The haemodynamic responses following induction of anaesthesia with ketamine and midazolam have not been determined previously. Twenty adult patients for elective myocardial revascularization were randomized to two regimens for induction of anaesthesia. Patients in Group I received ketamine, 2 mg.kg-1, and midazolam, 0.2 mg.kg-1 and those in Group II received ketamine, 2 mg.kg-1, and midazolam, 0.4 mg.kg-1. Measurements were recorded at baseline, 1 min post-induction, and at one, three, five and ten minutes after tracheal intubation. Tachycardia and hypertension (greater than 20% increases from awake baseline values) were treated with esmolol, 250 micrograms.kg-1. There were 11 patients in Group I and nine patients in Group II. There were no significant intergroup differences in demographic or haemodynamic variables. Both groups had decreases (P less than 0.05), in stroke volume, pulmonary capillary wedge pressure, and right ventricular end-diastolic volume at multiple study intervals following anaesthetic induction. None of these changes required clinical intervention. Five patients (all in Group II) had hypertensive responses to tracheal intubation. Preoperative hypertension (mean arterial pressure greater than or equal to 100 mmHg) was a predictor (P less than 0.05) of a hypertensive response to intubation, independent of the midazolam dose. Intravenous ketamine and midazolam was associated with a high incidence (25%) of haemodynamic responses to tracheal intubation. The higher dose of midazolam did not provide any haemodynamic advantage.