Objectives: To review the recent literature on possible causes of the increase in frequency of diagnosed autism reported from three countries, and to compare the medical diagnoses and drug therapy from a new series of autistic boys and their mothers with that of comparable nonautistic boys and their mothers. Design: Case-control evaluation. Participants: Members of over 250 general practices in the United Kingdom. Measurements and main results: Frequency of exposure to drugs and presence of preexisting clinical illnesses in autistic children and their mothers were compared with nonautistic children and their mothers over time. According to published studies, the incidence of boys diagnosed with autism rose dramatically in the 1990s. Numerous published studies have concluded that the measles-mumps-rubella vaccine is not responsible for the large rise in diagnosed autism. In our study, boys diagnosed with autism had medical and drug histories, such as vaccines, before diagnosis, that were closely similar to those of nonautistic boys, except that developmental and sensory disorders were far more common in autistic boys. No material differences during pregnancy were found between the mothers of autistic boys and those of nonautistic boys in relation to illness or drug therapy. In the early 1990s, boys with diagnosed developmental disorders were infrequently diagnosed with autism. In the later 1990s, such boys more often were diagnosed with autism. Conclusion: A major cause of the recent large increase in the number of boys diagnosed with autism probably is due to changing diagnostic practices. Comment in Study fails to establish diagnostic substitution as a factor in increased rate ofautism.