Objectives: Increasing numbers of young children referred for a differential diagnosis of autism spectrum disorders (ASD) necessitates better understanding of the early syndrome expression and the utility of the existing state-of-the art diagnostic methods in this population. Method: Out of 31 infants under the age of 2 years referred for a differential diagnosis, 19 were diagnosed with autism, and 9 with pervasive developmental disorder – not otherwise specified (PDD-NOS) when reassessed at 3 years. We examined 1) the symptoms of ASD in the second year and changes in the syndrome expression by the age of three; 2) relationship between expert-assigned clinical diagnosis and diagnostic classification based on Autism Diagnostic Observation Schedule-Generic (ADOS-G) and Autism Diagnostic Interview-Revised (ADI-R) in the second year; 3) the relationship between direct observation and parental report of ASD symptoms. Results: Symptoms of autism and PDD-NOS in the second year were pronounced and stability of the clinical diagnosis was high. The agreement between clinician-assigned autism but not PDD-NOS diagnosis and the ADOS-G was high. However, sensitivity of the ADI-R diagnostic classification of autism was poor. Comparison of concurrent parental report and direct observation revealed discrepancies in severity ratings of key dyadic social behaviors. Changes in communication reflected acquisition of language accompanied by the emergence of unusual language characteristics. Symptoms of social dysfunction were relatively stable over time, and so was the severity of stereotyped behaviors. Conclusions: The study provides support for stability of clinical diagnosis and syndrome expression in the second year and highlights advantages and limitations of the ADI-R and ADOS-G for diagnosing and documenting symptoms of ASD in infants.