A registry based study was conducted to assess the variation in first-onset stroke with weekdays and seasons, in relation to the effects of age. Between 1 December 1991 and 30 November 1998, 10,729 first-onset stroke patients aged 25 or more were registered in Toyama Prefecture, Japan. We compared the weekly and seasonal variation in first-onset stroke by a one-way goodness-of-fit chi(2)-test. The relationship between seasonal variation in stroke onset and age was also evaluated by the method of Kendall's tau-b R x C tables with ordered categories. The frequency of onset of all strokes and cerebral infarctions (CI) was significantly higher on weekdays than at weekends (P < 0.01). More men had strokes and CI on a Monday (P < 0.01), and more women had cerebral hemorrhage (CH) on a Monday and CI at the end of the week. Stroke incidence was higher in patients aged less than 60 years (20.6%) than in those aged 60 years or over (18.7%) on a Monday compared to the weekend. By chi(2)-test, comparing observed with expected numbers of stroke onsets, weighted by the number of days in each 3-month period, the incidence of all strokes, CI and CH was significantly higher in winter and spring than in summer. The seasonal variation in the onset of stroke declined with age: all strokes (P < 0.001) and CH (P < 0.001) in both genders; subarachnoid hemorrhage (P < 0.001) only in men. Our study shows that the onset of stroke is more frequent on weekdays than on weekends, and may be associated with changes in psychophysiological stresses between working days and the weekend. We also observed a clear negative dose response relationship between seasonal variations in occurrence and age. It may be speculated that younger people have more change to work outdoors, exposing themselves to the winter environment. Their lifestyle and physiological condition may be different from those of older people.