calculated for all possible pairs of bones of a given extremity. When all bones were present for an extremity, this produced 10 ratios for that extremity (for both length and weight) for that specimen. If individual bones were missing for a skeleton, length and weight were measured for the bones available, and subsequently, all the ratios that could be calculated were. Ultimately, however, when a specimen was missing more than one bone in an extremity, both hands (or both feet) were excluded entirely from all data analyses. Further, when a specimen was missing just one bone in an extremity (and thus missing three or four of the ten possible ratios for that extremity), the corresponding bone in the opposite extrem-

3 D. McFadden, M.S. Bracht / Hormones and Behavior 43 (2003) ity was also excluded. With this procedure, every included specimen contributed only to those average ratios to which it could contribute bilaterally. The same rules for inclusion and exclusion were applied independently for the hands and feet of each specimen. (As it turned out, the basic pattern of results was the same with and without these various exclusions.) Here, designations of the sort 2M:4M will be used to indicate the ratios of lengths or weights; in this example, the ratio of the 2nd to the 4th metacarpal (or metatarsal) of the same hand (foot). The emphasis here will be on descriptive statistics rather than statistical tests. Of primary interest will be effect sizes (Cohen, 1992) calculated to estimate the magnitude of the sex difference for various measures. These effect sizes were calculated by dividing the difference between the female and male means for the ratio of interest by the square root of the weighted mean of the variances for those two groups. (Thus, effect size provides an estimate of the degree of overlap between the two distributions for the ratio of interest.) For comparisons of this sort, Cohen (1992) has suggested that effect sizes of 0.2, 0.5, and 0.8 can be interpreted as small, medium, and large, respectively. In addition, whenever a comparison produced an effect size of about 0.4 or larger, a t test was calculated. The large number of t tests performed meant that the standard criterion for statistical significance was likely to be exceeded numerous times by chance. Accordingly, the results of those statistical tests should be interpreted as additional descriptive information only. Results Lengths and weights As expected, the male bones were generally longer than the female bones. The effect sizes for the sex difference in length ranged from about 3.0 to 3.7 for hands and from about 4.1 to 4.9 for feet. More interesting than this sexual dimorphism is that hands and feet exhibited different patterns of length for the 2nd to 5th metapodials. (For both hands and feet, the 1st metapodial was substantially shorter and lighter than all others.) For hands, the 2nd metacarpal was the longest for both sexes. Its length was about 46.5 mm for the left hand and about 46.4 mm for the right hand in females, with the lengths for males being about 9 mm longer than those for the females. Then, for females, average bone length gradually decreased from the 2nd to the 5th metacarpal, but for males, length was about the same for the 3rd, 4th, and 5th metacarpals. For feet, the 2nd metatarsal was the shortest (except for the 1st). Its length was about 54.0 mm for both feet in females, with the lengths for males being about 10 mm longer than those for the females. Then, for females, average bone length gradually increased by about 2.6 mm to the 4th metatarsal, before declining slightly, whereas for males, length increased monotonically by about 3.5 mm to the 5th metatarsal. These patterns were not identical for the weight data, in that (except for the 1st metapodials) the 3rd metapodials were the heaviest bones on both the left and right sides of the body. The weights were about 1.56 g for the left 3rd metacarpal and about 1.59 g for the right 3rd metacarpal in females, with the weights for males being about 1.4 g heavier than those for the females. The weight of the 3rd metatarsal was about 2.25 g for both feet in the females, with the weights for males being about 1.9 g heavier than for females. These substantial sex differences in metapodial weight produced effect sizes for the sex difference that ranged from about 3.0 to 3.8 for hands and from about 3.8 to 4.8 for feet. The correlations between length and weight ranged between 0.90 and 0.94 for the individual metacarpals (males and females combined), and between 