A disk agar diffusion method was developed for testing the susceptibility of rapidly growing yeasts in vitro. A totally defined, completely synthetic agar culture medium (synthetic amino acid medium, fungal) and clinical isolates of Candida spp. and Torulopsis glabrata were used. Turbidimetric adjustment of cell suspensions resulted in standard, reproducible inocula, which gave sharp, clear zones of inhibition when applied by an agar overlay method. Optimal disk loads were determined for amphotericin B, amphotericin B methyl ester, 5-fluorocytosine, clotrimazole, and miconazole. Disk potencies were stable over a 2-month period when stored in a vacuum desiccator at -30 degrees C. Using an error ratebounded classification, the zones of inhibition were correlated with both broth dilution and agar dilution minimum inhibitory concentrations (MICs). With amphotericin B and amphotericin B methyl ester, all isolates were susceptible, yielding zone diameters which clustered within 5 mm. Overall correlations between zone diameters and broth dilution MICs with 5-fluorocytosine, miconazole, and clotrimazole were 97, 96, and 82% (excluding T. glabrata), respectively; correlations of zone diameters with agar dilution MICs were 96, 92, and 88%, respectively. Disk diffusion susceptibility testing of yeasts appears to be generally applicable. However, when results are equivocal, quantitative test methods should be used.