278 JOURNAL OF THE SOCIETY OF COSMETIC CHEMISTS Originally the reported contaminants of toiletries and cosmetics covered a wide spectrum of organisms (1). However, more recent reports show that the typical product contaminant is in fact drawn from a much narrower range of organisms (2-7). The discrepancy between earlier and more recent reports is due to the fact that earlier investigators failed to differentiate between multiplying contaminants and those which were merely chance residents or transients. The most frequently reported con- taminants in recent years have belonged to such genera as Pseudornonas, Klebsiella, ,4chromobacter and ,41caligenes. These bacteria are common residents in water, both fresh (8-12) and distilled (13, 14) and it is now widely believed that the water used in the preparation of toiletry products is their likely source (11, 15, 16). In the UK mains water as supplied generally contains low numbers of bacteria (usually less than 300 ml -•) and if water were used in this condition it seems probable that it would contaminate only the more susceptible pro- ducts since it is likely that contaminants capable of multiplication in products represent only a small proportion of water flora. However, in the pro- duction of toiletries it is not always possible to use water directly from the main and it is frequently held in a storage tank before use. The ability of water-borne bacteria to multiply on storage is well known (9, 16, 17) and total counts of greater than 10 ? ml -• have been recorded (18). Indeed the readiness with which bacteria will multiply in stored water has led Cham- bers and Clark (11) to suggest that an absence of bacteria indicates that the water is toxic. Growth is probably at the expense of organic materials and salts present in the water (9) and perhaps materials dissolving in the water from the container (19). Growth may also be at the expense of dissolved volatile organic materials if these are used in the vicinity of the storage tank (20). Multiplication of bacteria will occur in pipelines where velocities are low (21) and it has been reported that polythene pipelines are prone to support- ing large growths of bacteria, sufficient oven to block the pipes (9), although Burman (22) has claimed that polythene piping has little effect on bacterial numbers. There is little question, however, that deionizing columns are sites at which bacterial numbers may increase (1 l, 23, 24). The bed concentrates Ca = + and Mg =+ from the water and at the same time collects organic materials by a process of sieving (24). The bacteria present in the water are either entrapped within the matrix of the bed or adsorbed on to the resin surface and so act as the inoculum. The numbers of bacteria in the effluent

WATER-BORNE BACTERIA AND SHAMPOO SPOILAGE 279 from the column vary with the frequency with which the column is used or cleaned and disinfected (23). There is now considerable practical evidence indicating a positive correlation between high levels of bacterial contamination in water used in manufacture and the incidence of contaminated toiletry products. This is presumably because the greater the number of contaminants in the water the wider the range of resistances to unfavourable conditions and the greater the probability of bacteria being present which are capable of surviving and multiplying in the product. The range of resistance is likely to be widest where the population is made up from a range of different species of bacteria rather than large numbers of only a small number of species. It would also be theoretically possible to have very heavily contaminated water presenting no threat to product stability because of the low resistance of the bacteria or, in reverse, water containing small numbers of bacteria which, because they are well suited to growth in the product, will give rise to contamination. In order to perform Challenge Tests on shampoos in the laboratory it is often found necessary to use very large inocula, usually of the order of 106-107 bacteria per gram of product. This is, however, much higher than would normally be found as an initial inoculum during manufacture in a well-ordered production unit and product contamination is known to occur even when the level of contamination in the plant is low and water of good bacteriological quality is used. MATERIALS AND METHODS Empicol ESB 3/S (sodium lauryl ether sulphatem27.3•o active) a specially prepared preservative-free batch supplied by the Marchon Division of Albright and Wilson Ltd. Lauryl Isopropanolamide--Marchon Division of Albright and Wilson Ltd. Nutrient Agar CM3--Oxoid Ltd. Tryptone Soya Agar CM131--Oxoid Ltd. Peptone Bacteriological Neutralized L34•Oxoid Ltd. Because of the wide variety of different ingredients used all experimental shampoos are atypical. It is, however, possible to produce a model system which is representative of a general shampoo formulation. For this work the following formulation was used.