Fear (e.g. associated with the threat of an electric shock) causes an increase in initial pupil diameter (IPD) and a decrease in the amplitude of the light reflex response. There is evidence for dissociation between the two responses to threat: only the reduction in light reflex response amplitude is sensitive to the anxiolytic drug diazepam. We examined the effects of peripheral sympathetic blockade with the alpha(1)-adrenoceptor antagonist dapiprazole on both responses to threat on the basis of the hypothesis that only the response of the IPD will be affected, whereas the response of the light reflex will remain unaffected. Twelve healthy volunteers (Experiment 1) and eight healthy volunteers with smaller pupils (Experiment 2) participated in one experimental session. Dapiprazole 0.5% (two drops of 20 microl, three times) was instilled in the subjects' right or left eye while the contralateral eye was treated with placebo eye drops (artificial tear, two drops of 20 microl, three times) according to a single-blind balanced design. Pupil diameter was monitored by infrared binocular television pupillometry. At the point of maximum dapiprazole-evoked miosis, the light reflex was elicited three times in each of three Safe blocks (no possibility of electric shock), alternating with three Threat blocks (possibility of electric shock). At the end of each Safe and Threat block, subjects rated their mood and feelings on the Visual Analogue Scales. In Experiment 1, dapiprazole caused significant miosis. Threat increased subjectively rated anxiety and inhibited the light reflex. The inhibition of the light reflex was unaffected by dapiprazole. The threat-induced increase in IPD was also unaffected by dapiprazole, probably due to a ceiling effect curtailing the threat-induced increase in IPD. In the smaller pupil group in Experiment 2, where the possible contribution of a ceiling effect was minimized, dapiprazole suppressed the threat-induced increase in IPD. The inhibition of the light reflex by threat is likely to reflect central parasympathetic inhibition and is unlikely to involve the peripheral sympathetic innervation of the iris. The threat-induced increase in IPD is likely to reflect mainly central sympathetic excitation. The different central autonomic mechanisms underlying the two pupillary responses to threat may explain the dissociation between the separate effects of threat on IPD and light reflex amplitude.