Biological affinities between the migrant groups of fishermen of Puri Coast Orissa India.код для вставкиСкачать
AMERICAN JOURNAL OF PHYSICAL ANTHROPOLOGY 74:407-415 (1987) Biological Affinities Between the Migrant Groups of Fishermen of Puri Coast, Orissa, India B. MOHAN REDDY, VIRENDRA P. CHOPRA, AND DEBA P. MUKHERJEE Institut fur Humanbiologie, Uniuersitat Hamburg, 2000 Hamburg 13, Federal Republic of Germany (B.M.R., VPC.); Department ofAnthropology, University of Calcutta, Calcutta-700 019, India (D.P.M.) KEY WORDS Anthropometry, Dermatoglyphics ABSTRACT Biological affinities between the three endogamous groups of marine fishermen of Puri are examined with the help of nine anthropometric measurements and 22 dermatoglyphic traits of fingers and palms. The results are interpreted in the light of their ethnic, geographical, and migrational backgrounds. Multiple discriminant analysis and Mahalanobis’s generalized distances suggest higher discrimination of populations with reference to anthropometric traits as compared to that in dermatoglyphic variables. While the nature of interpopulation distances conforms to the geographic pattern in anthropometrics, no clear picture emerges in the dermatoglyphic distances supporting either ethnic or geographic evidence. Several groups of marine fishermen live Jalari. The marital exchange even between along the 3,000-mile-longcoast of peninsular the VP and VV is found to be only 1%,while India. They number over a million individu- that of these two with the Jalari is totally als and constitute an important section of the nonexistent. Even in the parental places, Indian peasantry, supplying fish protein. there is no evidence of frequent marital conMost of them still depend primarily on sea tacts between VP and W. However, accordfishing with the help of traditional boats and ing to Rao (1980), the Jalari, along with nets and thus interact directly with the ma- Vadabalija and some other fishing groups, rine environment. Excessive population den- belong to a subcaste of the main caste Vada sity and consequent dwindling of resources Folk stories among them depict Vadabalija has often led these people to emigrate to dif- as boat makers of Jalari who were supposed ferent places in pursuit of better fishing. Such to be originally fishermen. This information a situation is evident in some parts of coastal certainly suggests a long and close associaAndhra Pradesh, and therefore migrant fish- tion between the two castes. However, diverermen from these areas are found in several gence of these two, from a common stock, at places, such as Andamans, Burma, and Puri. least in the remote past, cannot be altogether Over 13,000 marine fishermen live perma- ruled out. These groups present different migrational nently in Puri (Fig. 1) and are collectively called Nulia by the local Oriyas. They live in and geographical backgrounds as well. For four different settlements on the beach. All example, while the VP migrated from about the marine fishermen at Puri are migrants 48 villages of the East Godavari, West Godaand speak Telugu, a language spoken in the vari, and Visakapatnam districts of coastal neighbouring state of Andhra Pradesh. De- Andhra Pradesh (Fig. l), 30 years ago, the mographic investigations among these peo- W and J did so a century ago from 42 and ple (Reddy, 1984) suggest that they belong to 17 villages, respectively, from the Srikakuthree endogamous groups-namely, Vadabal- lam district of Andhra Pradesh and contigija of Penticotta (VP), Vadabalija of Vadapeta uous Ganjam district of Orissa. The W and OW), and Jalari (4, and number roughly 8,000,4,000, and 800, respectively. However, the VP and W belong to the same caste, Received May 14,1986; revision accepted May 12,1987. Vadabalija, which is socially different from 0 1987 ALAN R. LISS, INC. 408 B.M. REDDY, V.P. CHOPRA, AND D.P. MUKHERJEE J thus have geographically overlapping habitats in Puri as well as in their original places. Further, while