With the increase in the number of immunocompromised hosts, the number of fungal pathogens has increased markedly. Identification and classification, especially of yeast species and strains, is often difficult when based solely on phenotypic characteristics. Since it became clear that different fungal pathogens require specific treatment strategies, there is a need for simple, rapid and reliable methods to identify fungal isolates. Polymerase chain reaction (PCR) fingerprinting was successfully applied here to identify yeast isolates. Microsatellite [(GTG)5; (GACA)4] and minisatellite [(5'GAGGGTGGCGGTTCT 3'), derived from the core-sequence of the phage M13] specific primers were used as single primers in the PCR to amplify hypervariable interrepeat DNA sequences from over 200 European, American and Australian clinical isolates within the genus Candida. Each species, represented by its type strain, could be identified by a specific multilocus pattern, allowing for the assignment of all the isolates to the appropriate species. Intra-species variation in the multilocus profiles was about 20% compared to inter-species variation, which was up to 80%. Anamorph-teleomorph pairs could be identified by highly homologous PCR fingerprint patterns. PCR fingerprinting was more discriminatory when compared with routinely used biochemical tests (Vitek YBC and API ID 32C). PCR fingerprinting has proven to be a powerful tool for the identification of medically important yeasts. It is rapid, sensitive, reliable, highly reproducible, stable in vitro and in vivo, and applicable to large scale experiments. Potential applications include: yeast taxonomy, epidemiology, environmental surveys, and improvement of the diagnosis of mycotic diseases.