148 JOURNAL OF THE SOCIETY OF COSMETIC CHEMISTS Table V Derivation of Various Methods of Data Analysis a Control Collections Post-Treatment Collections Axilla 1 2 3 4 Mean i b 2 3 4 Mean R (T) CT CT CT CT CT•t PTT PTT PTT PTT PTT•t L (C) CU CU CU CU CU M PTU PTU PTU PTU PTUxt Ratio R R R R R•t R R R R Adiusted ratio AR AR AR AR AR M Procedure A 100 -- (PTT•/q- CT M X 100) •- % reduction B 100 -- (PTT•t .'-- PTU M X 100) = % reduction C 100 -- (AR X 100) ---- % reduction Symbol Identification. C = control collection T • treated axilla PT • post-treatment collection U = untreated control axilla R = nnadiusted ratio M subscript = mean values used in calculations AR = adjusted post-treatment ratio, post-treatment ratio mean control ratio First post-treatment co]lections were made following two applications. Table VI Comparative Results (Per Cent Reductions) Obtained in Three Methods of Analysis Procedure Test A B C 1 24.6 m 13.2 33.7 _ 3.7 30.7 --+_ 3.7 2 29.9 m 7.8 23.0 _ 5.1 21.5 +__ 4.8 3 60.6 _ 8.0 55.6 m 5.2 56.2 +__ 3.1 4 4.1 --+_ 10.4 6.1 _ 7.4 8.2 +__ 5.4 variation as evaluations are dependent on day-to-day variations in sweat rate of the individuals. Procedures B and C should yield comparable results provided the mean right-to-left sweating ratios of the panelists approach 1.000. Due to marked day-to-day variations in sweat output from the same axilla, tests carried out by Procedure A may indicate significant properspirant effects on one day and marked sweat reductions on the following day. A modification of Procedure C has proven to be particularly valuable in demonstrating significant differences of samples of similar efficacy. In this modification the normal control sweating ratios are determined as above. Both axillae are treated and shifts in the adjusted post-treatment ratios are determined. All ratios-controls and post-treatment-are made by assigning axillary treatment designations to correspond in a uniform manner, such as

ANTIPERSPIRANT EFFICACY 149 with Samples A and B to T and U, as shown in the above format, so ratios are calculated A/B for all panelists during control and post-treatment peri- ods. While this procedure is very sensitive in showing differences in antiper- spirant activity, which is its primary purpose, estimations of reductions of both samples can be also made by applying Procedure A to the data. PROCEDURE REPRODUCIBILITY AND PANELIST VARIABILYrY In compiling data which demonstrate the reproducibility of the outlined procedures, we frequently include a 10% w/w aqueous solution of aluminum chlorohydrol as a reference sample in cross-over studies. These evaluations have always been based on observed shifts in sweating ratios. Values from six 12-panelist studies for this reference are shown in Table VII. These were carried out as cross-over studies with formulated aerosol or lotion antiperspirants. The reference sample was applied by swabbing 0.5- ml aliquots over the designated axillae in Tests 1, 3, and 6, and by applying a 2-sec spray from a Preval©* aerosol unit in Tests 2, 4, and 5. Table VII Comparison of Values (Per Cent Reductions) i •and 2 Hours Following Sample Application Test No. Panelists No. of Applications 3 4 Post-Treatment Interval I hour 2 hours I 12 49.0 ___ 10.0 47.0 ñ 8.6 2 11 43.0 ñ 7.0 41.0 ñ 6.4 3 12 •3.0 ___ 6.8 37.0 ñ 8.1 4 12 33.3 ñ 10.0 31.3 ___ 11.5 5 12 46.4 ñ 7.6 . . . 6 12 37.3 ñ 7.2 32.9 q- 10.0 The variations in reductions in sweating observed in these six studies, 33.3% to 49.0%, are due to the comparative responsiveness of the panelists to appli- cations of aluminum chlorohydrol. This panelist-to-panelist variation is a problem we have not been able to overcome, but can be partially compen- sated for by carrying out evaluations on larger panels of 30 or more. This will overcome the variability to a large extent. It has been suggested that we exclude panelists who do not show a signifi- cant reduction in sweating. Our experiences indicate that this is not advisable. Failure of panelists to show reductions in sweating is not uniform for all test materials, particularly formulated samples. •Manufactured by Precision Valve Corp., Yonkers, N.Y.