Water pollution control policies generally direct sources (i.e., industry, agriculture) to reduce loadings of certain pollutants. Thus, evaluating the relative net recreation benefits of policies to improve water quality requires establishing a linkage between the sources, the resultant water quality degradation at the affected water bodies, and, ultimately, the effect on recreation behavior. This linkage is rarely present in the empirical literature which is, thus, deficient for water pollution control policy assessment purposes. In this paper, we estimate the relative recreational swimming benefits that may result from controlling point and nonpoint sources of pollution, respectively, in New Hampshire's lakes. We use a repeated discrete choice framework to model swimming behavior as a function of each lake's level of eutrophication, bacteria, and oil and grease. For each pollutant, at each affected lake, we identify which source is responsible for the pollution, and we conduct scenarios controlling each pollution source independently, and then, taken together. Seasonal benefit estimates are presented for each scenario. Coupled with information on the most cost effective means of generating the scenarios, these estimates provide a useful starting point for a quantitative assessment of the net recreation benefits of policies to improve the quality of New Hampshire lakes.