Out of a consecutive series of 130 potential probands with recent onsets of depression, 124 were deemed 'in scope' for a family interview study of social and genetic influences on depression. We were able to interview first-degree relatives of 83 probands (67% of those families 'in scope'). Unexpectedly, interviewed first-degree relatives had, on average, a significantly lower occupational status than the depressed probands. However, the rates of depression among the first-degree relatives were high. Based on a PSE-Catego classification, 17.2% of 244 relatives were current 'cases' of depression at the time of interview. This was significantly greater than the estimated population prevalence. Using a Past History Schedule in combination with the PSE, the lifetime prevalence of depressive illness, expressed as the proportion who had ever fulfilled PSE-Catego case criteria, was 38.9%. High rates of psychiatric treatment were also found in family members. The lifetime prevalence of any form of hospital treatment for depression in 315 first-degree relatives was 15.5%. This gives an age-corrected morbidity risk to age 65 years of 24.6%, as compared with a population estimate of 8.9%. In common with previous studies, significantly higher rates of depression were found among female than among male first-degree relatives for broader categories of disorder, but there was a small and nonsignificant excess of men among those relatives who had received in-patient treatment for depression. This apparent lessening of sex differences with more severe categories was not supported by an examination of 'lifetime ever' depressive symptoms, where severity was expressed as highest ever PSE-Catego-ID level.(ABSTRACT TRUNCATED AT 250 WORDS)