AMERICAN JOURNAL OF PHYSICAL ANTHROPOLOGY 5637-81 (1981) Brain Weight-Body Weight Relationships in 12 Species of Nonhuman Primates RODERICK T. BRONSON Department of Comparative Pathology, Harvard Medical School, New England Regional Primate Research Center, Southborough, Massachusetts 01772 KEY WORDS Brain weight, Body weight, Primates, Allometry ABSTRACT Necropsy data from a Primate Center were used in a study of the brain weight-body weight relationships of 1 2 species of nonhuman primates. The sample sizes ranged from six Cercopithecus aethiops to 163 Macaca mulatta. By plotting mean brain-mean body weight of each species on log-log paper, it was shown that the straight line fitting the plots of all species had a slope of 0.72. Slopes for three species of the genus Macaca, and for six species of the family Cebidae, were 0.61 and 0.81 respectively. Coefficients of determination of the three lines were greater than 0.90. Two species of the family Cebidae, Saimiri sciureus and Aotus trivirgatus, had equivalent body weights, but the former had a 30% larger brain than the latter. The results suggest that brain-body weight scaling characteristics of primate species can be studied effectively using necropsy data. Some statistically significant discrepancies between these and published data, however, show that more data are required to describe these characteristics with greater certainty. How brain weight varies with body weight between species is optimally studied in log-log plots of mean brain weight-mean body weight of large samples of individuals of each species (Jerison, 1976; Bronson, 1979). Thus species within a taxonomic class, plot along a line with slope of 0.67 (Jerison, 1976). Classes differ in that one class line, e.g., the Mammalia line, may be shifted above another, e.g., the Reptilia line (Jerison, 1976). How brain size varies with body size within and between taxonomic orders, families, and genera is not known, probably because the data have not been available. Brain-body weight scaling within species has been studied. Mean brain weight-mean body weight, for samples of dog breeds and cat breeds, have been plotted on intraspecies lines with slopes of 0.27 and 0.67 respectively (Bronson, 1979). This is a report of the brain-body weight relationships of 12 nonhuman primate species, for which data were available, from the pathology files at the New England Regional Primate Research Center. Mean brain weight and mean body weight for each species were calculated, and log-log plots of the data were analyzed. These data will be compared to those collected by Bauchot and Stephan (19691, from which they differ significantly in some respects. 0002-948318115601-0077502.00 Some real, and artifactual, reasons for variability in brain weight-body weight and relationships between them, will be considered. MATERIALS AND METHODS The monkeys for which weight data were studied, had been maintained in captivity for variable periods, and had died from a wide variety of natural and experimental diseases. The body and organ weights were recorded by many different people over a decade, on periodically adjusted scales, as part of routine necropsy procedure. The species, probable age, sex, brain weight, and body weight were recovered from each pathology report. Data were accepted into the study without regard to the pathologic diagnosis. Age was usually not known, since monkeys were caught in the wild. Weight data from monkeys known to be infants or juveniles, because they were born in captivity, or because their body weights were clearly too low to be appropriate for those of adults, were excluded from the analysis. The mean brain weight and mean body weight, and standard deviations for each sex, and for both sexes of each of 12 species, were 0 1981 ALAN R. LISS, INC. ~ _ _ _ Received February 17.1981; accepted June 22,1981 78 R.T. BRONSON T A B L E 1. Mean brain and body weights for 12pnmate species Female 1 Saguinus oedipus 2 Aotus trivirgatus 3 Saimiri sciureus 4 Cebus albifrons 5 Cebus apella 6 Ateles geoffroyi 7 Ateles paniscus 8 Cercopithecus aethiops 9 Macaca fasciculans 10 Macaca mulatta 11 Macaca arctoides 12 Papio anubis 47: 9.7 t 0.6 140: 17.2-1- 1.7 58: 23.1 -1- 1.8 21: 60.7 2 5.9 10: 62.6 2 6.4 10:107.7 f 10.1 17:101.2 t 11.5 6: 66.1 k 8.9 36: 61.9 f 7.9 132: 85.2 -1- 8.6 25: 99.8 f 9.8 5:148.0 2 9.7 47: 370k 85 143: 6 3 0 f 153 57: 5 8 0 f 85 21: 1,470 f 300 11: 1,810 f 260 11: 4,680 f 1,920 15: 5,220 t 2,210 6: 3,220 k 550 44: 3,360 f 1,150 163: 4,880 f 2,350 27: 7,450 -t 3,770 7:12,900 k 3,160 0.5 Male Body_weight, gm n: X t o Brain wzight. g m n: X t o Body weight, gm n: X + o I 5 58: 3 7 0 2 66 116: 720 2 180 33: 7 2 0 5 150 45: 1,770 k 640 7: 2,770 f 650 4: 4,760 f 1,620 6 5,040 k 2,730 10: 5,230 k 1,020 8: 4,740 t 1,680 125: 5,210 t 2,950 13: 7,810 t 4,010 13:16,060 3,770 + 10 Brain wAight, gm n: X-+o 59: 9.2 k 0.6 120: 17.2 & 1.5 32: 25.0 k 2.4 45: 65.2 ? 8.9 7: 70.8 -C 6.2 4:107.6 +. 16.5 6:106.5 % 12.4 10: 66.3 ? 5.6 7: 70.0 ? 11.6 93: 90.7 k 12.1 9:101.6 ? 12.0 15:174.7 ? 