A randomized, double-blind, double-crossover, placebo-controlled haemodynamic study was undertaken in patients with Grade II/III (NYHA) cardiac failure to examine the acute effects of intravenous alifedrine, 20 mg and 40 mg (17 patients), and oral 40 mg alifedrine (8 of these patients). Patients received single doses of alifedrine and placebo on separate days, with invasive monitoring. Alifedrine resulted in a significant (p less than 0.001), dose-dependent increase in cardiac output. The peak effect (+23% with 20 mg i.v., +42% with 40 mg i.v. and +29% with 40 mg orally) was seen approximately 1 hour after intravenous administration (with about half of these increases still apparent at 3 hours) but developed progressively over 3 hours after oral administration. There were significant reductions (p less than 0.001) in peripheral resistance (peak mean changes -21% with 20 mg i.v., -31% with 40 mg i.v. and -23% with 40 mg orally), but little (less than +/- 6%) observed change in arterial pressure. With intravenous alifedrine, there were significant increases in stroke volume (+19% with 20 mg, +35% with 40 mg, p less than 0.001) with little (5%) change in heart rate (+3% and +7%, respectively, N.S.). With the 40 mg oral dose, there was a small increase in heart rate (+12%, p less than 0.005) associated with a 19% (N.S.) increase in stroke volume. Peak haemodynamic responses to 40 mg alifedrine orally were 50% to 75% of those seen after administration of the same dose intravenously. When assessed 3 hours after administration, responses to the two routes of administration were similar. There were no clinically or statistically significant changes in arterial (non-invasive), pulmonary artery, pulmonary capillary or right atrial pressures with any dose of alifedrine. No significant arrhythmias were noted clinically with the doses studied. Alifedrine, therefore, is an interesting agent, available both orally and intravenously, which is well tolerated and appears to produce marked acute increases in cardiac output with little change in heart rate or blood pressure. Further studies should determine whether these effects are maintained during longer-term therapy and clarify the relative contributions of positive inotropic and peripheral vasodilator activity to the effects observed.