Previous studies have suggested a positive correlation between grand multiparity and the incidence of Down syndrome (DS). In order to study different parities as risk factors for DS, the Swedish health registries were used, and 2,615 infants with Down syndrome were selected from 2,184,590 infants born in 1973-1993. A statistically significant risk decrease for primiparas, and a significant risk increase for grand multiparas (5+), was found (age-adjusted odds ratios: 0.87 (95% CI: 0.80-0.96) and 1.40 (95% CI: 1.18-1.65), respectively. Potential confounders, such as the effect of truncated maternal 1-year age classes, citizenship, socioeconomic level, etc., were evaluated but were found to have only marginal effects. Evidence suggesting that the extension of prenatal diagnosis during the study period has decreased the incidence of DS among women of parity 1-4, but not among women of parity 5+, was found. The hypothesis that Swedish grand multiparas may have another attitude toward prenatal diagnosis than women of lower parities was confirmed when, in a data set containing information on 872 amniocenteses, a significantly lower rate of grand multiparity than expected was found. For the negative association between primiparity and DS, no obvious confounder was found.