This paper reviews evidence on two hypotheses about the relationship between cannabis use and psychosis. The first hypothesis is that heavy cannabis use may cause a "cannabis psychosis"-a psychosis that would not occur in the absence of cannabis use, the symptoms of which are preceded by heavy cannabis use and remit after abstinence. The second hypothesis is that cannabis use may precipitate schizophrenia, or exacerbate its symptoms. Evaluation of these hypotheses requires evidence of an association between cannabis use and psychosis, that is unlikely to be due to chance, in which cannabis use precedes psychosis, and in which we can exclude the hypothesis that the relationship is due to other factors, such as other drug use, or a personal vulnerability to psychosis. There is some clinical support for the first hypothesis. If these disorders exist they seem to be rare, because they require very high doses of THC, the prolonged use of highly potent forms of cannabis, or a pre-existing (but as yet unspecified) vulnerability. There is more support for the second hypothesis, in that a large prospective study has shown a linear relationship between the frequency with which cannabis has been used by age 18 and the risks over the subsequent 15 years of a diagnosis of schizophrenia. It is still unclear whether this means that cannabis use precipitates schizophrenia, whether it is a form of "self-medication", or whether the association is due to the use of other drugs, such as amphetamines, which heavy cannabis users are more likely to use. There is stronger evidence that cannabis use can exacerbate the symptoms of schizophrenia. Mental health services should identify patients with schizophrenia who use alcohol, cannabis and other drugs and advise them to abstain or to greatly reduce their drug use.