Cowpox virus is believed to be endemic in small wild rodents, the main hosts in Britain being bank voles, wood mice and field voles. The aims of the work described in this thesis were: • To develop novel serological and PCR-based techniques for detecting infection in wild and captive rodents. • To determine the epidemiology of cowpox in wood mice; in particular to study the interactions between wood mouse population dynamics and cowpox virus infection dynamics in wild woodland populations. • To determine the pathogenesis of cowpox in the wood mouse and bank vole. The work has involved the development of ELISAs to detect cowpox antibody in wood mice and bank voles, and the production of monoclonal antibodies against wood mouse immunoglobulins for possible use in these ELISAs. A sensitive, nested PCR, directed at the thymidine kinase gene, was developed and was used to study the tissue distribution of cowpox virus in experimentally-infected bank voles and wood mice, and to detect infection in wild voles and mice. The PCRpositive field samples were then tested in a further PCR directed at the fusion gene, and sequence analysis of the resultant amplicons used to confirm that the antibody detected in the field was, indeed, due to cowpox virus infection. Experimental infection of wood mice and bank voles confirmed that cowpox causes few clinical signs in these species. However, infection was found to reduce significantly fecundity — a potentially important finding since effects on fecundity might influence host population dynamics in the field. Futhermore, the use of the PCR enabled detection of virus in several tissues, including a viraemia, and some preliminary estimates to be made of infectious period. Wood mice were generally more resistant to infection with cowpox, and remained viraemic (and, by implication, infection) for less time than bank voles. However, the routes of transmission of cowpox virus amongst wood mice and voles remain unknown. Field studies involved extensive, longitudinal trapping of two woodland populations. Data are presented describing the population dynamics and infection dynamics of cowpox in wood mice over three years. The infection dynamics in wood mice, unlike in bank voles, do not appear to be dependent on wood mouse population size — although they may follow bank voles population size, suggesting that bank voles may be the source of much wood mouse infection. A survey of sera collected from Irish wood mice did not detected cowpox antibody. Cowpox has never been reported in Ireland, and no voles are found there. This further supports the hypothesis wood mice may be common 'accidental' hosts, rather than true reservoirs for the virus. Although the data sets were large, it was not possible to determine whether or not infection might affect population dynamics in the wild.