In some areas of Scotland, the prevalence of louping-ill virus has not decreased despite the vaccination of replacement ewes for over 30 years. The role of unvaccinated lambs in viral persistence was examined through a combination of an empirical study of infection rates of lambs and mathematical modelling. Serological sampling revealed that most lambs were protected by colostral immunity at turnout in May/June but were fully susceptible by the end of September. Between 8 and 83% of lambs were infected over the first season, with seroconversion rates greater in late rather than early summer. The proportion of lambs that could have amplified the louping-ill virus was low, however, because high initial titres of colostral antibody on farms with a high force of infection gave protection for several months. A simple mathematical model suggested that the relationship between the force of infection and the percentage of lambs that became viraemic was not linear and that the maximum percentage of viraemic lambs occurred at moderately high infection rates. Examination of the conditions required for louping-ill persistence suggested that the virus could theoretically persist in a sheep flock with over 370 lambs, if the grazing season was longer than 130 days. In practice, however, lamb viraemia is not a general explanation for louping-ill virus persistence as these conditions are not met in most management systems and because the widespread use of acaracides in most tick-affected hill farming systems reduces the number of ticks feeding successfully.