Sixty nulliparous women received epidural infusions in labour of either 0.125% plain bupivacaine or 0.0625% bupivacaine containing 2.5 mcg/ml fentanyl both starting at 12 ml/h and titrated to maintain a sensory block to T10. Those women who received low dose bupivacaine with fentanyl took significantly longer to reach full cervical dilation (P < 0.05). There was no statistical difference between the groups in the number of additional epidural bolus doses required during the infusions. Of the women receiving low dose bupivacaine with fentanyl, 77% required one or no additional top-up doses compared to 87% in the plain bupivacaine group. There was no reduction in the incidence of perineal pain in the group who received fentanyl. Significantly more women who received 0.125% bupivacaine had motor block after 2 h of the infusion (P < 0.05). The mode of delivery was similar in the two groups as was the satisfaction with epidural analgesia in both the first and second stages of labour and with labour overall. There were no significant differences in Apgar scores, umbilical cord blood pH levels or neurologic and adaptive capacity scores at 2 or 24 h. There was no significant difference in the incidence of symptoms 24 h after delivery.