Familial aggregation is a cardinal epidemiological feature of multiple sclerosis, but few investigators have systematically examined recurrence risks for relatives of affected individuals in the United Kingdom. As part of a cross-sectional study of multiple sclerosis in Cambridgeshire, pedigree details were taken on 674 probands. Sex-specific, crude and age-adjusted recurrence risks were assessed amongst relatives of probands, applying a statistical model based on the observed age at onset of affected individuals and providing risks for clinical counselling. Details on year of birth, present age or age at death, and disease status were available on 11 391 relatives of successive probands. Nineteen percent of patients reported an additionally affected relative; 128 non-proband affected relatives were identified and the highest risk was observed for sisters. There was a systematic reduction in relative risk with genetic distance from the proband and no preferential recurrence for maternal or paternal relatives.