Production of the odorous boar pheromone Sa-androstenone in the human axilla is inhibited by the topical application of bactericides (Bird & Gower, 1982), and apocrine secretion is odourless until incubated with bacteria (Shehadeh & Kligman, 1963). There is some evidence that Sa-androstenone and Sa-androst16-en-3a-01 are pheromonal in man (Cowley et al., 1977; Kirk-Smith & Booth, 1980; Benton, 1982; Filsinger et al., 1984), but as yet neither the bacteria which form these compounds, nor the steroids from which they are synthesized, are known. In bacteriological studies, Coryneform bacteria have been associated with typical axillary odour (Jackman & Noble, 1982), and so in our earlier attempts to establish the precursors of the pheromones this group of bacteria was studied exclusively (Nixon et al., 1984a,b). It soon became clear, however, that not only did different strains of Coryneforms vary widely in their metabolism of the trial substrate, testosterone (Nixon et al., 1986), but defined mixtures of bacteria had patterns of metabolism that could not always be predicted from the behaviour of the component strains when metabolizing this substrate in pure culture (A. Nixon, P. J. H. Jackman, A. I . Mallet & D. B. Gower, unpublished work). With this in mind, it was decided to examine the metabolism of testosterone by mixed bacteria, directly after removal from the axilla, without any attempt to separate the various bacterial types present. We also present a brief assessment of how the metabolism of testosterone by these 'wild' mixtures compares with that of the previously studied pure and defined cultures. Bacteria were collected by swabbing the axilla of subjects with sterile cotton wool soaked in 1% Tween 80 in sterile physiological saline. The swab was then agitated in 5 ml of the Tween 80/saline solution, and the resulting suspension used as inoculum in the incubation process described previously (Nixon et al., 1984~). A portion (0.1 ml) of the suspension was innoculated into Coryneform broth to which was added 0.1 ml of an ethanolic solution of testosterone [consisting of [''CItestosterone ( 5 mg), ['3C]testosterone (Smg) and 1OpCi of ['4C]testosterone per ml in redistilled ethanol]. The resulting cultures and controls (set up as cultures but without bacteria) were incubated for 21 days at 37°C. After incubation, the extraction procedure described earlier (Nixon et al., 1986) was used. Cultures and controls were extracted with ethyl acetate and the pooled extracts were passed down a combined sodium sulphate/florosil SepPak column (to remove, first, the water and then the Tween