The possibility of constructing a systematic theory of irreversible processes is surveyed in general terms, by utilizing some of the results established in later parts of the paper. Three assumptions underlying Gibbs' application of the second law to equilibrium problems are formulated in explicit but general mathematical form. It is shown that they restrict the equations governing irreversible changes. The theory of a general fluid mixture is developed in some detail, and is then applied to mixtures of ideal gases. It is shown that the usual equations for the velocity of chemical reactions are consistent with the second law provided that the departure from equilibrium is not too great. Mathematical complexities make it difficult to decide whether this is the case for larger deviations also. A somewhat general theory of diffusion and heat flow is considered and the requirements of the second law are formulated as the positive definiteness of a certain matrix whose elements depend on the diffusion coefficients, thermal conductivity, etc.

• Received 19 April 1940

DOI:https://doi.org/10.1103/PhysRev.58.269