Index Introduction Materials and Methods Subject macaques Somatometry and proportions in the extremities Statistical analyses Results Comparisons by size Principal component analysis Comparison of body size with reports from previous studies Discussion Acknowledgments References

Introduction

The rhesus macaque (Macaca mulatta), a member of the fascicularis species group (Fooden, 1976), has been studied intensively in diverse fields of biological research. The evolutionary history of macaques, however, has not been thoroughly elucidated (Delson, 1980; Fooden, 1980, 2000). Using tail length, Fooden (2000) classified living rhesus macaques (hereafter rhesuses) into three major groups: the eastern group, consisting of rhesuses in China and neighboring countries; the western group, consisting of those in India and its vicinity; and the southern group, consisting of those in the Indochinese peninsula. Although various divergence times, ranging from about 20 ka to 1.4 Ma, between the rhesus groups have been estimated from morphological and molecular evidence (Nozawa et al., 1977; Zhang and Shi, 1993; Hayasaka et al., 1996; Morales and Melnick, 1998; Fooden, 2000; Tosi et al., 2000, 2003), exact phylogenetic relationships have yet to be elucidated.

To clarify the obscure evolutionary history of rhesuses, detailed studies are needed that take into account biological variation related to geographical distribution patterns. Although some morphological evaluations have been conducted on rhesuses from various localities (Fooden, 1964, 1971, 1997, 2000; Jiang et al., 1991) and colonies (van Wagenen and Catchpole, 1956; Gavan and Hutchinson, 1973; DeRousseau and Reichs, 1987; Turnquist and Kessler, 1989; Clarke and O’Neil, 1999), detailed morphometric comparisons have not been made. In particular, only few studies have reported on the body size and proportions of macaques based on standardized morphometric methods (Iwamoto, 1971; Hamada et al., 1986, 1996).

In the present study, we describe and compare Indian- and Chinese-derived rhesuses reared at the Primate Research Institute, Kyoto University (hereafter PRI) by applying traditional methods of anthropometry (somatometry). We hereby present morphometric data of the PRI rhesuses, and offer some interpretations as a background to broader-based intra- and inter-specific comparisons.

Materials and Methods

Subject macaques

Rhesuses (M. mulatta) were sporadically introduced into PRI colonies, from India as of September 1969 and from China as of June 1972, until December 1979. Although the exact localities of origins are not known, the manager of animal-care at PRI had acquired written verification from the providers that the animals had originated from India and China. Results of genetic studies conducted on the PRI rhesuses (e.g. Nozawa et al., 1977) corroborate the stated regions of origin. The PRI rhesuses have been reared and bred separately according to country of origin. Subject individuals for this study were reared in corralled (open-enclosures of ca. 500 square meters) social troops of about 50 animals. During the autumn annual health check, we measured the subject rhesuses. As the number of subjects was limited, the somatometric data obtained from 1987 to 2002 were pooled. In this study, we report measurements on a total of 74 adult rhesuses (7.0–19.9 years of age), of which 23 were Indian-derived (17 females and 6 males) and 51 were Chinese-derived (36 females and 15 males).

Somatometry and proportions in the extremities

We used 17 somatometric measurements and indices for comparison. We followed established anthropometric methods (Martin and Saller, 1957) with slight modifications, such as the inclusion of tail length (for details see Iwamoto, 1971). As we did not measure all variables on all subjects, the number of subjects for each variable is listed in Table 1 and Table 2. The head index (%) was defined as (head breadth)/(head length), and relative tail length (%) was calculated as (tail length)/(crown–rump length). Proportions of extremities compared were as follows (in %):

brachial index = (forearm length)/(upper arm length),

crural index = (leg length)/(thigh length),

intermembral index = (upper arm length + forearm length)/(thigh length + leg length),

relative hand length = (hand length)/(upper arm length + forearm length),

foot index = (foot width)/(foot length),

relative foot length = (foot length)/(thigh length + leg length),

intercheiridial index = (hand length)/(foot length).

