182 JOURNAL OF THE SOCIETY OF COSMETIC CHEMISTS 2. TEST PROCEDURE A gravimetric procedure was used to compare the efficacy of four test antiperspirant products during a normal work day in each subject's normal work environment. Eight subjects (six females and two males) followed guidelines, described in a nine-week experimental protocol, at their own homes and jobs. The volunteers were asked to maintain normal bathing habits except for the axillary hygiene regimen throughout the test. The axillary hygiene regimen consisted of abstaining from non-test antiperspirant/ deodorant use, removal of axillary hair with a surgical depilatory Surgex* (Cooper Scientific Corporation, Watertown, Mass.), three days before each sweat collection period, and washing the area with Ivory* soap (Proctor and Gamble, Cincinnati, Ohio) and water only. Sweat was collected on tared absorbent 100% cotton Webril* (KendallFiber Products Division, Walpole, Mass.) pads taped snugly to the axillae with Dermicel* (Johnson and Johnson, New Brunswick, NJ.). A 4-ply Webril* pad was sewn onto a 2 x 4 inch piece of Ziploc* (Dow Chemical Company, Indianapolis, Indiana) to ensure retention of sweat on the pad and easy removal of tape without loss of padding. Each pad was sealed in a Ziploc* bag and weighed. All bags were labeled with the subject's name and the axilla and date each pad was to be worn. The tests were organized so that there was a two-week abstinence from antiperspirant use before the treatment sweat collections began. Subjects abstained from antiperspi- rant use for one week before the three pretest sweat collection readings v, ere made. At the start of each collection day, the subject washed both axiilae with Ivory* soap and water, and patted them dry. The appropriate pad, cotton side to the body, was taped Table II Pretest Sweat Collection Data Used for Blocking 4 by 4, Split Plot, Latin Square Sweat Collection Dates 4-4-78 4-5-78 4-6-78 Right Left Right Left Right Left Average Subject (mg) (mg) (mg) (mg) (mg) (mg) PR* B•D•(PR) 208.0 304.0 564.3 540.3 707.4 898.5 0.684 1.044 0.787 0.838 B•D2(PR ) 115.2 116.2 157.4 231.6 71.1 78.1 0.991 0.680 0.910 0.860 B2D•(PR) 237.2 192.2 56.3 34.2 555.4 262.2 1.234 1.647 2.118 1.666 B2D•(PR) 490.5 293.3 554.0 236.0 230.3 80.1 1.672 2.347 2.875 2.298 B3D•(PR ) 1,257.4 1,202.1 289.3 298.9 865.8 916.2 1.046 0.%8 0.945 0.986 B3D2(PR ) 836.0 630.0 468.1 710.2 %.5 119.6 1.327 0.659 0.807 0.931 B4D•(PR) 1,742.2 1,614.2 1,482.7 1,255.6 1,567.2 781.1 1.079 1.181 2.006 1.422 B4D•(PR ) 520.8 2%.4 1,375.6 970.4 1,592.3 918.2 1.757 1.418 1.734 1.636 Perspiration weight right axilla *PR = Perspiration Ratio = Perspiration weight left axilla

VEHICLE EFFECT ON ANTIoPERSPIRANT ACTIVITY 183 snugly to the axillae and worn throughout a normal day's conditions and environment. After a 10-hour collection period, the tape was removed, and the pads were returned to the designated bag and sealed. The difference in weight of the tared Ziploc © plus dry pad and Ziploc © plus pad following the 10-hour collection period was recorded as sweat collected. The weight was expressed in milligrams to the nearest tenth of a milligram. Pretest sweat data are listed in Table II. Pretest sweat ratios were calculated as Perspiration weight from right axilla Perspiration weight from left axilla -- Perspiration Ratio (PR) and averaged (15). These data were used as a blocking factor in designing the 4 x 4, split plot, Latin Square. Subjects were paired by their PR. One subject of each pair was treated on his heavy-sweating axilla, while the other subject was treated on his light-sweating axilla. For example, the PR 1.636 of subject B4D 2 and 1.422 of subject B4D • indicated both subjects sweated more on their right sides. subject B4D2's light-sweating or left axilla was chosen randomly as the side to be treated for the duration of the test therefore, subject B4D•'s heavy-sweating or right axilla was treated. Their opposite axillae were assigned as controls. No antiperspirant was applied to the control axillae during the test. The subjects were supplied with the appropriate test antiperspirants in tared sealed containers at the start of each treatment week. The two members of a pair applied the same test antiperspirant during a given week. The four pairs applied different treatments during a given week conforming to the Latin Square Design (Table III). The cream was applied directly to the axilla from glassine paper. The stick was applied Table III 4 by 4, Split Plot, Latin Square Design C• C2 C3 C4 B•Ds As A2 A• A 4 -.395* -.783 -.379 -.457 B2D s A2 A 3 A 4 A s -.543 --.959 --.384 +.006 B3Ds A3 A4 As A2 --.359 --.045 --.283 +.035 B4D s A 4 A• A• A 3 --.301 +.316 --.064 --.566 B•D: A• A• A 3 A 4 +.407 --.295 --.361 --.664 B2D2 A2 A 3 A 4 As --.252 --.812 --.658 +.118 B3D • A3 A 4 A• A• --.467 --.637 --.809 --.658 B4D 2 A 4 A• A2 A 3 +.015 +.554 --.497 +.079 A = Treatments: 1, stick 2, cream 3, lotion 4, hydroalcoholic base B = Subject Pairs, 1, 2, 3, 4 C = Weeks, 1, 2, 3, 4 D = Side Treated: 1, heavy-sweating side 2, light-sweating side *z = In if/x-) = In ([Ys + Y, + Y,l/[xs + x• + x,])