The results of previous studies in Nigeria indicate that 81% of the villages in Plateau and Nasarawa states probably qualify for the mass administration of praziquantel (PZQ) because of Schistosoma haematobium (SH) and/or S. mansoni (SM) infection. To determine the best strategy, relative costs were modelled for four different programmatic approaches to mass drug administration (MDA) at village level. The approaches considered were (1) village-by-village screening for SH (using dipsticks to test for haematuria), with MDA confined to those villages where at least 20% of school-aged children were found infected; (2) screening for both SM (using Kato-Katz smears) and SH, with MDA confined to those villages where at least 20% of school-aged children were found infected with SH or at least 10% of such children were found SM-positive; (3) the presumptive annual treatment of all school-aged children with PZQ (without village-by-village screening); and (4) the presumptive annual treatment of all eligible adults and children with PZQ. In the MDA in models 1 and 2, treatment is only given to children unless the prevalence of schistosome infection is >or=50%, when adults are also treated. As first-year 'assessment' costs were particularly high for the models that included screening, costs were projected over 5 years for all four models. The total 5-year costs, to cover a population of 30,000, were U.S.$18,673 for the model with screening only for SH, U.S.$36,816 for the model with screening for both SH and SM, U.S. $15,510 for the treatment of all school-aged children, and U.S.$68,610 for the treatment of the entire population. Although the presumptive treatment of school-aged children appeared to be the cheapest approach, it would exclude the community-wide treatment of highly endemic communities, the importance of which needs further study.