•24 JOURNAL OF THE SOCIETY OF' COSMETIC CHEMISTS ._• 2,0- .. • E [I- -. ... '.•__, ß • •0- "- ß •.o- T•.0- -o- .... -c- .... I I • I 5 15 25 35 •5 Temperature [øC] o--Hyp0thetica[ aqueous reference for/•.Oø/oW/v ch[or0creso[. Ratio oi[' water []=1.0 d-O'8 x-0'7 -0'6 ,-0.5 Fi#•re 8. Activity aõainst oe. coli o• 4.0% w/v overall conceDtmQon of ch]o•0c•eso] in •r•chis oiI--w•te• dispersions and chlorocresol in arachis oil/water dispersions has been examined (37,38) and is shown in Fig. 2 in which reference lines for aqueous solutions having the same overall concentration of bactericide are also shown. The aqueous reference for chlorocresol in Fig. 3 is, and can only be, a hypo- thetical line because the 4 per cent chlorocresol necessary to produce measurable activity in the dispersions exceeds by far the limit of aqueous solubility. It is immediately observable that at all temperatures at which observations were made, the dispersions of phenol in liquid paraffin and water are more active, and the dispersions of chlorocresol in arachis oil and water are less active than the aqueous reference. This difference is

THE i•CTIVITY OF ANTIBACTERIALS IN TWO-PHASE SYSTEMS a manifestation of the partition coefficients (Table I). The slopes of the regression lines of log extinction time on temperature are log temperature coefficients, and as the oil:water ratio is increased the temperature co- efficients are decreased. That is, the activities of oil:water dispersions containing preservatives are influenced less by temperature changes than is an aqueous solution of the same preservative. The temperature co- efficient for a dispersion containing a given overall concentration of pre- servative is related to that of the aqueous reference solution of the same concentration by the expression log 0 •-- log o s K= 0 where 0 a = temperature coefficient of aqueous solution o s --: temperature coefficient of dispersion 0 ---- oil:water ratio K = constant for each preservative/oil/water system , Since the temperature coefficients of the dispersions decrease with increasing oil:water ratio, it is possible by using the constant K to calculate a ratio at which the temperature coefficient is 1.0, and the antibacterial activity of the dispersion is independent of temperature. For the very few. systems studied so far, this ratio has been of the order of $.0 to and in two cases it was possible to verify this value experimentally. The reason for this unique ratio is as yet uncertain but it is possibly determined by the concentration of preservative at the oil/water interface.' It,. is hoped that further experiments will help to establish this. The conclu•sion which must be drawn from the experiments is that' it is utterly futile to spend time and labour formulating a cream, and as a final gesture add an ad hoc concentration of a preservative. The formu- lation must be approached ab initio from th? points of view of both physical and microbiological stability. The two aspects cannot be divorced one from the other, and any attempt to do so is an invitation to failure. (Received: 2nd Septembe r 1'964) REFERENCES (1) Bennett, E. O. in Developments .in Industrial Microbiology III 273 {1969.) {Plenum Pres•s, New York). {2) Kostenbauder, H. B. in Develot•ments in Industrial Microbiology III 286 (i962) (Plenum Press, New York). '-' (3) De Navarre, M.G. The Chemistry and Manufacture of Cosmetics I 257 {1962) ('D. Van Nostrand Company, Inc., Princeton) {4) Wedderburn, D. L. in Advances in Pharmaceutical Sciences I 195 ,{1964) (Academic ß Press, London).