A considerable amount of effort has been directed at finding methods for modifying the nonkeratinized sulcular epithelium on the assumption that a keratinized surface may offer a better barrier to antigens and bacterial products present in the gingival sulcus. It is argued here that keratinization in itself may not confer greater impermeability, for nonkeratinized epithelia also have been shown to resist the penetration of certain substances. Moreover, few workers have considered the role of junctional epithelium in the initiation of periodontal disease although experimental evidence suggests that this may be a permeable tissue. As formation of a surface with barrier properties seems to be a concomitant of epithelial differentiation while attachment is a property of relatively undifferentiated epithelial cells, attempts to induce junctional epithelium to differentiate could result in a loss of epithelial attachment to the tooth. It is suggested that attempts to keratinize the sulcular region, on theoretical grounds, may be unjustified.