This Article Appeared in a Journal Published by Elsevier. the Attached

This article appeared in a journal published by Elsevier. The attached copy is furnished to the author for internal non-commercial research and education use, including for instruction at the authors institution and sharing with colleagues. Other uses, including reproduction and distribution, or selling or licensing copies, or posting to personal, institutional or third party websites are prohibited. In most cases authors are permitted to post their version of the article (e.g. in Word or Tex form) to their personal website or institutional repository. Authors requiring further information regarding Elsevier’s archiving and manuscript policies are encouraged to visit: http://www.elsevier.com/copyright Author's personal copy ARTICLE IN PRESS

l l HOMO — Journal of Comparative Human Biology 59 (2008) 235–251 www.elsevier.de/jchb

Body form and nutritional status among adult males of different social groups in Orissa and Bihar States in India

S. Chakrabartya, M. Palb, S. Bharatic, P. Bharatia,Ã aBiological Anthropology Unit, Indian Statistical Institute, Kolkata 700108, West Bengal, India bEconomic Research Unit, Indian Statistical Institute, Kolkata 700108, West Bengal, India cSociological Research Unit, Indian Statistical Institute, Kolkata 700108, West Bengal, India

Received 12 March 2007; accepted 1 November 2007

Abstract

This paper aims to carry out a biological investigation of the body form and nutritional status of the major social groups of Orissa and Bihar States in India. For this, Cormic Index (CI) and Body Mass Index (BMI) have been computed using data on height, sitting height and weight, taken from adult males of age 18–62 years of various ethnic groups in these two states. The subjects have been classified on the basis of chronic energy deficiency (CED). It is found that a substantial proportion of the people with CED are in the grade II and grade III categories. ANOVA, t-tests, correlation and regression were carried out separately. The results reveal that in Orissa, Scheduled Tribes are shorter, lighter and have lowest mean values of BMI and Cormic Index compared to other groups, but in Bihar, though the Scheduled Tribes are shorter, Scheduled Castes are lower in weight and have the lowest mean values of BMI. There are significant differences in BMI as well as in CI between Scheduled Tribes of Orissa and Bihar. Scheduled Castes and Tribes of Bihar have the highest percentage of CED with 64.71% and 57.45%, respectively. Muslims of Bihar are also affected (52.95%), but overall prevalence of CED is lower in Orissa (49.11%) than in Bihar (54.62%). BMI and CI are highly correlated for each of the social groups in Bihar and Orissa. r 2008 Elsevier GmbH. All rights reserved.

ÃCorresponding author. Tel.: +91 33 25753210; fax: +91 33 25781834. E-mail addresses: [email protected], [email protected] (P. Bharati).

236 S. Chakrabarty et al. / HOMO — Journal of Comparative Human Biology 59 (2008) 235–251

Introduction

Anthropometric measurements have assumed a major role in exploring body form (body size and shape) and nutritional status of populations over time (Gorstein and Akre, 1988; Ruff, 2002). These methods are simple and inexpensive compared to other existing methods (FAO, 1996). Generally, the measurements of body size (height, weight) and relative indices (relative sitting height) reflect body shape (e.g. Body Mass Index). These are often used to measure the nutritional status of a population (WHO, 1995). There are various ways of measuring nutritional status of adults (Lohman et al., 1988), but Body Mass Index (BMI) is most widely used and is suitable for large-scale surveys (James et al., 1988; Ulijaszek and Kerr, 1999; WHO, 1995). It has been proposed that BMI can be used as a practical measure of chronic energy deficiency (Ferro-Luzzi et al., 1992; James et al., 1988). But BMI has several limitations (Banerji et al., 1999; Immink et al., 1992; Norgan, 1990, 1994). One of these limitations is that relative sitting height (sitting height/height) or Cormic Index (CI) indicate, that body proportions may overestimate or underestimate BMI in different populations (Norgan, 1995). Therefore, it is necessary to adjust BMI on CI to correctly understand the population differences based on body shape or BMI (Norgan, 1995). Information regarding population differences based on adult BMI (mean and distribution) can be used for assessing differences in the standard of living between population groups (Nube et al., 1998). That is so, because there is now substantial evidence that low BMI reflects poorer health and reduced physical capacity (James, 1994). A higher mortality rate has been found to be associated with low BMI (Harris et al., 1993). Morbidity and low BMI may be considered as ill health factors (Khongsdier, 2002). A cutoff point of 18.5 kg/m2 for BMI has been accepted for demarcating persons with less than normal nutritional status from persons with normal nutritional status. The poor demographic, economic, social and environmental conditions are also related to undernutrition, which is defined as BMI being less than 18.5 kg/m2 (Pryer and Rogers, 2006; Subramanian and Smith, 2006). Although human populations show a wide variation in height and weight as the major expression of different genetic make-up (Eveleth and Tanner, 1990; Majumder et al., 1990) at the population level, BMI may be considered as the major expression of nutritional consumption rather than genetic predisposition (Rolland-Cachera, 1993). Hence, ‘‘the usefulness of BMI as an anthropometric indicator of nutritional status in a country with diverse ethnic groups like India’’ (Khongsdier, 2001, p. 375) is beyond any question. Eastern parts of India, specifically the States of Orissa and Bihar, are considered to be the poorest States in India in the context of socio-economic and infrastructural development over time. These are manifested by high rate of undernutrition, particularly in rural areas. There are within and between State differences in nutritional conditions among various social groups as well as communities. Each and every social group has their unique body form with respect to size and shape, which may be manifested in the prevalence of undernutrition. It is well established that the Scheduled Tribes and Scheduled Castes in India are socio-economically Author's personal copy ARTICLE IN PRESS

S. Chakrabarty et al. / HOMO — Journal of Comparative Human Biology 59 (2008) 235–251 237 disadvantaged compared to the rest of the Indian population that is, to other castes and Muslims (Ghosh and Bharati, 2006). In the Indian context, large-scale studies have been conducted on South Indian (Ferro-Luzzi et al., 1992), Northeastern (Khongsdier, 2001) and Central Indian populations (Adak et al., 2006) regarding population differences of BMI and nutritional status among different social groups (castes and tribes). But these studies have not considered the relative inﬂuence of body proportions (CI) on body shape (BMI) in understanding population differences. Therefore, in the present study, an attempt has been made to make a general biological examination of body form among adult males of different social groups (Scheduled Castes, Scheduled Tribes, General and Other Castes and Muslims) in the States of Orissa and Bihar in Eastern India and, to show how this body form is manifested by nutritional status.