Statistical analyses

Basic statistics were calculated separately for males and females of Indian- and Chinese-derived rhesuses. Sex difference was expressed by the ratio (%) of (female size)/(male size). The two populations were compared by the ratio (%) of (Indian average)/(Chinese average). Statistical significance of mean differences between sexes and between Indian- and Chinese-derived rhesuses was examined by t-tests (Welch modified, two-sided, without supposing equal variances, using S-Plus 4.0, MathSoft Co. Ltd.) where significance level was modified by the Bonferroni method (Sokal and Rohlf, 1995).

We applied principal component analysis (PCA) to 11 somatometrical variables: crown–rump, head, upper arm, forearm, thigh, leg, hand (ulnar), and foot lengths and foot, head, and bi-zygomatic breadths. Measurements from a subset of 50 individuals, 33 Chinese-derived (23 females and 10 males) and 17 Indian-derived rhesuses (12 females and 5 males), were used in the PCA.

Results

Comparisons by size

Basic statistics are listed in Table 1 and Table 2 by geographic origin and sex. Age-related changes in body size and proportions were not evident among the subjects. This was true of body mass (BM) owing to considerable individual variation.

Ratios between average values of variables of Chinese- and Indian-derived female rhesuses, excluding body mass and tail length, ranged from 91.2% (radial hand length) to 100.6% (chest girth), with 13 out of 17 characters having ratios less than 100% (i.e. Indian rhesuses tended to be smaller) (Table 1). Ratios of crown–rump length (CRL) and anterior trunk length (ATL), which are representative linear measures of whole body size, were 95.8% and 96.1%, respectively. Similarly, in males, the size difference was not great, but the Indian-derived rhesuses tended to be consistently smaller than the Chinese-derived rhesuses as shown by the CRL (93.2%) and ATL (94.0%) ratios. The Chinese to Indian ratios ranged from 88.6% (foot breadth) to 100.2% (bizygomatic breadth), with 15 out of 17 characters having ratios less than 100%. Measurements of the extremities of the Indian-derived rhesuses, especially those of the hand and foot, were significantly smaller than those of Chinese-derived rhesuses. Sex differences were a little greater in the Chinese- than in the Indian-derived rhesuses.

The Indian-derived PRI rhesuses had significantly longer tails than the Chinese-derived PRI rhesuses, as shown by ratios of 117.1% and 114.7% in females and males, respectively. Relative tail length values were 44.2 ± 4.91% and 46.1 ± 4.06% in female and male, respectively, in the Indian-derived rhesuses, and 36.5 ± 2.85% and 37.8 ± 4.31% in female and male, in the Chinese-derived rhesuses (Table 2). Indices that express body proportions are listed in Table 2. In females, the Chinese-derived rhesuses have significantly larger values than the Indian-derived rhesuses in intermembral index and relative foot length. In males, similarly, relative hand and foot lengths tended to be larger in the Chinese-derived rhesuses, although this difference was not statistically significant.

Principal component analysis

The first three principal components explained 90.5%, 4.34 %, and 1.55% of the total variance, respectively (Table 3). The first component represents the ‘size’ factor, and the average component scores obtained were -36.35 in females and 20.65 in males of the Indian-derived rhesuses and -12.92 and 63.00 in females and males of the Chinese-derived rhesuses, respectively.

The second component expresses the proportion between trunk length and limb segment lengths, but showed no separation between either sex or country of origin. The third component expresses the proportion of hand and foot lengths relative to trunk and proximal limb segment lengths. This component showed a clear difference between countries of origin (Figure 1) as highlighted by average scores of 2.33 and 4.06 in the Chinese-derived females and males, respectively, and -5.45 and -5.73 in the Indian-derived females and males, respectively. The differences between mean scores of the Indian- and Chinese-derived rhesuses were significant (t-test, P < 0.05) in both sexes.

View Details

Figure 1. The second and third principal component scores of Indian- and Chinese derived rhesuses. Indian male (solid circles), Indian female (open circles), Chinese male (+), and Chinese females (×).

