A collaborative comparison of macro- and microdilution antifungal susceptibility tests was performed in five laboratories. MICs of amphotericin B, fluconazole, flucytosine, and ketoconazole were determined in all five centers against 95 coded isolates of Candida spp., Cryptococcus neoformans, and Torulopsis glabrata. A standard protocol with the following National Committee for Clinical Laboratory Standards Subcommittee on Antifungal Susceptibility Testing recommendations was used: an inoculum standardized by spectrophotometer, buffered (RPMI 1640) medium (pH 7.0), incubation at 35 degrees C, and an additive drug dilution procedure. Two inoculum sizes were tested (1 x 10(4) to 5 x 10(3) to 2.5 x 10(3) CFU/ml) and three scoring criteria were evaluated for MIC endpoint determinations, which were scored as 0 (optically clear), < or = 1 (slightly hazy turbidity), and < or = 2 (prominent decrease in turbidity compared with that of the growth control). Overall intra- and interlaboratory reproducibility was optimal with the low-density inoculum, the second-day readings, and MICs scored as either 1 or 2. The microdilution MICs demonstrated interlaboratory agreement with most of the four drugs higher than or similar to that of the macrodilution MICs. In general, there was good interlaboratory agreement with amphotericin B, fluconazole, and flucytosine; ketoconazole gave more variable results.