The purpose of this chronoepidemiologic study was to investigate the time-relationships between the yearly variations in occurrence of violent suicide in Belgium and the yearly variations in various biochemical, metabolic and immune variables in the Belgian population. The weekly mean number of deaths due to violent suicide for all of Belgium for the period 1979–1987 was computed. Twenty-six normal volunteers had monthly blood samplings during one calendar year for assays of plasma L-tryptophan (L-TRP), competing amino acids (CAA), and melatonin levels, maximal [3H]paroxetine binding to platelets, serum total cholesterol, calcium, magnesium, and soluble interleukin-2 receptor concentrations, and number of CD4+ T, CD8+ T and CD20+ B lymphocytes. The annual rhythm in violent suicide rate is highly significantly synchronized with the annual rhythms in L-TRP, [3H]paroxetine binding, cholesterol, calcium, magnesium, CD20+ B cells, and CD4+/CD8+ ratio; the mean peak (violent suicide, [3H]paroxetine binding) or nadir (all other variables) occurs around 3 May. There were significant inverse time-relationships between the time series of violent suicide rate and L-TRP, L-TRP/CAA ratio, total cholesterol, calcium and magnesium, CD4+/CD8+ T cell ratio and number of CD20+ B cells. Maximal [3H]paroxetine binding to platelets was significantly and positively related to the time series of violent suicide. An important part (56.4%) of the variance in mean weekly number of violent suicide rate was explained by the time series of L-TRP, cholesterol and melatonin.