A follow-up study involving 85 autistic boys and 35 girls, c. 5 1/2 years of age at initial evaluation and 12 years at follow-up, is presented and discussed in considerable detail. Measures, also applied to 26 non-psychotic subnormal controls, included speech, social, educational, and family adequacy ratings, IQ's, and neurological data. Most autistic children remained educationally retarded and 42% were institutionalized. Good agreement with 2 studies by other authors indicated the following prognosis in autism: 1–2% recovery to normal, 5–15% borderline, 16–25% fair, and 60–75% poor. The best predictor of functional capacity in a work/school setting was the child's rating at intake. Performance IQ and severity of illness were next to best predictors. Case histories of 20 children with the best outcome, including 2 functioning normally, are compared and analyzed. Also, etiological implications of results are outlined in support of theories linking the cause of autism to biological factors.