The Present State Examination was used to assess the psychopathology of 23 psychotic patients who were cannabis positive on urinary screening, and 46 matched drug-free controls. Cases and controls were indistinguishable in terms of psychopathology, DSMIII diagnoses, onset of recent illness, the proportion of first admissions, ethnicity, and socio-economic class, differing only in their histories of substance use. These data suggest that psychosis which develops or recurs in the context of cannabis use does not have a characteristic psychopathology or mode of onset, and is not restricted to a particular ethnic or socio-demographic group. There is thus little evidence to support the validity of 'cannabis psychosis' as a diagnostic entity.