The purpose of this work was to compare experimental diffusion among three gases with the solution given by Stefan's equations to understand better how this phenomenon can work in the multicomponent alveolar gas. Experiments were performed in a cylinder full of beads open at one end and closed at the other in which a mixture of oxygen with helium or argon or sulphur hexafluoride could diffuse with ambient air through the open end. We solved Stefan's equations for the non-steady state by a finite-difference method and applied them to our experimental conditions with diffusion coefficients we had measured in binary experiments. We then made experiments and calculations to show the influence of the beads on gas transport. Provided that diffusion is the only phenomenon, experimental and theoretical curves are very close together. Moreover beads nearly stop motions due to vortices or small differences of density. We conclude that: Stefan's equations should replace Fick's equations when more than two gases are involved. One should bear in mind the possible influence of gravity and devise diffusion experiments accordingly. In small spaces such as alveoli the influence of gravity must be negligible compared to diffusion.