A general mechanistic approach to the study of scaling reactions is proposed which dispenses with the conventional inert marker. Instead, radioisotopes of the reacting components are used to establish the nature of the diffusing species in the solid reaction products formed on metal surfaces at high temperatures. For investigating the mechanism of oxidation of metals, O TM is employed as a tracer and the stable isotope is subsequently activated by proton bombardment via the nuclear reaction O1S(p,n)F Is. The basis for the technique is outlined and its application to the scaling of iron and of zirconium is described. The experiments with iron establish conclusively that, at temperatures up to I050~ iron is the only diffusing component in the dense, adherent wi~stite layer which forms adjacent to the metal. Direct evidence is also presented for the vapor phase transport of oxygen by a dissociative mechanism during the growth of porous, non-adherent wfistite scales on iron in H~/H.~O atmospheres. It is shown that oxygen is mobile in thick ZrO2 scales formed on Zr in a water vapor atmosphere at 1200~ however, the unusual form of the O Is distribution in the scale indicates that open channels must have existed in the outer portion of the scale during oxidation which allowed direct penetration of oxygen to levels deep within the ZrO2 layer. The advantages and limitations of the present technique are discussed with particular reference to conventional marker methods. Major sources of difficulty in the interpretation of inert marker experiments are identified and a set of experimental requirements is given which, if satisfied, should permit reliable transport information to be obtained using conventional marker techniques.The mechanism of a heterogeneous chemical reaction, such as the oxidation of a metal, can be considered established, at least in principle, if the rate-determining step in the reaction is known and if the manner in which the reactants are transported to and from the various phase boundaries in the system can be specified. Information of this kind cannot generally be obtained from kinetic studies alone. In fact, even under conditions where diffusion-controlled kinetics are observed, it is often not possible to decide, unambiguously, whether the reaction is sustained by the outward migration of metal ions through the scale, or by the inward diffusion of the nonmetallic component. The dominant or faster diffusing species can, of course, be predicted if self-diffusion or ionic conductivity data are available for both components over the entire stability range of the product phase. Unfortunately, reliable data of this type currently exist for only a few oxide and sulfide systems (1-3).When the necessary transport data are not available, inert marker techniques are customarily employed to determine the nature of the mobile component in the growing scale. Marker methods were first introduced by Pfeil (4) in 1929 in his classic work on the scaling of iron, and they have been applied subsequently to a great many...