Page:Journal of the Optical Society of America, volume 30, number 12.pdf/20



ized conditions. The procedure need not be included here, since it is discussed in detail by Gibson and Nickerson in the paper following this one.

If Priest, Gibson, and McNicholas had investigated only five chromatic colors instead of fifteen, there would have been no choice in the selection of data upon which to base these calculations. However, since an arbitrary choice of data must be made, it seemed appropriate to select the five principal colors R5/5, Y5/5, G5/5, B5/5, and P5/5. Use is also made of the rule that a mixture of the five principal colors in equal proportions yields N5/, which according to this definition is nearly but not quite the color of a nonselective surface.

Reference to will disclose the fact that the five principal colors are not precisely of value 5 as intended by Professor Munsell. Instead, Y varies from a minimum of 0.2140 to a maximum of 0.2662. In view of the difficulties this paper may call attention to the fact that the involved in heterochromatic photometry, which were more serious then than now, it seemed reasonable to multiply each set of tristimulus principal color accurately to value 5. This procedure does not alter the hue of the principal colors, and it makes only a trifling change in the chroma. After this adjustment, the basic data relative to the principal colors are as listed in.

The data in indicate what the Munsell Atlas of Color would have been if, starting with the five basic colors as defined in, the extension of the system had been entrusted to the standard I.C.I. observer working under Illuminant C. The first column indicates the Munsell designation. This is followed by three columns giving the corresponding tristimulus specifications. Columns 5 and 6 list the trichromatic coefficients, x and y.

Professor Munsell devoted much attention to speculations concerning the shape of the color solid that results from the representation of hue, value, and chroma in cylindrical coordinates. His speculations had, perforce, to be based on the color gamut obtainable with pigments that were then available, and he was obviously unable to predict what new pigments might result from chemical research. By application of the work of MacAdam on “The maximum visual efficiency of colored materials,” it is now possible to establish the limits of the Munsell color solid which newly discovered pigments may approach but will never exceed. This gamut is expressed most conveniently in terms of the maximum chroma corresponding to a given hue and value, as in Table IV.

Since the Munsell Book of Color, which appeared in 1929, was intended to supplant the original Atlas of the Munsell Color System, it is not immediately obvious that the publication of this paper serves, at this time, a useful purpose. We believe that its publication is justified on the ground that it calls belated attention to the remarkable scientific insight of Professor A. H. Munsell. At a time when there was little to suggest such a procedure, he formulated rules for the construction of a psychophysical color system that could be used today without apology. We believe also that the publication of this paper may call attention to the fact that the psychophysical definitions of the terms hue, value, and chroma given in the Atlas of the original Munsell Color System differ from the specifications by a factor that would bring each purely psychological definitions used since the death of Professor Munsell.

The writers are indebted to Mrs. Blanche Bellamy, Manager of the Munsell Color Company, Inc., Baltimore, Maryland, for the loan of Professor Munsell’s Diary; to Dr. Walter Scott for valuable suggestions; and to Mrs. Arthur Howland for permission to review correspondence between her late husband and Professor Munsell. Miss Dorothy Nickerson has extended most gracious assistance in connection with the calculations.