Parallel air and water samples were collected in Lake Superior during August 1996 and May 1997, to determine the levels and air-water exchange direction of toxaphene. Concentration of toxaphene in water did not vary across Lake Superior or between seasons (averaging 918 +/- 218 pg/L) but atmospheric levels were lower in May (12 +/- 4.6 pg/m3) than in August (28 +/- 10 pg/m3). Two recalcitrant congeners, Parlar 26 and 50, also were determined. These congeners were enriched in the air samples, compared to a standard of technical toxaphene, but not in the water. Water-air fugacity ratios varied from 1.4 to 2.6 in August and 1.3 to 4.7 in May, implying volatilization of toxaphene from the lake. Estimated net fluxes ranged from 5.4 to 13 and 1.8 to 6.4 nm/m2d, respectively. The temperature dependence of toxaphene partial pressure (P) in air was log P/Pa = -3.291/T(a) + 1.67, where T(a) is air temperature. By using this relationship, the atmospheric levels of toxaphene, fugacity ratios, and net fluxes were estimated for the entire year. Fugacity ratios were highest in the winter and lowest in the summer; thus toxaphene was predicted to undergo net volatilization from the lake during all months. A net removal of approximately 220 kg/year by gas exchange was estimated.