PHARMACOLOGY OF CHEMICAL IRRITATION 167 subject. While the problem is vital to the study of analgesic drugs, we find that in the study of relative irritation, pain plays only a minor role. In practice the amount of pain produced in the experimental animal is recorded by such relative terms as mild, moderate, severe, extreme, depending upon the visible reaction of the animal as reflected in blinking, pawing, phonation, etc. A test, which has special sig- nificance for compounds coming into contact with mucous membrane is based on the ciliary activity in the mouth of the frog. Following suitable preparation, it is possible to measure the speed of normal ciliary motility by timing the passage of a small object such as a particle of cork along the roof of the mouth. Bathing the ciliary bed with a test solution may accelerate, retard, or stop the motility. High dilutions of a given compound may cause acceleration while more concen- trated solutions may permanently abolish the activity. It is a fas- cinating method and its implications are widespread. These are the basic tests for pharmacological evaluation of irrita- tion in the experimental animal. Each has been subjected to a great many variations and refinements in- volving intricate machines, elab- orately controlled conditions and ex- perimental designs, carefully de- scribed end points for the assign- ment of arbitrary values and statisti- cal evaluation of the results. All of these factors help immeasurably in the study of irritatiorl, but the original goal of simplicity consistent with reliability and quantitation is still far from sight. Each new chemical still presents its own prob- lem and rarely, if ever, will a single technique sut•ce to properly evalu- ate it with respect to other irritants. From a practical standpoint, how- ever, it is possible to outline a few general principles which will improve the reliability of the laboratory data. First consideration should be given to the appropriateness of the technique to be used, the speed and economy with which they can be per- formed, and the number of different techniques to be applied. It is usually possible to get a preliminary screening with the simpler and more rapid tests, after which the more quantitative tests may be per- formed. Since there are many variable factors involved in any of the tests, including animal variation and human fallibility, it is advisable to select a suitable reference stand- ard irritant wherever possible. This reference irritant might be one of the well-known irritant chemicals or, where product change or im- provement is the objective, it should preferably be the present product if such is a stable, reproducible item. The important thought in favor of a reference is that it not only improves the reliability of the individual data, but acts as a "season skipper" by integrating all the data collected over a period of time into terms of a common denominator. Interpretation of the data may usually be improved if it can be put

168 JOURNAL OF THE SOCIETY OF COSMETIC CHEMISTS into numerical values. , Two princi- pal approaches' are open to nu- merical evaluation, depending upon the circumstances. If only a limited number of compounds are to be studied, say six, they can be ar- ranged numerically under each symp- tom or test in order of increasing se- verity of action. Totaling the score for each compound will then give its relative standing, the smallest score being the least irritant. The pitfalls of such a method should be obvious to any scientist, biological or other- wise. In the first place, it puts an equal value on each of the observed end points, which has been shown to be invalid. Perhaps the greatest error arises, however, from the tendency to consider the numerical ratio as an index to the relative degree of irritant potency whereas its very basis limits it to' the relative standing of the compound with rela- tion to others of the series. A further disadvantage is that, even with a reference irritant, such a scheme does not integrate data from different series of tests. It will place the compounds in each test with relation to the reference but does not place the compounds in the different series with relation to each other. If any appreciable volume of irritation tests are contemplated a more practical method of numerical evaluation is to assign a weighted score fo'r each degree of irritation under each symptom or test. Such a method permits placing a total score for each observation which is in keeping with its relative sig- nificance. Totaling of the score again places the lowest score on the least irritating compound and per- mits integration of data obtained at different times either with relation to the reference or to the various test compounds. With the exception of sclerosing agents the degree of irritation should be suf•ciently great to produce clearly defined reactions but not severe enough to produce necrosis for, like other biological phenomena, minimal or maximal irritation can- not be evaluated on a graduated re- sponse basis. The data obtained, may be enhanced by using methods, animals, and scoring consistent with those in the literature This is not to condemn the search for new methods and improvements of the old, but to suggest that such efforts should be considered and en- couraged as research and that the results thereof be published as such. ' Observance of the above general principles, plus the exercise of good laboratory technique, should make it possible for the average laboratory technician to acquire considerable skill in the evaluation of relative irritant potential by routine meth- ods. When such tests are more uni- versally adopted as an essential part of new product development we may reasonably expect that there will be a concomitant improvement in the methods of testing, in the new products, and in the satisfaction that comes from having added one more important safeguard for the consuming public to whom we owe so much.