Dermatoglyphic patterns formed during fetal development reflect genetic or early developmental events. These patterns might provide a means for investigation of a biological basis for autism, but the results of prior studies are not conclusive. The authors undertook a study of 95 autistic children, defined by DSM-III criteria ascertained independently by three child psychiatrists. The dermatoglyphic patterns of these children were compared with several control groups, all age-matched, separated by sex, and comprising different ethnic compositions. Analysis of variance as well as chi-square methods of statistical analysis were applied. Consistency among hands, ATD angle values, total ridge counts, incidence and effects on ridge count of arches, double loops, and whorls were evaluated in this comparatively large, well-controlled group of autistic children, and all failed to support a value for dermatoglyphic analysis as a discriminant of autism.