To assess the influence of diurnal changes in the autonomic nervous system on the QT interval, 24 hour dynamic electrocardiographic monitoring was performed in six pacemaker dependent patients with normally innervated hearts, in six cardiac transplant patients with anatomically denervated hearts (but which respond to circulating catecholamines), and in nine diabetic patients with confirmed autonomic neuropathy. QT and RR intervals from hourly intervals were measured and Bazett's formula was used to correct QT intervals during sinus rhythm. All QT intervals were normalised by dividing by the mean QT for the 24 hours in each patient and were expressed as a percentage. There was pronounced diurnal variation of normalised QT in the patients with normally innervated hearts. QT intervals were longer during sleep than during waking hours (06.00 vs 18.00 h, 102.5% vs 97.8%). Diurnal variation was blunted in the transplant patients (101.3% vs 98.1%) and absent in the diabetic patients (100.0% vs 100.3%). In the normally innervated patients changes were most pronounced at the time of waking (06.00 vs 09.00 h, 102.5% vs 95.4%). There was no change in normalised QT in the transplant and diabetic patients at this time. There was no significant difference between normalised QT for the three groups during sleep, but this variable was shorter in innervated patients during waking hours (for example at 10.00 h, innervated 96.5%, transplant 100.7%, diabetic 100.7%). Diurnal changes of the QT interval may be pronounced in the innervated heart and are dependent on both variations in autonomic tone and concentrations of circulating catecholamines. These changes in repolarisation may be related to the reported diurnal pattern of ventricular arrhythmias.