566 JOURNAL OF THE SOCIETY OF COSMETIC CHEMISTS MALDODORANT: "MODIFIER:" ISOYALERIC ACID THRESHOLD MULTIPLES OF ACID: 10x ß VANILLIN ,+ LINALOOL MIX\• MIX 40x 60 x _/,MIX .., O Figure 6. Influence of odor modifier on malodor of isovaleric acid its usefulness at 60 x thresholds. More experiments could have established the limits of the efficacy somewhat closer. This procedure can be applied to any cosmetic treatment where realodors are controlled by fragrances. It is not necessary that the mixture of the mal- odorous emission with the fragrance exhibits an odor similar to the modifier only that the sensory complex of the realodor must be signifiantly modified. The odor of mixture may be not similar to the fragrance either, but, of course, it should be hedonically acceptable, and not only a different realodor. Hedonic Value The overall pleasantness/unpleasantness of the modified body odor is a function of its odor notes, but perhaps only one of these, "sickening" relates more directly to the degree of annoyance. Testing of hedonic shifts in pairs of "before" and "after" samples is an ex- tremely complex task. Some elementary considerations are given elsewhere (1). One is that large panels (50 to 100 panelists) must be used to judge the hedonic value. Another is that a familiarity with the formulations and brands is undesirable. ANALYTICAL EVALLITIONS Analytical evaluations of human odor control efficacy are based on mea- surements of decrease in emission of malodorants. In the case of applied

EVALUATION OF HUMAN BODY ODOR 567 fragrances, the rates of dissipation of fragrances from part of body also can be measured. As outlined in the discussion of odor intensity measurements, a twofold reduction in concentration of a malodorant, even if this were the only mal- odorant present, does not reduce the realodor intensity by a factor of 2, but rather by a fractional power of 2. A reduction in concentration by 20 to 30 per cent* usually will barely change the odor intensity sufficiently to notice that odor has weakened even if the "before" and "after" odors would be compared directly. This reduces the demands on the reproducibility of analytical data need- ed for odor-related interpretations. The most common form of analysis is gas-chromatographic, using hydro- gen-flame ionization detector. This detector responds approximately propor- tionally to the mass of the organic substance, but does not respond to water vapor, ammonia, Ha, CS2, HCOOH, and some other compounds that con- tain few CH links and several heteroatoms. To relate GC analysis data to odors, two conditions mttst be met. First, the gas-chromatographic sensitivity should be sufficient to obtain measurable GC peak areas even for those odorants which have a particularly low odor threshold and occur in the vapor sample at concentrations close to the thresh- old. Second, the GC peaks of those sample components that are most likely and significant contributors to the odor of the sample should be indicated. These two conditions can be satisfied by a preconcentration of the sample and by using a trained analyst's nose to assay the odors of the GC-separated components at their emergence from the GC column. This sensory assaying consists of noting which components, irrespective of their peak size, exhibit particularly strong odors and characteristic odor notes (26). Sample preconcentration should include collection of the vapors from or- ganic substances and their separation from excess water vapor, which occurs in most samples of human body vapors. If water is not separated, the re- quired GC sensitivity cannot be reached. As an example, analysis of the maximum possible amount of condensed perspiration yields only a few GC peaks. Collection of organic substances from several liters of perspiration vapors, without collection of water vapors, yields a sample size sufficient to obtain several dozen of GC peaks. Collection with the gross exclusion of water is accomplished using non- polar polymers organic polymers (Fig. 2A). It permits a several thousand- fold enrichment of organic substances from vapor phase. The lowest odor thresholds are in the concentration range of 10 © g/1., and such precon- centration brings even such components into GC recording range. *Differential threshold," related to Weber's ratio.