148 JOURNAL OF THE SOCIETY OF COSMETIC CHEMISTS ference. The significant correlation of skin surface pH and propionibacterial counts, in fact, corresponded to in vitro findings concerning the specific growth rate of propioni- bacteria and staphylococci at various relevant pH values (8). These recent experimental findings are consistent with the so-called acid mantle concept created by Marchionini and his group as early as half a century ago (9) and later by Arnold (2) and RiSckl (10). Facing the conclusions from our previous experiments, one might, however, argue that not the pH value of the different cleansing agents but the differing chemical composi- tion might influence the resident flora of the skin. To test this hypothesis, a corre- sponding investigation was performed using two synthetic detergent preparations with identical ingredients at pH 5.5 and 8.5. MATERIALS AND METHODS EXPERIMENTAL STUDY POPULATION Five male and five female healthy volunteers gave written informed consent to enter the study. While the age ranged from 24 to 35 years, the mean age amounted to 27.5 years. Half of the individuals were attributed to a group called A, starting defined washing procedures with the synthetic detergent preparation A characterized by pH 5.5, the other half with preparation B characterized by pH 8.5. CLEANSING AGENTS Both liquid cleansing preparations contained identical amounts of the following ingre- dients (stated in descending quantitative order): water, mipa-laureth-sulfate, and coca- midopropylbetaine potassium coco-hydrolyzed animal protein PEG-7 glyceryl-cocoate cocamidopropylaurylether PEG-10 olive oil sodium lactate, perfume, benzyl alcohol, and methylisothiazolinone hydrogenated cocoglyceride and tocopherol niacine and pyridoxine hydrochloride, biotin and amino acids disodium EDTA, dye C.I.47005 and dye C.I.61570 and BHT and citric acid and ascorbyl palmitate. The only difference between the two preparations consisted of the pH (5.5 with prep- aration A 8.5 with preparation B), which was adjusted by the addition of appropriate amounts of sodium hydroxide. Both preparations were kindly supplied by Sebapharma, Boppard, FRG, preparation A corresponding to a previous commercial formula (Sebamed © fliissig). pH DETERMINATION The pH of the skin surface was determined using the flat glass electrode for sur- face measurements developed by Ingold that was first used in the field of dermatology by Schirren (11) and now is commercially available (Glaselektrode 403-S7, Ingold-Mef3technik, Steinbach, FRG). This electrode was connected to a precision pH-mV meter (pH 125, WTW, Weilheim, FRG). At any point of time three consec- utive measurements were performed and the mean value calculated. By principle the rules for skin surface pH determinations established by Arbenz (12) were followed.

pH EFFECTS ON SKIN 149 DETERMINATION OF THE BACTERIAL SKIN FLORA Sampling of the cutaneous bacterial flora was based on the detergent scrub method as described by Williamson and Kligman (13,14). The specimens obtained were serially diluted and dispensed on solid media. For culturing staphylococci, Columbia agar was chosen (BBL, Heidelberg, FRG), to which 5% defibrinated sheep blood was added tryptic soy agar (BBL) served as medium for culturing propionibacteria. While the former medium was incubated at 37øC for two days in an aerobic atmosphere, the latter was incubated for seven days in an anaerobic setting (Gas Pak jars with nitrogen and carbon dioxide generator kits (BBL)) at the identical temperature. Coagulase-negative staphylococci were identified by Gram stain, plasma coagulase, test (BBL), and their biochemical reaction pattern (Api Staph, BioM•rieux, Nfrtingen, FRG) (compare 15). The identification of propionibacteria was--apart again from colony morphology-- based on Gram stain and a biochemical reaction pattern to be demonstrated by a ready-made kit (Api 20A, BioM•rieux) (16,17). CLEANSING PROCEDURE During the experiments every volunteer washed twice a day (in the morning and in the evening) the skin of the forehead (median line) and the proximal part of the flexor site of the forearm for a period of one minute. For washing, the synthetic detergent prep- arations were diluted in the way the individual was used to. Afterwards the washing solution was rinsed off with plain tap water. TIME COURSE OF INVESTIGATIONS On each of days 1-3 the volunteers still adhered to their previous washing habits. Both cutaneous pH and bacterial flora, however, were investigated at both sites. During the next eight weeks either preparation A or preparation B was used first and the corre- sponding one thereafter for four-week periods. Every seventh day (day 10 and so on) the same measurements were performed as at the beginning. This was done at about the middle of the application interval, i.e., at least six hours after the last cleansing procedure. MATHEMATICAL AND STATISTICAL ANALYSIS The figures included in the text, in general, show mean values, as well as maximum and minimum values the bars represent standard deviation symmetric to the mean value. The "=" sign is located at the median. The 95% confidence range for the median is symbolized by a blank space. This applies both to charts on pH values and bacterial counts. Time is generally given in days, pH values in units, and bacterial counts as a logarithm of colony-forming units (CFU). In general, in graphs data related to the application of preparation A are connected by continuous lines in connection with preparation B, dotted lines are used. Statistical analysis was based on sign test for tied pairs, on Student's t-test for the comparison of two independent samples, and on the determination of the correlation coefficient of Bravais and Pearson (18). p-Values are in