OBJECTIVES. The purpose of this study was to compare temperature and two windchill indices with respect to the strength of their association with daily variation in mortality in the Netherlands during 1979 to 1987. The two windchill indices were those developed by Siple and Passel and by Steadman. METHODS. Daily numbers of cause-specific deaths were related to the meteorological variables by means of Poisson regression with control for influenza incidence. Lag times were taken into account. RESULTS. Daily variation in mortality, especially mortality from heart disease, was more strongly related to the Steadman windchill index than to temperature or the Siple and Passel index (34.9%, 31.2%, and 31.5%, respectively, of mortality variation explained). The strongest relation was found with daytime values of the Steadman index. CONCLUSIONS. In areas where spells of cold are frequently accompanied by strong wind, the use of the Steadman index probably adds much to the identification of weather conditions involving an increased risk of death. The results of this study provide no justification for the wide-spread use (e.g., in the United States) of the Siple and Passel index.