Population

The Indian caste system is highly complex. There exists a traditional hereditary system of social stratification in India, in which social classes are defined by a hierarchical endogamous Chaturvarna system of social groups. For many years, the Chaturvarna system used to be the descriptive social superstructure of Hindu society in the Indian subcontinent. The Brahmins, a priestly caste, are at the top of this system. Kshatriyas and Vaishyas, rulers and businessmen, are in the middle position, and Sudras are at the bottom. Scheduled Castes and Scheduled Tribal communities have been accorded special status by the Constitution of India. These communities (now within Scheduled Castes and Scheduled Tribes) were previously considered ‘‘outcastes’’ and excluded from the Chaturvarna system. Traditionally, these communities had been regarded as the most menial labourers with no possibility of upward social mobility. They are the most disadvantaged group in India. The Scheduled Tribes have been unable to participate in the mainstream of the Indian society and thus, were deprived of any opportunity for integration with the rest of the society. The Scheduled Castes are also known as Dalits and Scheduled Tribes are referred to as Adivasis. After independence, the Constituent Assembly defined Scheduled Castes and Scheduled Tribes. This body made the President and Governors responsible for compiling a full listing of castes and tribes within these two groups, and also gave them the power to edit this listing later, as required (see Constituent Assembly, articles 341 and 342). The actual complete listing of castes and tribes was made via two orders, The Constitution (Scheduled Castes) Order, 1950, and The Constitution (Scheduled Tribes) Order, 1950, respectively. In terms of indices of development, the States of Orissa and Bihar are considered as disadvantaged states. Orissa has 36.80 million people comprising 16.5% from Scheduled Castes and 22.1% from Scheduled Tribes. The population of Bihar is more than double of that of Orissa with 82.87 million people, of which 15.7% are from Scheduled Castes and 0.9% are from Scheduled Tribes (Census of India, 2001). Author's personal copy ARTICLE IN PRESS

238 S. Chakrabarty et al. / HOMO — Journal of Comparative Human Biology 59 (2008) 235–251

According to the 1991 census of India, the Scheduled Tribal population of Bihar comprised 7.66% of the total state population. This difference between two censi is due to the fact that the major portion of the tribal population from Bihar is now in Jharkhond State, a new Indian state formed from some parts of former Bihar and Orissa territories.

Materials and methods

Scheduled Castes

Altogether 11 Scheduled Caste communities, i.e. Agaria, Ghasi, Keota, Pana and Teli of Orissa and Chamar, Teli, Kumhar, Bauri, Dusadh and Methar of Bihar are taken into consideration in the present study.

Scheduled Tribes

Similarly, altogether 13 Scheduled Tribes communities (seven of Orissa namely Bhuiya, Gond, Khond, Munda, Paroja, Santal, Savara and six of Bihar namely Bhuiya, Munda, Santal, Oraon, Ho, Tharu) have been chosen for the present study.

General and Other Castes

Other than Scheduled Castes and Tribes, a total of 15 General and Other Caste communities are considered for the study, in which Brahmin and Kahandayat communities in Orissa and Brahmin, Rajput, Goala and Yadav communities in Bihar occupy high positions in the social hierarchy.

Muslims

The Muslims have no caste hierarchy. They are taken as a distinct social group having their own Shia and Sunni subdivisions. The anthropometric data presented in this study were obtained from a baseline cross-sectional anthropometric survey conducted in the 1980s by the Anthropolo- gical Survey of India in Orissa (Bhowmik and Basu, 1989) and Bihar (Ghosh et al., 1988). The original data were examined and restructured by one of the authors. The study sample consisted of 9075 adult males, aged 18–62 years. As stated by Basu and Sreenath (1995), the ideal random sampling procedure was not feasible due to the large sample size of this anthropometric survey and hence no strict sampling design was attempted. However, efforts were made to see that in the total sample of the district all the various social and ethnic groups in that district were represented. On the basis of prior information from district headquarters and other available published and unpublished sources about the dispersion and concentration of population groups in a State, samples were drawn from well-deﬁned social groups of Author's personal copy ARTICLE IN PRESS

S. Chakrabarty et al. / HOMO — Journal of Comparative Human Biology 59 (2008) 235–251 239 different social ranks and from different locations of habitations in a district. The samples have been taken on a district level from both urban and rural areas but mostly from rural areas. Educational institutions, ceremonial and political gatherings have been avoided for taking measurements. Adult individuals who were available and cooperated to give measurements were included in the study. Subjects were not chosen on the basis of bodily structure and proportion. However, care was taken to avoid measurements of more than one individual belonging to the same family and to exclude persons with physical deformities (Basu and Sreenath, 1995). The data were collected using standard techniques (Martin and Saller, 1956) and by trained physical anthropologists. The trained investigators visited the sampled households and took physical measurements at the respondents’ residences. The data were collected giving appropriate representations of different social ranks such as castes, tribes and religious groups (Bhowmik and Basu, 1989; Ghosh et al., 1988). Efforts were made to take measurements from apparently normal adult males and to get samples free from any selection bias. In the present study, we have considered only the body height (cm), sitting height (cm), weight (kg) and age as on the last birthday. We computed relative sitting height or Cormic Index (sitting height/height) and also Body Mass Index (weight/height2). The subjects were classiﬁed on the basis of chronic energy deﬁciency (CED) grades (James et al., 1988; WHO, 1995) as follows:

Classiﬁcation Cutoff points (BMI ¼ kg/m2)

