In this case study of the Yale agro-pastoral zone in southern Burkina Faso, the sustainability of tsetse control as a local public good was shown to depend upon farmers' contributions to establish and maintain the traps and targets that attract and kill tsetse flies. Contingent valuation (CV) techniques were used to generate estimates of farmers' willingness to pay for tsetse control in money, labour, or both forms of payment. Of the 261 households that participated in the CV survey, these proportions were 23, 37 and 40 percent, respectively, indicating differentiation among the population and an overall preference for labour contribution. A comparison of predicted versus actual contribution of labour indicated that only 56 percent of households that said they would contribute actually contributed; 3 percent of households that said they would not contribute actually contributed. Major factors affecting contingent contributions of labour in discrete choice models were identified, as well as those to account for in any successful scheme for actual labour contribution. These factors include the age of household head, offtake of cattle, involvement in secondary activities, membership in rural organizations, current expenditure on drug therapy, and cash-on-hand. The results also indicate that full cost-recovery of the investment in targets--about US$8000--could not be achieved in the short run with the proposed contribution of US$0.90-1.00 per month per household. Contingent contributions of money were interpreted as maximum donations to expect of beneficiaries as part of the total cost of providing tsetse control.--abstract from Journal