QUANTITATIVE EYE IRRITATION TESTING 121 ditions which would resemble an accidental human exposure, e.g., splash- ing a substance into the eye. The degree of injury is then used to as- sess consumer hazard. For moderately and severely irritating ma- terials, there is no point of contention, since such substances are easily identified by almost any test procedure one might employ. A problem arises when we wish to test materials which fall in the nonirritating to mildly irritating class or to compare eye irritancy of related formulations, as in product development. It is under these conditions that all of the methods thus far developed prove to be unreliable. Two of the vari- ables which might influence the reliability of test procedures are dis- cussed below. A. Sample Drainage The rate at which a material drains from the eye is of paramount importance, since it greatly affects the duration of the exposure period this, in turn, exerts a profound influence on the response produced and the final score obtained. Drainage can also be altered by the movements of the animal's eyelids or through the •viping action of the nictitating mem- brane ( a structure which is not found in man). The influence of the rabbit's nictitating membrane on the results is probably not fully appreciated by many investigators. Differences in sample drainage are large enough to cause much of the eye-to-eye variability encountered in testing any individual material. The viscosity of the sample also has an effect on sample drainage and can, therefore, influence the results of the Draize test. In this labora- tory, using the Draize procedure, numerous instances were observed in which a highly viscous solution of 3 or 4% sodium hydroxide produced small areas of dense opacity in an otherwise normal eye. It is un- reasonable to assume that these variations were the result of hetero- geneity in the anatomical structure of the cornea, i.e., the cornea pos- sessed both chemically resistant and sensitive areas. The absence of opacity in these areas must be attributed to a nonuniform exposure due either to a lack of uniformity in instillation or, more likely, to the wiping action of the nictitating membrane. Apparently, a systematic study of the influence of viscosity on the Draize test results has not been reported. It would not be surprising to find a greater variance when testing viscous materials with the Draize test, since the most viscous samples tend to remain in the eye longer and are more likely to localize in one area. The amount of irritation or opacity would then, in large part, be related to the rate of drainage or removal from the eye and the rate

122 JOURNAL OF THE SOCIETY OF COSMETIC CHEMISTS of neutralization of any residue by the lacrimal fluid. One would ex- pect therefore less of a tendency to produce corneal and iridic irritation in a man than in a rabbit, because of the copious amounts of lacrimal fluid produced by the human eye when it is irritated. B. Tissue Response to Chemical Action An important point to be considered in testing for eye irritation is the qualitative differences in the response produced by various sub- stances. These may involve differences in penetrability, erosion and edema. Further, the rate of tissue regeneration (healing) of injured tissues may vary widely. Some substances, e.g., isopropyl alcohol, will produce a mild, superficial corneal opacity of grade 1.0 (Draize I-A). In most instances only the outer layer(s) of the cornea seem to be in- volved, since their removal (sloughing) usually leaves an optically clear but lusterless cornea which heals within forty-eight hours. A number of formulations will produce a severe dulling of the cornea after as brief an exposure as ten seconds, but little, if any, additional change when the exposure is increased to several minutes. Other substances, e.g., alkalies, produce a continuously increasing degree of corneal opacity with increases in concentration and/or exposure time. The product formed by the chemical action of the alkali on the corneal tissue does not prevent further alkali penetration. On the other hand, inorganic acids, e.g., hydrochloric acid, react with the corneal tissue and form products (coagulum) which tend to reduce further acid penetration. These differences between penetration of acids and alkalies have been re- ported (4, 7). II. METHODS Young adult albino rabbits, weighing 2-3 kg. each, were employed in these studies. The animals were anesthetized by injecting pentobarbital solution (50 mg./ml.) into the marginal vein of the ear. The depth of anesthesia for each animal was adjusted to iramobilize it completely and to permit the introduction of a corneal applicator. The use of anes- thesia eliminated the need for using large numbers of rabbit holders and reduced the incidence of self-inflicted injury to the eye by the animal during the first few but critical hours of testing. The eyes of each ani- mal were carefully examined before use, and all animals in which evidence of irritation was found were rejected. The animal was placed on its side and 0.1 ml. of the test sample was introduced into the eye directly or through a molded polyethylene applicator.