All patients presenting at the casualty department of King's College Hospital during the first six months of 1968 with deliberate self-poisoning or self-injury were followed up. Of 211 patients 204 (97%) were traced after a mean interval of 18 months (range one to two years). Despite official hospital policy, 22% had not been seen by a psychiatrist before discharge; these 44 untreated patients were compared with the remaining 160 who had received either brief (one or two interviews) or more prolonged psychiatric and social help.Subsequent suicidal attempts occurred significantly more often among untreated than among treated patients, prolonged treatment being associated with the best prognosis. The same trend was observed in respect of actual suicide, though the numbers were small and differences did not reach statistical significance. These findings held good when the untreated and treated groups were controlled for other variables which were found to be correlated with outcome. These results indicate that psychiatric intervention is associated with a significant reduction in subsequent suicidal behaviour.