0.89 and 0.95 for the individual metatarsals, with no obvious differences between the left and right hand or the left and right foot. The correlations between the lengths of the individual metacarpals of one hand with the corresponding metatarsals on the same side of the body were all between 0.94 and 0.97 (males and females combined), and the parallel hand/foot correlations for weight were all between 0.94 and Length ratios Although large, the above sex differences in bone length and weight were not of primary interest here; rather, the interest was in relative sizes of the bones of the hands and feet. The various ratios of bone length for pairs of ipsilateral metacarpals and pairs of ipsilateral metatarsals are shown in Fig. 1 (note the different ordinate values for the ratios involving the first metacarpal and for all the other ratios). As can be seen in the top panel of Fig. 1, the average ratios for females were generally smaller than those for males for the length ratios involving the first metacarpal (hereafter called the lower-order ratios), and generally larger than those for males for the ratios not involving the first metacarpal (hereafter called the higher-order ratios). Visual inspection reveals that the 2M:4M ratio did not exhibit the largest difference between the sexes (as does the 2D:4D ratio for human finger length see McFadden and Shubel, 2002a); the sex differences were noticeably larger for several of the other higher-order ratios. For the metatarsals (bottom panel of Fig. 1), the female length ratios generally were larger than those for the males, but there were some notable exceptions. The largest differences between the sexes were for the 4M:5M ratio for both feet. Note that, for the lower-order ratios, the values of the length ratios for the metatarsals were generally larger than the corresponding length ratio for the metacarpals (leftmost two panels), whereas for the higher-order ratios, the length ratios for the metatarsals were generally smaller than those for the metacarpals (rightmost two panels). The two right-

4 350 D. McFadden, M.S. Bracht / Hormones and Behavior 43 (2003) Fig. 1. Ratios of the lengths of various ipsilateral pairs of metacarpals (top two panels) and metatarsals (bottom two panels) in male and female baboons. Data were combined across two collections of animals and are shown separately for left and right hands and feet. The ordinate values shown on the left are appropriate for the four ratios involving the first metacarpal and first metatarsal; for all other ratios, the right-hand ordinate provides the correct values. Asterisks appear above pairs of bars for which the effect size for the sex difference was 0.5 or larger (see Table 1). hand panels in Fig. 1 also reveal that the 2nd to 5th metacarpals of male baboons generally were more similar in length (the ratios were closer to 1.0) than were the 2nd to 5th metacarpals of female baboons, but the 2nd to 5th metatarsals of males generally were more different in length (ratios farther from 1.0) than were the 2nd to 5th metatarsals of females. Two three-factor analyses of variance were calculated for the length ratios. The factors were sex (2 levels), side of body (2 levels), and ratio (10 levels), with repeated measures on the latter two factors. For metacarpals, the main effect for sex was not statistically significant [F(1,79) 1.63, P 0.21], the main effect for hand was not statistically significant [F(1,79) 0.35, P 0.56], the main effect for ratio was statistically significant [F(9,711) 10,462.1, P ], and the interaction between sex and ratio was statistically significant [F(9,711) 7.87, P ]. For metatarsals, the main effect for sex was statistically significant [F(1,76) 10.0, P 0.002], the main effect for foot was not statistically significant [F(1,76) 1.89, P 0.17], the main effect for ratio was statistically significant [F(9,684) 5,804.2, P ], and the interaction between sex and ratio was statistically significant [F(9,684) 4.43, P 0.01]. As a way of gaining perspective on the various differences between the sexes illustrated in Fig. 1, effect sizes were calculated for each pair of adjacent bars in those figures. For example, the difference between the mean for the 4M:5M metacarpal length ratio for the right hand of females (1.016 in Fig. 1) and the corresponding mean for the males (0.999) was divided by the square root of the weighted mean of the variances for those two samples, yielding an effect size of The latter value is shown half way down the rightmost column in Table 1. Effect sizes are shown separately for metacarpals and metatarsals (top and bottom halves of Table 1, respectively), for the left and right sides of the body, and for the two collections of skeletons as well as for the combined data that were shown in Fig. 1. A negative effect size indicates that the average ratio for the females was smaller than that for the males. Also, asterisks indicate which comparisons achieved which level of statistical significance, but, as noted previously, these outcomes must be interpreted only as additional descriptive information because of the large number of statis-

