It is argued that in their account of mood disturbances, psychiatrists have traditionally distinguished their concepts of depressive illness from distress on the grounds that social adversity provides a completely adequate explanation for the latter, but not for the former. We have already shown in our comparison of a psychiatric out-patient sample with a community sample from the same area that symptoms in the former were more likely to be multiple, severe and of diagnostic significance (J.K. Wing et al. 1981a). In this paper we test the hypothesis that disorders in the community are more likely to be cases of distress. We predict that the association between such cases and social adversity will be stronger than for psychiatric out-patients and that the difference will be accounted for by the differences in clinical picture between the groups. In general our data confirm these predictions.