Because anecdotal reports suggest that concentrations of atrial natriuretic peptide are raised during tachycardias, plasma immunoreactive atrial natriuretic peptide concentrations were measured in 34 consecutive patients when tachycardia was diagnosed and again five and 15 minutes after conversion to sinus rhythm. Plasma atrial natriuretic peptide concentrations were raised in all but four patients, and were higher in patients with known heart disease than in those without. The concentrations were higher with ventricular tachycardia than with atrial fibrillation or supraventricular tachycardia, and in acute versus chronic tachycardia. There was only a weak positive relation between ventricular rate and atrial natriuretic peptide (r = 0.31); but there was a closer inverse correlation between atrial natriuretic peptide and systolic arterial pressure (r = -0.60). Conversion to sinus rhythm was associated with a definite fall in plasma atrial natriuretic peptide concentrations. Despite very high baseline concentrations of atrial natriuretic peptide only two patients reported polyuria. It is likely that atrial pressure rather than ventricular rate determines atrial natriuretic peptide release during tachycardia. Despite the absence of polyuria in all but two patients in this study atrial peptides could still contribute to, or cause, the polyuria of tachycardias.