CED grade III (severe thinness) o16.00 CED grade II (moderate thinness) 16.00–16.99 CED grade I (mild thinness) 17.00–18.49 Normal 18.50–24.99 Overweight 25.00–29.99 Obese X30.00

The populations were classiﬁed into four major social groups namely General and Other Castes (GN), Scheduled Castes (SC), Scheduled Tribes (ST) and Muslims (MUS). There were 21 communities altogether comprising these four groups in both Orissa and Bihar. We calculated mean and standard deviation of heights, sitting heights, weights, BMI and CI for each community. The means and standard deviations were also found for each of the social groups. It was thought that age and CI may be an interfering variable for understanding social group differences based on BMI. It was thus necessary to remove the effect of age and CI from BMI. To do this, we regressed BMI on age and CI. If the effect of CI is removed from BMI then it should give a better picture about the variation of BMI within and between the communities and social groups. In addition to the adjustment of BMI by age, we can adjust BMI by CI. Within each community/social group there is more homogeneity in food habits, occupational patterns, marriage and other social customs compared to between Author's personal copy ARTICLE IN PRESS

240 S. Chakrabarty et al. / HOMO — Journal of Comparative Human Biology 59 (2008) 235–251 communities/social groups. It is customary among the different communities/social groups to arrange marriage within the communities/social groups. This has genetic implications. If this situation persists for a long time, it will be reﬂected in body form and general level of nutrition. Thus, we expect that variation between the groups will be larger than the variation within the group. To see within and between group variation of BMI, we have carried out ANOVA for BMI, age-adjusted BMI and age and CI adjusted BMI. It may be worth to regress BMI on CI for each community. Since the number of communities is too large, we have reported only the average R2 values that show the goodness-of-ﬁt of regressions. If this average is higher than the R2 considering all communities together then it will imply that the relation between BMI and CI is stronger within each group on the average than when all groups are considered together.

Results

Mean and standard deviation of height, sitting height, weight, Body Mass Index, and Cormic Index of four social groups from the States of Orissa and Bihar (21 communities from each state) were presented in Table 1. Within four major social groups (Scheduled Caste, Scheduled Tribe, General and Other Caste and Muslim) in Orissa, Scheduled Tribal groups were shorter (height), lighter (weight) and had lowest mean values of relative sitting height (CI) compared to other social groups. In Bihar, though the Scheduled Tribal groups were shorter (height) and had the lowest mean values of relative sitting height (CI), Scheduled Caste groups were lighter and had the lowest mean values of BMI compared to other social groups. Bhuiya of Orissa, were the shortest (155.87 cm) among the 42 communities of Orissa and Bihar States. The Yadav of Bihar, were the tallest (166.51 cm). Surprisingly, the next tallest group (165.84 cm) was the Agaria, Scheduled Caste of Orissa and they were also heavier (53.57 kg) than other communities of Orissa. Brahmins of Bihar were the heaviest (53.79 kg) out of all 42 communities. The Mehtar of Bihar had the lowest mean value of weight (41.60 kg) followed by Paroja tribes of Orissa (44.28 kg). The Paroja tribes were also very lean and thin (mean BMI 17.31). All the 42 communities were very lean and the mean BMI varied between 19.75 among Brahmin of Bihar and 16.39 among Bauri, Scheduled Caste of Bihar. The Nulia community of Orissa had comparatively high BMI (19.62) and was much closer to the Brahmin and Rajput communities of Bihar. The lowest value of Cormic Index was among the Ho Scheduled Tribes in Bihar (0.499) with a corresponding value of BMI (17.75), whereas the Brahmin of Orissa showed high Cormic Index (0.516) with the highest mean BMI (19.75). Table 2 shows the distribution of mean BMI by age and social group. A trend of increase in the mean level of BMI with age has been noticed in all cases except in the Scheduled Caste communities and in one case from the Scheduled Tribe communities. The mean BMI values were approximately similar in different age groups among Scheduled Castes and Scheduled Tribes of Orissa and Bihar. Author's personal copy

Table 1. Mean and standard deviations of anthropometric measurements and indices .Carbrye l OO—Junlo oprtv ua ilg 9(08 3–5 241 235–251 (2008) 59 Biology Human Comparative of Journal — HOMO / al. et Chakrabarty S. Population N Height SD Sitting SD Weight SD BMI SD Cormic Index SD (cm) height (cm) (kg)

Orissa Scheduled Caste 1013 162.51 5.88 83.33 3.56 49.94 6.74 18.88 2.17 0.513 0.015 Agaria 150 165.84 5.71 84.97 3.67 53.57 6.05 19.49 2.22 0.512 0.014 Ghasi 200 159.65 4.79 81.52 3.03 47.44 5.88 18.58 1.94 0.510 0.015 Keota 202 163.35 5.40 83.40 2.86 51.00 5.97 19.09 1.91 0.510 0.014 Pana 50 159.48 5.71 81.23 3.06 46.08 5.04 18.09 1.42 0.509 0.017

