Aim: This is the ninth report from the western-Swedish study of the prevalence and origin of cerebral palsy. Methods: A population-based study covering the 88 371 live births in the area in 1995–1998. Birth characteristics, neuroimaging findings and risk factors in children with cerebral palsy were recorded, prevalence was calculated, and aetiology was analysed. Results: The study comprised 170 children with cerebral palsy, i.e. a prevalence of 1.92 per 1000 live births. Excluding eight post-neonatally derived cases, the gestational age-specific prevalences were 77 per 1000 for children born before 28 wk of gestation, 40 for children born at 28–31 wk, 7 for children born at 32–36 wk and 1.1 for children born after 36 wk of gestation. Spastic hemiplegia, diplegia and tetraplegia accounted for 38%, 35% and 6%, respectively, dyskinetic cerebral palsy for 15%, and ataxia for 6%. For the first time, hemiplegia was now most common, due to the decline in preterm diplegia. There was a further increase in full-term dyskinetic cerebral palsy. The origin of cerebral palsy in children born at term was considered to be prenatal in 38%, peri/neonatal in 35% and unclassifiable in 27%, while in children born preterm it was 17%, 49% and 33%, respectively.