This paper evaluates three hypotheses about the relationship between cannabis use and psychosis in the light of recent evidence from prospective epidemiological studies. These are that: (1) cannabis use causes a psychotic disorder that would not have occurred in the absence of cannabis use; (2) that cannabis use may precipitate schizophrenia or exacerbate its symptoms; and (3) that cannabis use may exacerbate the symptoms of psychosis. There is limited support for the first hypothesis. As a consequence of recent prospective studies, there is now stronger support for the second hypothesis. Four recent prospective studies in three countries have found relationships between the frequency with which cannabis had been used and the risk of receiving a diagnosis of schizophrenia or of reporting psychotic symptoms. These relationships are stronger in people with a history of psychotic symptoms and they have persisted after adjustment for potentially confounding variables. The absence of any change in the incidence of schizophrenia during the three decades in which cannabis use in Australia has increased makes it unlikely that cannabis use can produce psychoses that would not have occurred in its absence. It seems more likely that cannabis use can precipitate schizophrenia in vulnerable individuals. There is also reasonable evidence for the third hypothesis that cannabis use exacerbates psychosis.