![]() |
![]() |
---|---|
The Metric Traits Hemphill and Lukacs (1993) used tooth crown diameters to compare Garasias, Bhils, and Vaghela Rajputs from Gujarat, as well as ten pan-Indian ethnic groups. Their research was designed to explore three main questions: do tooth sizes vary according to social groups in India? Can the variation be used to map genetic relatedness? Do the patterns of size variation correspond to evidence from other indicators of relatedness? Their research was placed within a broader biocultural perspective which sought to bridge the inconsistencies between the historical and archaeological records, to test theories on the origins and evolution of the caste system itself, as well as testing the relationship between tooth size and major subsistence classifications. Many of the available anthropometric and serological analyses have been seen as confirmation of traditional social divisions. Genetic distance has been interpreted as furthest between caste and tribal groups, Indo-Aryan and Dravidian language speakers, and finally between tribal groups associated with north and east versus south and west India. However, Srinivas (1956, 1968) redefined the observed disparity as that of high and low status, a distinction which is generally determined by the level of Hinduization accepted by the group and so may follow caste versus tribe divisions to an extent. The concept of Hinduization is important to the definition because it incorporates a more realistic version of a flexible caste system and mobility of groups. Hinduization acts as a vehicle for upward mobility and increased status but simultaneously ensures some degree of preservation of tribal genetic marker patterning by adherence to the Hindu system of endogamous marriage. For an introduction to the caste system and discussion of Hinduization please refer to the first two sections in the site: Background Three Caste Groups in Gujarat The samples for each study group consisted of 100 males and 100 females generally students between 10-25 years old. The female sample for the Rajputs was smaller because of purdah practices in that group. Seventeen anthropometric measurements, dental casts, and one odontometric control measurement were collected for each subject. The control measurement was compared with a measurement on the same dimension for each cast to assess any possibility of error being introduced by the casting method. Intraobserver error was assessed by repeating 30 random measurements 10 months after the originals were conducted and the mean difference was -0.06 mm (standard deviation of 0.18 mm). Univariate and inferential statistics were performed on the metric data by group and by sex. The left side of the dentition was used for the measurements except in cases where tooth sizes were significantly different between left and right antimeres, in which case the average for the two teeth was used. In cases where the left tooth was missing the right antimere was used. Individuals with dental pathologies and/or severe occlusal wear were eliminated from the sample. Comparison by social group tested for significant intergroup size differences, level of difference between Hindu and tribal groups, level of difference between high and low status groups, and accuracy of predictions from the historical record that Garasias are genetically intermediate between Rajputs and Bhils. The sex standardized samples were compared first with a cluster analysis to compare general patterns. Principle component analysis was then used to understand the influence of specific variables and to calculate the level of separation contributed. The results indicate that Garasia tooth size is indeed intermediate between Rajputs and Bhils though Garasia males are closer to the Bhils than the Rajputs. Intergroup differences were greater than intragroup sex dimorphism. The parental Bhil and Rajput groups were the furthest from each other within the Gujarat sample. Both sexes of all three Gujarati groups showed significant fluctuating asymmetry between left and right antimeres suggesting the value of using the average of the two sides. Principal components analysis suggests that over half the variance between Gujarati males and females is accounted for by 3 variables: overall size, difference between mesiobuccal and buccolingual dimensions, and contrast between anterior and posterior teeth. Intergroup similarity was compared using data collected by Bhasin, et al. (1985) for the Bodhs, Khatris, Jats, Ahirs, U.P. Brahmins, Dangis, Kunbis, Varlis, Bengali Brahmins, and Tamil Brahmins. To reduce interobserver error, the measurements were taken again from only the left side as Bhasin, et al. (1985) had done. The Bhils and Garasias were again the closest two groups, then the Rajputs. Similarity to the pan-Indian groups appeared to be correlated with geographic proximity. In the principle components analysis again three variables accounted for over half the difference between the thirteen groups: a difference in the variable with the largest dimension among the posterior teeth, the same type of dimensional difference existed in the anterior teeth, and in comparisons between anterior and posterior teeth. One interesting difference in the results obtained in this part of the analysis is that Gujarati males clustered closer to Rajput males than to Bhils. The three Gujarati groups were clustered most closely in the pan-Indian comparison. The three Gujarati groups showed the most resemblance to each other and the next closely related category was that of geographically proximal groups in Maharashtra and Haryana. Tribal Bhils showed more similarity to proximal caste groups than to other tribals suggesting long intermixing and Hinduization. The Brahmin groups were also distant from each other except for neighboring Uttar Pradesh and Tamil Nadu groups. The Bodhs and Varlis presented an exception to the general pattern. These two tribal groups are spatially distant but show evidence of relatedness. Three results of this analysis were surprising to Hemphill and Lukacs: Tamil Nadu Brahmins (high caste, southern India) clustered close to the Bodhs (tribal group, north- western India), the Varlis (another tribe from northwestern India) clustered close to Khatris and Uttar Pradesh Brahmins (two high status castes from northern India), and the Dangis (farmers from northeastern India) were close to Bengali Brahmins (high status caste also from northeastern India).