The present study investigated the effects of leadstimulus modality on modification of the acoustic startle reflex during three reaction time tasks. In Experiment 1, participants (N = 48) were required to press a button at the offset of one stimulus (task relevant) and to ignore presentations of a second (task irrelevant). Two tones that differed in pitch or two lights served as signal stimuli. Blink startle was elicited during some of the stimuli and during interstimulus intervals. Skin conductance responses were larger during task-relevant stimuli in both groups. Larger blink facilitation during task-relevant stimuli was found only in the group presented with auditory stimuli, whereas larger blink latency shortening during task-relevant stimuli was found in both groups. Experiment 2 (N = 32) used only a task-relevant stimulus. Blink magnitude facilitation was significant only in the group presented with tones, whereas blink latency shortening was significant in both groups. Experiment 3 (N = 80) used a go/nogo task that required participants to press a button if one element of a compound stimulus ended before the second, but not if the asynchrony was reversed. The offset asynchrony was varied between groups as a manipulation of task difficulty. Startle magnitude facilitation was larger during acoustic than during visual stimuli and larger in the easy condition. The present data indicate that startle facilitation in a reaction time task is affected by stimulus modality and by task demands. The effects of the task demands seem to be independent of lead stimulus modality.