716 JOURNAL OF THE SOCIETY OF COSMETIC CHEMISTS to demonstrate its attainment. For example if products are each con- taminated with 100 organisms which are inactivated by a factor of 107 at the recommended dose, then only one survivor would be expected in 105 products. Sterility testing is not sensitive enough to detect such a low frequency of occurrence of contaminated items. New approach Recent theoretical studies by Tallentire, Dwyer and Ley (10) have been devoted to investigating the influence of various parameters, such as the initial numbers of contaminating organisms and their resistance, on the relationship between the probability of observing positive samples and radiation dose. The results of these studies will be published shortly but it is relevant to refer to them here since they were based originally on the concept that a microbiological quality control procedure could be devised based on the use of 'sub-sterilizing' doses of radiation given to samples taken from normal production and followed by sterility testing. Such testing after low dose treatment would be expected to reveal a number of positive samples from which conclusions can be drawn concerning the expec- tancy of positives at much higher doses. This approach is well suited to the determination of a dose for sterilization of individual cosmetic preparations. It makes full use of the fact that radiation doses can be delivered precisely and homogeneously, and a dose treatment given in one cobalt-60 source can be reproduced exactly in another so allowing preliminary investigations to be made using experimental facilities. Identification of organisms res- ponsible for positive growth in a small number of samples after 'sub- sterilization' doses gives an excellent indication of inactivation which would be expected at high doses as well as indicating the type of con- tamination which should be avoided in the product preparation. Information pertinent to the establishment of dose requirement may be summarized as follows: (a) initial number of viable organisms prior to treatment, (b) types of organisms involved, (c) influence of the nature of the product on radiation resistance, (d) margin of safety required in respect of the product taking into account its subsequent use. The procedure of 'sub-sterilization' treatment in initial investigations on the actual product and eventually as a quality control procedure appears to be the most straightforward and revealing in relation to (a), (b) and (c) above. The dose required for the sterilization of cosmetic preparations probably lies within the range 0.25-2.5 Mrad.

GAMMA RADIATION FOR PRODUCT STERILIZATION 717 EFFECT ON MATERIALS Packaging Experience with medical products has shown that a wide range of packaging materials are suitable. These include polyethylene, polyvinyl- chloride (PVC), Cellophane, nylon, laminates, coated papers and metal foils. PVC tends to decompose with the release of hydrochloric acid but this is not significant up to a dose of 2.5 Mrad. Rubbers vary in sensitivity depending on compounding butyl and chlorinated rubbers are unsuitable. Glass, while not affected mechanically, darkens in colour according to com- position. This is caused by displacement of electrons out of their normal lattice position leaving ions which absorb light the darkening can be removed by heating. A radiation resistant glass, based on the incorporation of cerium, has been formulated but it is very expensive and its use would only be warranted in special circumstances. Polystyrene has a high radia- tion resistance owing to its highly aromatic character whereas polypropylene may be degraded, becoming brittle particularly after post-irradiation stor- age. A guide to the radiation stability of materials is available (11). Products The vast bulk of materials currently being radiation sterilized is made up of objects such as disposable plastic hypodermic syringes, needles, catheters and rubber gloves. Only a few irradiated pharmaceutical prepara- tions are marketed and these are mainly antibiotic preparations, particu- larly ophthalmic creams. The sterilization of talc is a distinct possibility but it appears that costs have limited this application. Experience of the effects of irradiation on cosmetic preparations is limited and results that do exist are in the hands of individual manufac- turers. However, many creams and other types of preparation would be expected to be undamaged after the comparatively low radiation doses which are probably required for adequate microbiological control. Glycerol esters for example are stable at the ester links although the fatty acid chains might suffer radiation induced oxidation. The unsaturated acids oxidize most readily but even so, only about 0.1% breakdown would be expected after 1 Mrad. Oxidation is more rapid in emulsions but emulsion breakdown is unlikely although the nature of the emulsifier used might be important. Proteins are generally very resistant to change. Essential oils have been