The duration of a cardiac cycle can be changed by a stimulus occurring within it. The degree of change depends on the cycle time of the stimulus (Lacey & Lacey, 1977). In three experiments, a warned, twochoice, reaction time task was used to evaluate this cycle time effect. Information delivered by warning and respond stimuli was manipulated by varying the degree to which the former stimulus predicted the latter. Stimuli were presented at delays of between 0 and 600 msec from the 'R' wave, and changes in interbeat interval and respiration phase were assessed for each stimulus. Subjects with slow heart rates showed a cycle time effect (increase in deceleration with delay) for both warning and respond stimuli presented during expiration. For inspiration, no trend was found for respond stimuli; but, for warning stimuli, a paradoxical increase in acceleration with delay was observed. Fast heart rate subjects showed a cycle time effect (decrease in deceleration with delay) for respond stimuli only, but for both respiration phases. Information about response choice, when given either by warning or respond stimuli, led to larger decelerations but this result was only significant when data for conditions associated with maximal deceleration were considered. Results are discussed in terms of the role of vagal activity in mediating the cycle time effect.