To determine the concordance rate for Type II (non-insulin-dependent) diabetes mellitus in monozygotic twin pairs, initially ascertained discordant for diabetes, we carried out a prospective study on 44 non-diabetic subjects, each of whom had a sibling twin with diabetes (21 men, 23 women, median age 55 years, interquartile range 47-65). The subjects were referred as discordant for Type II diabetes. The twin pairs were part of the British Diabetic Twin Study and ascertained between May 1968 and January 1998. These subjects underwent an OGTT at time of referral and periodically thereafter. The mean follow-up was 8 years (range 0-18 years) and data were collected until January 1996. The percentage of twins who developed Type II diabetes was assessed by standard actuarial life-table methods and the pairwise concordance rate, that is the proportion of concordant pairs over the sum of concordant and discordant pairs, was calculated. The observed rates of concordance for Type II diabetes at 1, 5, 10, and 15 years follow-up were 17, 33, 57, and 76%, respectively. The concordance rate for any abnormality of glucose metabolism (either Type II diabetes or impaired glucose tolerance) at 15 years follow-up was 96%. The concordance rate for Type II diabetes in monozygotic twins is very high even in twins initially ascertained discordant for diabetes.