Comparison of body size with reports from previous studies

Differences in methods of measurement (reference anatomical points) may affect comparisons, but fundamental measures, such as BM and CRL (proxy of head and body length, HBL), can be used for comparison. We also compared extremity lengths among studies, in cases where measurement methods appeared similar.

Fooden (2000: p. 26) compiled somatometrical data of museum specimens taken from collectors’ measurements. Grand averages of HBL for all specimens were 468.8 ± 49.1 mm (mean ± SD; range: 370–550) in females and 531.8 ± 55.2 mm (range: 410–660) in males. The BM grand average was 5.34 ± 1.34 kg (range: 3.00–9.98) in females and 7.70 ± 2.33 kg (range: 4.01–14.06) in males. The CRL values of the present study were greater than Fooden’s (2000) grand averages by 0.66 SD and 0.25 SD in the Indian-derived females and males, respectively, and by 1.11 SD and 0.97 SD in the Chinese-derived females and males, respectively. In considering geographical variation in HBL (Fooden, 2000: Figure 8), the CRL values of the present study correspond to the largest end of the range for both Chinese and Indian rhesuses (max. ca. 660 mm in males and 580 mm in females: Fooden, 2000).

Average BM and CRL (or its alternatives) values of rhesuses reared in some other facilities are listed in Table 4. Yale rhesuses at the age of ca. 7.0 years (Indian-origin: van Wagenen and Catchpole, 1956) had a CRL comparable with the PRI Indian-derived rhesuses, but the latter had relatively shorter thighs and longer legs. Cayo Santiago rhesuses (Indian-origin: Turnquist and Kessler, 1989), which were reared under semi-wild conditions, were slightly larger than the PRI rhesuses not only in CRL but also in limb segment lengths by up to 13%. Although the number of morphometric characters is limited and the way of measurement differs (e.g. they measured forearm length from the olecranon to the tip of styloid process ulna: Turnquist and Kessler, 1989), Cayo Santiago rhesuses could be compared with the PRI rhesuses in limb segment lengths. They have relatively (against CRL) shorter upper arm and foot like PRI Indian-derived rhesuses, but have relatively longer thighs. The Tulane rhesuses, both of Indian and Chinese origin (Clarke and O’Neil, 1999), had a smaller BM and a much smaller CRL than Indian- and Chinese-derived PRI rhesuses of the present study, respectively. Their limb segment lengths (Clarke and O’Neil, 1999: Figure 2), however, were slightly greater than those of the Chinese-derived PRI rhesuses.

Discussion

Before discussing the morphometric differences in the two populations, we must first consider some factors that influence morphometric traits. The rhesus subjects of the present study were reared separately by origin, and therefore their genetic properties have been maintained. Ontogenetic plasticity due to environmental conditions is, then, the first factor to be considered as a possible cause of morphological differences (DeRousseau and Reichs, 1987; Stearns, 1992; Bogin, 1999). Morphological change in response to large degrees of environmental change has been reported in Japanese macaques transferred to the USA (Paterson, 1996) and in rhesuses that matured in the USA (Clarke and O’Neil, 1999). DeRousseau and Reichs (1987) studied ontogenetic plasticity of morphometric traits in Cayo Santiago rhesuses. Although they found that body size fluctuated significantly with food quantity/quality and population size, the magnitude of such fluctuations was not high (2–4% between the largest and smallest 5-year cohort averages: see Table 2 of DeRousseau and Reichs, 1987). In the present study, all rhesus subjects were reared in comparable environments, in corrals with comparable population density and structure, and were fed a standardized diet.

