Methicillin resistance in Staphylococcus aureus can be difficult to recognize because it may be expressed in only a small portion of the population. The optimal reference to use in comparing different test systems is controversial. In this study MIC values of less than or equal to 4 mg/l as observed in two experiments, each including 12 test systems (24 h incubation), were used to define strains as methicillin-susceptible. This strict standard was used to compare 24 quantitative and 36 agar diffusion test systems applied to 61 strains. These test systems included six agar media (Mueller Hinton, Oxoid and BBL; and Iso-Sensitest, Oxoid; all with and without 5% NaCl), three broth media of the same type, two incubation temperatures (30 degrees C and 37 degrees C), two incubation times (24 h and 48 h) and five disc types [methicillin 10 micrograms (M 10), three types of methicillin 5 micrograms (M 5) and oxacillin 1 microgram (O 1)]. Standardized high inocula were used. It was concluded that 24 h incubation is preferable to 48 h; at 24 h, differences in recognition of methicillin-resistant Staphylococcus aureus (MRSA) among 12 quantitative test systems were not statistically significant; incubation at 30 degrees C gave slightly better results in diffusion tests. Reproducibility of quantitative data (Friedman test) was better on Mueller Hinton media (p less than 0.001) than on Iso-Sensitest medium in both dilution and agar diffusion test systems. The rank order of discs was M 5 greater than M 10 greater than O 1 (p less than 0.001). An inhibition zone of less than 13 mm with a M 5 micrograms (BBL) disc should be interpreted as MRSA.