Using the Swedish childhood diabetes register, a nationwide, case-referent study was performed from September 1, 1985 to August 31, 1986. Based on the information from a mailed questionnaire sent to all incident diabetic children and for each diabetic child - two referent children matched according to age, sex, and county, we have analysed perinatal events and aspects of the social environment as possible risk factors for Type 1 (insulin-dependent) diabetes in childhood. A significantly larger proportion of the mothers of the diabetic children were older than 40 years compared to those of the referent children (33% and 24%, p = 0.01 respectively). A smaller percentage of mothers of the diabetic children had a high educational level compared to mothers of referent children (10% and 15%, p = 0.03 respectively) and 39% of the fathers of the diabetic children were manual workers compared to 31% of the fathers of referent children (p = 0.03). Perinatal events did not differ between diabetic and referent children. In children 0-6 years, the duration of breast-feeding was significantly shorter in diabetic children than among referent children (median duration for diabetic children 5 months compared to 6 months for referent children p = 0.03). When considering the presence of Type 1 diabetes among relatives, maternal age over 40 years, low educational level of the mother, and the father being a manual worker as risk factors, the presence of 1 to 4 of any of these risk factors increased the relative risk for Type 1 diabetes cumulatively from 1.2-7.5.(ABSTRACT TRUNCATED AT 250 WORDS)