the status of W and J as local breeding populations is almost established, the VP has still active marital contacts with the parental places, and thus represents a more widely spread VadabaZGa population of coastal Andhra Pradesh. A marked technological variation in terms of boat size and fishing methods also exists between the VP on the one hand and W and J on the other. Puri thus presents a situation in which there are three endogamous groups of fishermen living in a similar environment but with different ethnic, migrational, and geographical backgrounds and with different degrees of reproductive isolation. In the light of these backgrounds, the present paper aims to examine the biological affinities of the people with the help of ecosensitive anthropometric traits and relatively stable dermatoglyphic variables. In the process, it also contributes to the knowledge of biological variation among a special occupational category, the marine fishermen of India, who are hitherto little known, anthropologically. MATERIALS AND METHODS In the years 1977 and 1978,473 adult males 18 to 65 years of age belonging to the three endogamous groups VP, W, and J were investigated. A set of nine anthropometric measurements-stature (St), sitting height (SH), head length (HL), head breadth (HB), nasal height (NH), nasal breadth (NB), biacromial breadth (BAB), bicristal breadth (BCB), and chest girth (CGt-were obtained following methods given in Martin and Saller (1957). All the measurements have been taken by one of us (B.M.R.). In addition to the anthropometric measurements, rolled finger and palm prints of 394 males 8 to 75 years old were collected by the ink and roller method (Cummins and Midlo, 1961).Twentytwo different quantitative variables, following Cummins and Midlo (1961) and Holt (19681, were scored. The variables are ridge counts on individual fingers, radial ridge count, ulnar ridge count, total number of triradii on fingers and of triradii on palms, main line index, and a-b ridge count. Left and right hands are considered separately. Due to unfavourable field conditions, the two sets of data could not be collected from the same individuals. However, there is a considerable overlap of the subjects. Analysis of variance was applied to study the interpopulation variation. The importance of these results is difficult to interpret as the variables included are correlated. Moreover, the human body is not made up of independent varying parts but is a complex, integrated system. Therefore, a multivariate approach was considered proper. The data have been subjected to multiple discriminant anlaysis by using SAS package programs. This method transforms original variables into a set of multivariate vectors which are a linear combination of independent variables. In this process of transformations, correlations between variables are taken into consideration, and the ratio of among-group variance to the total variance is maximized (Tatsuoka, 1971). Mahalanobis’s generalized distance (D2) has also been calculated between the groups. RESULTS Anthropometric measurements Means and standard deviations along with F-ratios and heritability estimates for each of the measurements are presented in Table 1. Except for stature and sitting height, all variables show significant population heterogeneity (P<.05). Although these results indicate that significant intergroup differences exist, the overall pattern of variation is difficult to interpret. The problem lies in the correlation between variables and their treatment one at a time. The data, therefore, were subjected to multiple discriminant analysis. The multivariate approach in this study is regarded as descriptive and heuristic as some of the assumptions made in the analysis may not be fulfilled. Our main concern lies in the pattern of relative differences in the distance between pairs of populations rather than in statistical significance. Table 2 and Figure 2 show the results of discriminant