The Brazilian Wilms’ Tumour Study Group carried out a hospital-based multicentre case–control study of potential risk factors for the disease between April 1987 and January 1989. The parents of 109 cases of Wilms’ tumour (WT) were interviewed when they were admitted to hospital for diagnosis and treatment. Also interviewed were the parents of two controls per case, matched for age, sex and interviewer, who were admitted to the same or nearby hospitals for treatment of non-neoplastic conditions. Odds ratios adjusted for family income and parental education were calculated by conditional logistic regression. Among cases diagnosed before 25 months of age there was a marked gradient of increasing risk of WT with increasing maternal age at the time of the child’s birth. There was no increased risk for cases diagnosed after 25 months of age. The effects of paternal age were less marked. Possible explanations for these results are discussed.