The authors investigated the validity of the Clinical Adaptive Test (CAT)/Clinical Linguistic and Auditory Milestone Scale (CLAMS) for developmental screening and early prediction in very low birth weight (VLBW) infants, using the Bayley Scales of Infant Development II (BSID-II) as a reference standard. A total of 808 VLBW infants discharged from neonatal intensive care units in Taiwan from January 1995 to December 1997 were enrolled and followed up at the corrected ages of 6, 12, 18, and 24 months. The CAT/CLAMS and BSID-II were performed separately during each visit. The CAT/CLAMS showed strong concurrent correlations (r = 0.51-0.86, p < .0001) with BSID-II mental developmental index (MDI), and had high conegativity scores (96.7%-100%) but low copositivity scores (30.0%-45.3%) for detecting developmental delay. At 24 months of age, the CAT/CLAMS had better predictive validity for significant (MDI <70) rather than borderline (MDI = 70-84) developmental delay. The authors conclude that the CAT/CLAMS is not an appropriate developmental instrument for screening and early prediction in VLBW infants in Taiwan.