Abstract : The experiment to determine the isotope effect for diffusion of Cd(109) and Cd(115m) in pure lead was performed as a critical test of the proposed interstitial-vacancy-pair mechanism of solute diffusion. The relative diffusivity of the two radio-isotopes was determined, after sectioning, by counting with a thin NaI detector system, using energy discrimination. Stability and reproducibility of the count rates were tested and verified, an upper limit on the effect of crossover shown to be negligible, and a zero isotope effect experiment performed successfully. The isotope effect for the volume diffusion of cadmium in lead was determined to be (D(109)/D(115m) - 1) = 0.0032 at 248C, corresponding to a solute correlation factor f = 0.12 for a mechanism involving single atomic jumps. It is then shown that this result, together with previous results for the solute and self-diffusivities D(0)2 and D(0)1 and the linear enhancement factor b11 for self-diffusion, are inconsistent with the simple vacancy, interstitial, interchange and ring mechanisms of solute diffusion, but agree with predictions of the model for solute diffusion by means of interstitial solute - vacancy pairs. The migration of associated vacancies results in a strong correlation of successive atomic jumps of bound interstitial solute atoms. (Author)