The reliability of the standardized 30-mug cephalothin disk and that of an experimental 30-mug cefaclor disk in predicting probable clinical susceptibility to cefaclor were compared. Quantitative determinations of cefaclor susceptibility were measured by the World Health Organization International Collaborative Study agar dilution procedure; diffusion disk tests were performed by the standardized U.S. Food and Drug Administration disk test. The cephalothin disk erred in predicting probable susceptibility in 52% of isolates of Enterococcus spp. resistant to 16 mug or less of cefaclor per ml; the cefaclor disk did not. The cephalothin disk erred in correctly predicting susceptibility in only 20% of cefaclor-susceptible isolates of Enterobacter spp.; the cefaclor disk correctly predicted susceptibility for 70%. These results indicate the need for further evaluation of a separate cefaclor disk for use in susceptibility testing with this new cephalosporin.