Diazepam (5 mg/kg) increased the number of shocks accepted by rats on two successive trials in the punished drinking test. Thus, the phenomenon of "one trial tolerance" to the anxiolytic effects of benzodiazepines in the elevated plus-maze does not generalise to this other animal test of anxiety. FG 7142 (20 mg/kg) and prior exposure to the odour of a cat had significant anxiogenic effects on two successive trials in the plus-maze. Thus the phenomenon of "one trial tolerance" does not generalise to these anxiogenic effects in the plus-maze. Furthermore, chlordiazepoxide retained its ability to counteract the anxiogenic effects in the plus-maze of prior exposure to cat odour, over successive trials. On the basis of these and previous experiments it is suggested that the state of anxiety generated on trial 2 in the plus-maze is close to a phobic state, against which benzodiazepines are relatively ineffective. Chlordiazepoxide (5 and 10 mg/kg) was also ineffective against the behavioural responses of rats during exposure to cat odour, another possible animal test of phobia. This contrasted with its efficacy against the anxiogenic effects of cat odour that subsequently generalised to and could be detected in the plus-maze.