Page:JOSA-Vol 06-06.djvu/28

 was established indirectly by use of the known positions of the solar white in each of the triangles, this correlation permitting the reduction of the two sets of values to three common elementaries. The red and green values of each set were now tested separately against the data of independent color matches published by Priest in connection with his investigations of the leucoscope and camouflage paints, which accidentally provided materials for checking the results. Both sets checked equally well, so that they were given equal weight in the computation of average values. The violets were tested by data on complementaries and that of Abney was rejected. These best values were then retransformed to terms of elementaries determined by a triangle based on extreme spectral red and violet, with its sides as closely tangent to the locus of the spectral colors as possible. Finally the excitation values were reduced to terms of an equal energy spectrum. Thus expressed, the areas under the three curves are equal for the energy distribution of average noon sunlight, the magnitudes or “weights” of the three elementaries having been so chosen as to yield the solar white with equal excitations.

When quite differently weighted, in terms of the relative powers of the three elementary processes to generate brilliance, the three chromatic curves should summate to yield the visibility curve. It is a well-known fact that in this summation the value of the violet or blue excitation is extremely small compared with that of the red and green. König and Dieterici give no data from which these specific visibility coefficients of the chromatic processes can be deduced. Abney, however, provides data of this sort leading to coefficients by which the ordinates of the excitation curves must be multiplied in order that all three curves should summate to yield his own visibility curve. This latter curve, however, as derived from Abney’s luminosity curve and carbon arc data, departs so widely from the average visibility function, as specified by the Standards Committee of the Illuminating Engineering Society, as to throw doubt upon the general validity of these values. Mr. Weaver and the writer have made