analysis. Multivariate test statistics (Wilk’s lambda) suggest significant differentiation (Table 2) among the groups. The first canonical variate explains the major portion of the total dispersion (94.3%)and is sufficient to differentiate VP from W and J. Standard canonical coefficients show that the variables bicristal breadth, stature, and shoulder breadth are of importance in discriminating the groups on this axis. Although the second canonical variate explains only about 5.6% of the total dispersion, it is also statistically significant. On this axis, W and J are on the two ends, while the VP is in BIOLOGICAL AFFINITIES BETWEEN FISHERMEN, PURI 409 Fig. 1. Map showing the geographical location of Puri and areas in coastal Andhra Pradesh and Orissa from which the fishermen migrated. the middle. Nasal height makes the largest contribution to this canonical variable. Contributions of sitting height and head length and breadth are rather low in discriminating the groups. However, a stepwise discriminant analysis was carried out to judge the discriminatory power of these variables. The order of their contribution in discriminating the groups is given in Table 3. Stature, which has a nonsignificant F-value in the univariate analysis of variance, is significant in the multivariate space. An overall measure of population dissimilarity is Mahalanobis’s generalized distance (D2). Table 4 presents results between pairs of the three populations. The largest distance is observed between the VP and J (6.54) and the smallest is found between the W and J (0.97). However, all the D2 values are statistically significant (P<.01). Here also the separation of VP from W and J is evident. Intergroup differences in a particular situation may be affected more or less by genetic or environmental components or by a combination of both. In order to assess the relative importance of the genetic component in the comparisons, we compared the heritability estimates of the measurements along with the F-ratios of within- and between-group variances. The heritability estimates used in the comparison are from the two populations of fishermen from the places of origin of the present groups. The estimates may, therefore, be considered as a close approximation to the population-specificvalues. Spearman’s rank correlation was computed between the 410 B.M. REDDY, V.P. CHOPRA, AND D.P. MUKHERJEE enced more by the environmental component. However, these results are to be interpreted with caution. Heritability estimates and Fratios are based on univariate analysis and therefore may not reflect precisely the role of variables in discrimination as observed in multivariate analysis, where the correlations between variables are removed. Nevertheless, as a rough approximation to the underlying trend, this analysis may be useful. Further, we divided data into <30- and 1.40 > 30-years of age groups, based primarily on the consideration that the VP migrated to this area only 30 years ago. This would mean that almost all the individuals in the younger age group are born and brought up in the same geographical region of h i and thus have shared the same environment. Discriminant analysis results based on Wilk's lambda (Table 2) and D2 matrix (Table 5) show relatively lesser discrimination for the younger generation when compared to the older one. However, the basic pattern 1 . I remains the same. This is what we would -1.40 -.70 0.0 .70 1.40 expect under the hypothesis of environmencanonical variate 1 tal convergence. But, if we look at the graphical representation of the group separation Fig. 2. Plots of centroids of the populations based on the discriminant analysis of anthropometric variables. on the canonical variates (Fig. 3), we find VP, VadabalGa of Penticotta; W, VadabalGa of Vada- that the change is in the VV group rather petq J , Jalari. than in the VP. This change is difficult t o heritability estimate and the magnitude of univariate F-ratio of the