The aetiology of mental retardation (MR) was studied in a population-based series of Norwegian children derived from 30 037 children born between 1980 and 1985. The study included 178 children, 79 with severe MR (SMR) (IQ<50) and 99 with mild MR (MMR) (IQ 50 to 70). Aetiology was divided into two main groups: biopathological and unspecified. The biopathological group comprised 96% of SMR and 68% of MMR, and was subdivided into prenatal (70% and 51%), perinatal (4% and 5%), and postnatal damage (5% and 1%), and a group of undetermined timing of the damaging event (18% and 11%). Single-gene disorders accounted for 15 of the 63 children with genetic disorders, including X-linked recessive in six. During the course of the study, at least 27 (15%) children had their aetiological diagnosis revised. Gestational age <32 weeks, birthweight <1500 g, and Apgar scores 0 to 2 at 1 and 5 minutes implied a significantly increased risk of MR, but contributed to only 4% of the children in the study. Decreased birthweight (1500 to 2499 g) and Apgar scores 3 to 6 at 1 and 5 minutes showed increased probability of MR. Despite extensive investigations, 4% of SMR and 32% of MMR were not identified with any biological markers and were considered as unspecified MR, several most probably representing the lower end of the normal IQ distribution in the population.