The growth patterns observed in the trailing wells when fluconazole is being tested may give rise to readings that suggest resistance or increased MICs for known susceptible strains. We conducted a multicenter study to evaluate the intralaboratory and interlaboratory reproducibilities of a method that uses agitation to disperse these types of growth. Ten strains of Candida albicans and five strains of Cryptococcus neoformans were tested against fluconazole, flucytosine, and amphotericin B by using a microdilution adaptation of the proposed reference method of the National Committee for Clinical Laboratory Standards for yeasts (M27-T). The endpoint criterion used before agitation was consistent with the M27-T recommendation, while a criterion of 50% or more reduction of growth compared with the control was used after agitation. The results of this study showed that use of agitation and the modified endpoint criterion both improved intralaboratory and inter-laboratory agreement and increased the frequency of interpretable MICs. The MICs obtained by this method were comparable to those obtained by the broth macrodilution M27-T method. Like M27-T, this method was not able to definitely distinguish amphotericin B-susceptible from -resistant strains, although the MICs for the resistant strains were consistently higher than those for the susceptible ones. The findings imply that agitation should be seriously considered when antifungal agents, particularly fluconazole, are tested in a microdilution format.