Teli 411 162.64 5.88 83.84 3.65 49.77 7.15 18.80 2.39 0.515 0.015 PRESS IN ARTICLE Scheduled Tribe 655 158.65 5.47 80.87 3.19 46.59 4.66 18.52 1.75 0.509 0.016 Bhuiya 50 155.87 4.55 79.63 3.11 47.18 4.89 19.41 1.84 0.511 0.014 Gond 99 160.63 5.14 81.42 3.13 46.69 4.16 18.11 1.51 0.506 0.013 Khond 100 156.27 4.69 80.73 2.73 46.77 4.58 19.17 1.94 0.517 0.016 Munda 50 157.22 5.22 78.87 3.65 47.26 4.95 19.11 1.68 0.502 0.019 Paroja 50 160.06 5.77 81.32 3.01 44.28 4.64 17.31 1.84 0.508 0.017 Santal 106 160.27 5.28 81.65 2.97 47.00 4.71 18.28 1.47 0.509 0.014 Savara 200 158.70 5.53 80.96 3.21 46.49 4.65 18.46 1.58 0.510 0.016 General and Other Caste 2030 162.11 6.14 83.19 3.64 49.41 7.04 18.79 2.37 0.513 0.016 Brahmin 657 164.21 5.99 84.95 3.62 51.07 7.95 18.93 2.69 0.517 0.016 Kumbhar 405 160.14 5.69 81.68 3.14 48.03 5.23 18.74 1.98 0.510 0.017 Chasa 100 162.20 5.63 82.64 3.20 49.65 6.45 18.92 2.65 0.509 0.015 Gaura 50 163.36 3.43 82.82 1.74 47.62 5.09 17.86 2.06 0.507 0.013 Karan 200 163.43 6.07 84.43 3.16 51.11 7.79 19.12 2.68 0.516 0.017 Kahandayat 158 162.77 6.02 83.99 3.50 49.87 6.82 18.81 2.23 0.516 0.015 Nulia 205 161.35 5.42 81.52 2.92 51.12 5.04 19.62 1.49 0.505 0.013 Tanti 255 158.70 5.97 81.23 3.09 44.54 5.63 17.68 1.92 0.511 0.013 Muslims 560 163.45 6.05 83.88 3.19 50.88 8.61 19.03 2.92 0.513 0.016 Total 4258 161.85 6.13 82.95 3.64 49.29 7.00 18.80 2.33 0.513 0.016 Author's personal copy 242 Table 1. (continued ) .Carbrye l OO—Junlo oprtv ua ilg 9(08 235–251 (2008) 59 Biology Human Comparative of Journal — HOMO / al. et Chakrabarty S. Population N Height SD Sitting SD Weight SD BMI SD Cormic Index SD (cm) height (cm) (kg)

Bihar Scheduled Caste 1292 160.64 6.41 81.64 3.65 46.69 6.75 18.07 2.17 0.508 0.017 Chamar 397 160.04 6.17 81.51 3.65 45.26 5.96 17.64 1.76 0.509 0.016 Teli 445 161.24 6.68 81.84 3.65 48.94 7.56 18.81 2.62 0.507 0.018 Kumhar 150 161.18 6.11 82.84 3.50 46.22 5.52 17.77 1.74 0.514 0.015 Bauri 50 163.22 5.00 82.15 3.84 43.62 4.08 16.39 1.58 0.503 0.017 Dusadh 200 160.31 6.63 80.83 3.57 46.91 6.38 18.23 2.06 0.504 0.017 RIL NPRESS IN ARTICLE Mehtar 50 157.09 5.23 80.09 2.92 41.6 4.69 16.83 1.40 0.510 0.016 Scheduled Tribe 940 160.09 5.80 81.09 3.72 47.00 5.44 18.33 1.77 0.507 0.016 Bhuiya 196 157.87 6.29 79.09 3.81 46.55 5.27 18.66 1.62 0.501 0.017 Munda 98 158.88 6.41 79.63 2.93 46.82 4.36 18.55 1.49 0.502 0.015 Santal 347 160.74 5.18 81.71 3.17 47.03 5.32 18.21 1.91 0.508 0.016 Oraon 199 160.50 5.42 81.70 3.28 46.26 4.91 17.94 1.59 0.509 0.014 Ho 50 160.21 4.43 79.95 3.28 45.64 5.39 17.75 1.66 0.499 0.016 Tharu 50 164.86 5.24 86.33 3.46 53.36 6.91 19.57 1.79 0.524 0.013 General and Other Caste 1788 163.86 6.81 83.59 3.89 51.16 8.02 19.08 4.19 0.511 0.026 Brahmin 545 165.88 7.51 85.45 3.52 53.79 8.37 19.75 6.69 0.516 0.040 Rajput 250 163.75 6.11 83.15 3.44 52.99 8.86 19.71 2.71 0.507 0.015 Tanti 100 158.96 5.10 80.66 3.56 44.39 6.34 17.55 2.24 0.507 0.015 Ahir 299 163.61 6.03 83.25 3.69 49.29 5.83 18.39 1.72 0.509 0.017 Kurmi 100 162.45 6.03 82.66 4.06 48.58 6.06 18.38 1.71 0.509 0.015 Kayastha 48 163.33 6.85 83.89 3.34 48.38 7.50 18.09 2.27 0.514 0.015 Goala 346 162.11 6.25 81.92 3.75 49.73 7.38 18.89 2.38 0.506 0.018 Yadav 100 166.51 5.99 84.96 3.18 53.47 7.56 19.25 2.18 0.510 0.015 Muslim 797 162.75 6.09 82.90 3.65 49.67 7.16 18.73 2.39 0.509 0.017 Total 4817 162.08 6.60 82.47 3.89 48.91 7.37 18.61 3.09 0.509 0.021 Overall total 9075 161.97 6.38 82.69 3.78 49.09 7.21 18.70 2.76 0.510 0.019 Author's personal copy ARTICLE IN PRESS

S. Chakrabarty et al. / HOMO — Journal of Comparative Human Biology 59 (2008) 235–251 243

Table 2. Age and social group wise mean BMI

State Social group Age group All ages

p25 years 26–50 years 450 years n Mean

n Mean n Mean n Mean

Orissa SC 169 18.42 765 18.98 79 18.93 1013 18.89 ST 121 18.27 492 18.57 42 18.68 655 18.52 GN 409 18.50 1515 18.82 106 19.43 2030 18.79 MUS 137 18.18 390 19.21 33 20.38 560 19.03 Total 836 18.40 3162 18.87 260 19.28 4258 18.80 Bihar SC 378 17.88 797 18.19 117 17.88 1292 18.07 ST 234 18.31 646 18.36 60 18.14 940 18.33 GN 473 18.51 1172 19.17 143 19.22 1788 19.00 MUS 256 18.46 473 18.66 68 20.33 797 18.74 Total 1341 18.29 3088 18.67 388 18.84 4817 18.58

SC, Scheduled Caste; ST, Scheduled Tribe; GN, General and Other Caste; MUS, Muslim.

