OBJECTIVE--To determine whether the high rate of forceps delivery associated with the use of epidural analgesia could be reduced through giving an intravenous infusion of oxytocin during the second stage of labour. DESIGN--A randomised, double blind, placebo controlled trial. SETTING--Delivery suites in three hospitals. SUBJECTS--226 Primiparous women with adequate epidural analgesia in whom full dilatation of the cervix had been achieved without prior stimulation with oxytocin. INTERVENTION--An infusion of oxytocin or placebo starting at the diagnosis of full cervical dilatation at an initial dose rate of 2 mU/min increasing to a maximum of 16 mU/min. MAIN OUTCOME MEASURES--The outcome of labour was assessed in terms of the duration of the second stage, mode of delivery, fetal condition at birth, postpartum blood loss, and the incidence of perineal trauma. RESULTS--Treatment with oxytocin was associated with a shorter second stage (p = 0.01), a reduction in the number of non-rotational forceps deliveries (p = 0.03), and less perineal trauma (p = 0.03) but was not associated with any reduction in the number of rotational forceps deliveries performed for malposition of the occiput. No adverse effects on fetal condition at birth or in the early puerperium were seen in association with the use of oxytocin. CONCLUSIONS--The use of an oxytocin infusion may reduce the high rate of operative delivery associated with epidural analgesia provided that the fetal occiput is in an anterior position at the onset of the second stage of labour but within the dose range studied does not seem to correct malposition of the fetal occiput.