MILDNESS ASSESSMENT TECHNIQUES 191 differences between these exaggerated methods is the implement used to apply product. The sponge is, by nature, rough, while a towel presents a smooth surface hence, the implements differ in their abrasive potential. To determine what impact, if any, implement abrasiveness had on the study outcome, a modified flex study was run in which a single product was applied in the usual fashion, or with a Masslinn © towel. Visual grades assigned during the course of this experiment are summarized in Figure 1. The implement effect is clearly visible in these data both the rate and magnitude of erythema development are greater when the product is applied with a sponge. In addition to visual grades, TEWL values were recorded throughout the implement study to provide a measure of the damage induced in the stratum corneum by the treatments. These data are summarized in Figure 2. To provide a more quantitative evaluation, the data were analyzed by approximating the change in mean TEWL value, with time for each treatment as a linear function and calculating least-squares regression coefficients for the fitted lines. The calculated intercepts provide an indication of the stratum corneum integrity at the start of the study, i.e., the baseline TEWL value. The intercepts calculated for the sponge and towel treatments are not significantly different (p • 0.05), showing that both treatment groups began the study with equal barrier function. The calculated slopes provide an indication of the rate at which each treatment induces stratum corneum damage, a higher slope indicating a greater rate of damage. 2.0 I 1.5- 1.0- 0.5- 0.0 0 5 10 15 Wash Cycle Number Figure 1. Mean erythema scores of subjects' arms washed with a sponge or towel in the flex test plotted as a function of wash cycle number. O = sponge O = towel.

192 JOURNAL OF THE SOCIETY OF COSMETIC CHEMISTS 18- 'E I.l.I 16- 14- 12- 10- l i I I I I I I 2 3 4 5 Day Figure 2. Mean daily transepidermal water loss values obtained from subjects' arms washed with a sponge or towel plotted as a function of time. The fitted lines were obtained by least-squares regression. O = sponge, y = 8.29 + 1.70X ß = towel, y = 8.33 + 0.33X. Although both treatments induce damage, the rate is greater when a sponge (slope = 1.70) is used to apply the product rather than a towel (slope = 0.33). The greater rate of stratum corneum damage induced in the flex wash method is not due solely to the abrasiveness of the implement. To demonstrate this, a test leg was included in the modified flex study in which no product was applied, i.e., the subjects' arms were "washed" with a moistened sponge to provide an indication of how much barrier damage was due to the sponge alone. Treating the TEWL data generated in this test leg as described above yields a line having a slope value of 0.39, which is only slightly greater than that observed when the test product is applied with a towel. Clearly, both the sponge and the product contribute to the greater rate of stratum corneum damage. The small amount of damage induced when the product was applied with a towel, however, indicates that the product contribution is not due to skin surface effects the contribu- tion must be due to product interacting with lower skin layers exposed after the stratum corneum is damaged by the sponge. This suggests that a two-stage mechanism is operating in the flex method: an initial stage in which the stratum corneum is damaged by the sponge, followed by a stage in which product comes in contact with lower layers of the skin, inducing an inflammatory response that contributes to further barrier breakdown. These findings point out a key difference between the forearm and flex wash methods. The forearm wash method provides an indication of product effects on the surface of the