The effects of stressful life events which cause depression in the elderly and the role of the social support network--a longitudinal study in Hokkaido prefecture It has been reported that various stressful life events experienced by the elderly increase the risk of depression, and that a support network mitigates the effects. However, reports in our country are still lacking. This research was a longitudinal study in a former coal mining area, conducting a baseline survey on 1991. Every three to four years, we followed-up the elderly in the area. The questionnaire included :1) base attributes 2) stressful life events 3) networks, 4) instrumental/emotional support, support provided themselves, 4) Zung's Self Rating Depression Scale, 5) subjective health/number of illness/hospital admission/body aches/vision and hearing/urinary incontinence/signs of dementia, 6) hobbies and motivation in life/ADL/IADL. Whether male or female, SDS scores after three years were significantly high in cases of poor health, body aches, and signs of dementia. Females who did not participate in social activities scored significantly high after three years. In cross-sectional analysis, both males and females who were not working had significantly high scores. The effect of networks on SDS scores in females was significantly recognized for items regarding children living separately, neighborhood, close friends/relatives, and groups. However, it was not significant for males. Thus, a gender difference was found. In females, the level of depression was low when there were supports, though it was not significant in males. The subjective health condition was significantly different from SDS scores in both male and female groups. Admission to hospital, existence of body aches and vision disturbance were significantly different in females. Nevertheless, those were not significant in the male group. For social activity, whether the persons possessed hobbies and motivation in life or not created a significant difference in SDS scores. For stressful life events, both males and females had significantly high SDS scores caused by aggravation of health condition. In the female group, moving against the will, illness of spouse/family, and increased borrowing showed significance, and life events experienced within a year and SDS scores were significantly related. In multivariate analysis using parameters of males in 1995, only subjective health condition significantly increased the level of depression. In females, when the number of experienced life events was large, the level of depression was higher even when adjusted for the subjective health condition. Moreover, even then when "home visits by children separately living" were rare, the level of depression tended to be higher. As future issues, (1) adequate measures for evaluation of social support networks and (2) early preventive action in each community with validated scales for stressful life events are important.