Fifty cases (24 males, 26 female) of gastroschisis were admitted to the Regional Neonatal Surgical Unit at the Royal Hospital for Sick Children in Yorkhill, Glasgow, between January 1983 and October 1993. The mean birthweight was 2.17 kgs and mean gestation was 35.8 weeks. Prenatal diagnosis was made by ultrasound scan in 39 cases (78%) and 33 (66%) were delivered in the Regional Obstetric Unit and Fetal Medicine Centre at the Queen Mother's Hospital, Yorkhill, Glasgow, immediately adjacent to the neonatal surgical unit. Spontaneous vaginal delivery occurred in 23 (46%) but Caesarean section was performed in 27 (54%). Sixteen Caesarean sections were emergencies and 11 elective (5 on account of dilated bowel loops). Associated bowel problems were noted in 11 (22%)--atresia (5), stenosis (2), ischaemia (1), duodenal perforation (1), ileal perforation (1) and jejunal band obstruction (1). Primary closure was achieved in 42 (84%) and there were 5 deaths (10%). Prenatal diagnosis and mode of delivery showed no direct correlation with mortality (10%) in this series. Emergency Caesarean section was necessary for fetal indications in 28%. The prediction of compromised or damaged intestine by prenatal ultrasound was unreliable and the majority of cases of atresia and stenosis escaped detection. However, it is concluded that prenatal diagnosis and delivery in a regional centre offers the best outcome for the fetus. Further study is needed to define "at risk" pregnancies and fetuses and to determine the role of intervention.