ABSTRACTAutonomic response measures are well suited for the study of preparation because they allow the analysis of covert aspects of performance. This is illustrated by an experiment in which task‐evoked cardiac and pupillary responses were compared during a disjunctive (Go/No Go) reaction task. The motoric demands of the task were varied by manipulating foreperiod length (4 and 8 s) and probability of response (25%, 50%, and 75%). Reaction time increased with foreperiod length and decreased with probability of response. The depth of anticipatory heart rate deceleration was affected only by foreperiod length. Analysis of the beats during, and directly preceding and following the imperative stimulus revealed that interbeat intervals increased with probability of responding and foreperiod duration. The effect of stimulus timing relative to the R‐wave of the ECG was also analyzed. Early occurring stimuli prolonged the cycle of their occurrence more than late occurring stimuli. The cycle time effect was somewhat more pronounced for No Go stimuli than for Go stimuli. The subsequent cycle was shorter for early occurring stimuli compared to late stimuli. This effect was stronger for Go compared to No Go trials. Both Go and No Go reactions elicited significant pupil dilations. The No Go dilation peaked earlier than the Go dilation and its amplitude was smaller. Probability of responding affected the latency of the No Go dilation but not that of the Go dilation. The current results justify an interpretation of preparation in terms of a timing mechanism (indexed by heart rate deceleration during the foreperiod) and a mechanism allocating processing resources to stimulus encoding (indexed by cardiac slowing just prior to stimulus occurrence) and response preparation (indexed by continued cardiac deceleration and pupillary dilati