The authors analyze a series of 20 patients seen over the past 4 years who have shown a dramatic improvement following the introduction of lithium carbonate to their therapy. The results indicate that these patients showed a consistent syndrome with the following features: a) anergic endogenous depression; b) positive family history in first degree probands; c) obsessional personality traits and symptoms; d) hypochondriasis and somatic symptoms; e) failure to respond to previous antidepressant therapy with tricyclic and MAOI compounds as well as ECT. A previous study by Gittleson showed that one third of a series of psychotic depressives admitted to the Maudsley Hospital, London, also displayed obsessional symptoms and hypochondriasis. These patients, however, seemed to do as well with standard antidepressant treatment as a control group of psychotic depressives without obsessional features. However, in this series, there was a 7 per cent residue whose obsessional symptoms worsened, even after recovery from their depression. The authors' group of patients represented approximately 3 per cent of all psychotic depressives seen over the 4-year period and could, therefore, coincide with Gittleson's residue. The mean age of onset of illness in the authors' depressive group was 45.5 years, and this finding, coupled with the high incidence of psychotic depression in first degree relatives, indicates that these patients were suffering from a psychotic depression modified by personality traits, rather than an atypical obsessional neurosis. The consistency of clinical features and specificity of response to lithium therapy appear to indicate that this is a clearly definable clinical syndrome worthy of further investigation.