To assess the effect of hemolysis on serum retinol concentrations determined by direct fluorometry, we assayed 196 blood samples from children 6-72-mo of age with various grades of hemolysis for serum retinol by both fluorescence and HPLC. Mean serum retinol concentrations determined by HPLC did not differ significantly according to hemolysis grade; however, fluorometric values did. Additionally, serum retinol concentrations obtained from HPLC and those obtained from direct fluorometry were significantly different in samples with severe hemolysis. Multivariate-regression analysis showed that hemolysis grade was a significant predictor of the difference in mean serum retinol values determined by the two methods. Although severe hemolysis interfered with determinations of serum retinol by direct fluorometry, this method is still a viable choice for field studies of vitamin A status.