OBJECTIVE: Second to alcohol, cannabis is the most frequently misused substance among patients with schizophrenia. The aim of this paper is to examine at early onset of psychosis whether the high comorbidity of schizophrenia and cannabis abuse is due to a causal relationship between the two disorders. Previous studies have mostly included chronic patients or samples with mixed stages of the psychotic illness. METHOD: In a German catchment area with a population of 1,500,000, a representative first-episode sample of 232 patients with schizophrenia was included in the Age, Beginning and Course of Schizophrenia Study. By means of a structured interview, the Retrospective Assessment of the Onset of Schizophrenia, the onset and course of schizophrenic symptoms and of substance abuse was systematically assessed retrospectively. Information given by relatives validated the patients' reports. RESULTS: Thirteen per cent of the sample had a history of cannabis abuse, which was twice the rate of matched normal controls. Male sex and early symptom onset were major risk factors. While cannabis abuse almost always preceded the first positive symptoms of schizophrenia, the comparison of the onset of cannabis abuse and of the first (prodromal) symptoms of schizophrenia differentiated three approximately equal groups of patients: group 1 had been abusing cannabis for several years before the first signs of schizophrenia emerged, group 2 experienced the onset of both disorders within the same month, and group 3 had started to abuse cannabis after the onset of symptoms of schizophrenia. CONCLUSIONS: The vulnerability-stress-coping model of schizophrenia suggests possible interpretations of these findings. Group 1 might suffer from the chronic deteriorating influence of cannabis reducing the vulnerability threshold and/or coping resources. Group 2 consists of individuals which are already vulnerable to schizophrenia. Cannabis misuse then is the (dopaminergic) stress factor precipitating the onset of psychosis. Group 3 uses cannabis for self-medication against (or for coping with) symptoms of schizophrenia, particularly negative and depressive symptoms. These patients probably learn to counterbalance a hypodopaminergic prefrontal state by the dopaminergic effects of cannabis. The implications of these very preliminary results include issues of treatment and prognosis, but replication studies are needed.