This study was concerned with the measurement of volumes of surface epithelial cells in human small intestine, both in disease-control subjects and, for comparison, in patients with gluten sensitivity. Four procedures were employed, of which two were geometrical, based on cylindrical or truncated conoid models. The third method evolved from the proportionality of area to volume, and required determination of cellular and nuclear profile areas, and an estimation of nuclear volume based on models conforming to (1) a prolate spheroid, or (2) a cylinder with hemispherical caps. This procedure appeared to underestimate enterocyte volumes and failed to reveal volume differences between controls and gluten-sensitive individuals. Finally, a fourth method was devised, based on traditional intraepithelial profile counts per hundred enterocyte nuclei, calculation of surface epithelial volume and of the absolute number of lymphocytes contained therein. Enterocyte volumes appeared to be overestimated twofold by this procedure compared with the first two geometric methods. The results of this study indicate that the cuboidal-type enterocyte profiles typical of the untreated mucosa in gluten sensitivity are a reflection of cells with a reduced volume. From the number of enterocytes and the absolute lymphocyte population present within a morphometrically defined volume of surface epithelium, the ratios of intraepithelial lymphocytes to enterocytes were found to be 50% less than conventional density-profile counts.