The purpose of this study was to describe the differences in mortality by marital status in the Netherlands in the period 1986–1990 for specific causes of death and to estimate the contribution of each specific cause to the differences in total mortality. We have used mortality and population data from Statistics Netherlands. Poisson regression was used to calculate relative risks of dying from the specific causes of death. The relative risks and the overall mortality rates were used to estimate the contribution of the specific causes of death to the differences in total mortality by marital status. For men the general pattern was that the divorced had the highest risks, followed by the never-married and that the widowed had risks closest to married men. For women the general pattern was that the divorced had the highest risks, while widowed and never-married women alternately had risks closest to married women. Important exceptions to these risk patterns were found for, among others, infectious and parasitic diseases among men and breast cancer among women. External causes of death in particular, contributed more to the excess mortality of the 3 unmarried groups of men and women than expected, while the contributions of malignant neoplasms and diseases of the circulatory system were lower than expected on the basis of the percentages of these causes of death in mortality in the married population. Since the causes of death that contributed disproportionately to the excess mortality of the unmarried almost all have unhealthy lifestyles as important risk factors, we argue that the majority of the mortality differences by marital status can be explained by social causation (marital status affects health through lifestyle differences). However, longitudinal data are necessary to rule out selection effects (effect of health on marital status), preferably controlling for sociodemographic confounders such as socioeconomic status and taking into account living arrangements.