A hypothesis of a psychosocial origin of depression in diabetic persons (mainly patients with non-insulin-dependent diabetes mellitus, NIDDM) was tested. The population consisted of all the 1008 persons born in 1935 and living in a Finnish city (Oulu) on 1 October 1990. Data were collected in two phases. A screening for NIDDM and a collection of back ground data were carried out in the first phase. Oral glucose tolerance tests (OGTTs) were performed in the second phase in order to detect those with previously undiagnosed NIDDM, and depressive symptoms were measured with the Zung Self-Rating Depression Scale (ZSDS). Participants who scored 45 raw sum points or more were considered as suffering from depression. All those who were aware of their diabetes before the second phase were considered as previously diagnosed diabetics. Depression (raw sum score 45 points or more on the ZSDS) was associated with previously diagnosed diabetes mellitus, while it was not associated with undiagnosed diabetes. The sum scores on the ZSDS did not differ significantly from each other in the groups classified according to diabetic status. This finding was explained by the sum score of the ZSDS among the previously diagnosed diabetics, which had a skewed distribution towards higher sum scores. Among the previously diagnosed diabetic patients, all those who were depressed were on sick leave or retired, while the corresponding proportion for non-depressed patients was 52% (P = 0.009). The number of diagnosed diseases was higher among the former than the latter group (P = 0.025). Severe depressive symptoms were not associated with a more serious metabolic disease, but they tended to be associated with a great number of diagnosed diseases and with being on sick leave or retired. The results showed that the impact of diabetes mellitus itself on depression was not strong. Depression was not connected to elevated fasting blood glucose levels, but instead to a great number of diagnosed diseases and unfavourable social factors, such as being on sick leave or retired, which suggests a psychosocial origin of depression in NIDDM patients.