JOURNAL OF THE SOCIETY OF COSMETIC CHEMISTS PROF. M. DOg,ROW: The materials dealt with are all of a semipolar nature and might be expected to dimerise in the oil phase, so that the partition coefficient would not be constant. This does introduce the problem of calculating partition coefficients •or each concentration and each phase volume which is being studied. T•. L•c•u•R: In the aqueous phase one is likely to have dissociation, and association is likely to occur in the oil phase. If association were occurring in the oil then our partition coefficient would have been wrong and therefore our estimate of the phenol in our aqueous phase would have been wrong because a greater quantity would have gone into the oil. We would have picked this up by the position of our time survivor curves because the kill would have been very much slower than we had anticipated. In fact, our estimates of extinction time came very close to the observed extinction times. As for dissociation in the aqueous phase, we know that the pKa of these phenols is in the region of 10. We were working at •bout pit 6.5--? so there was not much dissociation. MR. A. H. F•oN: Have you considered the addition of a material such as propylene glycol which, when added to the aqueous phase, would increase the con- centration of the preservative? T• L•CTUR•R: We have examined a number of systems in which additives both increase and decrease the partition coefficient. The activity goes roughly as one would expect. We have looked at glycols and at the effect of salts, because the latter is of interest in sterilisation. The process of heating with a bactericide depends in part upon the shift of the partition coefficient in fayour of i•creased concentration of bactericide in the bacterial cell. MR. J. T. R•s: You state that cream formulations must be approached, from the start, from the points of both physical and microbiological stability. We have seen that the partition coefficient effect is only one of several factors which will influence preservative efficiency. These other factors may in fact be stronger than the above partition coefficient effect. The method of packaging, and the type of pack •employed, will also influence the rate of microbial growth. A point is reached, however, when the preservative of choice must be tested. Such testing should relate to conditions closely approximating those found in practice. To be of use to the development chemist the tests should also yield the required information in a short time. The method should also be simple to operate. Serial dilution methods have been described by several authors. Conditions here, however, can hardly be described as relating to those found in the cream itself. The agar cup-plate method is also well known the cream, containing the preservative, is placed in a central cavity in a seeded agar plate. These conditions again do not approximate closely to the true ones. A water soluble preservative may diffuse from the cream into the surrounding agar. The oxygen balance of the system must also differ markedly from that found in a normal cream jar or collapsible tube. Other authors have proposed inoculating the sample itself with the test organisms, incubating, and later plating the samples for colony counts. This method is both lengthy and complicated. I agree that cream development should embrace both physical and microbiological testing from the start but I am not sure that present test methods are always appropriate.

THE ACTIVITY OF ANTIBACTERIALS IN TWO-PHASE SYSTEMS 29 THE LECTURER: There are some guiding rules when one starts looking for a suit- able preservative. If you plot phase volume ratio against the reciprocal of the fraction of preservative in the aqueous phase you get a linear regression line the slope of which depends upon the partition coefficient. If the partition coefficient is low, a change in oil/water ratio will not materially alter the proportion of preservative in the aqueous phase. If you have selected a preservative with a very low slope you can assume that you are going to have some activity in your product. But if you went to the other extreme, and selected a preservative with a high partition coefficient, as you change your phase volume ratio the proportion of preservative in the aqueous phase falls very rapidly. I presume you would start off in selecting a preservative with the aid of a graph and then calculate the con- centration in the aqueous phase because the activity resides there. Your final product can be tested by inoculating and including a dye, which changes colour with changes in metabolic by-products from the organism, or just with a change of oxidation reduction potentials. After 18 hr some change in the indicator colour in the system could be seen long before it is possible to detect growth by inspection. With a semi- solid cream it might be perhaps several weeks before one detected mould growth. MRS. D. WEDDERBURN: Short term product inoculation tests can be misinter- preted, for although they will reveal those preservatives which are not likely to be active, they cannot be used to predict those which will be effective. I say this because it is known that over a long period of storage certain organisms, particularly the Pseudomonads, can adapt themselves to their environment. I have experience of this. On one occasion, during a prolonged product inoculation test, for the first four or five months the preservative seemed to be holding the inoculation in a quiescent state, but after six months vigorous growth began and deterioration of the emulsion set in. For this reason I see no substitute for long term tests which provide the oppor- tunity for adaptation to be detected. DR. J. B. M. COPPOCK: May I draw attention, from the food field in synthetic creams and even margarines and butter, that freeze-dried organisms do not necessarily behave, indeed often multiply faster, than organisms in their natural habitat. Thus I would not accept tests on o/w, or other systems, using freeze-dried organisms alone but would still require the more classical "suck it and see" experiment more related to practical conditions. MR. G. SYKES : I agree wholly with Mrs. Wedderburn. This is one of the occasions in which the microbiologist must not be rushed. Short term tests will give an approximate result, but this is not good enough for commercial products. Adapta- tions can take place rapidly in a generation or two, or very slowly during several weeks, and it is the latter which one has to watch. It can only be controlled by long term observations. There is also the question of the types and strain of organism which may be involved. MR. I). W. POXON: It does strike me that of the factors you were talking about, the interfacial phenomena were of considerable importance. If your micro-organisms are absorbed at the interface, surely the antibacterial, which was also strongly absorbed at the interface would be most effective. THE LECTURER: This is correct.