DescriptionAutism spectrum disorders (ASD) are complex neurodevelopmental disorders that are characterized by deficits in communication, social impairment, and the presence of restricted and repetitive behaviors. The work presented in this dissertation aims to reduce the genetic heterogeneity of samples ascertained for ASD by developing communication phenotypes for use in two genetics studies. Communication impairments in ASD can include impairments in speech or language and, like all traits in ASD, can range in severity from person to person. The first study involved a genome-wide linkage analysis in a sample of multiplex autism families for two non-verbal motor speech (NVMSD) phenotypes: NVMSD:ALL including nonverbal and minimally verbal subjects and NVMSD:C where there is behavioral evidence that language comprehension is relatively intact. Evidence for linkage was identified on several chromosomes: 1q24.2, 3q25.31, 4q22.3, 5p12, 5q33.1, 17p12, 17q11.2, and 17q22 for NVMSD:ALL and 4p15.2 and 21q22.2 for NVMSD:C. Genome-wide analysis and fine mapping of candidate genes did not produce strong evidence for association. The second study identified language (LI) and reading (RI) impairment phenotypes in a dataset ascertained for autism and specific language impairment (SLI) in the same family. These families were extensively phenotyped with a comprehensive testing battery where all language measures were found to be heritable. In addition to LI and RI, social impairment and obsessive-compulsive behavioral phenotypes were identified in these families using well-respected assessments (SRS and Y-BOCS, respectively). Genome-wide linkage analysis yielded evidence for linkage on 13q21.2 (YBOCS), 14q32.31 (SRS), 15q25.1 (LI), 15q26.2 (SRS), and 16p12.3 (RI). Genome-wide analysis and fine mapping of candidate genes did not produce strong evidence for association. The identification of non-overlapping loci for each phenotype supports the hypothesis that these phenotypes successfully identify unique communication and social impairment loci in ASD. Furthermore, as the second study was conducted in families ascertained for autism and for SLI, these results support the hypothesis that some individuals with ASD and those with SLI without ASD may have some shared genetic etiology. The lack strong evidence for association suggests that rare and/or multiple variants may play a role in the etiology of ASD.