We have compared the haemodynamic effects of ephedrine alone with ephedrine and colloid for the treatment of hypotension produced by subarachnoid anaesthesia in 30 patients aged 60-90 yr with fractures of the neck of femur. Group one received ephedrine as an initial bolus dose of 0.2 mg kg-1 followed by an infusion of 0.5 mg kg-1 h-1. Group two received ephedrine and colloid (polygeline, Haemaccel) 8 ml kg-1. If necessary, up to three rescue bolus doses of ephedrine (0.1 mg kg-1) and then colloid solution (8 ml kg-1) were given to maintain systolic arterial pressure (SAP) at > 75% of baseline. Arterial pressure was measured by automated oscillotonometry, central venous pressure (CVP) by a manometer and cardiac index (CI), stroke index (SI) and heart rate (HR) by transthoracic electrical bioimpedance. Systemic vascular resistance index (SVRI) was derived. In patients receiving ephedrine only, SVRI, CVP and SI decreased and HR increased (P < 0.0001). Five patients in this group required colloid, the effect of which was to restore CVP, increase CI and SI, and decrease HR (P < 0.02). In patients receiving ephedrine and colloid solution, SVRI decreased and CI, SI and HR increased (P < 0.0001). Ephedrine was not a potent arterial vasoconstrictor and SAP was maintained mainly by increases in SI and HR.