The hot corrosion of Ni induced by molten alkali carbonates has been studied to elucidate the mechanism of attack and in particular to determine whether the Rapp-Goto theory of fluxing is applicable to this system. Two principal ratedetermining processes were considered in some detail: transport of oxidant from the gas phase through the corrosion products, including the molten salt phase, to the site of active oxidation, and dissolution/reprecipitation of oxide within the corrosion product layer owing to a solubility gradient. To help evaluate these processes, we measured the solubility of NiO in various carbonate salts and we experimentally characterized the NiO scales with respect to average porosity and tortuosity. The kinetics of hot corrosion were measured by TGA under conditions expected to affect oxidant transport and/or fluxing. Experimental rates of oxidation were then compared with calculated maximum rates according to the two rate-determining processes to evaluate the feasibility of rate control by one or the other process. It was concluded that fluxing was very likely to be rate controlling in the early stages of hot corrosion, but later, when the actual hot corrosion rate had increased considerably, oxidant transport was controlling. The distribution of salt within the pores of the scale changes with time, and this is a key issue in determining the rate of reaction. * Electrochemical Society Student Member.