The monthly rates of completed suicides in France from 1978 until 1982 were analyzed. The seasonal variations of environmental (daylight and sunlight durations, mean temperature, geomagnetism), sociological (unemployment, deaths of all causes, birth and conception rates), and biological (melatonin, cortisol and serotonin circannual rhythms) factors were compared to the seasonal patterns of suicides. A clear seasonal variation (with peaks in May and September) in suicidal behavior was detected. These patterns tended to differ as a function of age (bimodal in young, unimodal in old people). The component analysis clearly pointed out that seasonal patterns of suicides may be considered as the sum of two components, unimodal and bimodal. Almost similar covariations were found between the main seasonal (unimodal) component of suicides and environmental (daylight duration and mean monthly temperature) or sociological factors whereas the secondary component was more correlated to variations in environmental factors and, to some extent, to biological parameters.