The intensity of putative benzodiazepine withdrawal symptoms was evaluated as part of a double blind placebo-controlled study of benzodiazepines and behaviour therapy in the management of agoraphobia. During the first phase of that study, some subjects were switched from low dose diazepam to placebo, and others remained on diazepam. Symptoms were evaluated in long-term benzodiazepine users (n = 30) and non-users n = 32) when they first entered the study and 4 weeks later, after both groups had been randomized to either diazepam or placebo, using eight analog rating scales measuring commonly reported withdrawal symptoms. At baseline, both users and non-users reported a substantial number of symptoms, with higher levels in the more anxious patients, but no differences between groups. After the transition to either diazepam or placebo, the users switched to placebo reported significantly higher levels of symptoms than users switched to diazepam and non-users combined. Multiple regression analysis suggested that the increase in symptoms was associated with the increase in anxiety, which was higher in the withdrawing group than the other three groups. These results raise questions concerning the extent to which withdrawal symptoms uniquely characterize benzodiazepine withdrawal. They confirm the common assumption that an increase in symptoms often accompanies withdrawal from benzodiazepines, but suggest that such symptoms are not so intense as to make withdrawal excessively difficult, at least in low-dose users.