Cross-sectional serological surveys of 13,006 small ruminants were conducted in 2003 to establish baseline levels of seropositivity to brucellosis and of 12,102 again in 2009 to evaluate the efficacy of controls based on biannual conjunctival vaccination with Rev 1. Seroprevalence dropped by 80 per cent in five years in eight pilot programme districts where vaccination was generally well implemented, and the prevalence of households with evidence of infection in their animals dropped from 25.1 to 7.5 per cent. Seroprevalence was reduced by 40 per cent in 10 districts where vaccination was intermittent with low coverage during some seasons. There were no changes in 19 districts where no vaccinations were carried out. Sheep vaccinated one or more times were found to be 2.5 times more likely to be serotest positive than were non-vaccinated sheep, whereas vaccinated goats were 6.4 times more likely to be serotest positive than non-vaccinated goats.