STUDY OBJECTIVE--Seasonality of coronary heart disease (CHD) was examined to determine whether fatal and non-fatal disease have the same annual rhythm. DESIGN--Time series analysis was carried out on retrospective data over a 10 year period and analysed by age groups ( < 45 to > 75 years) and gender. SETTING--Data by month were obtained for the years 1962-71. The Registrar General provided information on deaths and the Research and Intelligence Unit of the Scottish Home and Health Department on hospital admissions. SUBJECTS--In Scotland, between 1962 and 1971, 123 000 patients were admitted to hospital for CHD, of whom 29 000 died. There were a further 97 000 CHD deaths outside hospital. These two groups were also examined as one (coronary incidence) - that is, all coronary deaths and coronary admissions discharged alive. STATISTICAL ANALYSIS AND MAIN RESULTS: Where there was a single annual peak, the sine curve was analysed by cosinor analysis. When there were two peaks the analysis was by normal approximation to Poisson distribution. In younger men (under 45 years) admitted to hospital there was a dominant spring peak and an autumn trough. A bimodal pattern of spring and winter peaks was evident for hospital admissions in older male age groups: with increasing age the spring peak diminished and the winter peak increased. In contrast, female hospital admissions showed a dominant winter/summer pattern of seasonal variation. In male and female CHD deaths seasonal variation showed a dominant pattern of winter peaks and summer troughs, with the winter peak spreading into spring in the two youngest male age groups. CHD incidence in women showed a winter/summer rhythm, but in men the spring peak was dominant up to the age of 55. CONCLUSION--The male, age related spring peak in CHD hospital admissions suggests there is an androgenic risk factor for myocardial infarction operating through an unknown effector mechanism. As age advances and reproduction becomes less important, the well defined winter/summer pattern of seasonal variation of CHD is superimposed, and shows a close relationship with the environment, especially temperature, or the autumn and early winter fall in temperature.