The natural distribution of rhesus macaques (Macaca mulatta) extends from South to East Asia, with substantial morphological variation among populations from different geographic locations. In the following report we compare morphometric measurements from rhesus macaques from Bangladesh to measurements from free-ranging rhesus in Nepal and captive rhesus populations originating in China and India. Our data indicate that Bangladeshi rhesus are morphologically similar to populations in South Asia, particularly India, and distinct from rhesus macaques originating in China. Our results also indicate that relative to the South Asian population samples, the rhesus macaques originating from China are distinct morphometrically.
Rhesus monkeys are among the best studied of the nonhuman primates owing to their wide use as models for studying human physiology and disease. Their natural range extends from Afghanistan in the west to the eastern-most edge of China and south through central Thailand, and includes a variety of ecological niches. Rhesus macaques are among the handful of nonhuman primate species capable of thriving in human-altered environments, including densely populated urban areas. Research on Bangladeshi rhesus monkeys has provided data on their distribution, conservation status (Green 1978; Gittins and Akonda 1982; Khan and Ahsan 1986; Feeroz et al. 1995a; Feeroz 2001), ecology, behavior (Feeroz et al. 1995b; Taslima 2002; Sarker et al. 2005) and genetics (Feeroz et al. 2008). Feeroz et al. (1995a) described coat color variation in different local populations of rhesus macaques. To date, however, no morphological analysis of Bangladeshi macaques has been published. Morphometric analyses of other rhesus populations (see Smith et al. 1987; Smith and Scott 1989; Smith 1994; Clarke and O'Neil 1999; Hamada et al. 2005; Taylor and Schillaci 2008) provide a context for the present comparative intra-species study.
Between 2006 and 2008, 47 rhesus macaques (Macaca mulatta) from five sites in Bangladesh were trapped in conjunction with a large study on simian retroviruses in Asia (Jones-Engel et al. 2008) (Fig. 1). The sex and age-class distribution for the sample population is shown in Table 1. A number of morphological measurements were collected while the monkeys were under sedation (Table 2). Monkeys were placed in one of four age categories, based on their pattern of dental eruption and the occlusal wear of the adult third molars (Table 3). We compared these measurements to measurements taken on Nepalese macaques in 2005 (included in Table 1) and to previously published morphometric data from Indian and Chinese rhesus macaques (Hamada et al. 2005).
Univariate comparisons of adult morphometric variables were conducted using t-tests, and the nonparametric Mann-Whitney U-test. Growth patterns were assessed by plotting raw values by age category. A multivariate assessment of morphometric variation was achieved through principal components analysis of adult means, and through a cluster analysis of principal component scores.
The results from our univariate comparison revealed significant differences between the Bangladesh adult female rhesus macaques and the adult female rhesus from Nepal in bizygomatic breadth and bifrontal breadth (Table 4).
Maximum sample sizes for age- and sex-specific groupings for the free-ranging Bangladesh and Nepal sample populations.
Morphometric variables used in the study.
A significant difference in foot length between adult females from Bangladesh and India was detected, as was a difference in tail length between adult females from Bangladesh and adult females from China. The adult males from Nepal and China differed significantly from the adult males from Bangladesh in trunk and foot lengths. The Chinese adult males also differed significantly from the Bangladesh males in body mass.
Among the populations included in the analysis, the Bangladesh sample exhibited the lowest level of sexual dimorphism. The Bangladesh sample also showed the least variation in sexual dimorphism across traits. A similar condition was observed for the sample from India. The population samples from Nepal and China both exhibited a level of sexual dimorphism approximately 1.12 times greater than that observed for the Bangladesh sample.
Our evaluation of male and female growth for the Bangladesh sample revealed similar patterns of growth until age category 3 (approximately 3 to 4 years of age), at which time females exhibit slower rates of size increases (Fig. 2). A comparison of the growth patterns of the Bangladesh and Nepal samples revealed similar patterns with some minor differences apparent for both males and females (Fig. 3). Small sample sizes for some age categories, however, limit interpretation of small pattern differences.
Age categories based on dental eruption patterns.
