Two studies are reported demonstrating the reliability, validity, and clinical utility of the Language Development Survey (LDS; L. Rescorla, 1989) as a screening tool for the identification of expressive language delay in toddlers. In Study 1, 422 children (ages 24-26 months) were screened with the LDS in an epidemiological survey. The LDS manifested excellent concurrent validity with a brief direct screening measure of expressive vocabulary. Using the Rescorla (1989) Delay 3 cutoff of fewer than 50 words or no word combinations, 9.7% of the sample were language delayed (32 boys, 9 girls). In Study 2, 33 children identified as "at-risk" by the LDS Delay 3 cutoff and 33 typically developing children, matched on age, socioeconomic status, and gender, were seen for in-depth follow-up assessment approximately 3 weeks later. The LDS test-retest reliability was .97. The LDS correlated highly with Reynell Receptive and Expressive Language Scale scores, Bayley Mental Development Index, and Vineland Adaptive Behavior Composite. Sensitivity, specificity, positive predictive value, and negative predictive value between the screening LDS and the follow-up Reynell Expressive Language Scale were generally impressive. Finally, the at-risk group scored significantly lower than the LDS-identified typically developing group on all follow-up measures except the Child Behavior Checklist/2-3.