BACKGROUND: Studies of season of birth or season of conception can provide clues about etiology. We investigated whether certain months or seasons of conception are associated with increased risk of autism spectrum disorders, for which etiology is particularly obscure. METHODS: The study population comprises 6,604,975 children born from 1990 to 2002 in California. Autism cases (n = 19,238) were identified from 1990 through 2008 in databases of the California Department of Developmental Services, which coordinates services for people with developmental disorders. The outcome in this analysis was autism diagnosed before the child's sixth birth date. The main independent variables were month of conception and season of conception (winter, spring, summer, and fall). Multivariate logistic regression models were used to estimate odds ratios (ORs) with their 95% confidence intervals (CIs) for autism by month of conception. RESULTS: Children conceived in December (OR = 1.09 [95% CI = 1.02-1.17]), January (1.08 [1.00-1.17]), February (1.12 [1.04-1.20]), or March (1.16 [1.08-1.24]) had higher risk of developing autism compared with those conceived in July. Conception in the winter season (December, January, and February) was associated with a 6% (OR = 1.06, 95% CI = 1.02-1.10) increased risk compared with summer. CONCLUSIONS: Higher risks for autism among those conceived in winter months suggest the presence of environmental causes of autism that vary by season.