116 JOURNAL OF THE SOCIETY OF COSMETIC CHEMISTS (c) The electrode reaction of the type E-2 glass electrode may be improved by soaking the electrode in pH 10 buffer solution prior to use. (d) The string may be prepared in the following manner. A ball of cot- ton twine is placed in a beaker, covered with distilled water, and the water heated to boiling. Heating is continued for several minutes, then the water is replaced with fresh distilled water and the operation repeated. This is done a number of times until visually the water is free of sizing or other material which may be leached from the cotton twine. This twine is then cut into appropriate lengths and stored under a saturated KC1 solution. (e) Each piece of string must be used only once. Reference Beckman General Instructions, "Type E-2 Glass Electrodes," Bulletin No. 361, May, 1955, Beckman Division, Beckman Instruments, Inc., Ful- lerton, Calif. We have also made a start on two other problems we originally set 'for ourselves, that of shampoo viscosity or consistency. Here we have pro- gressed only to the point of asking the members of the subcommittee for sug- gestions and ideas. As might be expected there is not the same meeting of minds as on the pH procedure thus far, we have received suggestions for use of viscosimeters of various kinds, as well as flowmeters and penetrom- eters. A cursory examination of the problem leads to the opinion that we may need not one but several methods for viscosity measurements, perhaps one for cream or paste shampoos, one for thin liquids and one for viscous liquids. Additional problems become evident on further con- sideration. For example, shall we consider the initial resistance to flow or the resistance after application of a shearing stress? Which experimental technique will tell us the most about the flow characteristics of the sham- poo? Perhaps the most cogent observation which has been made thus far is that we consider which attributes of the complex quantity we call "viscosity" we are most concerned with before we attempt to set up any procedures for its measurement. Standardized procedures for pH, viscosity, and other physical proper- ties of cosmetic products are extremely desirable so that we may, in dif- ferent laboratories, be able to speak the same language and have the same frames of reference. I believe that we have already considered some of the necessary benchmarks for such procedures. Among the most im- portant are reproducibility, simplicity, and use of readily available equip- ment. Important as such procedures are I feel that the need is even greater for standard evaluation techniques for the performance properties of shampoos and related products. I say that the need is greater because

EVALUATION OF SHAMPOO 117 there is so much less background and so much greater divergence of opinion on performance evaluations than on evaluation of physical properties. And, after all, our proclucts are purchased by the consumer chiefly for their performance and not for their physical appearance--or at least we like to think so. Recognized and valid procedures for evaluation of foaming power, detergency, curl strength--to name just a few properties important in shampoo evaluation--would go a long way toward achieving more ob- jective screening and more scientific study of raw materials and finished products. Reproducibility of a procedure is not suffident when we are dealing with performance characteristics, although it is, of course, an essential property. In the case of standardized procedures for physical properties we are setting up specifications for an inanimate object, the product itself. By the very definition of performance properties for cosmetics, we must concern our- selves with the interaction of the product and the human body. And this most decidedly introduces complications. We all know, of course, that no two human beings are identical and that the effect of a preparation on one person may be quite different on another. This is merely another way of stating that we must be sure that any performance test which is developed must be based on the use of sufficiently large groups of subjects so that the results may be considered statistically valid for the population at large. Other considerations come to mind, as well. A test procedure for per- formance evaluation must be not only reproducible but must be able to predict with at least fair accuracy the results which will be obtained under actual consumer usage conditions of the product. Whenever possible, the performance tests should be set up so as to study the interaction of the product with the portion of the body to which the product is normally applied. Thus studies of shampoos should be made using human hair studies of deodorants and antiperspirants should be made on the axillae, etc. Only when it is impossible to develop such direct measurement techniques should we consider other test procedures. It is possible, very often, to develop procedures using other than human hair or skin as the substrate but in all such cases steps must be taken to prove that data obtained in such a procedure can be correlated with results on the appropriate portion of the human anatomy. Not only must a procedure not based on actual usage conditions give the same relative order of effectiveness for a group of products as obtained under usage conditions but also the order of magni- tude of the effects should also be the same. Otherwise, erroneous inter- pretations may be drawn from such data. To summarize very briefly what I am saying, the function of a perform- ance evaluation test is to enable us to work under standardized conditions to improve the performance of our products the closer we keep to actual