HYGIENIC MANUFACTURE AND PRESERVATION 723 equipment all contribute materially to the problem. It is impossible, however, to define in a single document exactly how every product should be manufactured and tested, if only because the contamination hazard is closely related to the scale of manufacture which differs greatly from one firm to another and from product to product. In the process of verifying compliance with standards or of monitoring the microbiological condition of production arrangements, attention will have to be paid to the question of representative sampling. When there is a moderately high level of contamination in the material being sampled, whether or not the product is distributed in separate small units (e.g. the finished packs) and the contamination evenly spread throughout, the sampling problem is unlikely to be highly critical since precise knowledge of the level of contamination is seldom essential. If the contamination is at a low level and unevenly distributed, representative sampling may be more important, especially when its purpose is to detect the presence or absence of pathogenic micro-organisms. Another aspect of sampling is the extent of replication necessary when carrying out the appropriate tests on a particular sample. Guidance on statistical sampling methods may be obtained from various publications such as: Lowe, C. W. Industrial statistics 1 259-273. (1968). (Business Books Ltd., London). Thatcher, F. S. and Clark, D. S. Micro-organisms in food. 52-58 and Appendix V. (1968). (Toronto Press Toronto). Cowell, N. D. and Morisetti, M.D. Microbiological techniques--some statistical aspects. J. Sci. Food Agric. 20 473. (1969). Sampling theory is perhaps of greatest importance when the purpose of testing is to examine material on a pass-or-fail basis. Microbiological quality is preferably studied continuously as a preventive measure and the intelligent application of experience may provide a better safeguard than reliance on a rigid scheme of sampling. For example, a single manufacturing unit may be involved in virtually continuous production of some high- volume toiletry lines such as shampoos and in the preparation of batches of decorative make-up items at quite protracted intervals e.g. once per month. The sampling problems are hardly comparable but an experienced microbiologist will be able to devise testing procedures appropriate to the circumstances and still within the capacity of the available laboratory services. For exceptional cosmetic preparations where sterility may be the un- qualified aim, a recognized routine of sterility testing should be applied. Suitable methods are given in the Therapeutic Substances (Manufacture and Interpretation) General Regulations 1963 The United States Pharma-

724 JOURNAL OF THE SOCIETY OF COSMETIC CHEMISTS copoeia, 18th Revision: British Pharmacopoeia 1968 and in the British Pharmaceutical Codex 1968. For most cosmetic products a criterion of sterility is not appropriate and suitable testing procedures are described in various sections of this monograph. Generally, the microbiologist will not seek to identify every contaminant found but it is advisable to undertake detailed identification if there is any reason to suspect the presence of pathogenic micro-organisms. If the presence of potentially harmful micro-organisms is confirmed, the product should either be rejected as unfit for sale or treated in such a way that such micro-organisms are no longer detectable on further testing. Test methods for verifying sterility and for enumerating viable micro- organisms should ideally be fully reproducible in different laboratories and should preferably be simple and quickly carried out. Unfortunately there are no available techniques capable of satisfying all these criteria. For example, the microbial count is liable to vary according to factors such as (i) the size of sample and frequency of sampling, (ii) the specified counting technique, (iii) whether inactivation of any antimicrobial preservative is carried out, (iv) the diluent and recovery medium used and the methods employed in mixing with the prcd]ct, and (v) the temperature and duration of incubation. Appendix B gives a selection of recommended procedures from which an appropriate scheme for testing a range of cosmetic products may be chosen. As indicated above, the sampling procedure in particular manu- facturing conditions can only be properly determined on the spot and deciding on the most suitable tests to apply similarly demands consideration of formulae and manufacturing techniques actually in use. For example, the criterion of excluding known pathogens might suggest the need for an elaborate programme of microbial identity tests on every batch of product. For a suitable testing programme in a given set of circumstances, however, it will probably suffice to devise a scheme capable of monitoring the overall product quality over the course of time this is preferable to a prohibitively expensive and elaborate check on every batch, providing that a compre- hensive study of product preservation has been undertaken at the research stage. If good provisions for hygienic manufacture are made and the preserv- ative system is shown to be capable of dealing with likely levels of con- tamination, routine control can safely be restricted to a limited range of testing.