Table 3. Signiﬁcance levels (t-statistics) of differences of means of Cormic Index (above diagonal) and BMI (below diagonal) for various ethnic groups of Orissa and Bihar States in India

Population Orissa Bihar

SC ST GN MUS SC ST GN MUS

Orissa SC – 3.84** 0.74ns 0.61ns 6.49** 8.55** 2.57* 4.39** ST 3.60** – 4.77** 3.78** 1.86ns 3.89** 4.77** 0.44ns GN 1.10ns 2.66** – 0.07ns 8.37** 10.35** 3.97** 5.56** MUS 1.09ns 3.73** 2.00* – 5.79** 7.55** 0.07ns 4.15** Bihar SC 8.84** 4.55** 8.73** 7.72** – 2.36* 2.47* 1.43ns ST 6.14** 2.11* 5.26** 5.74** 2.99** – 3.99** 3.45** GN 1.37ns 3.30** 2.68** 0.27ns 7.89** 5.22** – 0.96ns MUS 1.39ns 1.91ns 0.52ns 2.01* 6.47** 4.03** 2.16* –

*po0.05; **po0.01; nsp40.05. SC, Scheduled Caste; ST, Scheduled Tribe; GN, General and Other Caste; MUS, Muslim.

The statistical tests of signiﬁcance for differences of mean of BMI as well as Cormic Index between pairs of four social groups (within and between States) are presented in Table 3. It was noticed that there were high differences in CI whenever there were high differences in BMI. It was also found that between-State differences were greater than within-State differences. Apart from a few exceptions, there were Author's personal copy ARTICLE IN PRESS

244 S. Chakrabarty et al. / HOMO — Journal of Comparative Human Biology 59 (2008) 235–251 significant differences in BMI and Cormic Index between most of the pairs. The differences were more prominent between Scheduled Castes and Scheduled Tribes with other groups. Scheduled Castes and Scheduled Tribes of Bihar had the lowest BMI values. There would naturally be very highly significant difference of BMI of Scheduled Castes and Tribes of Bihar if compared with all other communities in Bihar and all communities in Orissa including Scheduled Caste and Tribe communities. General and Other Castes and Muslims were very close with respect to both BMI and CI. In general, we may put Scheduled Castes and Tribes in one group and General and Other Caste and Muslims in another group. The one-way analysis of variance (ANOVA) showed that the variation within and between communities and social groups in Orissa was smaller than those in Bihar (Table 4a). Since it was thought that a part of variation in BMI may be due to variation in age, we have eliminated the effect of age from BMI and performed the ANOVA for the age-adjusted BMI. The results were almost the same as those of unadjusted BMI (Table 4b). Thus, age was not an interfering factor. Since BMI and CI have been related, it was also necessary to further adjust BMI by CI. The result of ANOVA of age and CI adjusted BMI was given in Table 4c. This time, between- group variation had decreased in all cases. Within-group variation had also slightly decreased. As a result, between-group variation compared to within-group variation had decreased in all cases except for communities in Orissa. The connection between CI and BMI prompted us to consider correlation and regression between CI and BMI (Table 5a). The correlation was nonnegative for all the cases and these were found to be significantly different from zero for all cases based on F-test of R2. The fact that the correlation coefficients were positive could be seen from the regression coefficients associated with CI being positive in all the cases. The correlation was the highest among the Scheduled Tribes of Orissa and the lowest among the Scheduled Castes of Bihar. The correlation coefficient was calculated to see whether BMI and CI were linearly related. The regression between the two should also give the same picture. The correlation and regression results have been found to be similar also for age-adjusted BMI (Table 5b). The nutritional status of the social groups will be clearer if we consider the percentage distributions of BMI (Table 6a). Scheduled Castes and Scheduled Tribes of Bihar have the highest number of cases of chronic energy deficiencies (CED) with 64.71% and 57.45%, respectively. Muslims of Bihar are also affected (52.95%). The position of Orissa is much better, for all communities except Scheduled Tribes with a rate of more than 50% of CED, compared to Bihar except for General and Other Castes of Bihar. Thus, it also signifies that there are greater diversities of nutritional status in Bihar. Though the prevalence of CED in the General and Other Caste groups of Orissa is slightly higher (48.37%) than in the General and Other Caste groups of Bihar (46.58%), the overall prevalence of CED of Orissa (49.11%) is lower than that of Bihar (54.62%). Table 6b shows the age and social group distribution of CED. Percentage of CED is seen to decrease as age increases among General and Other Caste and Muslim social groups in both Orissa and Bihar States. The findings corroborate those seen with the mean BMI values given in Table 3. Author's personal copy ARTICLE IN PRESS

S. Chakrabarty et al. / HOMO — Journal of Comparative Human Biology 59 (2008) 235–251 245

Table 4. One-way analysis of variance for BMI (a), age-adjusted BMI (b), and age and CI adjusted BMI (c) among different communities and social groups

Source Degrees of Sum of Mean F-ratio Signiﬁcance freedom squares square level

(a) BMI Orissa Between communities 20 938.34 46.92 9.00 0.000 Within communities 4237 22088.42 5.21 Total 4257 23026.76 Bihar Between communities 20 2110.82 105.54 21.53 0.000 Within communities 4796 23515.25 4.90 Total 4816 25626.07 Orissa Between social groups 3 87.56 29.19 5.41 0.001 Within social groups 4254 22939.21 5.39 Total 4257 23026.76 Bihar Between social groups 3 731.49 243.83 47.14 0.000 Within social groups 4813 24894.58 5.17 Total 4816 25626.07 (b) Age-adjusted BMI Orissa Between communities 20 951.27 47.56 9.21 0.000 Within communities 4237 21890.90 5.17 Total 4257 22842.16 Bihar Between communities 20 2127.64 106.38 21.82 0.000 Within communities 4796 23381.41 4.87 Total 4816 25509.04 Orissa Between social groups 3 91.99 30.66 5.73 0.001 Within social groups 4254 22750.17 5.35 Total 4257 22842.16 Bihar Between social groups 3 717.62 239.21 46.44 0.000 Within social groups 4813 24791.42 5.15 Total 4816 25509.04 Author's personal copy ARTICLE IN PRESS

