66 JOURNAL OF THE SOCIETY OF COSMETIC CHEMISTS is also not consistent with information already in the literature regarding clinical testing of surfactants (2). Finally, the results of the appliance test reported in the Keswick article are very con- fusing and again inconsistent with the arm wash data presented in the same article. Although the product used was not identified, it may be inferred from the erythema score that the product is similar in irritation potential to product B. However, when using the towel, no irritation was observed--a lesser insult than that produced by the "milder" arm wash methodology. Keswick et al. do not attempt to explain these glaring inconsistencies. The description given of their consumer use model certainly cannot account for such inexplicable differences. In summary, the results of the study of Keswick et al. do not support any known models of surfactant-induced irritation, and the proposed model for irritation produced in a flex wash is simplistic and inaccurate. Not only are the data internally inconsistent, but they are contradicted by published information. These major discrepancies must be ad- dressed. REFERENCES (1) P. T. Sharko et al., Arm wash with instrumental evaluation--A sensitive technique for differentiating the irritation potential of personal washing products. J. Dermatol. Clin. Eval. $oc., 2, 19-27 (1991). (2) P. J. Frosch, "Irritancy of Soaps and Detergent Bars, in Principles of Cosmetics for the Dermatologist, P. Frost and S. N. Horwitz, Eds. (C.V. Mosby Co., St. Louis, 1982), pp. 5-12. Richard I. Murahata Unilever Research U.S. Gary L. Grove KGL, Inc. TO THE EDITOR: This responds to comments made regarding our article, "Comparison of exaggerated and normal use techniques for assessing the mildness of personal cleansers," which recently appeared in your journal. Specifically, this addresses criticisms that the article provided insufficient experimental detail and that the results presented are inconsistent. As we noted in our article, the forearm and flex wash methods examined were based on methods previously published in this journal (1,2), and the procedural details requested are available in the cited works. Neither the reviewers nor the editor deemed it necessary to repeat these details in our article. For those who may not have read the original articles, the forearm and flex wash studies were graded on seven-point and four-point scales, respectively. The forearm wash studies used 30-40 subjects for each product comparison, the flex wash implement study used 13 subjects for each treatment com- parison, and the flex wash studies used 40 subjects for each product comparison. These sample sizes met or exceeded the sample sizes specified in the original papers. The home-use studies were scored using the same grading scales as the forearm wash studies, with between 80 and 90 subjects randomly assigned to use each test product.

LETTERS TO THE EDITOR 67 The point made by our article is that exaggerated wash procedures differ in their ability to predict skin effects induced by personal cleansing products under conditions of actual consumer use. Specifically, our results indicate that the forearm wash method we used yields results that are more consistent with the skin effects experienced by consumers under home-use conditions than does the flex wash method. The correspondents selec- tively criticize these findings, not by presenting data that show the flex wash method to be predictive of skin effects observed under conditions of consumer use, but on the basis of results generated by a different forearm wash method and traditional theories of surfactant irritation. Neither of these arguments establishes a correlation between the product mildness predicted by the flex wash method and that observed when the product is used by consumers. In fact, the forearm wash results cited by the correspondents failed to differentiate the products in terms of their ability to induce dryness, a product attribute that is very noticeable to consumers. Further, the traditional arguments pre- sented by the correspondents do not hold for systems employing new technologies, specifically personal cleansing bars incorporating cationic polymers. The skin protectant properties of these materials are well-documented in the literature, and we have per- formed numerous studies that have demonstrated the clinical mildness benefits of this technology over simple, surfactant-based systems. The flex wash method differs from the arm wash method we used in that it involves a greater amount of mechanical manipulation (of the skin) during the washing process. The correspondents acknowledge that the flex wash method is an aggressive procedure, and the results from our implement experiment show that applying product with a sponge causes greater disruption of the stratum corneum barrier than does applying product with a towel. Overly aggressive wash methods may fail to demonstrate positive skin effects introduced by new technologies and, thus, fail to predict skin effects observed by consumers, because of the greater amount of barrier damage they induce. We believe that the disagreement noted by the correspondents casts doubt, not on the validity of our data, but on the ability of overly aggressive wash methods and traditional surfactant-irritation theories to account for skin mildness benefits afforded by new technologies. A second criticism offered by the correspondents is a lack of internal consistency of the data presented for the various products. This criticism is based on erroneous comparisons of absolute erythema scores generated by different wash protocols and scored on different grading scales and, hence, is unfounded. REFERENCES (1) M. F. Lukacovic, F. E. Dunlap, S. E. Michaels, M. O. Visscher, and D. D. Watson, Forearm wash test to evaluate the mildness of cleansing products, J. Soc. Cosmet, Chem., 39, 355-366 (1988). (2) D. D. Strube, S. W. Koontz, R. I. Murahata, and R. F. Theiler, The flex wash test: A method for evaluating the mildness of personal washing products, J. Soc. Cosmet. Chem., 40, 297-306 (1989). B. H. Keswick K. D. Ertel M. O. Visscher