The long-term prognosis of 192 surviving children with the syndrome of infantile spasms was evaluated. The children had been admitted to three paediatric hospitals in Helsinki at the time of initial diagnosis. The aetiological factors of the syndrome were carefully studied in each case. ACTH therapy was employed in 162, usually for about six weeks. The follow-up study 3-19 (mean 10.4) years later was made at the Children's Hospital, University of Helsinki. The rate of mortality was 19.6 per cent. Normal development was seen in 12 per cent and slightly subnormal in 10.4 per cent of the surviving children. Psychiatric disorders were seen in 27.6 per cent of the survivors. Sensory defects were also common. Severe cerebral palsy was seen in 4 per cent. Other seizures after cessation of the infantile spasms were seen in 60 per cent. Serial EEG studies showed that the temporal lobe was the most common site of abnormality. Abnormalities in the temporal lobes were seen frequently in children with symptomatic neonatal hypoglycaemia as a probable cause of the spasms. Prognostically favourable factors were "idiopathic" aetiology, normal development and not other fits prior to the spasms, short treatment lag, good response to ACTH and short duration of the spasms. In this study early treatment seemed to be of great importance even with regard to mental development. The factors connected with a bad outcome were: symptomatic aetiology (especially brain malformations, early infections and tuberous sclerosis), slow development before spasms, other seizures before infantile spasms, early onset of the spasms, long treatment lag, long duration of the spasms and other later occurrence of myoclonic-astatic seizures (Lennox-Gastaut). Large doses of ACTH (120-160 units) were not associated with a better prognosis than the smaller doses (20-40 units). The benefit of long versus short treatment schedules could not be evaluated in this study. The relapse rate here was 32 per cent.