To investigate the epidemiology of premature delivery, in particular regarding potential seasonal influences, a retrospective medical-record study was conducted in 3345 women who delivered prematurely over a six-year period (1988-1993). The control group was composed of 53,162 women who carried their pregnancies to term. Each study subject had delivered at least one liveborn baby, prematurely or at full term. The premature delivery rate (ratio of premature over term deliveries) was highest in May through July and decreased significantly in January and February (p0.05). Premature delivery rates were similar in primiparas and multiparas (p0.05) and were not influenced by the gender of the fetus (p0.05). In conclusion, premature delivery rates vary across seasons, suggesting an influence of weather conditions.