Eighty-nine patients attending neurology clinics with transient non-focal neurological symptoms were studied by routine electrocardiography and 24-hour monitoring of the electrocardiography and 24-hour monitoring of the electrocardiogram. In comparison with 109 control subjects there was no significant overall excess of arrhythmias (age-adjusted odds ratio 1.7, kappa2 = 2.67) except in the subgroup of patients under the age of 30 (odds ratio 11.6, p less than 0.05). Bradyarrhythmias, but not tachyarrhythmias, were significantly more common in the patients (odds ratio 7.4, p less than 0.001),. Since patients can rarely be studied while they are having symptoms a working diagnosis must be based on a balance of probabilities: arrhythmias in young patients, or bradyarrhythmias in any patient, are likely to be clinically relevant. Ambulatory electrocardiographic monitoring contributed to the diagnosis in at least 25 of the patients, Nevertheless, the extent to which further investigations are pursued, and the form of treatment ultimately adopted, must also be influenced by the frequency and severity of the patients' symptoms.