At least seven models have been proposed for the epithelial transport of IgA, and each model presents particular features which are not generally appreciated. Much of the confusion in this field has been caused by the many conflicting reports about the cellular origin of the secretory component (SC) and the mode in which it is expressed by secretory epithelial cells. The transport model proposed in 1973-74 on the basis of test-tube experiments and immunohistochemical studies has now gained considerable support from observations made on both normal and neoplastic living epithelial cells According to this model, the J ("joining') chain and SC represent "the lock and key' in the selective external translocation of both dimeric IgA and pentameric IgM through serous-type secretory epithelial cell. Incorporation of J chains into these two Ig isotypes during their production in gland-associated immunocytes induces a configurational fit (binding site) allowing them to combine by specific non-covalent interactions with SC in the plasma membrance of the epithelial cell. After being formed on the basolateral surface of the cell, the SC-IgA and SC-IgM complexes are transported in cytoplasmic vesicles to the gland lumen along with some free SC. Covalent stabilization of human secretory IgA during this process depends on unique possibilities for disulphide-exchange reactions and is not an inherent feature of the transport model.