The purpose of this study was to examine the efficacy of the oxacillin disk diffusion test and the methicillin and oxacillin agar screen tests as predictors of oxacillin resistance as defined by the reference broth microdilution method. A total of 444 clinical isolates of Staphylococcus aureus collected from individual patients over a four-year period were tested by (1) the oxacillin disk diffusion test, with particular attention to the presence of fine growth of a resistant subpopulation within the zone of inhibition (Ox Grow interpretive criteria); (2) the agar screen method using agar plates with 4% NaCl and either 6 mg/L oxacillin or 10 mg/L methicillin; and (3) the oxacillin and methicillin broth microdilution test methods with 2% NaCl supplementation. Overall, 62 (14%) isolates were resistant and 382 (86%) isolates were susceptible to oxacillin with the use of the reference broth microdilution system. The results indicate that the disk diffusion test with the use of the Ox Grow criteria had a high sensitivity (94%) and negative predictive value (98%) but a low specificity (67%) and positive predictive value (32%) when compared with the reference broth dilution test. Similarly, the agar screen tests had a high sensitivity (95-97%) and negative predictive values (99%) but low specificity (64-74%) and positive predictive values (30-37%). These data indicate that the agar screen tests and the oxacillin disk test with the use of the Ox Grow interpretive criteria may be useful as screening tests for detecting resistance to the penicillinase-resistant penicillins but that all resistant isolates should be confirmed by the reference broth dilution method because of the large number of false-resistant screening test results.