Living rhesuses range over a wide area in Asia (Fooden, 1980), and are expected to exhibit substantial geographical variation in morphology. Fooden (2000) found latitudinal variation in size, in accordance with Bergman’s rule. Significant longitudinal variation was found in tail length. On the basis of relative tail length (versus HBL, sex combined), Fooden (2000) classified rhesus macaques into three major groups: the eastern group, ranging through China and neighboring countries, with the shortest tails (about 30%); the western group, ranging through India and its neighboring countries (west of ca. 95° E longitude), with tails of medium length (46.7 ± 7.0%, range: 31.9–62.0, n = 68); and the southern group, ranging in the Indochinese countries, with the longest tails (53.0 ± 4.9%, range: 47.5–60.8, n = 8). Although values would differ slightly due to the difference in the denominator variable (CRL in this study and HBL in Fooden, 2000), the Indian- and Chinese-derived PRI rhesuses showed relative tail lengths comparable to those of the western and eastern groups, respectively (Table 2).

This longitudinal variation was not found in the other measures of body size and proportions (Fooden, 2000), and the size difference between the Chinese- and Indian-derived rhesuses reared at PRI was relatively small. The latter difference, at only 4–7%, is much smaller than the latitudinal variation found in Japanese macaques (Macaca fuscata) (Iwamoto, 1971; Hamada et al., 1996), where the average ATL of the smallest troop (Koshima) is about 88% of that of the largest troop (Shiga-highland males and Hakusan females). The limb proportions, however, showed differences between the Indian- and Chinese-derived PRI rhesuses. The Indian-derived Cayo Santiago rhesuses appeared either similar to the PRI Indian-derived condition or distinct (relatively longer thigh), and the Tulane rhesuses also have unique proportions. We must wait for future research to elucidate the factors responsible for morphometric variation in rhesuses.

Morphometric similarity between the Indian- and Chinese-derived rhesuses can be considered from the viewpoints of phylogeny and adaptation. Although an evolutionary scenario for rhesuses has been proposed (Delson, 1980; Fooden, 2000), the precise time of divergence between the eastern and western groups has not been ascertained. A quite recent divergence between the Indian and Chinese populations, ca. 23 ka, was suggested by a study based on blood protein polymorphisms (Nozawa et al., 1977). Molecular phylogenetical studies based on mitochondrial DNA, however, have suggested divergence times as remote as about 0.75 Ma (Morales and Melnick, 1998) or even ca. 1.4 Ma (Zhang and Shi, 1993). Analysis of the nucleotide sequence of mitochondrial DNA has shown that the rhesus is a paraphyletic species (Hayasaka et al., 1996). One of Hayasaka et al.’s (1996) rhesus subjects (probably from India) clustered with the Japanese and Taiwanese macaques, and then the other two subjects (origins not specified) clustered with that cluster. Similar results have been reported by other researchers (e.g. Tosi et al., 2003). The above summarized discrepancies in estimations of divergence time and complications encountered in phylogenetic reconstructions are due to social structure (female philopatry and male dispersal in macaques), introgression, and lineage sorting (Avise, 2000; Tosi et al., 2003), and we must await data from future studies based on local populations using appropriate genetic markers.

In the meantime, regardless of the depth of divergence, the general morphological similarity between the Indian- and Chinese-derived rhesuses may be explained by adaptation to similar habitats. Ecological competition (niche separation) has been suggested for rhesuses and other macaque species that have undergone range expansion (Delson, 1980). Competition may have restricted the habitat available for rhesuses, and consequently, their morphological diversity might have also been restricted. Rhesuses are known as “non-evergreen broad leaf forest dwellers” (Fooden, 1982) that also exploit disturbed or secondary forests as “weed” or “opportunistic” species (Richard et al., 1989), which other sympatric species of macaques would not share. The fact that insular rhesus populations, e.g. the Hainan Dao population, tend to show characteristic morphology (Jiang et al., 1991; Fooden, 2000) may reflect the fact that they are free from such ecological competition.

Acknowledgments

The authors express sincere thanks to all current and former staff of the Morphology Section, Primate Research Institute, Kyoto University, for their help and suggestions. We are also indebted to the Human Evolution Modeling Research Center of PRI, Kyoto University for their support in measuring rhesuses. Finally, we thank Dr Suchinda Malaivijitnond and Dr Gen Suwa (the editor-in-chief) for their valuable comments on the manuscript.