We determined the effects of single intravenous (10 mg) and oral (80, 160 mg) doses of verapamil in 8 digitalized patients with chronic atrial fibrillation. The time course of drug effect was analyzed with computer assistance by considering several measures in atrial fibrillation, including average R-R interval (ARR), shortest R-R interval (SRR), longest R-R interval (LRR), and variability of R-R intervals. Peak plasma concentrations of verapamil were observed immediately after intravenous verapamil (mean elimination half-life of 3.3 h) and 1 hour after oral verapamil (mean elimination half-life 3.4 h, 80 mg; 3.1 h, 160 mg). In contrast to previous studies, we observed the maximum bradycardic effect of intravenous verapamil to occur at one-half to 1 h, and this effect lasted for 2-4 h. Following oral verapamil the peak effect occurred at 3-4 h and lasted for 5-8 h. Analysis of the time course of changes in APR, SRR, LRR, and variability of R-R revealed two distinct "patterns" of ventricular response. In one group (4 of 8 patients), verapamil caused an increase in SRR but a decrease in LRR. R-R intervals "regularized" in this group. In the remaining patients, verapamil increased the SRR but did not change or increase the LRR. This enhanced the observed increase in ARR intervals. We postulate that the decrease in LRR intervals after verapamil is due to reflex adrenergic discharge as a result of the vasodilator effect of the drug while the increase in SRR is a direct effect. These two opposing effects result in regularization of R-R intervals in many patients. Patients demonstrating an increase in LRR intervals after verapamil may not get reflex adrenergic discharge or may be incapable of responding to the latter due to conduction disease; these patients may experience bradycardic complications after verapamil.