Unlike most bacterial infections, where diagnosis is by identification of the causal organism, diagnosis of infection by Borrelia burgdorferi (Lyme's borreliosis) relies mostly upon indirect techniques. This situation has some short-comings. As long as no technology permits a microbiological diagnosis of this infection, controversy will exist as to the clinical symptoms and the criteria for the cure of the disease. Despite the lack of consensus upon both the clinical definition and the treatment of Lyme's borreliosis, it is widely agreed that the affection is best understood if regarded as a progressive general infectious disease. Indeed, following a bite with local infection, there occurs a fairly rapid dissemination of the spirochaetes. In vivo therapeutic trials have shown the potential effectiveness of beta-lactams and tetracyclines, but no treatment is considered universally effective. Most of the first trials were empirical, as antibiograms were not used. Antibiotic concentrations reached with some oral therapies are too low for the protection of certain sites such as the central nervous system. In vitro studies conducted on various strains of B. burgdorferi both in the US and in Europe are very enlightening. Among the more perplexing results of some of these studies, it is worth noting the high resistance rate of some B. burgdorferi strains to penicillin, reported by Johnson et al. and by Preac Mursic et al. Therapy for Lyme's borreliosis is discussed in light of both the in vivo and in vitro studies.