5 D. McFadden, M.S. Bracht / Hormones and Behavior 43 (2003) Table 1 Effect sizes for the sex differences in various ratios of length of the metacarpals and metatarsals in two collections of baboon skeletons a Ratio Anubis collection Mixed collection Combined collection Number Effect size (f-m) Number Effect size (f-m) Number Effect size (f-m) Left Right Left Right Left Right Hands 1M:2M 13/ / / b M:3M 13/ / / M:4M 13/ / / M:5M 12/ / / M:3M 15/ / / M:4M 15/ b 26/ / M:5M 14/ b 25/ b b 39/ c c 3M:4M 15/ b 26/ b / b b 3M:5M 14/ b 25/ d d 39/ d d 4M:5M 14/ / c d 39/ c d Feet 1M:2M 14/ / / M:3M 14/ / / M:4M 14/ / / M:5M 14/ / b / c M:3M 16/ b b 24/ / b M:4M 16/ / / M:5M 16/ / d c 42/ d c 3M:4M 16/ / / b M:5M 16/ / d c 40/ b M:5M 16/ / d d 42/ d d a f designates female; m designates male; M designates metacarpal or metatarsal; negative sign ( ) indicates male ratio larger than female ratio. Results from unpaired t tests: b 0.05 P 0.01; c 0.01 P 0.001; d P. tical tests performed. Table 1 reveals that the (presumably genetically more homogeneous) Anubis collection exhibited a somewhat different pattern of results from the presumably more heterogeneous mixed collection. At this time, there is no way to know the correct explanation for this difference in patterns. Accordingly, this discussion will concentrate on the results for the two collections combined. [Note that Tague (2002) compared metapodial lengths for wild and captive animals of 17 primate species and found no significant differences.] The two rightmost columns in Table 1 contain a substantial number of medium and large effect sizes for sex difference [defined by Cohen (1992) as values of about 0.5 and 0.8, respectively]. The largest effect sizes were obtained for the higher-order ratios (those not involving the first metacarpal or first metatarsal), and for those largest effect sizes, the sex differences in length were generally larger for the right hand than for the left, and larger for the left foot than for the right. For the lower-order ratios, the effect sizes were mostly larger for the left hand and left foot than for the right hand and right foot. The sex difference in the higher-order ratios for human finger length also is larger in the right hand than the left (study 2 of Manning et al., 1998; McFadden and Shubel, 2002a; Williams et al., 2000; but compare George, 1930). In the only study known to us on human toe length, the effect sizes for sex difference were all small, with no apparent lateral asymmetry (McFadden and Shubel, 2002a). Because the 2D:4D ratio for finger length is the measure typically studied in humans, the 2M:4M ratios for baboons deserve special attention. For the combined collection of baboon hands at the far right in Table 1, the 2M:4M ratio exhibited only a small sex difference in both hands; much larger sex differences existed for the length ratios 3M:5M and 4M:5M. (Note, however, that the sex differences for the 2M:4M ratio were medium to large for both hands in the Anubis collection, suggesting that this ratio might have been especially affected by whatever genetic or developmental differences may have existed between the two skeleton collections.) Tague (2002) reported that only the metacarpal ratio 3M:4M exhibited a significant sex difference in length. Table 1 also suggests the absence of a sex difference in the 2M:4M ratio for metatarsals. Weight ratios Ratios of the weights of all 10 possible pairs of metacarpal bones and all 10 possible pairs of metatarsal bones were calculated in a manner exactly parallel to that used for the length ratios. Those weight ratios are shown in Fig. 2. For both metacarpals and metatarsals (top and bottom panels of Fig. 2, respectively), the vast majority of the weight ratios were smaller for the males than for the females. Note that the weight ratios covered a larger range than any of the sets of length ratios shown in Fig. 1, indicating that individual metatarsals were generally rather different in weight.