measurements. The correlation coefficient (rs= -50; .05 <P < .lo) indicates that the measurements with lower heritability tend to contribute more to the intergroup differences. This would suggest that the pattern of variation is influ- I TABLE 1. Mean and standard deuiation of the nine anthropometric measurements along with F-ratios for intergroup heterogeneity, and heritability estimates' Variable (mm) Stature Sitting height Head length Head breadth Nasal height Nasal breadth Biacromial breadth Bicristal breadth Chest girth VP (N = 208) Mean f SD VV (N = 200) Mean k SD J ( N = 65) Mean k SD 1,603.89 f 50.86 816.78 +_ 29.73 188.47 5 6.79 144.53 5.11 46.75 k 3.19 36.80 k 2.60 382.14 + 17.52 268.62 E 15.65 875.72 k 44.70 1,608.18 k 55.67 817.37 f 30.03 186.72 ? 6.60 143.98 f 4.60 45.51 f 2.89 37.59 k 2.65 369.21 17.05 247.36 13.48 858.24 ? 49.24 1,618.17 k 49.32 817.05 k 30.68 184.55 8.07 142.74 k 4.78 46.65 k 2.98 37.24 k 2.67 361.02 + 15.18 242.23 12.76 862.79 f 46.52 + + + F-ratio h2 1.83 0.02 0.65' 0.45' 0.643 0.563 0.483 0.463 0.46' 0.33' 0.304 8.81** 3.39* 9.09** 5.14** 50.55** 145.94** 20.25** ' V P , Vadabalija ofpenticotta; W, Vadabalija of Vadapeta;J , Jalari. 'Devi and Reddy (1983). 'Bernard et al. (1980). 4Kapooret al. (1985). *P < .05. **P < .01. TABLE 2. Multivariate test statistics for differentiation among the groups Wilk's lambda F-value df P< a) Anthropometrics < 30 years > 30 years Pooled 0.503 0.368 0.452 0.778 18 by 450 18 by 452 18 by 924 44 by 740 .001 ,001 b) Dermatoglyphics 10.25 16.31 25.04 2.25 Group .001 ,001 411 BIOLOGICAL AFFINITIES BETWEEN FISHERMEN, PURI TABLE 3. Results on the step wise selection of variables that make significant contributions to the group discrimination Sequence 1 2 3 4 5 6 7 Anthropometrics Dermatoglyphics Bicristal breadth Stature Biacromial breadth Nasal breadth Nasal height Chest girth Head length a-b ridge count (L) Triradii on palm (L) Finger ridge count (L2) Finger ridge count (L1) Main line index (R) Main line index (L) Radial ridge count (L) Finger ridge count (R3) Ulnar ridge count (R) Finger ridge count (R5) 8 9 10 m2) TABLE 4. Mahalanobis's distance between fishing groups based on the nine anthroDometric variables VP vv .T - 3.6843* 6.5361* 0.9727" VP vv - J - *P < .05. TABLE 5. Mahalanobis's distance D2)between fishing groups in the young (< 30 years) and adult (> 30 years) generations (Anthropometrics) > 30years VP w VP < 30years w - 2.6879* 5.2126* 6.9130* J - 1.6741* J 6.4317* 1.2395* - *P < .05. 1.40 explain and does not support the assumption of environmental convergence. Since the intermarriage frequency between the VP and (Reddy, 1984), the hypothVV is as low as 1% esis of gene flow from VP to VV causing such a shift can also be safely negated. .70 hl 0) CI m a > 'L 0.0 -m :. C -.70 0 c m 0 -1.40 -1.40 -.70 0.0 .70 1.40 canonical variate 1 Fig. 3. The centroids of the groups in two age groups (circles for <30- and crosses for > 30-years of age category) based on anthropometric variables. Dermatoglyphics The three groups of fishermen of Puri were also investigated for dermatoglyphic traits of the fingers and palms. Procedures for statistical analysis of these data were the same as those used for anthropometric traits, enabling direct comparison of results. Means and standard deviations, heritability estimates and the results of analysis of variance are presented in Table 6. Only ten of 22 variables studied show significant heterogeneity. However, multivariate treatment of the data shows significant discrimination milk's lambda, 0.778; F =2.25, 3Mukherjee(1966). *P < .05. **P< .01. L R a-b ridge count (a-b RC) 'Das Chaudhury and Chopra (1983). 