We have compared two methods of reducing hypotension during spinal anaesthesia in elderly patients, 6% hetastarch and crystalloid or methoxamine 10 mg i.m., in terms of haemodynamic stability and requirements for additional vasopressors. Sixty-two patients (aged 60-97 yr) undergoing surgical fixation of fractured neck of femur were allocated randomly to receive 6% hetastarch (Hespan) 500 ml followed by Hartmann's solution 500 ml (group HS, n = 32) or a bolus injection of methoxamine 10 mg i.m. (group MX, n = 30), 10 min before induction of spinal anaesthesia with 0.5% hyperbaric bupivacaine 2.25-3.0 ml. Arterial pressure was measured non-invasively by an oscillotonometer at 2-min intervals from 0 to 40 min and at 5-min intervals thereafter. Methoxamine 2 mg i.v. was given if systolic arterial pressure (SAP) decreased to < 100 mm Hg. Hypotension was defined as a 25% decrease from baseline SAP or mean arterial pressure (MAP). Patient data, sensory level and blood loss were similar in the two groups. SAP and MAP increased initially from baseline until induction of spinal anaesthesia and then decreased for 30 min in both groups, but remained higher in group MX (P < 0.05). Heart rate (HR) decreased from baseline in group MX (P < 0.05) and was less than in group HS at all times from 2 to 60 min (P < 0.01). The incidence of SAP hypotension (47% vs 75%; P = 0.03, odds ratio (OR) = 3.43) and MAP hypotension (47% vs 67%; P = 0.09, OR = 2.51) was less in group MX than in group HS. Requirements for rescue methoxamine i.v. (27% vs 53%, P = 0.04, OR = 3.11) was less in group MX than in group HS but the dose of rescue methoxamine given (mean 6.3 (95% confidence intervals 3.0-9.6) vs 8.9 (5.6-12.2) mg) and time to onset of hypotension (20.7 (14.5-26.7) vs 17.3 (11.4-23.1) min) were similar in groups MX and HS, respectively. We conclude that methoxamine 10 mg i.m., given 10 min before induction of spinal anaesthesia in normovolaemic elderly patients, reduced subsequent SAP and MAP hypotension, HR and requirements for rescue vasopressor therapy compared with a combination of 6% hetastarch 500 ml and crystalloid 500 ml. The previously reported benefit of such volume administration may not extend to the elderly.