To study clonal polymorphism of Borrelia hurgdorferi antigens in the course of an experimental infection sequence, the low-passage tick isolate ZS7 was cloned by two rounds of agar subsurface plating. The resulting clones showed a variable pathogenic potential after experimental infection of C.B-17. scid mice. The test clone 4.2.II, selected for virulence by two passages in immunodeficient scid mice, failed to establish a successful infection in immunocompetent AKR/N mice, indicating the loss of pathogenicity traits required for evasion of the specific immune response. Cloning of natural or clinical B. burgdorferi isolates is a prerequisite for analyzing genetic and antigenic variation of the pathogen. However, the inevitable propagation in artificial media during cloning may lead to a loss of pathogenic features rendering the subsequent experimental infection of animals impossible. A combined procedure of in vitro cloning and in vivo selection also does not solve the dilemma because B. burgdorferi variants arise by recombinatorial processes in the pathogen's dynamic genome during the course of infection. Consequently, the resulting bacterial isolates from infected animal tissues represent again non-clonal, heterogeneous B. burgdorferi populations. In principle, cloning of a B. burgdorferi population is the appropriate method to analyze the polymorphism of individual molecules during infection. As a caveat, however, one has to envisage that during propagation of individual clones in vitro and in vivo independent genetic variations