246 S. Chakrabarty et al. / HOMO — Journal of Comparative Human Biology 59 (2008) 235–251

Table 4. (continued )

Source Degrees of Sum of Mean F-ratio Signiﬁcance freedom squares square level

(c) Age and CI adjusted BMI Orissa Between communities 20 931.33 46.57 9.35 0.000 Within communities 4237 21097.10 4.98 Total 4257 22028.43 Bihar Between communities 20 2001.73 100.09 20.84 0.000 Within communities 4796 23032.64 4.80 Total 4816 25034.37 Orissa Between social groups 3 66.14 22.05 4.27 0.005 Within social groups 4254 21962.29 5.16 Total 4257 22028.43 Bihar Between social groups 3 631.85 210.62 41.54 0.000 Within social groups 4813 24402.52 5.07 Total 4816 25034.37

Discussion

The four major social groups of Orissa and Bihar States in Eastern India differ much in their body proportions (CI) and body shape (BMI). The results are in agreement with the ﬁndings of Adak et al. (2006) who studied social groups in Central India. They have also found that the Scheduled Tribes are shorter than other social groups of the same area. The shortness in height and Cormic Index (CI) may result in lower mean BMI among Scheduled Tribes. It is interesting to note that the mean BMI of adult males of the Munda tribe (18.65) of the peri-urban region of West Bengal State of India (Ghosh and Bharati, 2006) is higher compared to the Munda tribes of Bihar (18.55) but lower than the Munda tribes of Orissa (19.11). These variations may be due to the impact of socio-economic and nutritional conditions across the States in India among the Munda tribes. The mean BMI values are important in predicting morbidity. Khongsdier (2002) found the mean BMI of the War Khasi tribes of Northeast India to be 19.18, among those who reported illness and 20.06 among those who reported no illness. The mean BMI in Orissa (18.80) and in Bihar (18.61), irrespective of their social group, have been found to be lower than that of the Northeastern population (19.14) and adult Indian males Author's personal copy ARTICLE IN PRESS

S. Chakrabarty et al. / HOMO — Journal of Comparative Human Biology 59 (2008) 235–251 247

Table 5. Linear regression of BMI (a) and age-adjusted BMI (b) on CI by social group

State Social group Regression coefﬁcients Goodness-of-ﬁt

Intercept Significance Coefficient Significance R2 Significance of CI

(a) BMI Orissa SC 4.85 0.047 27.36 0.000 0.032 0.000 ST 2.69 0.218 31.06 0.000 0.075 0.000 GN 3.68 0.033 29.44 0.000 0.036 0.000 MUS 6.14 0.134 25.10 0.002 0.017 0.002 Total 3.88 0.001 29.11 0.000 0.036 0.000 Bihar SC 11.54 0.000 12.85 0.001 0.008 0.001 ST 8.77 0.000 18.87 0.000 0.030 0.000 GN 8.84 0.000 19.92 0.000 0.018 0.000 MUS 9.13 0.001 18.85 0.000 0.016 0.000 Total 8.85 0.000 19.12 0.000 0.018 0.000 (b) Age-adjusted BMI Orissa SC 4.55 0.062 27.90 0.000 0.033 0.000 ST 2.73 0.210 30.96 0.000 0.075 0.000 GN 3.57 0.038 29.65 0.000 0.037 0.000 MUS 6.02 0.139 25.38 0.001 0.018 0.001 Total 3.74 0.002 29.36 0.000 0.037 0.000 Bihar SC 11.30 0.000 13.35 0.001 0.009 0.001 ST 8.61 0.000 19.21 0.000 0.030 0.000 GN 8.53 0.000 20.53 0.000 0.019 0.000 MUS 8.42 0.001 20.27 0.000 0.019 0.000 Total 8.53 0.000 19.77 0.000 0.020 0.000

SC, Scheduled Caste; ST, Scheduled Tribe; GN, General and Other Caste; MUS, Muslim.

(18.90) in the years 1988–1990 (Naidu and Rao, 1994). In fact, means of BMI in Orissa and Bihar are lower than those of most other communities in India (Bharati, 1989; Rao et al., 1990, 1995; Reddy, 1998). These values are, however, slightly higher than those of Central Indian population (18.43) and higher than those of South Indian male adults (17.70) (Ferro-Luzzi et al., 1992). Wang et al. (1996) argued that Asian males had a smaller BMI (23.3) than other White, Black and Peurto Rican social groups. It was suggested that BMI is positively correlated with Cormic Index (Norgan, 1994), and the present results also show this relation. The same relations are also found in caste populations in Nepal (Strickland and Tuffrey, 1997). In the Asia Pacific region, the Australian Aborigines (0.475) also show a low value of Cormic Index and BMI and these are positively correlated (Norgan, 1995). Although the influence of nutrition on BMI is more plausible than genetics (Rolland-Cachera, 1993), the significant positive correlation may also be due to Author's personal copy ARTICLE IN PRESS

248 S. Chakrabarty et al. / HOMO — Journal of Comparative Human Biology 59 (2008) 235–251

Table 6a. Percentage distribution of BMI of different communities and social groups of Orissa and Bihar in India

Population CED Grade CED Grade II CED Grade I Normal Overweight III (o16.0) (16.0–16.99) (17.0–18.49) (18.50–24.99) (X25.00)

