Data from a computerised linkage between the Swedish Cancer Registry and the Fertility Registry were used to study the influence of reproductive history on the risk of colorectal cancer in women aged less than 60, in the period 1960 to 1984. A "nested" case-control design was used with 1,415 colon cancer cases and 733 rectal cancer cases, and 7,073 and 3,665 age-matched controls, respectively. Conditional logistic regression was employed to study the independent effects of parity, age at first birth and years since last birth on the risk of colorectal cancer. High parity was associated with reduced risk in women older then 50 years at diagnosis, both for the colon as a whole and for colon subsites. Women having had 5 children or more had an odds of colon cancer of 0.44 (95% confidence interval 0.24-0.82) relative to that of nulliparous woman. Contrary to the hypothesis advanced by McMichael and Potter (J. Nat. Cancer Inst., 75, 185-191 [1985]), which predicted a strong protective effect of parity in the proximal compared to the distal colon, we found the reverse. Increasing years since last birth seemed to be associated with a reduced risk of rectal cancer. A significant positive trend with increasing age at first birth was suggested with respect to distal colon cancer and rectal cancer risk. The mutually adjusted effects of the reproductive variables under study were similar to the unadjusted effects. Our data suggest that increasing parity is associated with a decline in risk of colon cancer but not rectal cancer, although the association with colon cancer is restricted to women aged 50 or more. The effects of age at first birth and years since last birth are not consistent or pronounced.