HAIR GLOSS 309 were the suppliers of the hair material. As a basic material we used Central European hair without chemical pretreatment. In preinvestigations we had to realize that gloss could not be improved by products such as gloss sprays on undamaged dark brown hair. So for a great number of tests, these strands were damaged by both perm treatment and hair bleaching. The treatments were done with commercial products, as for example: ß Stylewave X (Clynol) and ß Igora Brillantblond R/Oxigenta Lotion 6% (1:2) (Schwarzkopf GmbH) The perm, however, should not make the hair frizzy. For this reason, the strands were moistened with the perm lotion for 15 minutes with the hair lying fiat. After rinsing at 35øC under running water, the hair was neutralized for 10 minutes with a 2% hydrogen peroxide solution. The hair strand was then bleached by applying the bleach- ing pulp evenly onto the dry tress. After a processing time of 45 minutes, it was again rinsed out under running water at a temperature of 35øC, and finally the strands were left in a bath of 1% citric acid for 30 minutes. After rinsing out again, the strands were dried in the air overnight. In spite of all the efforts to prepare strands of uniform patterns and to fix them, the measurements display individual characteristics of various hair preparations again and again. Exact comparisons with marginal changes of the scattering indicatrix can there- fore only be obtained from the same strand. As indicated in the last chapter, the evaluation of the same tress part always has to be ensured. This refers to the testing of sprays particularly if the hair strands are measured several times after each spray appli- cation. A tress sample that will maintain its arrangement of single hair fibers in a strand before and after a treatment (for example, a hair wash) has to be fixed on both ends--the distal and the proximal side. The embedding material must not only allow fast pro- cessing but also an acceptable elasticity of the single hair fibers apart from sufficient stability. Of all tested materials, Xantopren L (Bayer Dental) used primarily in den- tistry, proved to be most suitable. It is a silicone-based condensation curing elastomeric precision impression material of low viscosity. As an activator we used Optosil- Xantopren liquid, also a product of Bayer Dental. The embedding of the single strands in a silicone rubber was performed in several steps. First the hair fibers of the strand were straightened out, placed exactly in a parallel direction, and clamped between two Teflon blocks over a length of 2 cm in the measuring area. Both ends were glued on either side with silicone rubber. These gluings were made in a chronological order and required 3-5 minutes for drying each one. Strands prepared in this way showed a high degree of stability even if treated with aqueous or alcoholic solutions, which normally cause considerable changes in both length and diameter. The air-conditioned room was set at 23 ø - 2øC and 45% --- 5% R.H. For measuring the scattering indicatrix, the prepared strand samples are fixed in sample holders in order to achieve a quick changing on the goniophotometer and a good reproducibility. The sample holders, constructed by us, are mainly made of two metal sheets (100 mm x 100 mm x 1 mm), one of them containing a hole to illuminate the samples. VISUAL EXPERIMENTS AND COMPARISONS To confirm the relationship between photometric measurements and visual gloss sen-

310 JOURNAL OF THE SOCIETY OF COSMETIC CHEMISTS sation (see VISUAL GLOSS EVALUATION above), some visual experiments were performed. For this purpose a gloss-matching booth was constructed, which allowed a "paired comparison" of two samples under identical conditions. The choice for using "paired comparisons" was governed by the fact that estimating absolute magnitudes in gloss is too difficult at present. The visual comparisons corresponding to the objective measurements were performed either on identical strands or on strands prepared from the same hair quality with identical treatment conditions. Thus more stable and reproducible test conditions were achieved than by evaluating directly on a model. To be in accordance with the goniophotometric measurements with respect to the sample illumination, an incident angle of ½• -• 30 ø was chosen. Three types of light sources were available, fluorescent lamps in warm light (nearest color temperature 2700 K) or in cold light (5000 K), producing convergent beams, as well as a 35-mm slide projector (3200 K) illuminating the samples with nearly parallel light via a mirror through a hole in the ceiling of the booth. To avoid unwanted light scattered from the walls, the whole interior of the booth was painted mat black. The "paired comparisons" were performed by presenting two hair strands lying side by side as shown in Figure 10. Five persons judged these strands in a blind test as a double operation with the question: better, worse, or the same? An agreement of at least four judgments characterized the sample as "visually identifiable." It turned out that the gloss evaluation of all test persons correlated with the earlier mentioned indicatrix parameters. RESULTS AND DISCUSSION Hair strands show a different reflection behavior depending on the orientation of the strands in relation to the direction in which they are illuminated. An example of the change of the respective indicatrix of the same strand of dark brown hair shape is given in Figure 11. The ordinate shows the indicatrix I in arbitrary units--a measure for the reflected light•nd the abscissa the corresponding receptor angle. The dashed lined indicatrix represents the strand that has been placed parallel to Figure 10. Two hair strands prepared for pair comparison in the gloss-matching booth. Gloss is remark- ably increased in the right strand.