Fluxes of the biogenic sulfur gases carbonyl sulfide (COS), dimethyl sulfide (DMS), methyl mercaptan (MeSH), and carbon disulfide (CS2) were determined for several freshwater and coastal marine tundra habitats using a dynamic enclosure method and gas chromatography. In the freshwater tundra sites, highest emission, with a mean of 6.0 nmol m-2 h-1 (1.5--10) occurred in the water-saturated wet meadow areas inhabited by grasses, sedges and Sphagnum mosses. In the drier upland tundra sites, highest fluxes occurred in areas inhabited by mixed vegetation and labrador tea at 3.0 nmol m-2 h-1 (0--8.3) and lowest fluxes were from lichen-dominated areas at 0.9 nmol m-2 h-1. Sulfur emissions from a lake surface were also low at 0.8 nmol m-2 h-1. Of the compounds measured, DMS was the dominant gas emitted from all of these sites. Sulfur emissions from the marine sites were up to 20-fold greater than fluxes in the freshwater habitats and were also dominated by DMS. Emissions of DMS were highest from intertidal soils inhabited by Carex subspathacea (150--250 nmol m-2 h-1). This Carex sp. was grazed thoroughly by geese and DMS fluxes doubled when goose feces were left within the flux chamber. Emissions were much lower from other types of vegetation which were more spatially dominant. Sulfur emissions from tundra were among the lowest reported in the literature. When emission data were extrapolated to include all tundra globally, the global flux of biogenic sulfur from this biome is 2--3¿108 g yr-1. This respresents less than 0.0001% of the estimated annual global flux (~50 Tg) of biogenic sulfur and <0.01% of the estimated terrestrial flux. The low emissions are attributed to the low availability of sulfate, certain hydrological characteristics of tundra, and the tendency for tundra to accumulate organic matter. ¿American Geophysical Union 1992