Factors influencing the prevalence and severity of dysmenorrhoea were assessed longitudinally in a representative sample of young women born in 1962. The prevalence of dysmenorrhoea was lower (P less than 0.01) at 24 years of age than at 19 years of age. At 24 years of age, 67% of the women still experienced dysmenorrhoea; 10% reported dysmenorrhoea which limited daily activity. The severity of dysmenorrhoea (linear analogue scale) was lower (P less than 0.001) at 24 years of age (3.4, SD 2.8) than at 19 years (4.1, SD 3.2). The prevalence and severity of dysmenorrhoea were reduced (P less than 0.05) in women who were parous in 1986 and nulliparous in 1981, but was unchanged in women who were still nulliparous or women who had had a miscarriage or abortion. Dysmenorrhoea was reduced (P less than 0.001) in oral contraceptive users. The severity of dysmenorrhoea was significantly associated with the duration of menstrual flow, menarcheal age and cigarette smoking. The severity of dysmenorrhoea was not associated with age as an isolated factor, nor with height, weight, length of menstrual cycle or frequency of physical exercise.