The random amplified polymorphic DNA (RAPD) assay and the restriction enzyme analysis of PCR amplified rDNA are compared for the identification of the common spoilage yeasts Zygosaccharomyces bailii, Z. rouxii, Saccharomyces cerevisiae, Candida valida and C. lipolytica. Both techniques proved to be adequate tools for yeast identification. Since the RAPD does provide less stable patterns than restriction enzyme analysis of PCR amplified rDNA, and a large amount of data had to be compared without data reduction, Principal Component Analysis (PCA) was applied successfully for clustering the RAPD patterns. The success of PCA is highly influenced by the primer used in RAPD and the amount of reference samples. A large amount of reference samples improves the performance of clustering in PCA. The primer of choice was shown to be important with respect to the discriminatory power of the RAPD method. Some primers used enabled discrimination on the subspecies level. The results collected with both typing methods justify the conclusion that the present typing system can be applied for taxonomical purposes.