'Mueller and Chopra (1982). L R L Main line index (MLI) R L R L No. of triradii on fingers (Tr.F) Palm No. of triradii on palm (Tr.P) R L R R1 R2 R3 R4 R5 L1 L2 L3 L4 L5 Ulnar ridge count (URC) Radial ridge count (RRC) Finger Finger ridge counts (FRC) Variables 5.86 5.91 9.28 7.79 36.39 37.69 18.51 11.83 13.37 16.38 13.93 16.84 11.07 14.07 16.68 14.18 71.39 70.36 28.62 27.35 6.93 6.84 * 1.06 1.04 1.92 2.19 6.26 6.34 5.91 5.93 4.73 5.50 4.48 6.01 6.17 5.31 5.06 4.31 19.89 21.60 25.55 24.93 1.83 1.87 VP(N = 160) Mean SD 5.91 6.10 9.13 8.20 38.37 40.10 18.26 13.72 13.93 17.07 14.19 16.72 13.79 14.74 17.46 14.78 73.38 74.48 36.69 33.29 7.38 7.26 Mean ~~ 1.06 1.29 1.89 2.00 6.04 5.85 6.91 6.14 5.53 5.55 4.66 7.07 6.06 5.76 5.73 4.56 23.82 23.92 26.62 28.26 1.76 1.92 * SD w (N = 102) 5.52 5.58 8.59 7.53 38.38 40.04 18.49 11.92 12.98 16.65 13.78 16.53 11.67 14.45 17.50 14.30 70.39 72.26 33.85 28.33 7.25 7.00 0.80 0.81 2.05 2.12 5.39 5.06 5.51 6.26 5.80 5.57 4.98 ... 5.45 6.49 6.01 5.71 4.76 22.49 21.83 27.01 26.99 1.83 1.75 J ( N = 132) Mean i SD 5.99** 7.87** 4.67** 2.86** 5.36** 7.91"" 0.06 3.49* 0.93 0.48 0.22 0.10 6.15** 0.45 1.04 0.57 0.55 1.07 3.20* 1.68 2.20" 1.55 F-ratio TABLE 6. Mean and SD o f the 22 dermatoglyphic variables, F-ratios for intergroup heterogeneity, and heritability estimates 0.3B3 0.50' 0.70' 0.65' 0.55' 0.353 0.682 0.70' 0.602 0.582 0.49' 0.46' 0.50' 0.581 0.54' 0.48' 0.65' 0.60' 0.53' 0.521 0.53' 0.56' h' 5m 3! 55 g 3 "k P 3 8 0 ct6 -4 d Eu 5 m 413 BIOLOGICAL AFFINITIES BETWEEN FISHERMEN, PURI 1.40 5200. UEOO. .?O vv cv 0 t., I Q + .% 0.0 m > -m .uC VP -.?O uuoo. 4000. 3600. 0 C m u 3200. -1.40 -1.40 -.70 0.0 .70 canonical variate 1 1.40 2 800. Fig. 4. The centroids of the three fishing groups, based 2400. I I on dermatoglyphic traits. 2000. P<.Ol) between the groups. The group means on the two canonical variates are shown in Figure 4. This and the D2 matrix (Table 7) summarizing overall differences show that the groups are more or less equidistant from each other. There is no clear grouping between the groups reflecting either ethnic or geographical similarity, although the D2 values are all significant. The two canonical variates explain 59% and 41%of the total dispersion. Palmar variables, a-b ridge count, and main line index show maximum influence on the first canonical variate, while the number of triradii of the palm is most important on the second variate. These results show the greater importance of palmar variables, compared to fingers, in the observed intergroup variation. The results of stepwise discriminant analysis (Table 4)support this conclusion. Here also, the a-b ridge count of the left hand is the most important variable in discriminating the groups. We also compared the F-ratios from the analysis of variance for the set of dermatoglyphic traits with the heritability estimates of the related populations. The correlation coefficient (rs) was found to be -.05 and statistically insignificant. There is no definitive pattern to suggest an effect of relative heritability of the traits on population discrimination. As in anthropometric traits, we did not attempt to divide the dermatoglyphic data into age categories for the following reasons: 1600. 1 200. 800. 1100. +--IVP 1 VV J r I VG JV VE Fig. 5. Dendrogram based on the Hiernaux’s distance, computed by using five of the nine anthropometric variables between six samples of the Vudubuliju and Juluri groups living in Puri and Coastal Andhra Pradesh. VG, Vadabalija of Gangavaram; JV,Jalari of Visakapatnam; VB, Vadabalija of Bhimudipatnam. (1)The role of gene flow between the groups is considered negligible, as the rate of intermarriage is observed t o be only 1%between the VP and W and nonexistent between these groups and the J. (2) Environmental convergence is not likely, for the effect of postnatal environment is known to be absent on dermatoglyphic traits. DISCUSSION Multivariate analysis of the data suggests considerable discrimination among the three fishermen groups of Puri (Table 2). The discrimination for the anthropometric variables is greater than that for dermatoglyphic ones. 4 14 B.M. REDDY, V.P. CHOPRA, AND D.P. MUKHERJEE TABLE Z Mahalanobids D2, based on 22 quantitative dermatoglyphic variables between the fishing groups VP vv VP vv J - 0.8703” 