The aim of the study was to investigate whether proximity to the nuclear installation at Sellafield, in Cumbria, North West of England, increases the risk of stillbirth in the resident population. The cohort consisted of all 256066 live and 4034 stillbirths to mothers usually domiciled in Cumbria, 1950-1989. The study was a retrospective cohort analysis allowing for year of birth, social class and birth order using: (i) Poisson probability mapping, (ii) comparison of cumulative observed and expected numbers of stillbirths by distance from Sellafield, (iii) logistic regression of stillbirth risk in relation to distance and direction from Sellafield. Poisson probability mapping of stillbirths within 25 km of Sellafield provided no evidence to suggest that proximity to Sellafield increased the risk of stillbirth, either overall or in any specific direction. Comparison of the cumulative observed and expected numbers of stillbirths also showed no increased risk with proximity to Sellafield. Logistic regression analysis of all Cumbrian births supported these results, showing, in particular, that distance from Sellafield did not significantly influence stillbirth risk (P = 0.30). Although there was significant variation in stillbirth risk with direction (P = 0.0004), this was due to stillbirths in areas much further than 25 km from Sellafield. There was no significant effect with distance from Sellafield within any of six directional sectors (P > 0.05). There was no evidence to suggest that proximity to Sellafield increases the risk of stillbirth in the resident population.