NACREOUS AND INTERFERENCE PIGMENTS 165 reflectance curves.) The light from a tungsten source first passes through the monochromator before it is allowed to fall alternately on the sample and a reference standard which are located adjacent to one another and in the same plane (5, 6). Two adjustments are made to set the angle of incidence and the angle of viewing. Restrictions on the settings are that the total angle between the incident beam and the viewed beam cannot be less than 30 ø nor more than 90 ø . Thus, the values of the angles i/v, i.e., angles of incidence and viewing, respectively, must be between --15ø/15 ø and --45o/45 ø for specular reflectance. Diffuse reflectance, however, can be measured over a greater range of angles, e.g., up to --15ø/75 ø. In either case, the reflectance is recorded as a function of wavelength from 400 to 700 nm. The reference standard for these experiments was a barium sulfate pressed cake which was chosen as a white diffuse reflector. Because the reflectance at specular angles from nacreous surfaces would be con- siderably greater than specular reflectance from the BaSO4 standard, a neutral density filter transmitting 10% of the light was placed between the nacreous sample and the light detector for angles close to specular. This filter reduced the intensity of the reflected light sufficiently to ob- tain a balance against the BaSO4 cake without distorting the color char- acteristics of the reflected light. For readings far from specular, e.g., m15ø/45 ø to --15ø/75 ø, where diffuse reflectance by nacreous samples is much less than by BaSO4, the filter was either removed or placed in front of the BaSO4. Reflectance readings made with the filter were mul- tiplied or divided by 10 to make all readings comparable. The reflec- tance readings are relative to the reflectance of the BaSO4 standard under the same angular conditions, and do not indicate the absolute reflectance of the nacreous pigment sample. Some of the nacreous pigments on which data are reported are not at present commercially available. RESULTS AND DISCUSSION Nacreous Luster The essentially specular nature of nacreous luster has been men- tioned. A first impression of the reflectance of nacreous pigments can be obtained by examining specular reflectance with the trilac gonio- spectrophotometer. Figure 2 demonstrates the reflectance of natural pearl essence plates (crystals approximately 30 X 6 X 0.07 3t from fish scales) and natural pearl essence "needles" (a mixture of plates and the

166 JOURNAL OF THE SOCIETY OF COSMETIC CHEMISTS I/v : -15ø/15 * o 88 • Pearl Es5ence Plates o 88 • Pe3rl Essence Needles Clear Lacquer 0 88 • EOe 550 ' 65o - WAVELENGTH (rim) Figure 2. Spectrophotometric curves at specular reflection of pearl essence plates and needles very narrow platelets, approximately 30 X 1 X 0.07 /•, from fish skin). Detailed descriptions of pearl essence and other nacreous pigments are given elsewhere (1, 3). Both curves are based on drawdowns of 0.88% pearl essence crystals. (All concentration figures refer to weight per cent in the wet drawdown lacquer.) A curve for the reflectance of a clear lac- quer fihn is shown as a blank. Finally, to illustrate the difference be- tween a nacreous and a nonnacreous pigment, a curve is shown for a conventional TiO=dispersion, also at 0.88% pigment in the lacquer. Figure 2 was made at --15ø/15ø, that is, --15 ø is the angle of inci- dence and +15 ø is the angle of viewing. The known higher luster of pearl essence plates in comparison with needles is shown by the higher specular reflectance of the plates. The fact that all the curves are so close to the horizontal indicates that the colors are essentially "white." Nevertheless, the plate pearl essence coating appears very slightly greenish yellow at specular reflection in comparison with the TiO,,, as is evident in the peak of the spectrophotometric curve for plates at approxi- mately 520 ran. This very weak color arises from the fact that pearl essence plates are interference films. The optical thickness of the plates, or the geometrical thickness multiplied by the refractive index (1.85), is close to one-quarter of 520 rim, causing the plates to have maximum reflectivity at this wavelength.