17.4 50 MEAN BODY WElGHT g x103 Fig. 1. Log-log plots of mean brain weight-mean body weight data for 12 species of primates. The numbers refer t o species listed in Table 1. The ellipse around each point represents one standard deviation in each dimension. Four lines are fitted to the appropriate points by the method of least squares; their formulae appear in Table 2. The average mammal line is from Jerison (1976). and is equivalent to the line fitting data for domestic cats la) and dogs (b) IBronson, 1979). 79 BRAIN-BODY WEIGHTS IN PRIMATES calculated. To analyze how brain weight varied with body weight for various groups of primates, power curves were fitted to the data by the method of least squares.Thecurves had the general formula, y = axb, where y = brain weight, x =body weight. The coefficient of determination (r2),was calculated for each curve. The groups so analyzed were: (1)males of 12 species; (2)females of 1 2 species; (3) animals of both sexes of 12 species; (4) each sex and combined sexes of s ecies of the enus Macaca (mulatta, fascicufks, arctoide5; ( 5 ) each sex and both sexes of six species of the family Cebidae (Cebus albifrons, Cebus apella, Ateles paniscus, Ateles geoffroyi, Saimiri sciureus, and Aotus trivirgatus); and (6) individuals of each of three species, M. mulatta, C. albifrons, and C. apella. These were chosen arbitrarily as examples of species represented by large, intermediate, and small sample sizes. RESULTS AND DISCUSSION The weight data are presented in Table 1. The data for males and females combined are plotted on log-logpaper in Figure 1. The ellipse around the point for each species in Figure 1 represents the standard deviations in both dimensions. The power curve fitting data for 12 species, and for Macaca and Cebidae, are straight lines in the figure. Formulae and coefficients of determination for these and other lines are presented in Table 2. The Carnivora line in the figure is from Bronson (1979). Variability in body weight is greater than variability in brain weight in most species studied. This was especially pronounced in such species as macaques, which are prone to obesity (Walike et al., 1977).Variation in body weight may also be due to disease factors. Saguinus oedipus, for example, often lose considerable weight before death (Chalifoux and Bronson, 1981). Inadvertent inclusion of data from immature animals also increases variability, and may seriously bias the data, since immature animals have larger brain weight-body weight ratios than adult animals. A consequence of the large variability in body weight, is that power curves fitted to brain weight-body weight data for individuals of the same species have almost zero slope, and very low r2 values (Bronson, 1979). This finding, repeated for three primate species, is presented in Table 2. Clearly intraspecies brainbody weight scaling cannot be studied usefully from individual animal data. In studying brain-body weight scaling of species within a taxonomic group, four potential sources of error must be considered. First, the location of the point for each species is subject to error as a result of the intraspecies variation in both parameters. This effect can be severe if the sample is small. Second, the validity of the line fitting a group of species depends on the number of species. When small numbers are studied, the effect of inclusion or exclusion of a single species can be severe, particularly if it plots at one end of and away from the regression line fitting the data. For example, if mean brain and body weight for nine Tupaia glis, 3.1 k 0.5 gm and 129 k 19.7 gm respectively, are included in the regression analysis of the 12 species studied here, the slope of the line is 0.81 instead of 0.72. A third source of error is expected when the groups plotted have similar body weights. This is especially true when large standard de- TABLE 2. Formulae ofpower curves, y = axb, fitting brain weight fy) body weight (x) data from various samples of primate species Sample a b Coefficient of determination,r2 0.20 0.21 0.17 0.72 0.7 1 0.73 0.91 0.91 0.91 0.43 0.51 0.45 0.61 0.59 0.61 0.90 0.95 0.71 0.12 0.16 0.08 0.12 34.40 7.76 26.5 0.81 0.77 0.85 0.67 0.11 0.28 0.12 0.91 0.92 0.94 12 species of primates M & F F M ~3 species of Macaca M & F F M 6 species of Cebidae M & F F M Average mammal (Jerison, 1976) 92 M M. mulatta 17 M & F C. aaella 61 M & F C. aibifrons - 0.06 0.48 0.06 80 R.T. BRONSON viations are also present. Thus the Primates line, including species ranging in body weight from 370 gm to 16,060 gm, is less affected by individual animal variation than the Macaca line. The fourth source of error results from inclusion of both male and female animals of sexually dimorphic species, or of adults and juveniles. The mean body weight of a sample consisting primarily of female or juvenile animals might be considerably less than that of one consisting of male animals. Since brain-body weight scaling of females was similar to that of males of the three taxonomic groups studied (Table Z), sexual dimorphism was probably not a source of error here. In spite of the