Women who had participated in a randomised controlled trial of policies of restricted (10%) versus liberal (51%) episiotomy during spontaneous vaginal delivery were recontacted by postal questionnaire three years after delivery. Altogether 674 out of 1000 responded, and there was no evidence of a differential response rate between the two trial groups. Similar numbers of women in the two groups reported further deliveries, almost all of which had been vaginal and spontaneous. Fewer women allocated to restrictive use of episiotomy required perineal suturing after subsequent delivery, but this difference was not significant. Pain during sexual intercourse and incontinence of urine were equally reported in the two groups. The similarity in incontinence rates persisted when severity, type of incontinence, and subsequent deliveries were taken into account. Liberal use of episiotomy does not seem to prevent urinary incontinence or increase long term dyspareunia.