Page:Journal of the Optical Society of America, volume 33, number 7.pdf/61

 about the middle of this range, but reference to the corresponding part of the ovoid developed in Adams’ space showed that it was flattest there. Furthermore, the ovoid would need not only to be more flat, but actually concave in order to produce agreement with visual observation. The plots of the data in the Adams’ space originally had suggested concavity in this one limited region, but the evidence had seemed insufficient to justify such a drastic localized departure from what the Adams’ space was believed to represent. The low chromas indicated the type of curve first used (dotted line, Fig. 11) but higher chromas indicated concavity. In the original smoothing, the curve was pushed to the limit of the data in this region in order to afford some agreement with the rest of the curve. Now, however, visual check on the result provided reason for giving the strong-chroma data more weight, and so the ovoid pattern was reduced to virtually a straight line (Fig. 11) in the region under discussion. Transformation to the I.C.I. (x, y)-diagram now resulted in chroma loci which partially correct the observed discrepancy; and the corresponding revisions were made in Figs. 1 to 9 and Table I. But the loci still are not really satisfactory in this region. Why, one might ask, should very sudden transitions be required in this particular region and nowhere else? This exception may be due, not to the Munsell samples or to the Adams’ conversion, but to the I.C.I. system itself. The chromatic data, on which the standard observer is based, were taken with a 2° field centrally fixated. The luminosity data probably represent

TABLE IV. The I.C.I. (x, y) equivalents of the theoretical pigment maxima for 40 hues on nine value levels.