uncertainties intrinsic to this analysis, some general conclusions about brain-body weight scaling in primates can be drawn safely. First, scaling within the order Primates, the family Cebidae, and the genus Macaca, occurs along lines with slopes approximating those of taxonomic classes, 0.67, rather than that of breeds of dogs, 0.27 (Bronson, 1979). This finding for the genus Macaca contradicts that of Sholl (1948),and the assumption by Pilbeam and Gould (1974)that scaling within a genus occurs along lines with slopes of between 0.20 and 0.40. Sholl studied various ways of fitting curves to brain-body weight data of individual macaques of several species and ages, including what must have been juveniles, judging from the weights. He found that the line fitting log-logplots had a slope of around 0.25. This line was probably an artifact of summation of data for which the interspecies line would have a slope of around 0.67, and for which the intraspecies line would have virtually no slope, as we have seen. Another interesting finding is that S. sciureus has a 30%larger brain than A. trivirgatus, even though the two species have similar body weights. A likely explanation for the weight difference is that S. scuireus has a much larger occipital lobe than A. trivirgatus (Fig. 2). How this might relate to neurological or behavioral differences is unclear. This pair of species seems a reasonable one in which to explore the poorly defined concept of “encephalization” (Holloway,1968;Krompecher and Lipak, 1966; Pilbeam and Gould, 1974).That Suimin’isdiurnal, and Aotus is nocturnal, might make comparisons difficult, however. Other conclusions about these data are more tentative, because of uncertainties arising from the four sources of error previously mentioned. Of greatest interest is the difference in Fig. 2. Skulls of Sairniri sciureus (top),and Aotus triuir. (bottom). The distance from the posterior edge of the foramen magnum t o the occipital pole is 1.4 cm in the former. 0.9 cm in the latter. X 1.29. gatus slopes for the macaques and the Cebidae. Whether it is real or artifactual can be resolved with certainty only by including more data and more species. There appears to be no statistical test available to test the significance of differences in slopes of two regression lines fitted to data points that themselves represent the means of individual data points. In comparing the data presented here with those collected by Bauchot and Stephan (1969) for the same species, the validity of the difference in the slopes for Cebidae and Macaca become still less certain. Whereas mean brain weight for none of the species differed significantly between the two samples when tested by Student’s t test, body weights for five species did: C. apella, A. geoffroyi, A. paniscus, C. aethiops, and M. mulatta. The mean body weight of all these, other than A. geoffroyi, was lower in the published samples, even after excluding those individuals stated to be immature, or whose weights clearly indicated they were. In A. geoffroyi, Bauchot and Stephan (1969)reported very high mean body weights, 7,159 gm, n = 98, compared to that of this sam- BRAIN-BODYWEIGHTS IN PRIMATES ple. There is no clear explanation for the discrepancy. Significantly, if the scaling line of the Cebidae is recalculated using this value the slope is 0.71, not 0.81. This study, like a previous one (Bronson, 1979),has shown that brain-body weight scaling within taxonomic groups can be studied effectively by plotting mean weights of large samples of necropsy specimens. In the future, as more data are collected, it may be possible to map many primate species onto a log-log plot with a reasonable degree of certainty. At that time, scaling characteristics of species within families and genera can be looked at again. ACKNOWLEDGMENTS This work was supported by NIH grant NO. RR00168, from the Division of Research Resources, Bethesda, Maryland. 81 LITERATURE CITED Bauchot. R. and Stephan H. (1969)Encephalization et niveau evolutif chez les simiens. Mammalia 33:225-275. Bronson, RT (1979) Brain weight-body weight scaling in breeds of dogs and cats. Brain Behav. Evol. 16227-236. Chalifoux, LV, and Bronson, RT (1981) Colonic adenocarcinoma associated with chronic colitis in cotton top marmosets, Saguinus oedipus. Gastroenterology 80:942 -946. Holloway, RL (1968) The evolution of the primate brain: Some aspects of quantitative relations. Brain Res. 2121 172. Jerison, H J (1976) Paleoneurology and the evolution of mind. Sci. Am. 23490-101. Krompecher, ST, and Lipak. J (1966) A simple method for determining CerebraliGtion brain weight and intelligence. J. Comp. Neurol. 122113-120. Pilbeam, D, and Gould, SJ (1974) Size and scalingin human evolution. Science 186892-901. Shall, D (1948) The quantitative investigation of the vertebrate brain and the applicability of allometric formulae to its study. Roc. R. Sac. Lond. Ser. D. 135243-358. Walike, BD, Goodner, CJ, Keorker, DJ, Chideckel, EW, and Kolnasy. LW (1977) Assessment of obesity in pigtailed monkeys (Macacu nemestrinu). J . Med. Primatol. 6 1 5 1 162.