Abstract:The relationship between body mass index (BMI) and percent body fat (PBF) with central adiposity measures varies between ethnic groups and it has not been much studied in low socio-economic groups in India. Therefore, this study was undertaken to test the relative efficacy of waist circumference (WC), hip circumference (HC), waist hip ratio (WHR) and conicity index (CI) to predict BMI and PBF among Bengalee Hindu male slum dwellers. A total of 465 adult (aged 18-72 years) male slum dwellers of Bengalee Hindu ethnicity were included. Standard anthropometric techniques and formulae were used. WC showed the strongest significant (p < 0.001) partial correlation with BMI and PBF (0.82 and 0.77, respectively). Stepwise multiple linear regression analyses of BMI and PBF with the central adiposity measures revealed that WC had the strongest impact on BMI and PBF. On the other hand WHR, HC and CI had weaker effect. Moreover, WC alone accounted for about 67 and 60 per cent of the variations, respectively, of BMI and PBF. The models involving only WC and CI, explained almost similar proportions of variations (adjR 2 = 94.1, and 69.4, respectively).Furthermore, the WC was found to best predict BMI (R 2 Change = 0.666; F Change = 921.32, p < 0.001) and PBF (R 2 Change = 0.597; F Change = 686.66, p < 0.001). Even after controlling for each other, i.e., BMI and PBF (results not shown) WC had the strongest significant impact on these two measures. Therefore, in this population, WC may be preferred over other measures of central adiposity in studies dealing with obesity and cardio vascular disease risk factors.