Objectives-To examine the rate of cerebral palsy in twins and triplets in births from 1980 to 1989 in Western Australia and to identify factors associated with increase in risk.Design-Pluralities for all births in Western Australia were identified through the standardised midwives' notification system, and cases of cerebral palsy were identified from the Western Australian cerebral palsy register.Main outcome measures-Multiple births, cerebral palsy, excluding postneonatal cause.Results-The prevalence of cerebral palsy in triplets, of 28 per 1000 survivors to 1 year (95% confidence interval 11 to 63) exceeded that in twins (7.3; 5.2 to 10) and singletons (1.6; 1.4 to 1.8). Although twins and triplets were more likely than singletons to be low in birth weight, their risks of cerebral palsy if low in birth weight were similar. In contrast, in normal birthweight categories twins had a higher rate of cerebral palsy (4.2; 2.2 to 7.7) than singletons (1.1; 1.0 to 1.3). The prevalence of cerebral palsy was similar in twins of unlike sex pairs, all of whom are dizygotic, and in like sex pairs. A twin pair in which one member died in utero was at higher risk of cerebral palsy: 96 per 1000 twin pairs (36 to 218) compared with 12 (8.2 to 17) for twin pregnancies in which both survived. There was a similar but non-significant trend for death of one triplet to be associated with increased risk of cerebral palsy in the survivors of the set.Conclusion-Triplet pregnancies produced a child with cerebral palsy 47 times more often than singleton pregnancies did and twin pregnancies eight times more often. Eighty six per cent of cerebral palsy in multiple births was in twins. As multiple births are increasing mainly because of personal and medical decisions the increased risk of cerebral palsy in multiple births is of concern.