To investigate the nature of the genetic component controlling susceptibility to leprosy and its subtypes, 402 nuclear families were ascertained through a leprosy patient followed at the Dermatology Hospital in Ho Chi Minh City, Vietnam; 285 families were of Vietnamese origin and 117 were of Chinese origin with a higher proportion of lepromatous forms among Chinese patients. Segregation analyses were conducted using the model developed by Abel and Bonney [(1990) Genet Epidemiol 7:391-407], which accounted for variable age of onset and time-dependent covariates. Three phenotypes were considered: leprosy per se (all forms of leprosy together), nonlepromatous leprosy, and lepromatous leprosy. For each of this phenotype, analyses were performed on the whole sample and separately on the Vietnamese and the Chinese families. The results showed that a single Mendelian gene could not account for the familial distributions of leprosy per se and its two subtypes in the whole sample. However, these results were different according to the ethnic origin of the families. In the Vietnamese subsample, there was evidence for a codominant major gene with residual familial dependences for the leprosy per se phenotype, and borderline rejection of the Mendelian transmission hypothesis for the nonlepromatous phenotype. In Chinese families, strong rejection of Mendelian transmission was obtained in the analysis of leprosy per se, and no evidence for a familial component in the distribution of the nonlepromatous phenotype was observed. For the lepromatous phenotype, the discrimination between models was poor, and no definitive conclusion could be reached. Referring to immunological data, we suggest that these results could be explained by a heterogeneity in the definition of the lepromatous phenotype. It is likely that progress in the understanding of the genetic components involved in the expression of leprosy will come from a better definition of the phenotype under study, and immunological studies are ongoing in this population to investigate this hypothesis.