A retrospective study of the minor morbidity diagnoses of 775 workers during two 10-year periods was undertaken. During the first, they worked an 8-hour rotating shift, during the second, a 12-hour shift system. Data were abstracted from well-kept medical records from the plant medical department. Illnesses and symptoms indicative of stress-related complaints were abstracted. By age-sex standardization and the calculation of morbidity ratios and confidence intervals, it was shown that the introduction of the 12-hour shift system was accompanied by a statistically significant fall in the incidence of stress-related complaints such as headaches, gastrointestinal upsets and alcohol-related complaints. No reduction was found in the incidence of other illness complaints, considered to be unrelated to stress. Similar but even more marked reduction in the incidence of stress symptoms was found in a cohort of 247 workers who had 10 years of experience in each shift system.