618 JOURNAL OF THE SOCIETY OF COSMETIC CHEMISTS Infected scales inoculated onto plates of Dixon's medium and incubated at 37øC for 5-7 days readily grew identifiable colonies of P.ovale, with complete inhibition of micrococci and other common scalp bacteria. At least four different colony types which could be regarded as P.ovale variants were observed (Fig. 6), the most prevalent being opaque, circular, 1-4 Inm diameter, light tan in colour, convex elevation, sometimes with a central nipple developing on further incubation, finely granular structure, fairly smooth surface and with an entire edge. The actively growing cultures had a pleasant, slightly fruity odour having some resemblance to aidehyde C.14 y-undecalactone. The colonies were rather difficult to emulsify and did not form homogeneous dispersions in water. Clumping was apparent microscopically, the individual yeasts being rounded to oval in shape, with a small terminal bud some pleomorphism was seen, but no multiple budding or development of hyphae. The other colony types (which will be reported fully elsewhere) often showed spherical cells morphologically similar to P.orbiculare, Gordon (5,6). One variant showed the distinctive elongated "bottle shape" cell of the P.ovale reference culture No. 14521 held in the American Type Culture Collection although this differed from our usual type in the shape of the cells, the colonies were closely similar macroscopically. SAMPLING TECHNIQUE The ideal method for obtaining a quantitative record of levels of infection Figure • Applicator and method of inoculation.

THE INVESTIGATION OF DANDRUFF 619 would be to develop a standardized method of taking samples, followed by a viable count procedure. Both steps involve serious practical difficulties. The area liable to show dandruff is considerable and the distribution of scaling is uneven, whilst the presence of hair interferes with the sampling. Furthermore, owing to entrainment in the horny scale and fatty matter, and the clumping tendency of the organisms, a true viable count is virtually out of the question. We therefore attempted to devise a reproducible, "semi-quantitative" assessment of the level of infection. A sterile plastic applicator (Fig. 7) was used to brush the scalp of each subject six times in a standardized manner. Samples of scale adhering to the teeth of the applicator were then trans[erred with light pressure to the culture medium in a 35 in petri dish. Two plates, representing two separate applicators, were obtained from each subject. The inoculated plates were then placed in a plastic box with a close-.fitting lid to prevent excessive drying of the culture medium, and incubated at 37øC for 5-7 days. Figure 8 Typical incubated plate showing units of growth of P. ovale.