Modest additions of sodium acetate, nitrate, iodide, or benzoate to MgC12 test solution boiling at 130~ are found to increase resistance to or inhibit stress corrosion cracking of 18-8 stainless steel. The critical applied potential in MgC12 solution (--0.145v) above which, but not below, cracking occurs is shifted in the noble direction by extraneous salt additions. When the shift exceeds the corrosion potential for 18-8 in the same solution, cracking is apparently inhibited. On the other hand, salt additions, e.g., FeC13, which shift the corrosion potential in the noble direction may induce or accelerate stress corrosion cracking.The critical potential is interpreted as that value above which but not below C1-ions adsorb on imperfection sites of plastically deforming metal in amount adequate to cause failure (stress sorption cracking). The present data do not support an electrochemical mechanism of stress corrosion cracking based on anodic dissolution of metal ions at the tip of a crack, nor the mechanism dependent on continuous cracking of a surface oxide film.The phenomenon of failure of ductile metals by cracking when stressed in tension and exposed to specific anions is still not well understood. Several theories have been proposed since the time stress corrosion cracking (scc) was first observed, the most important of which are: (i) electrochemical, based on dissolution of metal ions at the base of a notch or crack acting as anode in contact with metal elsewhere acting as cathode; (ii) progressive cracking of a brittle surface oxide film; (iii) reduction of surface energy or weakening of metal bonds by adsorption of specific anions at the root of a notch or crack. There is supporting evidence for each theory, but no general consensus as yet that any one of them can adequately account for the general phenomena of scc (1).The mechanism by which extraneous salts, when added to damaging solutions, act as inhibitors of scc, and in some instances as accelerators, challenges each of the currently discussed theories and provides the basis for the present paper. The work reported at present began with our observation that a few per cent of a salt, e.g., sodium acetate or sodium iodide, added to the usual boiling concentrated MgC12 test solution for 18-8 stainless steel effectively inhibited scc. This fact excited curiosity as to how such additions could in principle affect any supposed electrochemical process occurring within the growing crack, particularly when the added salt had no readily discernible major effect on cathodic or anodic polarization behavior, and certainly had little effect on the conductivity of an already good conducting solution [K ~ 0.27 ohm -1 cm -1 at 146~ (2)]. The cracking of an oxide film, if one construes the passive film to be an oxide, did not seem to be involved in view of the fact that polarization measurements of stainless steels in MgC1._, solution indicated absence of passivity (2, 3). Visible patches of surface oxide, if any, that may form during tests in MgCI., are equally evide...