EVALUATION OF ANTIPERSPIRANTS UNDER NORMAL CONDITIONS 19 To adapt the gravimetric method to be suitable for assessing anti- perspirant preparations under normal conditions of use, Wooding et al. (6) used silica gel drying tins of the type used in analytical balances. These were strapped to the backs of subjects and had the advantages of being easy to handle, reusable and efficient. From consideration of the different types of procedure described above, it was clear that the gravimetric method was the most adaptable for assess- ment of antiperspirant effect under normal conditions of use, if the ratio technique was employed. As to the method of sweat collection, it was found, after experimentation, that the moisture absorbing tins of Wooding et al. could be strapped to the body side of the axilla with waterproof tape, without greatly inconveniencing the tester, and that they gave more reliable and accurate results than the various pad and cup arrangements tried. The experimental method described in this paper is, therefore, a gravimetric one, whereby silica gel moisture absorbing tins are used to collect axillary perspiration and the change in ratio of sweat production between axillae for an individual, when only one axilla is treated, is a measure of anti- perspirant effect. TEST PROCEDURE The particular moisture absorbing tins selected were Silica Gel Air Dryers*, size No. 2, 60 mmx 9 mm (Fig. 1). These are perforated disc- shaped containers of silica gel, which are capable of absorbing 8-9 g of moisture. They are supplied in aluminium outer containers. The rate of loss of moisture by evaporation, from a used tin, is negligible for a tin in its container left on the bench for half an hour. Thus, during the time that the tin is removed from the axilla, placed in its container and weighed, there is negligible evaporation of sweat. The waterproof tape used is a micro- porous, plastic, self-adhesive plaster, Elastooe1ast Airstrioe?, which was supplied in rolls 1« inch wide. The tins and containers were numbered, and were used randomly during a test. ß At approximately 09.00 every working morning, the tins were weighed in their containers to the nearest milligram. Subjects were asked to dry their axillae with tissue paper, and the tins were removed from their * Silica Gel Limited, London,. ? T. J. Smith and Nephew Ltd, Hull and Welwyn Garden City.

2O JOURNAL OF THE SOCIETY OF COSMETIC CHEMISTS containers and placed in position on the body side of the axilla, as close to the apex of the axilla as possible without causing discomfort. They were completely covered and held in position by three horizontal strips of plaster (Fig. 2). Three hours later the tins were removed, placed in their respective containers and re-weighed. The tins were dried in the afternoon in a vacuum oven at 80-90øC for about 2 h, when they returned virtually to their original weight each time. Application of test materials was carried out by indivdual panelists at home. One arm of each subject was arbitrarily designated as the 'treatment arm' and the other as the 'untreated or control arm'. The ratio of the weight of sweat obtained from the treated arm to that obtained from the untreated arm, as explained above, was taken as a measure of the efficacy of the treat- ment and this had the advantage that such diverse factors, as air tempera- ture, humidity and the activity of the subject, that affect the level of sweat- ing, are automatically compensated for. In addition, the ratio is independent of the time for which the absorbing tins are in place, so any discrepancies in collecting time do not affect the results. Mixed panels of approximately twelve subjects were used, and panel members were asked to shave the axillae a few days before the test. They were also requested to shave during the test when necessary. The panellists were asked to abstain from using an antiperspirant 2 weeks before the start of the test, and unperfumed soap was supplied for use during the test. The panellists were also supplied with a mild aerosol deodorant containing a very low level of cationic germicide, which they were allowed to use during control periods, if desired. The antiperspirant treatment was applied after washing in the morning and in the evening. No attempt was made to control the quantity of application of the test product by the panellists, as: (a) evaluation under normal conditions of usage was required (b) the surface area of one axilla to the next for different panellists varied very considerably (c) individual panellist accuracy in using aerosol products was very variable. The experiment was divided into three phases pre-treatment, during which neither arm received any treatment treatment, during which the pre- designated treatment arm received the specific treatment and finally a post- treatment phase, during which no treatment was given. The inclusion of a post-treatment phase, was to observe the carry over effect of any of the