j. $oc. Cosmet. Chem., 32, 1-14 (January/February 1981) The variation of skin color in different areas of the human body in a caucasian population in CIE 1976, L*, u*, v* color space ROBERT P. VAN OORT, D.D.S., Department of Prosthetic Dentistry, Department of Oral Surgery, University of Groningen, The Netherlands, JAAP J. TEN BOSCH, PH.D., and PETER C. F. BORSBOOM, Laboratory )•r Materia Technica, University of Groningen, The Netherlands. Received April 14, 1980. Synopsis The prosthetic rehabilitation of patients with a defect in the facial region requires color matching of the prosthesis to the adjacent skin. The investigation of the skin color and the relation between skin colors of different easily accessible skin regions was carried out using a subtractive colorimeter. This method was verified using a spectrophotometer. The objective measurement of the color of the skin is possible if the diameter of the viewing field is taken into account. The data from the spectrophotometer with 5-mm viewing field agreed closely with the results of the subtractive colorimeter. Mean and standard deviations of the color indices in three investigated areas of a population sample are given. The inter-regional correlations of measurements of the color indices for the three measured regions palm, cheek and forearm were weak. The correlations between the color indices measurements were also weak, except for the inner side of the forearm. By means of factor analysis an overall characterisation of the human skin color is presented. To apply the skin color measuring method to a color matching system for facial prothesis, only the measurement in the relevant skin region would be reliable. INTRODUCTION One of the objectives of the rehabilitation of patients with facial defects is the inconspicious reconstruction of this de:fect as soon as possible after operation or trauma. The color match of the prosthesis to the surrounding skin is one of the criteria involved in achieving this objective and, ideally, should be non-metameric. The prosthetic replication of the skin color requires a color system, i.e., a procedure of adjusting colorant mixture until all visually apparent differences in skin color are eliminated. Many color matching procedures for facial prostheses have been devel- oped recently, mainly based on artistic procedures and reproducible color shade guides (1,2). Evaluations of the results of different color matching procedures are not available in the literature. In practice such non-quantitative systems prove to be unsatisfactory in producing an adequate match, especially when taking into consider- ation length of treatment time and costs. Also evident is the metamerism problem due to the differences in spectra of skin and prosthesis. These differences are present

2 JOURNAL OF THE SOCIETY OF COSMETIC CHEMISTS because the natural skin colorants are instable in vitro and therefore cannot be used as a prosthetic material. To improve the color matching procedure and to quantify such a system the purpose of the present study involves two objectives: Firstly, the instrumental and quantitative assessment of the variation distribution, difference and correlation of skin color of different parts of the body in a sample population consisting of Caucasian males and females and secondly, the assessment of the absolute values of skin color from a spectrophotometric measurement of the skin of eleven subjects. Several methods have been employed to study the color of the skin. Edwards and Duntley (3) performed spectrophotometric measurements of the skin and its pigments in 10 subjects of differing races. This work was concerned with the biophysical and biochemical analytic properties of human skin. Buckley and Grum (4) did spectrophotometric measurements on the region of the cheek and converted the mean reflectance curve of 10 white subjects into CIE 1931 color specifications. Weiner and Lasker (5,6) introduced two different photovoltmeters in the anthropological field studies, which were directed towards inter- and intrapopu- lational comparisons of skin color for small wave length regions. Since then Lontz (7) has used the Hunterlab D25 tristimulus colorimetric method for this work. However the latter investigators used too few samples to be of great value in estimating data for a population. The limited applicability of the tristimulus colorimetric method was discussed by Billmeyer, et al. (8). The main conclusion of this paper was that colorimetric methods in general can only be applied in measurements of color differences. For "absolute" color measurements a spectrophotometric method is required. On the other hand, if one demands easy and fast operation and a reasonable price, colorimeters are to be preferred. For our purpose of comparing the skin colors of various parts of the body and in different individuals in the population we selected a colorimetric method. For the assessment of the mean and the spread of skin color in a given population we used a spectrophotometer for absolute color measurement. To facilitate relating the physical specifications of color stimuli to the visual perceptions that arise from them, all data were transformed to coordinates L*, •*, v*, of the approximately uniform color space (CIE, 1976). MATERIALS AND METHODS For the colorimetric measurements we used a Lovibond MK III (Tintometer Ltd, Salisbury, G.B.) with a movable measuring head connected with fiber optics of 2 m in length. We obtained serial No. AF 751-5271. All observations were done by the first author, his color vision was found to be normal by the Ishihara, the H-R-R test and the 100 Hue Farnsworth-Munsell test. The light source was specified as a CIE Illuminant C (approximately). The skin is illuminated at an angle of 45 ø to the surface. The measuring head prevents the leakage of light from the surroundings. The light reflected from an area of 25 mm 2 is received by a fiber optic perpendicular to the surface of observation (Figure 1 and 2). A contact was maintained between the measuring head and the skin. The pressure never exceeded 13 mm Hg, or 1.7 x 103N/m 2, i.e., below the subcutaneous capillary