Autism spectrum disorder is an inherently complex phenomenon requiring large studies of many different types to further understanding of its causes. The National Database for Autism Research (NDAR) is being constructed to aid in this effort by providing a means for researchers to share and integrate data. An autism ontology drafted by a group at Stanford is being incorporated for use by NDAR to allow semantic data integration. The architecture upon which NDAR is built - the UCSD developed data integration environment - supports the use of this autism ontology, including annotation of data with ontological concepts and ontology enhanced queries on databases, both central and federated.