Regulation of non-point emission problems such as pesticide and nitrogen pollution of lakesand ground water is a major policy challenge. The emissions-based instruments that economistsusually advocate for cost-effective pollution control are not feasible since emissions are unobserv-able. Among the policy instruments suggested by the theoretical literature on non-point manage-ment, the tax schemes applied to ambient concentrations (“ambient taxes” for short) have drawnparticular interest. While the relative efficiency of ambient based taxes has been demonstratedexperimentally, the instrument may raise serious acceptability problems. In particular, individualswho take costly actions to improve their environmental performance could find themselves subjectto larger rather than smaller penalties due to environmental shirking on the part of others. Ourobjective in this paper is to assess the acceptability of the ambient tax. Concretely, we ask subjectsto choose between (A) an ambient tax and (B) an individual tax system. In case (A), they actuallyparticipate in a game in which their payoff depends on all participants’ decisions and on naturalvariability as would be the case in the real world if an ambient tax was implemented. In case(B) they simply earn a sure payoff, which is supposed to reflect their maximal profit under theindividual tax system. We take the percentage of agents preferring the ambient tax to a given surepayoff level as an indicator of the acceptability of the ambient tax given this sure payoff level. Ourexperimental results mitigate the common belief that ambient taxes are totally unacceptable. Ifthe “sure” alternative to the ambient tax policy is very costly for the polluters, for example becauseit involves high inspection costs, polluters might eventually prefer being liable to an ambient tax.