In South Africa, two biotypes of type species 1 of the Lyssavirus genus are maintained independently among the members of the Herpestidae and Canidae families, respectively. Canid rabies is a relatively new addition to the African subcontinent, having been introduced from infectious cycles, which had existed among dogs in Angola, in the early 1940s. Two epidemics, believed to have originated from dog endemic regions which had existed in the southern Maputo district of Mozambique since 1952, have occurred among domestic dogs in the KwaZulu Natal province in recent years. The first of these epidemics started in 1964, and ended by 1968, while the second epidemic which started in 1976, has proven to be intractable, despite the concerted efforts which have been implemented to bring it under control. In order to contribute to the understanding of the molecular epidemiology of rabies in the KwaZulu Natal province, and to thereby assist in future surveillance and control efforts, we conducted a molecular sequence analysis of representative panel of viral isolates which were obtained from the province during the year 2003. A 591 nt. sequence encompassing the G-L intergenic region and glycoprotein cytoplasmic domain was sequenced for 128 viral isolates, which were obtained from the different magisterial districts and affected host species of the province, and was subsequently used to characterize these viruses phylogenetically. Characterization of the KwaZulu Natal variants, and comparison of the obtained sequence data, to sequences data which was obtained from rabies endemic regions from elsewhere in South Africa and Zimbabwe, in general supported the pattern of spread which led to the introduction of rabies into the province, as was previously suggested from the literature. The phylogeny which was established from the analyses, indicated that the viral isolates from the province were highly related to each other, and could be divided into two groups, which although belonging to the canid biotype, were in general distinguishable from canid rabies virus isolates which were obtained from elsewhere in South Africa and Zimbabwe. The observation that these subfamilies showed a low genetic divergence, as well as that they shared a unique recent common ancestor, suggested that they were introduced recently into the northern reaches of the province, probably from the same geographical region (i.e. southern Mozambique). logenetic characterization of the KwaZulu Natal isolates further suggested that at least three enzootic fronts are currently responsible for the introduction of rabies into the northern and southern regions of the province. The first of these fronts was hypothesized to have spread directly across the southern Mozambique border (possibly via southeastern Swaziland), into the northeastern coastal regions of KwaZulu Natal, while the second front represented the south-eastwards spread of synergistic dog-jackal cycles from southeastern Mpumalanga, into the northern and northeastern regions of the province. The third front on the other hand, represented the possible spread of a remnant infectious cycle, left over from the 1964-1968 epidemic, from the northern region of the Eastern Cape, into southern KwaZulu Natal. logenetic characterization further proved useful for identifying the distribution of viral variants, and allowed us to propose a pathway by which the disease might have spread throughout the province. The proposed pathway of spread suggested that viral variants may have been translocated over long distances, and highlighted the role that major routes of human transportation may play in the dissemination of the disease. The regional characterization of viral variants from KwaZulu Natal, further demarked the location where the identified viral variants circulated in individual host populations, allowing us to place the current epidemic into an epidemiological framework which attempts to explain the long term persistence of the disease. This provided clues as to the intractability of the second epidemic, and allowed us to develop a proposal as to how current control strategies may be altered, in order to contain the current outbreak in the province. initial phylogeny which was established from the study provides an epidemiological framework, which will play an important role in determining the origin of future human spillover cases, and for tracking the spread of viral variants throughout the affected regions of the province. It is further envisaged that the data which was generated during the course of the project will be utilized in future surveillance efforts, targeted to the evaluation of the efficacy of potentially implemented control campaigns.