In this third report from the collaborative study of children with bilateral spastic cerebral palsy born between 1975 and 1986, aetiology was analysed. Evidence for a prenatal aetiology increased with gestational age, whereas evidence for a peri-/neonatal aetiology decreased. The largest subgroup, the leg-dominated subtype, showed the same distribution of aetiology as the total group. A prenatal aetiology was found mainly among term and moderately preterm children with a four-limb-dominated subtype; a peri-/neonatal aetiology among very preterm children with a three- or four-limb-dominated subtype or among term children with a dyskinetic-spastic subtype. The findings support the hypothesis generated from the authors' epidemiological results of a peri-/neonatal aetiology being predominant among preterm, and a prenatal aetiology among term, children with bilateral spastic cerebral palsy.