One hundred and eighty one primigravid women delivering vaginally in July and August 1982 in the Rotunda Hospital, Dublin, were randomly allocated to one of two groups. Patients in one group were to undergo episiotomy. Those in the other group were not to undergo episiotomy unless it was considered to be essential. The outcome was compared with that of the clinical practice over the previous six months at the hospital. Of the 92 patients allocated not to undergo episiotomy, seven (8%) had one done for medical reasons compared with 507 (89%) in the previous six months. First degree tears occurred in 23 (25%) and second degree tears in 43 (47%). Nineteen (21%), however, retained an intact perineum compared with only 35 (6%) of the women who had delivered in the preceding six months. Assessments of perineal pain, bruising, swelling, and healing and records of ingestion of analgesics were made for the first four days after delivery, and again at a check up six weeks after delivery, in patients who had had spontaneous vertex deliveries. Forty patients who underwent episiotomy and 37 who sustained a second degree tear formed two comparable groups. There was no difference in outcome between them. Data were also evaluated for 19 women who retained an intact perineum, 22 who sustained a first degree tear, and 11 who underwent episiotomy and epidural anaesthesia; all 52 of these women had spontaneous vertex deliveries. Despite severe soft tissue injury in two patients those who fared best were those who retained an intact perineum. First degree tears were associated with symptoms similar to those associated with second degree tears. Those who fared worst were women who underwent episiotomy after epidural anaesthesia. The value of routine episiotomy in primigravid patients is questioned, but the final decision can be made only by the accoucheur at the time of imminent delivery.