Data on alcohol use and smoking habits was available from the 1975 questionnaire of the entire cohort. Prior to pairwise analyses, the data of individuals was compared to that of age-sex matched groups of pairs reared together. The early separated twins had a higher alcohol consumption, while for smoking only slight differences were observed compared to twins reared together. Probandwise concordance rates were computed from smoking status (ever smoker/never smoker), alcohol use (user/nonuser) and "heavy" drinking (half-bottle of spirits on one occasion at least once a month). The following results were obtained in those pairs with the environmental dissimilarity score greater than 15: (table; see text)