Orissa Scheduled 6.02 10.86 31.19 50.35 1.58 Caste Agaria 2.00 4.00 32.67 58.67 2.67 Ghasi 8.00 12.00 31.00 48.00 1.00 Keota 4.46 8.42 26.24 60.40 0.50 Pana 2.00 16.00 50.00 32.00 0.00 Teli 7.79 13.38 30.90 45.74 2.19 Scheduled 5.34 12.98 34.20 47.18 0.31 Tribe Bhuiya 2.00 4.00 24.00 68.00 2.00 Gond 7.07 13.13 44.44 35.35 0.00 Khond 5.00 9.00 21.00 64.00 1.00 Munda 2.00 12.00 20.00 66.00 0.00 Paroja 18.00 28.00 34.00 20.00 0.00 Santal 3.77 14.15 45.28 36.79 0.00 Savara 4.00 13.00 36.00 47.00 0.00 General and 9.31 12.36 26.70 49.95 1.67 Other Caste Brahmin 8.98 14.46 25.27 48.40 2.89 Kumbhar 7.41 9.14 30.12 52.35 0.99 Chasa 12.00 11.00 22.00 51.00 4.00 Gaura 22.00 12.00 32.00 34.00 0.00 Karan 8.50 15.50 24.00 49.50 2.50 Kahandayat 9.49 13.29 24.05 52.53 0.63 Nulia 0.00 2.44 20.49 77.07 0.00 Tanti 17.65 17.65 34.51 29.80 0.39 Muslims 9.64 16.43 23.57 46.07 4.29 Total 7.96 12.64 28.51 49.11 1.78 Bihar Scheduled 14.01 16.41 34.29 34.29 1.01 Caste Chamar 16.37 17.88 38.29 27.46 0.00 Teli 8.76 11.91 31.46 45.62 2.25 Kumhar 16.00 22.67 28.00 33.33 0.00 Bauri 32.00 30.00 32.00 6.00 0.00 Dusadh 11.50 13.00 37.50 36.50 1.50 Mehtar 28.00 26.00 36.00 10.00 0.00 Author's personal copy ARTICLE IN PRESS

S. Chakrabarty et al. / HOMO — Journal of Comparative Human Biology 59 (2008) 235–251 249

Table 6a. (continued )

Population CED Grade CED Grade II CED Grade I Normal Overweight III (o16.0) (16.0–16.99) (17.0–18.49) (18.50–24.99) (X25.00)

Scheduled 7.02 14.26 36.17 42.13 0.43 Tribe Bhuiya 4.59 10.71 32.65 51.53 0.51 Munda 3.06 11.22 34.69 51.02 0.00 Santal 9.51 13.54 39.19 36.89 0.86 Oraon 7.04 20.10 39.70 33.17 0.00 Ho 14.00 22.00 34.00 30.00 0.00 Tharu 0.00 8.00 20.00 72.00 0.00 General and 7.77 11.07 27.74 50.73 2.68 Other Caste Brahmin 6.97 9.17 24.77 54.50 4.59 Rajput 3.60 7.60 22.40 61.20 5.20 Tanti 27.00 21.00 22.00 29.00 1.00 Ahir 6.35 13.71 35.79 44.15 0.00 Kurmi 7.00 12.00 35.00 46.00 0.00 Kayastha 14.58 16.67 33.33 33.33 2.08 Goala 8.67 11.27 27.46 50.87 1.73 Yadav 2.00 8.00 30.00 58.00 2.00 Muslim 8.78 14.68 29.49 45.17 1.88 Total 9.47 13.72 31.43 43.72 1.66

Table 6b. Percentage of CED (BMI o18.50 kg/m2) by age and social group

State Social group Age group All ages

p25 years 26–50 years 450 years n %CED

n %CED n %CED n %CED

Orissa SC 169 50.9 765 47.6 79 46.8 1013 48.1 ST 121 62.0 492 50.0 42 54.8 655 52.5 GN 409 51.1 1515 48.1 106 41.5 2030 48.4 MUS 137 60.6 390 47.2 33 33.3 560 49.6 Total 836 54.2 3162 48.2 260 44.2 4258 49.1 Bihar SC 378 69.6 797 61.6 117 70.1 1292 64.7 ST 234 56.8 646 57.0 60 65.0 940 57.4 GN 473 53.9 1172 43.9 143 44.1 1788 46.6 MUS 256 58.2 473 53.7 68 27.9 797 52.9 Total 1341 59.7 3088 52.7 388 52.3 4817 54.6

SC, Scheduled Caste; ST, Scheduled Tribe; GN, General and Other Caste; MUS, Muslim. Author's personal copy ARTICLE IN PRESS

250 S. Chakrabarty et al. / HOMO — Journal of Comparative Human Biology 59 (2008) 235–251 other environmental and genetic factors (Khongsdier, 2001). Most likely, it is a result of particular body form, possibly a somatotype. Individuals with large trunks and short legs will have different values of both indices from those individuals whose trunks are small and legs long. A higher correlation between BMI and CI possibly implies that each community maintains a separate identity from other groups. Lean and thin body with a low weight to surface area ratio is one of the general characteristics of the people living in tropical and subtropical climates (Schreider, 1968). This may be one of the reasons for the high prevalence of formally classiﬁed CED grade I in most of the communities in Orissa and Bihar States in India as mentioned by others (Adak et al., 2006; Khongsdier, 2001). A substantial proportion of CED grade II and III males of Orissa and Bihar States, however, indicate the high prevalence of undernutrition, speciﬁcally in the Scheduled Castes and Scheduled Tribal groups, which may be due to poor socio-economic conditions.

