(Difco control no. 544244) was autoclaved and adjusted to pH 7.9. To bottles containing 98 ml of the melted agar cooled to 50 C was added 2.0 ml of a well-mixed 18-hr Trypticase Soy Broth (BBL) culture of S. aureus ATCC 6538P. Viable-colony counts of this inoculum, determined on 10 separate occasions, ranged from 2.4 X 10U to 1.8 X 109 colony-forming units (CFU)/ml (mean, 6.6 X 108 CFU/ml). Portions of 9 ml of the seeded agar, containing an average of 1.2 X 107 CFU/ml, were pipetted into plastic petri dishes (100 by 15 mm) and allowed to harden; the plates then were stored at 5 C for a maximum of 5 days.Preparation (no. 740E, Schleicher & Schuell Co.), were saturated with each of the dilutions of gentamicin and placed on the surface of the seeded agar (each disc absorbs approximately 0.08 ml of solution). The solution containing 2.0 ,4g/ml was selected as the reference standard. Two discs saturated with the reference standard were placed on each assay plate opposite each other, and two other discs saturated with 1, 4, or 5 ,g/ml were placed in the other two quadrants.These concentrations were replicated 20 times. All plates were incubated for 4 hr at 37 C. The diameters of the zones of inhibition of the reference standard discs were measured by use of a dark-field Quebec bacterial colony counter with a millimeter scale on the glass; the mean of these values was taken as the correction point. Mean diameters of inhibition zones at 1, 4, and 5 jAg/ml were corrected according to the deviation from the correction point of the reference standard tested on the same plate. The corrected zone diameters of 5.0, 4.0, and 1.0 ,ug/ml with the correction point of 2.0 jug/ml were plotted against concentration on semilogarithmic paper.Assay procedure. Serum samples containing unknown concentrations of antibiotic were diluted 1:2, 1:4, and 1:5 in sterile pooled human serum (previously tested for antibacterial activity against S. aureus ATCC 6538P). Four seeded agar plates were used, each with two discs saturated with the reference standard and two discs saturated with the patient's undiluted or diluted serum. For maximal accuracy, duplicate plates were set up with discs saturated with diluted serum. The plates were incubated for 4 hr at 37 C. Zone diameters of the reference standards and of the patient's serum were measured and averaged. Correction of the zone diameter was accomplished by adding or subtracting the difference between the mean zone diameter of the reference standard obtained in the assay plate and the zone diameter of the correction point of 2.0 ,ug/ml in the standard