58 JOURNAL OF THE SOCIETY OF COSMETIC CHEMISTS usually has the reversed problem of beginning with a source of light such as daylight, and having the prob. lem of adding a dyestuff or several dye- stuffs to an object to remove part of the white light. This is the princi- ple of subtractive colorimetry, namely, to start out with white light and selectively remove parts of it by means of dyes or pigments. Color matching is still an art rather than a science. By far the great majority of all color matches produced in industry are produced visually rather than by means of in- struments. In practicing this art, the cosmetic color matcher will find the publications of Mr. W. H. Pea- cock particularly helpful. (Refer- ence is made specifically to Calco Technical Bulletin No. 573 entitled "The Practical Art of Color Match- ing" by William H. Peacock and Calco Technical Bulletin No. 715, :'The Application Propei'ties of Certified Coal Tar Colors" by William H. Peacock.) A useful con- cept in visual color matching is the color triangle which has red, yellow and blue at the points of the tri- angle. These are the primary col- ors. Equidistant between these three points, in the middle of the triangle, is the spot which is labeled "black." This is formed by mixing all three primaries together. Equi- distant between any two primary points are three more points on the side of the triangle the one mid- way between red and yellow is "orange" the one mid-way be- tween yellow and blue is "green" and the one mid-way between blue and red is "purple." This indicates that mixing a red and a yellow dye or pigment together will produce an orange color. This concept will be found very helpful in color match- ing. For instance, if someone has a standard red shade, and a sample which looks different from the standard, the question to ask is--in what manner 'does the sample differ from the standard? Suppose, for instance, that the standard is a red. With this color triangle in mind, the color marcher should ask himself if the sample differs from the standard red by being more of an orange or if it differs by being more of a pur- ple. Whichever way it differs, the sample can be brought to a match by adding the opposite color. Thus if the sample is tending toward an orange, then a purple pigment or toning color must be added to cor- rect the shade. It is clear from this triangle that dark colors, such as greys and browns, can differ from the standard in at-least three different directions, tending toward the three different primaries. Once this direction of divergence is estab- lished, the opposite color must be added to bring it back to shade. A small percentage of color match- ing is done spectrophotometrically. Such color matching can be directed toward either one of two objectives. One objective might be to duplicate exactly the spectrophotometric curve of a standard. This is highly desirable, because in this case, the sample and standard will look alike to any observer under any condition of illumination. It is also possible,

COLOR MATCHING IN THE COSMETIC INDUSTRY though more complicated, to d,.rect the color matching toward produc- ing a visual color match to the standard observer under a standard illumination, although the spectro- photometric curves do not have to be alike. As an illustration with transmission data, the transmission curve of a blue glass filter is given by the spectrophotometric curve relat- ing transmission to wave length. Similarly, the spectrophotometric curve gives the same kind of data for a yellow filter. If these two filters are in series, that is if the light passes first through one and then through the other, then the resultant curve may be obtained by multiply- ing the transmission of one filte•: by the transmission of the other. For instance, let the light pass first through •he blue and then through the yellow filter. Suppose that at 500 millimicrons, 70% of the inci- dent light is transmitted by the blue filter. Now let us further suppose that the transmission of the yellow filter at this wave length is 40•. Forty per cent is then the fraction of the blue transmitted fraction, which is transmitted by the yellow filter.. The product of these two numbers, or 28%, is the portion of the initial light which passes through both filters. By multiplying the trans- mission curves of the filters together, wave length by wave length, the re- sulting transmission of the combina- tion is found. It does not matter whether the light passes first through the yellow filter and then through the blue, or in the reverse order. In preparing a colored solu- tion which must be green, it is pos- sible to add first a yellow dye to the solution, and second a blue dye to the solution. It does not matter whether the light passes first through the yellow and then through the blue, or whether it passes through both of them simultane- ously. Thus the color of the solu- tion can be predicted from the color of the individual dyes used to make up the solution. A similar situation holds for re- flectance work, only it is more com- plicated. In reflectance work, there are two processes that affect the light one is absorption and the other is scattering. In order to get a high reflection, it is necessary to. have a material of low absorption and high scattering power, such as one of the white pigments. On the other hand, in order to get a dark sample, it is necessary to have a low scattering power, and a high absorp- tion. Some prediction of colors of reflectance samples have been made, calculated on the basis of the reflect- ance of the individual pigments in- volved• but this kind of work is still in the experimental stage. There are a number of factors in- volved in the judging of color effects, which the cosmetic color marcher must keep in mind. These factors are too numerous to elucidate here, and anyone interested is referred to the Calco Technical Bulletin No. 573 by W. H. Peacock, which lists and discusses twenty-two of these factors. One of the most important of these factors. discussed by Pea- cock is partial solubility.