tions. By means of a table of random numbers, the two lesions that would receive an application from tube A were determined (two of lesions 1, 2, 3, and 4). The other two lesions received an application from tube B. 15. After the two lesions that were to receive an application from tube A were determined, one was randomly selected to be a biopsy site; likewise, the two lesions that were to receive an application from tube B were selected randomly. Therefore, one of the two lesions from each treatment was scheduled to be biopsied but the code (tube A and tube B) was not broken until after biochemical analyses were completed. The two remaining lesions, one having received a glucocorticoid and one a control cream treatment, were not biopsied but were left to be visually evaluated 2 days later. 16. J. J. Voorhees, E. A. Duell, M. Stawiski, E. R.Harrell, Adv. Cyclic Nucleotide Res. 4, 11717. The surgically removed treated or control tissue used as starting material in this report is designated by several terms which are meant to be synonymous. These terms are lesional epithelium,-lesional epidermis, lesional tissue, and diseased tissue. By frozen section histology we estimate that keratinocytic epithelium (epidermis) occupies 80 to 95 percent of the specimen volume. The 80 to 95 percent range is the result of surgical technique, vertical lesional configuration, and the presence of non-keratinocytes (inflammation-associated cells, endothelial cells, fibroblasts, and collagen). Therefore, although 5 to 20 percent of the specimen volume is occupied by collagen and nonepithelial cells, the latter of which probably contain the arachidonic acid transformation cascade (Fig. 2), it seems probable that the data in Fig. I are derived mainly from lesional epithelium. 18.