References

Adak, D.K., Gautam, R.K., Bharati, S., Gharami, A.K., Pal, M., Bharati, P., 2006. Body mass index and chronic energy deficiency of adult males of central Indian populations. Hum. Biol. 78, 201–218. Banerji, M.A., Faridi, N., Atluri, R., Chaiken, R.L., Lebovitz, H.E., 1999. Body composition, visceral fat, leptin, and insulin resistance in Asian Indian men. J. Clin. Endocrin. Metab. 84, 137–144. Basu, A., Sreenath, J., 1995. Anthropometric Variations in Assam, Bihar and Orissa. Anthropological Survey of India, Calcutta. Bharati, P., 1989. Variation in adult body dimensions in relation to economic condition among the Mahishyas of Howrah district, West Bengal, India. Ann. Hum. Biol. 16, 529–541. Bhowmik, D.C., Basu, A., 1989. All India Anthropometric Survey, North Zone, Basic Anthropometric Data, Vol. 3, Orissa, Calcutta. Anthropological Survey of India, Calcutta. Census of India, 2001. Registrar General and Census Commissioner. Controller of Publication, New Dehli. Eveleth, P.B., Tanner, J.M., 1990. Worldwide Variation in Human Growth, second ed. Cambridge University Press, Cambridge. Ferro-Luzzi, A., Sette, S., Franklin, M., James, W.P.T., 1992. A simplified approach of assessing adult chronic energy deficiency. Eur. J. Clin. Nutr. 46, 173–186. Food and Agriculture Organization (FAO), 1996. Sixth World Food Survey. Food and Agriculture Organization, Rome. Ghosh, R., Bharati, P., 2006. Nutritional status of adults among Munda and Pod populations in a peri urban area of Kolkata City, India. Asia Pac. J. Public Health 18, 12–20. Ghosh, G.C., Bose, M.P., Kumar, G.D., Mody, G.S., Bose, D.K., Huq, F., Basu, A., 1988. All India Anthropometric Survey, North Zone, Basic Anthropometric Data, Vol. 2, Bihar. Anthropological Survey of India, Calcutta. Gorstein, J., Akre, J., 1988. The use of anthropometry to assess nutritional status. World Health Stat. Q. 41, 48–58. Harris, T.B., Ballard-Barbasch, R., Madan, J., Makuc, D.M., Feldman, J.J., 1993. Overweight, weight loss, and risk of coronary heart disease in older women. The NHANES I Follow-up Study. Am. J. Epidemiol. 137, 1318–1327. Immink, M.D.C., Flores, R., Diaz, E.O., 1992. Body mass index, body composition and the chronic energy deficiency classification of rural adult populations in Guatemala. Eur. J. Clin. Nutr. 46, 41–49. James, W.P.T., 1994. Introduction: the challenge of adult chronic energy deficiency. Eur. J. Clin. Nutr. 48, S1–S9. James, W.P.T., Ferro-Luzzi, A., Waterlow, J.C., 1988. Definition of chronic energy deficiency in adults. Eur. J. Clin. Nutr. 42, 969–981. Khongsdier, R., 2001. Body mass index of adult males in 12 populations of northeast India. Ann. Hum. Biol. 28, 374–383. Khongsdier, R., 2002. Body mass index and morbidity in adult males of the War Khasi in northeast India. Eur. J. Clin. Nutr. 56, 484–489. Author's personal copy ARTICLE IN PRESS

S. Chakrabarty et al. / HOMO — Journal of Comparative Human Biology 59 (2008) 235–251 251

Lohman, T.G., Roche, A.F., Martorell, R., 1988. Anthropometric Standardization Reference Manual. Human Kinetics Books, Chicago. Majumder, P.P., Shanker, B.U., Basu, A., Malhotra, K.C., Gupta, R., Mukhopaddhyay, B., Vijayakumar, M., Roy, S.K., 1990. Anthropometric variation in India: a statistical appraisal. Curr. Anthropol. 31, 94–103. Martin, E., Saller, K., 1956. Lehrbuch der Anthropologie, vol. 3. Fisher, Stuttgart. Naidu, A.N., Rao, N.P., 1994. Body mass index: a measure of the nutritional status in Indian populations. Eur. J. Clin. Nutr. 48, S131–S140. Norgan, N.G., 1990. Body mass index and body energy stores in developing countries. Eur. J. Clin. Nutr. 44, S79–S84. Norgan, N.G., 1994. Relative sitting height and the interpretation of body mass index. Ann. Hum. Biol. 31, 79–82. Norgan, N.G., 1995. Body mass index and nutritional status: the effect of adjusting body mass index for the relative sitting height on estimates of the prevalence of chronic energy deﬁciency, overweight and obesity. Asia Pac. J. Clin. Nutr. 4, 137–139. Nube, M., Asenso-Okyere, W.K., van den Boom, G.J.M., 1998. Body mass index as indicator of standard of living in developing countries. Eur. J. Clin. Nutr. 52, 136–144. Pryer, J.A., Rogers, S., 2006. Epidemiology of undernutrition in adults in Dhaka slum households, Bangladesh. Eur. J. Clin. Nutr. 60, 815–822. Rao, K.V., Balakrishna, N., Thimmayama, B.V.S., Rao, P., 1990. Indices and critical limits of malnutrition for use among adults. Man in India 70, 351–367. Rao, K.V., Balakrishna, N., Shatrugna, V., 1995. Variations in forms of malnutrition in well-to-do adults and the associated factors. Man in India 75, 241–249. Reddy, B.R., 1998. Body mass index and its association with socioeconomic and behavioural variables among socioeconomically heterogeneous populations of Andhra Pradesh, India. Hum. Biol. 70, 901–917. Rolland-Cachera, M.F., 1993. Body composition during adolescence: methods, limitations, and determinants. Hormone Res. 39, S25–S40. Ruff, C., 2002. Variation in human body size and shape. Ann. Rev. Anthropol. 31, 211–232. Schreider, E., 1968. Ecological rules, body heat regulations and human evolution. In: Bresler, J.B. (Ed.), Environments of Man. Addison-Wesley, Reading, MA, pp. 55–66. Strickland, S.S., Tuffrey, V.R., 1997. Form and Function: A Study of Nutrition, Adaptation and Social Inequality in Three Gurang Villages of the Nepal Himalayas. Smith-Gordon, London. Subramanian, S.V., Smith, G.D., 2006. Patterns, distribution, and determinants of under- and overnutrition: a population-based study of women in India. Am. J. Clin. Nutr. 84, 633–640. Ulijaszek, S.J., Kerr, D.A., 1999. Anthropometric measurement error and the assessment of nutritional status. Br. J. Nutr. 82, 165–177. Wang, J., Thornton, J.C., Burastero, S., Shen, J., Tanenbaum, S., Heymsﬁeld, S.B., Pierson, R.N., 1996. Comparisons for body mass index and body fat percent among Puerto Tacians, Blacks, Whites and Asians living in the New York City area. Obes. Res. 4, 377–384. World Health Organization (WHO), 1995. Physical status. The use and interpretation of anthropometry. WHO Technical Report No. 854. WHO, Geneva.