0.8227” J - 0.7194* - *P < ,051 Another important observation emerging from these results is that the anthropometric variability among the groups is in agreement with the geographic criterion; the smallest D2 value is between the two sympatric groups, W and J, and the highest value is between VP and J, who belong to different castes and also live in different regions. For the dermatoglyphic traits, the distances are much smaller than those for anthropometric D2 values. Also, the relative differences in the distances between different pairs of populations are too small to reflect clearly either ethnic or geographical patterns. Several investigators (Chai, 1972; Neel et al., 1974; Friedlaender, 1975; Rudan, 1978; Jantz and Chopra, 1983) have earlier observed different patterns of variation for anthropometric and dermatoglyphic traits. This is generally thought to be explained by the relatively smaller influence of the environment on dermatoglyphic traits. Dermatoglyphic characters are generally believed to be selectively neutral. With the exception of Babler’s (1978)findings on prenatal selection of dermatoglyphic patterns the few attempts that have been made to provide evidence contrary to this belief have failed (Van Valen, 1963; Loesch and Wolanski, 1985). If we accept that dermatoglyphic traits are relatively stabIe environmentally, we may assume that at the level of subcastes, with a relatively recent history of separation, the groups differ little genetically. Therefore the distances are small and not much can be said about recent ethnic differentiation. The influence of geographical rather than ethnic proximity on anthropometric variation has been reported in a number of studies (Majumder and Rao, 1960; Hiernaux, 1966; Neel et al., 1974; Rudan, 1978). In this connection, it is of interest to know the nature of variables that contribute most to the intergroup differentiation. Stepwise discriminant analysis (Table 3) reveals that besides stat- ure, breadth measurements are the main contributors t o discrimination. These breadth measurements have relatively lower heritability values than stature and other vertical measurements (Table 1).Though stature has high heritability, it is very sensitive to environmental conditions such as nutrition and socioeconomic factors. On these grounds, we may conclude that the environmental component is of importance in determining the pattern of variation in the present groups. No such pattern was evident in case of dermatoglyphic traits. It is also possible that the large differences in breadth measurements between the VP and the other two populations were, at least in part, a result of occupational selection, since the VP work with boats three to four times bigger than those of the other groups. The VP go much farther out to sea and remain longer than the W and J are equipped to do. This pattern can be traced back to ancestral differences. It would be of interest to know the biological relationship of the VP, VV, and J with their parental groups. Unfortunately, sufficient data from representative segments of the parental groups are not presently available. However, mean values for five (St, HL, HB, NH, and NB) of the nine measurements used in the present study were available for two samples of VadabalGa and a Jalari sample of coastal Andhra Pradesh from where the groups of present study migrated. We have, therefore, resorted to Hiernaux’s method (Hiernaux, 1965) t o compute distances between the six samples. The dendrogram (Fig. 5) conforms to a grouping of populations essentially on the basis of geographical homogeneity. The VP, W, and J, the three Puri groups, cluster closely compared to the Vadabalija of Gangavaram, (VG), Vadabalija of Bhimudipatnam (VB), and Jalari of Visakapatnam (JV), who are living in coastal Andhra Pradesh. By caste, however, the JV and J belong to the Jalari, while the other four belong to the Vadabalija BIOLOGICAL AFFINITIES BETWEEN FISHERMEN, PURI CONCLUSIONS