We hypothesized that maternal-fetal incompatibility at the Rh or ABO loci may contribute to the risk of autism. There are biologically plausible reasons to believe such effects may play a role, and two previous epidemiologic studies have provided suggestive evidence. To further test this hypothesis, we genotyped the Rh and ABO loci in a sample of 389 independent case-parent trios from the AGRE repository and analyzed the data using a modification of the log-linear model for case-parent trios in which the effects of maternal-fetal genotype incompatibility are modeled jointly with the effects of the affected child's or mother's genotypes. We did not find any evidence that incompatibility at the Rh or ABO loci increases the risk of autism. Furthermore, we did not find any evidence for the presence of a high-risk susceptibility allele at or near these two loci operating either through the mother or child.