Three experiments investigated normal subjects' ructions to pictures of fear-relevant stimuli (snakes, spiders, guns) and control stimuli (e.g., flowers, rabbits) under b u t of electric shock First-interval electrodcmal responses (RRs) and shock expectancy ratings were recorded. Experiment 1 demonstrated larger RRS and exptctlncy ratings to fear-relevant stimuli. with and without threat of e l e c~c shock. In Experiment 2 , trait anxious subjects showed elevated expectancy ratings that were additive with the bias associated with feu-relevant stimuli. Experiment 3 demonstrated that reactions to fear-relevant stimuli were not an artefact of the expectancy rating task. There were no substantial differences between biological and technological feu-relevant stimuli. Overall, the experiments confmn that fear-relevant stimuli elicit larger reactions in the absence of any specific manipulation, such as conditioning. The d t s are most consistent with cognitive modch that emphasix learning rather than genetic factors, and that attribute elccaodermal reactions to elevated expectancy of rversive events.It has long been recognised that specific fear reactions in nonclini d samples, and phobias in clinical subjects, tend to be supported by a limited range of stimuli (Agras, Sylvester, & Oliveau, 1969; Kirkpatrick, 1984). Such "fear-relevant" stimuli include animals (e.g., dogs, snakes, spiders), people (e.g., crowds, being watched or evaluated), and inanimate objects and situations (e.g., heights, water, enclosed spaas, distance from home). One obvious explanation for fear reactions to these stimuli is that they are based on an objective analysis of danger, in that the stimuli signal the possible ocarrencc of negative physical and social outcomes.However, this does not seem to be the full story, since fear is often excessive relative to objective danger, and other stimuli that are also dangerous do not appear to elicit the same quality and intensity of fear and avoidance (e.g., cars, power tools, guns). The most common account of what is "special" about fear-rele-VMI stimuli appeals to genetic factors. Since not all individuals fear all fear-relevant stimuli, the genetic component is usually taken to consist of some sort of predisposition. For example, Seligman (1971) proposed that fear-relevant stimuli are genetically "prepared" to enter into fear conditioning.