susceptible-infected-recovered-susceptible (SIRS) model of the epidemiology of feline leukemia virus is formulated and analysed. The dynamics of the disease are dramatically different in no-risk, low-risk and high-risk subpopulations of asocial, free roaming, and multiple cat household cats. Among low risk ( < 1% prevalence) free roaming cats, the model predicts than an effective immunization rate of 4% year−1, or an effective removal rate of 8% year−1are adequate to control the disease completely. Under higher risk (10% prevalence) conditions, an effective immunization rate of 23 –72% year−1or a removal rate of 60 –145% year−1are required for control. At very high (30%) prevalance rates, even heroic measures may not suffice to substantially reduce disease prevalance: a vaccination rate of 100% year−1even if attainable, would only slightly reduce disease prevalence from 30% to 29%. We conclude that the current estimated effective feline leukemia virus immunization rate of 11 –19% of the general population is inadequate to provide herd immunity in the subpopulation of cats which are genuinely at risk of infection. A substantial increase in the vaccination rate and/or intensification of test and removal efforts in the at risk population would be required to attain an effective level of protection.