Because strains of Staphylococcus aureus that are resistant to penicillinase-resistant penicillins may be difficult to detect in the clinical laboratory, a variety of changes in methodology have been suggested to increase their detection. In 1984, the West Los Angeles Veterans Administration Medical Center experienced an increase in clinically significant strains of oxacillin-resistant S. aureus. To insure that such strains would not be missed by the disk diffusion test methods employed for routine testing, changes in methodology were insituted. These included interpreting zone diameters around oxacillin disks at 48 h of incubation. We collected 139 isolates from patients thought to have oxacillin-resistant S. aureus based on these test results and later retested the isolates using microdilution MIC testing. Only 85 isolates (61%) had microdilution oxacillin MICs of greater than or equal to 8.0 micrograms/ml, whereas 54 (39%) had oxacillin MICs of less than or equal to 2.0 micrograms/ml. A review of medical records revealed that in 1 year there were 98 patients with isolates appearing resistant by disk diffusion but not confirmed by microdilution MICs; many patients were placed in isolation and treated with specific antimicrobial agents. We conclude that incubation of oxacillin disk diffusion tests for longer than 24 h in conjunction with disregard for resistance to other classes of antimicrobial agents may result in an unacceptably high degree of false resistance results. Because the resistance of S. aureus has important therapeutic and infection control implications, it is necessary to recognize problems that may result in ambiguous or inaccurate susceptibility results.