The results from the multivariate analysis suggest that the Chinese macaques are distinctive when all variables are considered simultaneously. A plot of the first two principal components that explain over 89% of the total variation revealed separation of males and females along the first principal component, and separation of the Chinese male and female samples from the South Asian samples along the second principal component (Fig. 4). All positive eigenvector loadings for the first principal component indicate that it describes size variation, while subsequent components describe variation in shape (Table 5). A plot of second and third principal components describing over 92% of the shape variation again shows that the Chinese samples are distinct (Fig. 5). The eigenvector directions and magnitudes mapped onto the bivariate plot of shape components indicates the second component is defined by variation in tail length relative to body mass and foot length. The Nepal and Indian samples are separated along the third principal component which is defined primarily by variation in bizygomatic breadth and trunk length.
A UPGMA cluster analysis of the second and third principal components describing shape variation grouped the samples from India and Bangladesh together. The Nepal samples joined the India and Bangladesh grouping to form a larger cluster to the exclusion of the Chinese samples (Fig. 6). Two primary groups are therefore apparent: one South Asian and one East Asian.
Comparison of adult body size variables among populations of M. mulatta. Shaded values differ significantly (α=0.05) from those observed for the population sample from Bangladesh.
To date there has been no systematic characterization of the morphology of rhesus macaques from Bangladesh. The results of our study suggest that the rhesus macaques from Bangladesh are more similar morphometrically to rhesus macaques from South Asia, particularly India. The rhesus macaques of Chinese origin are distinct. Short tail length seems to be the primary trait distinguishing the Chinese rhesus from the South Asian samples. Fooden (2000) indicated that tail length in M. mulatta does vary slightly with latitude (also see Fooden and Albrecht 1999), consistent with Allen's rule (Allen 1877). More interesting, perhaps, is Fooden's observation of a strong negative correlation between tail length and longitude. The Chinese rhesus in the present study exhibited the shortest tails.
The female macaques from Bangladesh exhibited a mean body weight (7.73 kg, SD = 1.43) generally comparable to the other populations in the study, but much larger than the species average reported by Fooden (2000) (5.34 kg, SD= 1.34). A similar condition was observed for male body weight. In our study, the male rhesus macaques exhibited the lowest mean body weight estimate (8.66 kg, SD=2.66) among the geographic populations. This estimate, however was greater than Fooden's (2000) reported mean male body weight for the species (7.70 kg, SD=2.33). The reason for this difference is not immediately apparent but may be a product of provisioning or access to human food sources (e.g., crops or garbage) for our Bangladesh and Nepal populations.
A comparison of male and female growth patterns for the Bangladesh sample revealed divergent trajectories after age category 3 (approximately 3 to 4 years of age), eventually leading to observed dimorphism in adult dimensions. This sexual dimorphism is more pronounced for cranial traits than for body weight and trunk length. Interestingly, the overall level of sexual dimorphism in the Bangladesh sample is low relative to the other population samples. The observed differences in sexual dimorphism invites speculation that there may be variation within the species in the magnitude of male-male competition for mates.
In conclusion, the results of our comparative morphometric study of the rhesus macaques from Bangladesh indicates this population is broadly similar to other South Asian populations, but differs, like other South Asian populations, from the rhesus macaques of Chinese origin.
Sexual size dimorphism scores for adult Macaca mulatto1.
Eigenvector loadings and eigenvalues from the principal components analysis of adult means.
We thank Vice Chancellor Prof. Khandaker Mustahidur Rahman of Jahangirnagar University for his support of this research project. We are also most grateful to H. and L. Engel, M. Adnan, M. Irfan and J. Heidrich for their outstanding logistic support and expert assistance with the health assessment of the rhesus macaques in Bangladesh; R. Liszanckie and E. McArthur for their excellent assistance with the preparation of the manuscript and the figures; the Wildlife Rescue Centre (WRC), Jahangirnagar University, for facilitating fieldwork; and The Forest Department of Bangladesh for its assistance with the acquisition of permits. This research is supported by NIH-NIAID R01AI078229 and R03AI064865 and NIH-NCRR 5P51RR000166.