Controlled studies have revealed that neuroleptic treatment alone has a significant effect in schizophrenic patients and that with parenteral treatment with depot neuroleptics the relapse rate after 1 and 2 years is significantly lower than with oral neuroleptic medication. About 20% of schizophrenic patients do not have relapses even without treatment. The effectiveness of neuroleptic long-term medication with regard to the prevention of relapses was proved by discontinuation trials for a period of up to 3 years. Up to now, however, there is no proof that drug therapy influences the long-term prognosis beyond the 3 years. The results of the Bonn schizophrenia study show the extraordinary variability of the courses and outcomes. Although some anamnestic and clinical prognostically favourable or unfavourable factors could be found, a reliable individual prognosis at the onset of the disease is not possible. Initially phasic courses with complete remission of the early psychotic features can later nevertheless lead to residual syndromes; on the other hand, chronic persistent psychoses can show a permanent remission leaving only slightly, non-specific residual states even as late at the 2nd to the 4th decade of the disease independent of the therapy. Some findings of the Bonn study, indicating a favourable influence of psychical treatments on the long-term course are described. They include the so-called catastrophic schizophrenia which at a rate of 4% occurs less frequently than in the past, the significantly poorer long-term prognosis of those patients who had not been treated in the beginning of the disease (taking into account subgroups with peracute, acute and subacute onsets on the initial psychotic manifestations), the significantly less favourable long-term prognosis of patients whose disease began prior to the era of psychopharmacological drugs, and the significantly more frequent occurrence of complete remissions of patients of the subgroups with peracute and acute onsets of the psychosis who had started treatment within 1 year after the onset of the disease (including the prodromal stage). Early diagnosis and treatment of the prodromal symptoms can probably improve the chance of a complete remission. The finding of a higher rate of patients with mixed residual states at the expense of the typical schizophrenic defect psychoses within the partial subgroup of the specific residual states with long-term neurolepsy could suggest that long-term medication is responsible for the ‘partially pharmacogenic change in symptomatology’ of schizophrenia in as far as it causes or favours a shift from the typical syndromes to the mixed residual states mainly brought about by a potential reduction in sympt