In situ mesocosm experiments were performed to examine dissolved gaseous mercury (DGM), mercury volatilization, and sediment interactions in a frozen freshwater fluvial lake (Lake St. Louis, Beauharnois, QC). Two large in situ mesocosm cylinders, one open-bottomed and one close-bottomed (no sediment diffusion), were used to isolate the water column and minimize advection. Mercury volatilization over the closed-bottom mesocosm did not display a diurnal pattern and was low (mean = -0.02 ng m(-2) h(-1), SD = 0.28, n=71). Mercury volatilization over the open-bottom mesocosm was also low (mean = 0.24 ng m(-2) h(-1), SD = 0.08, n=96) however a diurnal pattern was observed. Low and constant concentrations of DGM were observed in surface water in both the open-bottomed and close-bottomed mesocosms (combined mean = 27.6 pg L(-1), SD = 7.2, n=26). Mercury volatilization was significantly correlated with solar radiation in both the close-bottomed (Pearson correlation = 0.33, significance = 0.005) and open-bottomed (Pearson correlation = 0.52, significance = 0.001) mesocosms. However, DGM and mercury volatilization were not significantly correlated (at the 95% level) in either of the mesocosms (significance = 0.09 in the closed mesocosm and significance = 0.9 in the open mesocosm). DGM concentrations decreased with depth (from 62 to 30 pg L(-1)) in the close-bottomed mesocosm but increased with depth (from 30 to 70 pg L(-1)) in the open-bottomed mesocosm suggesting a sediment source. DGM concentrations were found to be high in samples of ice melt (mean 73.6 pg L(-1), SD = 18.9, n=6) and snowmelt (mean 368.2 pg L(-1), SD = 115.8, n=4). These results suggest that sediment diffusion of mercury and melting snow and ice are important to DGM dynamics in frozen Lake St. Louis. These processes may also explain the lack of significant correlations observed in the DGM and mercury volatilization data.