410 JOURNAL OF THE SOCIETY OF COSMETIC CHEMISTS Although these theses have not been universally accepted among perfumers, they have had considerable influence on perfumery practice. A few studies have been reported to test their validity. Paukner (4a) found that adding increasing concentrations of an "erogenous complex of civet, synthetic ambergris, and synthetic musk" to a lavender-type perfume brought its position in a three factor semantic differential analysis increasingly close to the position of "the ideal erogenous fragrance." Steiner (4b) presented, in very brief flashes (V•70 to •/250 sec.), ambiguous slides in which a picture of a nude woman and a cityscape were superimposed, to a group of 25 males aged 20-39 years. In half of the cases the subjects viewing the pictures were also exposed to a perfume which they had previously selected as being sexually stimulating. In the presence of the perfume, significantly more respondents reported seeing the woman than in the absence of perfume. The purpose of the present study was to study the reactions of untrained subjects to the odors of several aromatic chemicals related to human body odor, when these chemicals were presented under conditions where they might be taken for actual human body odors. Using actual human body odors and 5c•-androst-16-ene-3-ol (I) as comparison stimuli, we hoped to establish in what ways the synthetic chemicals were similar and how they differed from these natural stimuli. We adopted the design of M. Schleidt's T-shirt experiment (5) but replaced part of the shirts that had actually been worn by test subjects by clean shirts that had been treated with androstenol (I) or with the aromatic chemicals we wanted to study, at various concentrations. We selected the following chemicals: 2-Methyl-1(2,6-exo-tricyclo [5.2.1.02'6] dec-4(3)-ene-8-yl)-pent- 1-ene-3-ol (III), synthetic aromatic chemicals first prepared in the laboratories of Dra- goco, with odor descriptions similar to those of androstenol (6) and cyclopentadecano- lide (IV), a macrocyclic lactone found in ambrette seed oil and closely related in struc- ture to key ingredients of natural musk and civet tincture, which has been used for many decades in fine perfumery as a musk-type odorant (Figure 1). We were particularly interested in exploring the following questions: Is it at all possible to fool respondents into believing that the odor of chemical- treated shirts is human body odor? Schleidt (5) had found that sex attribution was possible and that her subjects could correctly identify the sex of the wearer of a given shirt to a statistically significant degree. Is sex attribution possible also for the aromatic chemicals studied? If so, to what sex are they attributed? Schleidt also found that strong body odors were predominantly judged as unpleasant and that they were predominantly attributed to male wearers. Do the same patterns prevail with the aromatic chemicals studied? METHOD Test subjects were 30 men and 29 women, aged 18 to 30, all student volunteers at the University of Utrecht. For participating in the test, they received a modest monetary reward (Fl. 35,--), or some academic credit. They were not prescreened in any way. There was only one among them who had trouble smelling many of the stimuli his ratings were counted along with those of the other participants.

HUMAN BODY ODOR 411 5o•-An d rost- 16-e n e-3o•-o I (Androstenol) II 2- M ethyl- 1 (2,6-exo-tri cyclo [5.2.1.0 2,•] dec-4(3)-ene-8-yl)- pent-l-ene-3-one Oh • -0"'/ ' 0 III IV 2-Methyl-1 (2,6-exo-tricyclo Cyclopentadecanolide (CPD) [5.2.1.0 2,6] dec-4(3)-ene-8-yl)- pent-1 -ene-3-ol Figure 1. Structures of materials used in these studies. The subjects were provided with yellow cotton T-shirts and were instructed to wear these on five consecutive nights. During this period they were instructed not to use any deodorant and to wash only with a very faintly perfumed soap* with which they were provided. They were also given odorless polyethylene bags in which to store the shirts between and after wearing them. At the end of the 5-day period, the test subjects participated in two smelling sessions which were conducted on two consecutive days. In these sessions, they were presented with 30 T-shirts of the kind they had been wearing and were asked to rate these for odor intensity (four replications) and for pleas- antness of odor (three replications) and to try and identify the sex of the wearer of each shirt (three replications). The respondents were randomly divided into two groups. The sequence of the tasks performed by the two groups is shown in Table I. The order of presentation of the shirts was randomized. This was done (1) within tasks, * Penaten soap, the same brand as was used in Schleidt's experiments. The experimental design of this test replicated in all essential aspects that used by Schleidt.