6 352 D. McFadden, M.S. Bracht / Hormones and Behavior 43 (2003) Fig. 2. Ratios of the weights of various ipsilateral pairs of metacarpals (top two panels) and metatarsals (bottom two panels) in male and female baboons. Data were combined across two collections of animals and are shown separately for left and right hands and feet. The ordinate values shown on the left are appropriate for the four ratios involving the first metacarpal and first metatarsal; for all other ratios, the right-hand ordinate provides the correct values. Note the substantial difference in ordinate units between Fig. 1 and 2. Asterisks appear above pairs of bars for which the effect size for the sex difference was 0.5 or larger (see Table 2). Examination of the ordinate values in the two rightmost panels of Fig. 2 reveals that the male ratios for metacarpal weight were generally closer to 1.0 than those of the females (top panel), but that the male ratios for metatarsal weight were sometimes closer to, and sometimes farther from, 1.0 than those of the females (bottom panel). The correlations between the length and weight ratios for each of the metacarpal ratios ranged from 0.27 to 0.64 (males and females combined), with the correlations being generally similar for the left and right hands. The correlations between the length and weight ratios for each of the metatarsal ratios ranged from 0.00 to 0.56 (males and females combined), and again, the correlations were generally similar for the left and right feet. (Note that these correlations between length and weight ratios were much lower than the correlations between the physical measures of bone length and weight reported above.) Two three-factor analyses of variance were calculated for the weight ratios. The factors were sex (2 levels), side of body (2 levels), and ratio (10 levels), with repeated measures on the latter two factors. For metacarpals, the main effect for sex was statistically significant [F(1,79) 12.25, P ], the main effect for hand was not statistically significant [F(1,79) 0.81, P 0.37], the main effect for ratio was statistically significant [F(9,711) 3,661.4, P ], the interaction between sex and hand was statistically significant [F(1,79) 4.75, P 0.03], and the interaction between sex and ratio was statistically significant [F(9,711) 12.03, P ]. For metatarsals, the main effect for sex was statistically significant [F(1,76) 36.26, P ], the main effect for foot was statistically significant [F(1,76) 10.15, P 0.002], the main effect for ratio was statistically significant [F(9,684) 1,180.7, P ], and the interaction between sex and ratio was statistically significant [F(9,684) 10.68, P ]. Table 2 contains the effect sizes for the sex difference for all of the ipsilateral weight ratios calculated for both hands and both feet. Just as in Table 1, effect sizes are shown separately for the two collections of skeletons as well as for the combined data that were presented in Fig. 2. As was true for the length ratios (Table 1), there were differences in the patterns of the data for the two collections of skeletons, but, in ignorance of the origins of these differences, we will concentrate on the results for the two collections combined.

7 D. McFadden, M.S. Bracht / Hormones and Behavior 43 (2003) The two rightmost columns of Table 2 contain a considerable number of medium to rather large effect sizes for sex difference in the weight ratios. As was true for the length ratios (Table 1), the largest effect sizes were shown by the ratios not involving the 1st metacarpal or 1st metatarsal, but the trend toward those effect sizes being larger for the right hand and left foot was not as evident for weight ratios as for length ratios. Again, the ratios showing the largest sex difference were other than the 2M:4M ratio whose counterpart (2D:4D) is so important in human research (e.g., Manning et al., 1998). However, for the first time in these data, a medium effect size was achieved by a 2M:4M ratio for the combined collections the 2M:4M weight ratios for the two feet. Discussion In summary, substantial sex differences existed for many of the length and weight ratios for both metacarpals and metatarsals of the baboon. In accord with the pattern seen for human finger length (e.g., Manning et al., 1998; Mc- Fadden and Shubel, 2002a; Williams et al., 2000); the length ratios tended to be smaller in male than in female baboons. For the length and weight ratios showing the largest sex differences, the effect sizes tended to be larger for the right hand than for the left (as they are for human fingers), and larger for the left foot than for the right. Many of the effect sizes for the weight ratios were substantially larger than those for the corresponding length ratio. These differences between the ratios for length and weight suggest that future investigators should strive to obtain multiple measures when possible. Apparently, measures of weight have the potential to capture aspects of a bone s size that are not measured as simply as is length. The length and weight ratios showing the largest sex differences were not 