Five baculovirus strains were introduced in 1984–1985 into virus-free island populations ofOryctes rhinoceros(L.) in the Maldives, either singly or in combination. Pest populations were monitored by palm damage surveys for up to 4 yr on 22 islands, some of them deliberately untreated. The method used allowed estimates from a single survey of the number of beetle attacks during each of 18 mo (3rd–20th month before the survey date). Such surveys gave continuous records forO. rhinocerosattacks during 1983–1987. Most islands where the virus was released showed a highly significant reduction in palm damage. The prerelease pest density on a given island influenced the efficacy of the virus introduction. Adjustments were made for this effect when the performances of the different virus strains were compared. Strain X2B gave consistently better reduction in pest populations and more virus-infected beetles compared with strain V2/3B; the remaining strains gave intermediate control. These results agreed with earlier laboratory tests, in which the virulence of both strains was compared by infectingO. rhinoceroslarvae. Two islands where three strains were released simultaneously also experienced a high damage reduction; here, X2B predominated eventually over strains V2/3B and S2A.