Two groups of patients who requested extradural analgesia were studied in a within-patient controlled trial and received either 0.5% bupivacaine with adrenaline 5 mug/ml or 0.5% bupivacaine plain as the analgesic agent (first group) or 0.5% bupivacaine or 2% lignocaine both with adrenaline 5 mug/ml (second group). Arterial pressure, central venous pressure (CVP), maternal and foetal heart rate and uterine contractions were monitored continuously. There was no significant difference in any of the cardiovascular measurements when solutions of 0.5% bupivacaine (with or without adrenaline 5 mug/ml) were used. The second group had a statistically significant increase in CVP during the study period in which 2% lignocaine was used. The position of the patient did not affect the cardiovascular measurements in either group. Since the addition of adrenaline 5 mug/ml to bupivacaine solutions did not confer any apparent advantages, it is concluded that plain solutions of 0.5% bupivacaine should be used except in the longest labours.