The objectives of this study were to assess the feasibility of administering the CHecklist for Autism in Toddlers (CHAT) at the 18-month developmental check, estimate the prevalence of screening positive for autism at the first and second administrations of the CHAT and estimate the prevalence of diagnosed cases of autism. A cross-sectional study design was utilised and data was collected at child developmental screening clinics in counties Cork and Kerry. The sample group consisted of infants attending the routine 18-month developmental assessment, who were broadly representative of infants in the catchment area. The main outcome measure was a medium or high-risk score following two administrations of the CHAT screening instrument and a positive diagnosis of autism after clinical assessment. The CHAT was administered to 2117 infants (79% of those approached) of whom 29 were scored at medium or high risk at first screening, resulting in a prevalence rate of 137 per 10,000 (95% CI: 87-187). A total of 7 of the 29 first screen positive infants were positive (medium or high risk) at second screening, 12 were low risk and 10 parents refused to participate. On subsequent clinical assessment of the 7 infants screening positive on first and second assessment and assessment of 5 of the 10 infants whose parents declined second screening, 7 children received a diagnosis of autism. Thus the overall prevalence of clinically diagnosed autism following this screening exercise was 33.1 per 10,000 (95% CI: 13.3 to 68.0). The CHAT instrument is a useful tool to help identify childhood autism among infants. Routine use of this instrument at 18-month developmental assessment merits consideration.