Ten laboratories tested 18 isolates of Candida spp. and two of Cryptococcus neoformans against amphotericin B, flucytosine, fluconazole and itraconazole on two occasions by the Etest method. Two individuals read each set of results. Of the 18 isolates of Candida spp., five were duplicated, but the participants were not told this. In 40 of the 60 drug-organism combinations studied, at least 80% of the Etest MICs fell within a five-concentration range corresponding to the modal MIC +/- 1 log2 dilution. In five combinations, >50% of the Etest MICs fell outside this five-concentration range. In 17 (85%) of the 20 drug-organism combinations tested in duplicate, at least 80% of the paired Etest results fell within two concentrations of each other (corresponding to one log2 dilution). Overall, 88.5% of the paired Etest results for amphotericin B agreed to within two concentrations, as did 94% of results for flucytosine, 92% for fluconazole and 79% for itraconazole. The broth microdilution MICs of the four antifungal agents for the 15 isolates were measured on five occasions in the Mycology Reference Laboratory, Bristol. In each case, the results fell within a three log2 concentration range. In 24 (40%) of the 60 drug-organism combinations tested, at least 80% of the Etest results fell within the broth microdilution test MIC range, but 27 (45%) showed <50% exact agreement. In 33 (73%) of 45 drug-organism combinations involving flucytosine, fluconazole or itraconazole, at least 80% of the Etest results fell within the same class (susceptible, resistant, or susceptible dependent upon dose) as the broth microdilution results. With fluconazole, the Etest method misclassified three susceptible isolates of Candida spp. as resistant in 1.5-15% of tests. With itraconazole, the Etest misclassified seven susceptible isolates of Candida spp. as resistant in 5-62.5% of tests. The Etest also misclassified both C. neoformans isolates as resistant to flucytosine, fluconazole and itraconazole in 7.5-65% of tests. Our results suggest that the Etest is suitable for routine use with Candida spp. and amphotericin B or flucytosine. It is less reliable for the azoles, and isolates that appear to demonstrate acquired resistance should be retested with well-established reference methods.