The widely accepted kinetic model for the gaseous reduction of hematite is critically examined. The formation of the wustite layer during internal reduction cannot be explained in terms of solid-state diffusion. Present evidence indicates the existence of a micropore system in the wustite adjacent to the magnetite with the gas/solid type of reaction occurring not only at the iron/wustite interface but also at the wustite/magnetite interface. During the partial reduction of dense hematite pellets, in a controlled atmosphere, porous magnetite and wustite are formed as products. The mean pore size, as measured by mercury injection, increases with time at constant temperature and with temperature increase over the range 6500C to 950C. The sizes of the pores indicate that Knudsen diffusion is the principal gaseous transport mechanism across thick layers of magnetite or wustite.The kinetics of the gaseous reduction to iron are consistent with a mixed control mechanism involving the interaction of gaseous diffusion effects with a first order reversible chemical reaction at the iron/wustite interface. Because of the larger pore sizes in the iron layers, ordinary molecular gaseous diffusion must be the principal transport mechanism. The usual experimental methods used for establishing the validity of chemical rate ~ontrol are incapable of differentiating between the mixed control- mechanism and exclusive chemical control.Some implications of mixed control kinetics are developed, including a predicted independence of rate on reducing gas pressure at higher pressures.INTRODUCTIONAn extensive literature review of the reduction of iron oxides has been compiled recently by Themelis and Gauvin (1962). From the large number of papers published and the widely divergent opinions of workers in this field, it appears that a simple unified quantitative analysis of reduction kinetics is not possible. The aim of the present paper is to show that a major cause of confusion is due to the attempts of many workers to interpret experimentally observed reaction rates in terms of a single rate controlling mechanism without due regard to other rate processes which must occur simultaneously.Some of the differences in behaviour reported for particles...