On the hypothesis that sleepiness and alcohol interact to increase the risk of alcohol-related traffic fatalities, the percentages of alcohol-related fatal traffic crashes were assessed for the entire state of New Mexico for the years 1989-1992, for each of the seven days that preceded the changes to and from Daylight Savings Time and for each of the 14 days which followed the changes to and from Daylight Savings Time. Consistent with our hypothesis the percentage of alcohol-related fatal crashes increased significantly during the first seven days after these changes in Daylight Savings Time.