Maternal cardiovascular changes and acid-base status were in 29 women undergoing lower Caesarean section under spinal anaesthesia. The patients were allocated randomly to one of three groups to receive an i.v. infusion of one of the following: ephedrine 1 mg min(-1) (group E1: n = 10), ephedrine 2 mg min(-1) (group E2: n = 9), or phenylephrine 10 mu g min(-1) (group P: n = 10). Invasive arterial pressure was monitored continuously and if hypotension occurred (defined as a 20% decrease from baseline, taken after i.v. preload administration), bolus doses of either ephedrine (6 mg in groups E1 and E2) or phenylephrine (20 mu g in group P) were given. Only four patients became hypotensive in group E2, compared with eight patients in group E1 and nine patients in group P. The total time that the patients remained hypotensive was greatest in group P (P < 0.005), less in group E1 and least in group E2. Neonatal Apgar scores and acid-base profiles were similar in all three groups. In this study, an infusion of phenylephrine 10 mu g min(-1) with bolus doses of 20 mu g was shown to be significantly less effective in maintaining systolic arterial pressure within 20% limits of baseline compared with an infusion of ephedrine 1 or 2 mg min(-1) with bolus doses of 6 mg