In formulating a theory of anxiety incubation, Eysenck (1968) has cited evidence from infrahuman and human aversive conditioning studies to suggest that nonreinforced CS exposure may, under certain conditions, lead to increases in fear. In several studies, repeated or prolonged exposure to symbolic representations of feared stimuli has been found to elicit unexpected increases in autonomic responses, whether these stimuli were visual (Borkovec and Glasgow, 1973), verbal descriptions (Boulougouris, Marks and Marset, 1971), or self-induced thoughts (Rankin et al., 1964; Bresnitz, 1967). While these studies support the existence of an incubation process, recent studies suggest that duration of CS exposure may be an important variable in producing this effect. Rohrbaugh and Riccio (1970) and Rohrbaugh, Riccio, and Arthur (1972) investigated the effect of varying the temporal parameters of extinction following Pavlovian fear conditioning and conditioned suppression, respectively. In both studies, short duration exposure to the CS subsequent to conditioning resulted in greater fear relative to zero or long exposure conditions. Furthermore, in support of an incubation effect, the latter study found increases in fear over post-conditioning base levels among brief exposure Ss. Miller and Levis (1971) using a procedure analogous to Rohrbaugh and Riccio (1970) obtained similar results with human Ss. Subjects who refused to touch a live snake were assigned either 0. 15, 30 or 45 min exposure to the snake. Consistent with analogue outcome research employing pre-posttest design. all groups showed greater approach scores on posttest. However, the approach of the 15 min exposure group was significantly less than that of the zero exposure group, while the 45 min condition did not differ from zero exposure. Miller and Levis (1971) suggested that the paradoxical enhancement of fear (Rohrbaugh and Riccio, 1970) exhibited by their brief exposure Ss relative to both zero and long exposure Ss was a function of the failure of redintegrated fear-eliciting cues to extinguish. In contrast, Staub (1968) has suggested that the reduced efficacy of brief exposure may be attributed to a failure to complete a cognitive re-evaluation of the probability of adverse consequences. The primary purpose of the present study was to replicate and extend the Miller and Levis study. Analogue phobic Ss were exposed to a snake for 0, 15 or 45 min. To assess whether fear enhancement suggested by Miller and Levis (1971) or fear re-evaluation offered by Staub (1968) might mitigate the paradoxical enhancement phenomenon, Ss in the brief (15 min) exposure condition were divided into three groups. During the 30 min subsequent to 15 min phobic stimulus exposure, Ss in one group were directed to write an essay concerning the most terrifying experience imaginable involving a snake. Ss in a second group were directed to identify and evaluate the logic of assumptions underlying their fear of snakes (similar to Meichenbaum, 1972). Ss in a control condition engaged in activities unrelated to fear or snakes. Of secondary interest was the relationship of subject characteristics to the occurrence of paradoxical enhancement. Eysenck has suggested that incubation is more likely to occur in Ss scoring high on anxiety/emotionality indices. Consequently, snake phobic Ss in each treatment group were divided on the basis of Mandler. Mandler, and Uviller's (1958) Autonomic Perception Questionnaire into high and low perceivers to represent high and low emotionality groups.