An experimental and analytical program examining sodium/sulfur chemistry has been conducted in a series of fuel-rich and -lean H-2/O2/N2 flames, with and without added sulfur, and covering a wide range of temperatures and stoichiometries. Fluorescence measurements of OH and Na downstream profiles and sodium line reversal temperatures provided a broad data base for kinetic modeling. Analysis indicated NaSO2 to be the only significant sodium/sulfur product formed in the lean flames. Even so, its concentrations remain an extremely small fraction of the total sodium. The more important perturbation of the distribution of sodium over its molecular forms results from the catalytic effect of sulfur on the flame radical concentration levels rather than from the formation of additional species. A bond dissociation energy of D-degrees 0(Na-SO2) = 197 +/- 20 kJ mol-1 is derived assuming a nonplanar structure or 210 +/- 20 kJ mol-1 if the molecule is planar. NaOS is dominant in the rich flames, coupled with small contributions from NaSO2, NaSH, and NaS2. Together, these can constitute from about 10 to 20% of the total flame sodium and do represent in this case an enhancement of molecular formation. Preliminary data in fuel-rich C2H2/O2/N2 flames are consistent with this model. This further illustrates the general insensitivity of alkali-metal chemistry to fuel type. When coupled to other available data, the low levels of NaSO2 estimated for the present fuel-lean flames tend to suggest that Na2SO4 formation in the gas phase appears to be unlikely under most practical conditions. However, this still cannot be rigorously ruled out for all situations and requires further study to resolve the critical dependences that control high-temperature Na2SO4 corrosion.