In conclusion, it may be said that the results of our study suggest significant differentiation between subgroups of the fishermen community of Puri. Anthropometric measurements demonstrate higher intergroup variation than the dermatoglyphic ones. The anthropometric pattern of variation documents the importance of geographic proximity of microdifferentiation under the Indian social setup. ACKNOWLEDGMENTS B.M.R. is grateful to the authorities of the Anthropological Survey of India and Indian Statistical Institute, Calcutta, for providing necessary facilities in the initial stages of this work, and to Prof. K.C. Malhotra for inspiration. We thank Mr. Hermann Mueller for computational assistance. Thanks are due to the Alexander-vonHumboldt Foundation for awarding a research fellowship to B.M.R., during which this paper is written. LITERATURE CITED Babler, W J (1978) Prenatal selection and dermatoglyphic patterns. Am. J. Phys. Anthropol. 48:21-28. Bernhard, VW, Brauer, G, Chopra, VP, and Hancke, A (1980) Quantitative genetical analysis of morphological characters of human head and face. J. Hum. Evol. 9t621-626. Chai, CK (1972)Biological distances between indigenous populations of Taiwan. In JS Weiner and J Huizinga (eds): The Assessment of Population Affinities in Man. Oxford: Clarenden Press, pp. 182-210. Cummins, H, and Midlo, C (1961) Finger Prints, Palms and Soles. New York: Dover. Das Chaudhury, AB, and Chopra, VP (1983) Factoral analysis of dermatoglyphics in an Indian Family sample. J. Indian Anthropol. SOC.18t147-153. Devi, MR, and Reddy, GG (1983) Heritability of body measurements among the Jalari population in Visakapatnam. Ann. Hum. Biol. 10:483-485. Friedlaender, JS (1975) Patterns of Human Variation. Cambridge: Harvard University Press, pp. 1-24. Hiernaux, J (1965) Une nouvelle mesure de distance anthropologique entre populations, utilisant simul- 415 tanement des frequences geniques, des pourcentages de traits descriptifs et des moyennes metriques. C. R. Hebd. Seances. Acad. Sci., Paris 26Ot1748-1750. Hiernaux, J (1966) Human biological diversity in Coastal Africa. Man 1.287-306. Holt, SB (1968) The Genetics of Dermal Ridges. Springfield: Charles C. Thomas. Jantz, RL, and Chopra, VP (1983) A comparison of dermatoglyphic methodologies in population studies. Am. J. Phys. Anthropol. 60:61-67. Kapoor, S, Kapoor, AK, Bhalla, R, and Singh, IP (1985) Parent-offspring correlation for body measurements and subcutaneous fat distribution. Hum. Biol. 57t141150. Loesch, DZ, and Wolanski, N (1985) Dermal ridge patterns and fertility in a Polish rural sample. Ann. Hum. Biol. 12463-477. Majumder, DN, and Rao, CR (1960) Racial Elements in Bengal: A Quantitative Study. Calcutta: Asia Publ. House. Martin, R, and Saller, K (1957)Lehrbuch der Anthropologie. Stuttgart: Gustav Fisher. Mukherjee, DP (1966) Inheritance of total number of triradii on fingers, palms, and soles. Ann. Hum. Genet. 29.349-353, Mueller, S, and Chopra, VP (1982) Inheritance of ridge counts: Radial and ulnar counts, asymmetry and diversity. Modern Man, Anthropos/Brno/22:49-55. Neel, JV,Rothhammer, FF, and Lingoes, JC (1974) The genetic structure of a tribal population, the Yanomama Indians: X. Agreement between the representatives of village distances based on different sets of characteristics. Am. J. Hum. Genet. 26281-303. Rao, SN (1980) Socioeconomic and cultural profile of a coastal fishing village in Andhra Pradesh. Man in India 60:221-234. Reddy, BM (1984) Demographic structure of the migrant groups of fishermen of Puri Coast, India. J. Biosoc. Sci. 16t385-398. Rudan, P (1978) Biological structure of the Istrian population-study of distances by anthropometric traits, dermatoglyphic properties and blood group gene frequencies. Coll. Anthropol. 253-58. Tatsuoka, M (1971) Multivariate Analysis: Techniques for Educational and Psychological Research. New York John Wiley. Van Valen, L (1963) Selection in natural populations: Human finger prints. Nature 2OOt1237.