JOURNAL OF THE SOCIETY OF COSMETIC CHEMISTS It is seen that, on the average, a conclusion is not likely to require more than about 500 tests and in most cases--particularly those when the product is rejectable--many less. Reference to the Dodge and Romig technique will show that the sequential method has advantages. The required number of tests curve, although showing that on the average a smaller number of tests is required by the sequential technique than by the Dodge and Romig scheme, still requires a very large number of tests. In normal circumstances the time taken by the evaluation, the personnel required to carry it out and the expense of organising such a large number of tests on subjects selected at random from the possible users of the product, make it impracticable, and it is necessary, therefore, to devise a more compact prophetic patch test to replace the present ones. However, the above discussion has been useful as an example of sampling techniques, as the latter have other applications in the cosmetic industry. For example, the Dodge and Romig scheme can be used when examining any consignment of containers or other goods which can be judged in any of their qualities by a "go--no go" method, and is useful wherever automatic filling and packing machines are used which require a supply of accurately controlled materials. Sequential analysis is not so suitable for this purpose, since it means continual random sampling from the consignment of goods as testing proceeds, which is possible only when the tests can be carried out quickly at the site -of the goods the Dodge & Romig scheme calls for a fixed number of samples drawn at random which can be inspected elsewhere at a convenient time. Sequential analysis can be applied to the examination of two different products to decide which is to be preferred: this is useful in market research and will be considered in a later article on that subject. PROBIT ANALYSIS Substances which give rise to a dermatitis are classified into primary irritants and secondary, or sensitisation, irritants. Primary irritants are substances which produce an irritant effect on first contact with the skin of any individual: secondary irritants produce irritation only after the substance has at some previous time been in contact with the skin of a susceptible person. Primary irritants are not specific, but act by partial or complete destruction of the epidermis: secondary irritants probably produce an antibody in the dermis after the application of the substance and give rise to the dermatitis on the next application of it. This reaction may be delayed. •, 3,, The reaction produced by a primary irritant is dependent on its concen- tration. Sulphuric acid, caustic alkalies, phenol, alkaline thioglycollates and inorganic sulphides are typical primary irritants. Secondary irritants :.. 232

!}77½':4 STATISTICAL METHODS IN THE COSMETIC INDUSTRY i' •:5•'Undergo some reaction with the prickle cells of the basal layer, which results }?in the liberation of histamine, this giving rise to the dermatitis. (Local }i :[:' application of an antihistamine will frequently give rapid relief from the {?•iMtation.) Typical irritants are flower pollen, orris root, linalool, methyl •:[•.:•:[ :heptine carbonate, terpineol, hydroxycitronellal, rosin, penicillin, sulphona- • '?'mides, paraphenylenediamine, chromium salts, chlorxylenols, bromoeosine, bergamot oil, etc., the last two being photosensitisers activated by U.V. [:•{[:light to give reactions, and the penultimate two primary irritants at high i.:•: ' concentrations. Everybody is probably sensitive to some substance or other, and it is •(•:•:-sometimes found that a person suddenly becomes sensitised to a substance which has harmlessly been in frequent contact ruth the same area of the .:•:. sk• for years. ::/•:' The prophetic patch test is concerned with secondary irritants, primary irritants being easily eliminated from consideration after a dozen or so tests. The only way to discover whether a product used at a very low concentration :::is likely to be a secondary irritant to more than a very few people is to use one of the techniques discussed above on a random sdection of possible users of the product. However, a way can be found by which the number of the •sensitised reactions obtained from a substance can be increased and the effect of a new substance assessed relatively to that of a substance whose use is well established. Schwarz, • when testing a deodorant soap, carried out patch tests at various concentrations to determine the concentration at which a solution of a well-kno• (presumably widely sold and accepted as safe) toilet soap would not cause a reaction on most of his ten test subjects. tt. F. Smyth, et al,* found that such substances as glycerin, lanolin, petroleum jelly and cocoa butter, can give rise to sensitisation reactions and are secondary i•itants to an extent of gi•ng between 1 to 3 positive sensitisation reactions among 100 test subjects. The materials also gave reactions on the first application to the skin and, since they are not primary irritants, it must be concluded that those persons so reacting had pre•ously been sensitised, thus bringing the number of reactions out of 100 persons to 4 (glycerol, cocoa butter, petroleum jelly) and to B 0artolin). It is common knowledge that dilute solutions of these substances do not give rise to anything like this per- centage of secondary reactions, and therefore it must be concluded that con- centration of substance has some bearing on the matter. If the prickle cells of any person's skin include a cert ain proportion w•ch Mll react •th a given sub- stance to liberate histamine, and if the distribution of those persons with each proportion of such cells follows the normal Gaussian curve, there •11 be a few people who Mll be sensitised by a weak concentration, more people who •11 be sensitised by a higher concentration and so on. Increasing the concentration and/or the period of contact should increase the number of sensitised reactors. 233