Insulin treatment of schizophrenia is of definite value, but is not as effective as the earlier enthusiastic reports led one to believe. Impressive results are obtained only in patients who have not been ill more than one or two years. The number of treatments and the depth of coma bear no constant relation to the results obtained. Patients who, after twenty-five or thirty treatments, do not show definite improvement will not become sufficiently well after more treatments. Gralnik believes that the specific 'insulin-treatment-situation', the transference, is of definite importance.