2M:4M, which best corresponds to the ratio showing the largest sex difference for human finger length (e.g., Manning et al., 1998; McFadden and Shubel, 2002a; Williams et al., 2000). Rather, the 3M:5M ratio exhibited the largest sex difference for metacarpal length, metacarpal weight, and metatarsal weight, and 4M:5M was largest for metatarsal length (Tables 1 and 2). Worth noting, however, is that recent data from human skeletons have revealed that the largest ratio for human metacarpals (and metatarsals) also typically is not 2M:4M (McFadden and Bracht, 2002). Following the submission of this article, Tague (2002) reported length measurements of the metacarpals and metatarsals in 17 species of primates. Although length ratios were used in some of the statistical analyses, Tague s interests were different from those here, so results for the full array of length ratios were not presented. Tague did mention that a significant sex difference was found in baboons for the metacarpal length ratio 3M:4M. Curiously, here that ratio exhibited a smaller sex difference than several other metacarpal length ratios. It may be relevant that Tague s collection of baboons consisted of 31 males but only 4 females. The patterns of results here were not identical in the two collections of skeletons measured, but it is unclear what factors may have contributed to those different patterns. One collection consisted entirely of animals born and bred in the wild and living in a relatively localized geographic region at the time of death. The other consisted of primarily hybrid animals born and bred in captivity. Harris et al. (1992) have reported that the lengths of human metacarpals can change during adulthood, and they suggested that some of this effect may be attributable to the mechanical stresses to which these bones are exposed. So perhaps differences in the demands of life in the wild and in captivity are responsible for some of the differences observed in our two collections. Current thinking is that the sex difference in relative digit length in humans is attributable to androgens modulating the influences of certain homeobox genes during prenatal development (Kondo et al., 1997; Manning, 2002). The present results suggest that similar, but not identical, modulation processes occur during development in baboons. Although this similarity should come as no great surprise, greater knowledge about the similarities and differences between and among various species in those underlying developmental mechanisms could be informative about a large number of physiological and behavioral topics. Because it appears that acquisition of knowledge about those underlying developmental mechanisms can be guided by additional knowledge of the relative lengths of the bones in the extremities in other species, we regard the study of other species to be an especially worthwhile endeavor. In another study on nonhumans, Brown et al. (2001a) found a sex difference in the 2D:4D length ratio in the rear paws of mice, but only on the right side, and ratios other than 2D:4D were not measured. Only collection of similar data on other primates and mammals can reveal what patterns exist across species. Of particular interest should be which ratios show large sex differences, which ratios show the largest sex difference for each species, and which ratios show lateral asymmetries in the two sexes of each species. We admit to having considerable curiosity about the extremities of early hominids because of the possibility that some of those species might exhibit patterns of ratios similar to those of modern humans and other species might not. Elsewhere (McFadden and Shubel, 2002b, 2003) we will show that another measure also believed to be modulated by exposure to prenatal androgens otoacoustic emissions is uncorrelated with the 2D:4D ratio, suggesting that otoacoustic emissions and relative bone size in the extremities are separate windows onto different periods or processes during prenatal development. No behavioral data are reported here, but the work of Manning (2002) suggests that the mechanisms responsible for the sex differences in relative digit length in humans are also involved in the production of various differences in

8 354 D. McFadden, M.S. Bracht / Hormones and Behavior 43 (2003) Table 2 Effect sizes for the sex differences in various ratios of weight of the metacarpals and metatarsals in two collections of baboon skeletons a Ratio Anubis collection Mixed collection Combined collection Number Effect size (f-m) Number Effect size (f-m) Number Effect size (f-m) Left Right Left Right Left Right Hands 1M:2M 13/ / / M:3M 13/ / / M:4M 13/ b / / M:5M 12/ d c 24/ b / d b 2M:3M 15/ / b 41/ b 2M:4M 15/ / / M:5M 14/ c b 25/ b 39/ c c 3M:4M 15/ / / c M:5M 14/ d b 25/ d d 39/ d d 4M:5M 14/ b b 25/ c d 39/ d d Feet 1M:2M 14/ / / M:3M 14/ b 21/ / b b 1M:4M 14/ c 23/ c b 37/ d d 1M:5M 14/ c d 23/ d d 37/ d d 2M:3M 16/ / / M:4M 16/ / b b 42/ b c 2M:5M 16/ b 26/ d d 42/ d d 3M:4M 16/ / c d 40/ c d 3M:5M 16/ / d d 40/ d d 4M:5M 16/ / d d 42/ d d a f designates female; m designates male; M designates