The aetiology of colorectal cancer was studied in a cohort of 11,888 residents of a retirement community. After four and one-half years of follow-up, 58 male and 68 female incident colorectal cancers were identified. Daily alcohol drinkers experienced nearly a two-fold increase in risk (2 sided P = 0.002). Colorectal cancer was also positively associated with Quetelet's index and inversely associated with avocational physical activity. The results were consistent for both sexes but were statistically significant only for males. With the exception of dietary vitamin C, none of the nutrients under study (i.e., vitamins A and E, dietary fibre, calcium, and beta carotene) showed a significant association with colorectal cancer. An inverse relationship between colorectal cancer and dietary vitamin C was observed in females, but there was no association with either vitamin C from supplements or with total vitamin C intake. Males and females who had 3 or more children showed a significantly reduced risk of colorectal cancer (RR = 0.45, 95% CI = 0.2, 0.9), but those with no children did not show the highest risk.