Aims: To assess the ef®cacy of numerical analysis of PFGE-DNA pro®les for identi®cation and differentiation of Campylobacter fetus subspecies. Methods and Results: 31 Camp. fetus strains were examined by phenotypic, PCR-and PFGE-based methods, and the 16S rDNA sequences of 18 strains compared. Numerical analysis of PFGE-DNA pro®les divided strains into two clusters at the 86% similarity level. One cluster contained 19 strains clearly identi®ed as Camp. fetus subsp. venerealis. The other cluster comprised 12 strains, of which 10 were unambiguously identi®ed as Camp. fetus subsp. fetus. The remaining two strains were identi®ed as Camp. fetus subsp. venerealis by either phenotypic or PCR methods, but not both. At higher similarity levels, clusters containing isolates from each of two countries were identi®ed, suggesting that certain clones predominate in certain geographical regions. Conclusions: Numerical analysis of PFGE-DNA pro®les is an effective method for differentiating Camp. fetus subspecies. Signi®cance and Impact of the Study: Critical comparison of PFGE, PCR, 16S rDNA sequencing and phenotypic methods for differentiation of Camp. fetus subspecies was attained. Novel phenotypic markers for distinguishing subspecies were identi®ed. Evidence for dominant clones of each subspecies in certain countries was provided.