The current development of potent antifungal drugs necessitates reliable methods of in vitro susceptibility testing of clinical isolates. Other laboratories have noted with some antimicrobials that conventional tube dilution visual end points for determining minimal inhibitory concentrations (MIC) vary markedly with the inoculum. With seven isolates (four species) we compared the conventional method with one developed with a spectrophotometer. Yeasts were diluted to inocula of 10(2) to 10(5) cells/ml and incubated at 37 degrees C with 5-fluorocytosine or miconazole or without drug in a synthetic defined medium. At frequent intervals MICs were determined by the lowest drug concentration without visible turbidity. Concurrently, the percent transmission of each tube was measured and an inhibitory concentration (IC((1/2))) was calculated from the turbidimetric measurements. MICs of both drugs for all strains varied with the inoculum. Intrastrain differences averaged 128-fold for miconazole and greater than 8-fold with 5-fluorocytosine. This variability in MIC occurred when readings were made after 12 to 48 h of incubation. In contrast, IC((1/2)) values of all isolates varied less than two-fold with the same range of inocula. Most isolates showed this independence when measured at 48 h. Fast-growing yeasts showed greater IC((1/2)) variability unless the turbidimetric measurements were taken earlier while the drug-free growth was still in a rapid growth phase. This method is reproducible, inoculum independent, rapid, free from subjectivity and observer variability, and suitable for adaptation to a clinical setting.