metacarpal or metatarsal; negative sign ( ) indicates male ratio larger than female ratio. Results from unpaired t tests: b 0.05 P 0.01; c 0.01 P 0.001; d P. physiology and behavior, some of which are related to susceptibility to disease. Further, Tague (2002) summarized arguments that some of the differences in metapodial length seen across primate species are correlated with the forms of locomotion used, and he discussed how the actions of certain homeobox genes might produce these morphological differences. Also, Harris et al. (1992) suggested that changes in the lengths of hand bones may be attributable in part to the uses to which the hand is put. The implication is that the relative sizes of the bones of the hand and foot eventually will prove to be associated with numerous behavioral measures. Finally, we wish to emphasize that skeleton collections offer a number of attractive advantages over live animals when the goal is measurement of the extremities. Not the least of these is that no surrogate for length need be adopted. One disadvantage is the difficulty in reassembling the phalanges to obtain estimates of length or weight for the individual fingers and toes, but skeleton collections may exist in which that information was preserved. Acknowledgments This work was supported by research grant DC from the National Institute on Deafness and other Communication Disorders (NIDCD). We thank C.A. Bramblett for generously providing us access to the collections of baboon skeletons, and for numerous helpful discussions. We thank J.C. Loehlin, J.W. Kappelman Jr., A.D. Gordon, and E.G. Pasanen for assistance and discussions. E. Shubel assisted with data analysis and with the figures. J.C. Loehlin, E.G. Pasanen, and E. Shubel commented on a previous version of this article. References Bramblett, C.A., Pathology in the Darajani baboon. Am. J. Phys. Anthropol. 26, Brown, W.M., Finn, C., Breedlove, S.M., 2001a. A sex difference in digit length ratio in mice [abstract]. Horm. Behav. 39, 325. Brown, W.M., Finn, C.J., Cooke, B.M., Breedlove, S.M., Differences in finger length ratios between self-identified butch and femme lesbians. Arch. Sex. Behav. 31, Brown, W.M., Hines, M., Fane, B., Breedlove, S.M., 2001b. Masculinized finger length ratios in humans with congenital adrenal hyperplasia. Horm. Behav. 39, Cohen, J., A power primer. Psychol. Bull. 112, George, R., Human finger types. Anat. Rec. 46, Glassman, D.M., Coelho Jr., A.M., Carey, K.D., Bramblett, C.A., Weight growth in savannah baboons: a longitudinal study from birth to adulthood. Growth 48, Harris, E.F., Akshanugraha, K., Behrents, R.G., Metacarpophalangeal length changes in humans during adulthood: a longitudinal study. Am. J. Phys. Anthropol. 87,

9 D. McFadden, M.S. Bracht / Hormones and Behavior 43 (2003) Jolly, C.J., Baboon nomenclature. Retrieved April 22, 2002, from University of Wisconsin, Wisconsin Regional Primate Research Center website: Kondo, T., Zakany, J., Innis, J.W., Duboule, D., Of fingers, toes and penises. Nature 390, 29. Manning, J.T., Digit Ratio: A Pointer to Fertility, Behaviour, and Health. Rutgers University Press, Piscataway, NJ. Manning, J.T., Baron-Cohen, S., Wheelwright, S., Sanders, G., The 2nd to 4th digit ratio and autism. Dev. Med. Child Neurol. 43, Manning, J.T., Bundred, P., The ratio of 2nd to 4th digit length: a new predictor of disease predisposition? Med. Hypotheses 54, Manning, J.T., Scutt, D., Wilson, J., Lewis-Jones, D.I., The ratio of 2nd to 4th digit length: a predictor of sperm numbers and levels of testosterone, LH and oestrogen. Hum. Reprod. 13, Martin, S., Manning, J.T., Dowrick, C.D., Fluctuating asymmetry, relative digit length and depression in men. Evol. Hum. Behav. 20, McFadden, D., Bracht, M.S., Sex differences in length ratios from the extremities of humans, gorillas, and chimpanzees [abstract]. Horm. Behav. 41, 479. McFadden, D., Shubel, E., 2002a. Relative lengths of fingers and toes in human males and females. Horm. Behav. 42, McFadden, D., Shubel, E., 2002b. Lack of correlation between otoacoustic emissions and relative lengths of fingers in humans (the 2D:4D ratio) [abstract]. Horm. Behav. 41, 480. McFadden, D., Shubel, E., The relationships between otoacoustic emissions and relative lengths of fingers and toes in humans. Horm. Behav. (In press). McKay, K.P., nd and 4th digit ratio of children with ADHD. Undergraduate honors thesis, University of Texas, Austin, June Phelps, V.R., Relative index finger length as a sex-influenced trait in man. Am. J. Hum. Genet. 4, Robinson, S.J., Manning, J.T., The ratio of 2nd to 4th digit length and male homosexuality. Evol. Hum. Behav. 21, Tague, R.G., Variability of metapodials in primates with rudimentary digits: Ateles geoffroyi, Colobus guereza, and Perodicticus potto. Am. J. Phys. Anthropol. 117, Williams, T.J., Pepitone, M.E., Christensen, S.E., Cooke, B.M., Huberman, A.D., Breedlove, N.J., Breedlove, T.J., Jordan, C.L., Breedlove, S.M., Finger-length ratios and sexual orientation. Nature 404, Wilson, G.D., Finger-length as an index of assertiveness in women. Person. Individ. Diff. 4,