Secondary alterations in Descemet's membrance were noted in corneal specimens from 20 patients undergoing transfixing keratoplasty, after examination by optical, and in some cases, by transmission or scanning electron microscopy. The most frequent anomaly, thickening of Descemet's membrance, was observed in cases of prolonged or recurrent corneal affections, particularly interstitial keratitis, advanced forms of keratoconus, cicatricial leukomas, and keratoconus. Thickening was due to juxtaposition, with a normal Descemet's membrance, of repair tissue made up of collagen interspersed with zones of cell necrosis and isolated fibrils. These different layers could be interpreted as successive deposits of reparative endothelial cells, the ultrastructural appearance of each layer being characteristic of the type of cell from which it was derived. In this way, the history of each cornea can be reconstituted, from the original normal endothelial cell to the final endothelial repair cells, passing through all the intermediary types of fibroblastic cells. This relationship between the cells which manufacture descemetic tissue and the structure of Descemet's membrance is particularly evident in Groenouw's macular dystrophy type II in which mucopolysaccarides are deposited. Descemetic verrucosities were observed in interstitial keratitis, of syphilitic origin or not, in one case of keratoconus, one case of buphthalmia, two cases of corneal leukoma, one case of vitreocorneal contact from aphakia, and in cases of Groenouw's type II dystrophy. The lesions are either isolated or spread across the posterior surface of the cornea. They differ from the verrucosities of primary cornea guttata by their irregular disposition, their frequent coalescence, the persistence of an endothelial regenerative process, their ultrastructural characteristics close to those of Hassal-Henlè's bodies, and more particularly by the association of a granular material with narrow and wide interval striated collagen. Various hypotheses are discussed which could provide an explanation why certain cells lay down Descemet's membrance in the form of tuberosities, and why these may present in alignments, but no satisfactory definite conclusion can be reached.