Electrodermal activity (EDA) was investigated in 162 neurotic outpatients who were randomly assigned to 4 weeks of treatment with various combinations of placebo, amitryptiline, and diazepam. During the measurement of EDA, all patients heard a series of tones. Ninety-one patients experienced this habituation series with no accompanying task (Condition 1), and 71 performed a simple forewarned reaction time task to which the tones were irrelevant (Condition 2). Amitriptyline produced a consistent reduction in skin conductance level and in response magnitudes in both conditions. Diazepam produced a marked reduction in responding in Condition 1, but in Condition 2 responses to the tones were reduced by the attention-engaging reaction time task, and diazepam did not produce a further reduction. Subjects treated with diazepam were more likely to press the RT button erroneously in response to the tones. These subjects had the lowest levels of skin conductance activity. On the other hand, the subjects who made such errors although not treated with diazepam had the highest levels of skin conductance activity. This interaction between the effects of diazepam and task performance in Condition 2 suggests that diazepam affects EDA at a relatively central level, since its effects interact with the attentional demands of the situation.Anxiety states and neurotic depression are the commonest of neurotic illnesses and are usually treated with benzodiazepines and/or tricyclic antidepressants. However, the effects of these drugs on psychological functions remain little understood. At one level, these illnesses have clear associations with extreme arousal states. That is, anxious patients are characterized by overarousal and depressed patients are characterized by underarousal. Electodermal responses and, in particular, the habituation of the electrodermal response have frequently been used as indices of arousal. On this basis, it would seem useful to examine the effects of the two types of drugs on electrodermal responding. A number of studies, most notably those by Lader and Wing (1964), have shown that anxious patients are characterized by high skin conductance activity and a slow rate of habituation. These authors also showed that amylobarbitone and chlordiazepoxide both reduced skin conductance activity and increased the rate of habituation. The evidence on depressed patients is less clear, but two studies have found that skin conductance level and responsivity is abnormally reduced inThe authors are very grateful to the patients for their cooperation with this study. They also wish to thank T. J. Crow for his encouragement and advice in the design and conduct of the study. They are grateful to him and to