SUMMARY Fixed coupled ventricular premature depolarizations (VPDs) are usually considered reentrant; recent experimental models have demonstrated that parasystolic rhythms may also appear in fixed coupled patterns. To analyze the mechanisms of fixed coupled VPDs, 60 exercise tests were chosen to evaluate the response of VPD coupling intervals to changes in cycle length of the dominant supraventricular rhythm. Selection criteria included the presence of frequent, unifocal VPDs that were fixed coupled (variation < 80 msec) at any one cycle length, with the persistence of VPDs at several different cycle lengths. Three patterns of response of coupling intervals to changes in cycle length were noted: 1) 32 patients with a direct linear relation (r > 0.9) of coupling intervals to cycle length; 2) 16 patients with coupling intervals fixed, independent of cycle length; and 3) 12 patients with no consistent relation over a wide range of cycle lengths. Two patients in group 2 and four in group 3 fulfilled criteria for parasystole with interectopic intervals that remained constant at different cycle lengths. These results suggest that VPDs in group I are reentrant, while some patients with group 2 or 3 responses have evidence for parasystolic rhythms. We conclude that 1) fixed coupling of VPDs is not diagnostic of reentry, and 2) changes in cycle length induced with exercise may be useful in the analysis of mechanisms of VPDs.