HUMAN CUMULATIVE IRRITATION TEST 445 posure was made and a score of 3 was carried forward for all scoring days for the remainder of the test. The following statistical analyses were performed: 1. Analysis 1 was performed on the data from the 21 consecutive days. 2. Analysis 2 was performed on the first 14 consecutive days from the 21-day data. 3. Analysis 3 was performed on the data from the 14-day every-other-day patching scheme. 4. Analysis 4 was performed on the first ten consecutive days from the 21-day data. 5. Analysis 5 was performed on the first seven consecutive days from the 21-day data. 6. Analysis 6 was performed on the first four consecutive days from the 21-day data. The transformed patch test scores for each test article were summed across all test days for each test subject. The overall test article scores were ranked within each subject and then analyzed using the Friedman rank sum test (4). The hypotheses for this test are shown below: Ho: The rank sums of the test articles were identical. Ha: At least two of the rank sums differed. Multiple comparison testing was performed using Fisher's LSD test (4) if significant differences (p -- 0.05) were observed with the Friedman rank sum test. NOTE: For the 21-day test the maximum summed score for a subject for any test article is 63, which would indicate the development of a strong reaction on test day 1 (score of 3 for 21 days = 63). The most irritating test article in this study, test article D (2.5% lactic acid), had a mean total score of 51.95, indicating that most subjects achieved the maximum score of 3 by test day 3 or test day 4. RESULTS Forty-one subjects, 18-50 years of age, completed the test. ANALYSIS 1 (21 CONSECUTIVE PATCH APPLICATIONS) The Friedman rank sum test indicated significant differences among the total reaction scores for the test articles (p = 0.0001). Followup multiple comparison tests of the data yielded the following result: © The six test articles were found to be statistically significantly different from each other (Table I). ANALYSIS 2 (UTILIZING THE FIRST 14 DAYS OF THE 21-CONSECUTIVE-DAYS DATA) The Friedman rank sum test indicated significant differences among the total reaction scores for the test articles (p = 0.0001). Followup multiple comparison tests of the data yielded the following result: ß The six test articles were found to be statistically significantly different from each other (Table II).

446 JOURNAL OF COSMETIC SCIENCE Table I Analysis #1 21 consecutive -atch a-•lications Si nificant Com arisons (0.05) N=41 Mean Total B /r A C ' ' E Reaction Score 3.41 4.46 8.46 20.24 47.78 LSD Grouping • D 51.95 •Means joined by the same grouping are not significantly different. Table II Analysis #2 Utilizing the first 14 days of the 21 consecutive days data Significant Comnarisons (n0.05 N=41 Mean Total "'/r F , A C ' •. ' F ..:..i..•D Reaction Score 1.80 2.73 5.20 10.88 28.02 31.90 LSD Grouping • •Means joined by the same grouping are not significantly different. Table III Analysis #3 14-day data -- patch applications every,,9ther day Significant Comparisons (n0.05) N=41 Mean Total :•'B .... •.F •]-•i•.. : .•E' "•• ReactionScore [1.12. I ,1.90 I 2.63 I 3.04 I 10.10 I 11.88 LSD Grouping • •Means joined by the same grouping are not significantly different. ANALYSIS 3 (14-DAY DATA--PATCH APPLICATIONS EVERY OTHER DAY) The Friedman rank sum test indicated significant differences among the total reaction scores for the test articles • = 0.0001). Followup multiple comparison tests of the data yielded the results shown in Table III: © Test articles B and F were found to be statistically significantly different from test articles A, C, E, and D. © Test articles A and C were found to be statistically significantly different from test articles E and D. © Test article E was found to be statistically significantly different from test article D.