Measurements of air-sea gas exchange rates are reported from two deliberate tracer experiments in the southern North Sea during February 1992 and 1993. A conservative tracer, spores of the bacterium Bacillus globigii var. Niger, was used for the first time in an in situ air-sea gas exchange experiment. This nonvolatile tracer is used to correct for dispersive dilution of the volatile tracers and allows three estimations of the transfer velocity for the same time period. The first estimation of the power dependence of gas transfer on molecular diffusivity in the marine environment is reported. This allows the impact of bubbles on estimates of the transfer velocity derived from changes in the helium/sulphur hexafluoride ratio to be assessed. Data from earlier dual tracer experiments are reinterpreted, and findings suggest that results from all dual tracer experiments are mutually consistent. The complete data set is used to test published parameterizations of gas transfer with wind speed. A gas exchange relationship that shows a dependence on wind speed intermediate between those of Liss and Merlivat [1986] and Wanninkhof [1992] is found to be optimal. The dual tracer data are shown to be reasonably consistent with global estimates of gas exchange based on the uptake of natural and bomb-derived radiocarbon. The degree of scatter in the data when plotted against wind speed suggests that parameters not scaling with wind speed are also influencing gas exchange rates. ¿ 2000 American Geophysical Union