The Mississippi Delta Management Systems Evaluation Area (MD-MSEA) project was established in 1994 in three small watersheds (202 to 1,497 ha) that drain into oxbow lakes (Beasley, Deep Hollow, and Thighman). The primary research objective was to assess the implications of management practices on water quality. Monthly monitoring of herbicide concentrations in lake water was conducted from 2000 to 2003. Water samples were analyzed for atrazine, cyanazine, fluometuron, metolachlor, and atrazine metabolites. Herbicide concentrations observed in the lake water reflected cropping systems of the watershed, e.g., atrazine and metolachlor concentrations were associated with the level of corn and sorghum production, whereas cyanazine and fluometuron was associated with the level of glyphosate-sensitive cotton production. The dynamics of herbicide appearance and dissipation in lake samples were strongly influenced by herbicide use, lake hydrology, rainfall pattern, and land management practices. The highest maximum concentrations of atrazine (7.1 to 23.4 microg L(-1)) and metolachlor (0.7 to 14.9 microg L(-1)) were observed in Thighman Lake where significant quantities of corn were grown. Introduction of s-metolachlor and use of glyphosate-resistant cotton coincided with reduced concentration of metolachlor in lake water. Cyanazine was observed in two lakes with the highest levels (1.6 to 5.5 microg L(-1)) in 2000 and lower concentrations in 2001 and 2002 (<0.4 microg L(-1)). Reduced concentrations of fluometuron in Beasley Lake were associated with greater use of glyphosate-resistant cotton and correspondingly less need for soil-applied fluometuron herbicide. In contrast, increased levels of fluometuron were observed in lake water after Deep Hollow was converted from conservation tillage to conventional tillage, presumably due to greater runoff associated with conventional tillage. These studies indicate that herbicide concentrations observed in these three watersheds were related to crop and soil management practices.