428 JOURNAL OF COSMETIC SCIENCE 80 ._.70 Piedmont •6o t • 30 Green (sp) 2o lO o , 360 C 410 460 510 560 610 660 710 760 810 Wavelength (rim) Figure 2. (Continued from previous page) A B Figure 3. Pictures of colored hair: (A) Coloration of Piedmont hair after 45 minutes of dyeing with semipermanent blue, green, and red colors. (B) Coloration of Piedmont hair with CIAR-4, CIAO-4, and CIAB-25 dyes under specific conditions. with the actual luster change measured with the goniophotometer. Additionally, the total color differences for different colors for the same dyeing time are very different, whereas the experimental luster values are found to be very similar under white-light illumination (see discussion in the last section). This shows that assembling a luster scale from the measured color parameters by this method is obviously not possible, except in specific cases. LUSTER MEASUREMENTS Luster by instrumental detection depends on spectral reflectance of the hair, spectral power distribution of the light source, and eventually on spectral response of the de- tector. For simplicity, GP measurements were done using a monochromatic illumination source, and it serves as a sensitive probe of fiber absorptive and scattering properties.

EFFECT OF HAIR COLOR ON LUSTER 429 Luster values as a function of dyeing time under monochromatic illumination are shown in Figure 4. For all dyes the increase in luster with dyeing time was low. Most changes in luster occurred between zero and five minutes dyeing time, similar to changes observed in reflectance spectra. During this time basic dyes penetrated the cuticle and deposited on the fiber surface. From five minutes to 45 minutes the change in luster was smaller. Even though the HC dyes penetrated into the cortex during this time, they were not sufficient to reach sorption equilibrium (15). Thus, changes in luster with dyeing time are similar to those of reflectance spectra and are related to the different affinities and different time-based dyeing characteristics of semipermanent dye components. For all colors we noticed that with increasing dyeing time the diffuse peaks were weaker and the resolution of the specular and diffuse peaks was better, which resulted in increased luster values. The increase in luster with increase in coloration, and thus with dye concentration in the fiber, can be explained by a simple optical model for light reflection from hair fiber, shown in Figure 5. When light with an intensity of I o impinges on a hair fiber with an angle of incidence of 45 ø , several processes such as reflection, absorption, and scattering occur. The conservation of light can be expressed as follows: •0 = IR q- IAB q- ISC q- IOUT (2) where I R is the intensity of reflected light, lAB and Isc are intensities of absorbed and internally scattered light, and lou T is the intensity of transmitted light, i.e., lost from the back side of the fiber. The first term in equation 2, IR, contains reflections with different origins: = Ix. + Io,FF(x) + IO,F(,NT) (3) 28 26 24 õ 2o 18 16 14 12 Red Green Blue 0 10 20 30 40 50 Dyeing time (min! Figure 4. Luster calculated from goniophotometric curves for various dyeing times for different semiper- manent colors. Measurements were performed under He-Ne laser beam illumination (k=632 nm).