A survey of the mental morbidity of an urban group and two rural groups was made with the same method, same operational definition of a case and by the same team. The aim was to find out the nature and extent of the difference in their rates of morbidity and to identify the psychosocial variables associated with this difference. It was found that the rate of total morbidity was significantly higher in the urban group than in the rural groups. Psychosis was, however, commoner among the Brahmins, a rural group. The wide difference in the rates of mental morbidity between the urban and rural groups was mainly due to the difference in the rate of neurosis (165.3/1000, 51.6/1000 and 1.5/1000, respectively). The rate of neurosis in turn was considered to be positively correlated with certain psychosocial characteristics irrespective of urban or rural residence of the group concerned.