The paper presents a detailed reaction model for the corrosion of III–V compounds which quantitatively represents the experimental data (mainly Tafel plots of current vs. band displacement), presented in Part I J. Electroanal. Chem. Such data allow the discrimination between different reaction schemes. From a physical point of view, the model shows clearly that the differences observed between GaAs and InP arise from their different electronic surface properties. From a chemical point of view, a chemical signature of the corrosion states of GaAs is proposed: the states EC −0.98 eV and EC −1.15 eV, previously derived from photocapacitance spectroscopy, Ber. Bunsenges. Phys. Chem. 92 (1988) 895, could be associated with As sites and Ga sites, respectively. An analogous description holds for InP, with different positions of states. Within the framework of these original data, corrosion can be described by a first hole capture process on the cation site which becomes an anion site by the liberation of the cation. A chemical reaction step occurs before the second hole capture proceeds to dissolve the anion-related defect. Such a description is predicted by thermodynamical data and is further sustained by experimental facts such as the well-known As enrichment of GaAs surface in acidic media.