Diquat herbicide and rhodamine WT dye were applied in a lake to three 1.6 ha plots either with a polymer, which reportedly aids in sinking and confinement of aquatic herbicides, or without a polymer. Diquat and dye concentrations were measured at three different depths in the water column within the plots during the first three hours after application to determine vertical distribution of diquat and dye, and in composite samples at fixed distances from the plot up to 168 hours after application to determine movement out of the treated plots. Diquat and dye were homogeneous in the water column when no polymer was used, but were concentrated near the surface when polymer was used. This distribution may have resulted from temperature stratification. Polymer did not affect movement of diquat or dye out of the plots. The half‐lives of diquat within the plots were 25 (SE=6.2) hr, 39 (SE=4.3) hr, and 25 (SE=2.0) hr. Forty‐six percent of samples collected at the edges of the plots did not contain detectable diquat residues and only 66 percent of those samples with detectable diquat contained greater than the potable water tolerance (10 ppb). Diquat was not found in any samples 168 hours after application 61 m or farther from the edge of the plots. Dye and diquat concentrations were weakly correlated within and outside the plots. Dye half‐lives were consistently higher than diquat, which suggests that the herbicide was removed from the water by plants and sediments more rapidly than dye.