The last ten years of research on the genetics of infantile autism were critically reviewed. Epidemiologic findings have shown that autism is a rare disorder with a prevalence of two to five per 10,000, a male-female ratio of 3:1, and an association with mental retardation (66% to 75% of autistic subjects have full-scale IQ scores ). Autism is familial, as reflected in an empiric sibling recurrence risk of 3% and pooled monozygotic and dizygotic concordance rates of 64% and 9%, respectively, which are much greater than the population prevalence of 0.02% to 0.05%. Genetic heterogeneity is pronounced with potential genetic subgroups, including autosomal recessive inheritance, X-linked inheritance, and sporadic chromosomal anomalies. Studies of subclinical markers in autism have elucidated potential markers at various levels of phenotypic expression from the DNA to the behavioral level. Linkage and cytogenetic studies point to two chromosome regions as putative markers, 9q34 and Xq27. Results of family studies support a putative biochemical marker, low levels of plasma dopamine-beta-hydroxylase, and a putative cognitive marker, ie, normal visuospatial but low verbal functioning, in autism. The frequency of minor physical anomalies and presence or absence of mental retardation are two dimensions of the physical and behavioral phenotype that may demark etiologically distinct subgroups. Genetic heterogeneity is offered as one explanation of the observed sex difference in the prevalence of autism. Directions for potentially fruitful research should be considered.