Page:Journal of the Optical Society of America, volume 30, number 12.pdf/67



are fairly apparent from inspection

Comparison with the 1929 notation of a given estimate of any attribute, is accomplished by looking along the appropriate row of the table. Wherever a difference is found it means that the consensus of the observers placed that particular sample at the new indicated position in Munsell color space. Thus G 5/6, as viewed on the white ground, appears somewhat misplaced in all dimensions and is better located by the notation 5.1G 4.7/6.4, as entered in the respective “white” columns. The notation of this sample on the gray ground was found to fall at 5.0G 4.6/6.4, as taken from the respective “gray” columns; and the notation for the black ground is 5.1G 5.8/6.3. The general indication from the table as a whole is that the averaged estimates usually approximate rather closely the 1929 Munsell notation and so support the validity of the latter. Still, minor adjustments of many samples may be expected to affect the final smoothing.

Inter-comparisons of the mean estimates from the different viewing grounds are facilitated by the arrangement of all relevant data in three adjacent columns. Thus in the case of G 5/6, the hue estimates from the three grounds are adjacently tabled in the same row, the figures being 5.1G, 5.0G, and 5.1G. Similarly, and further to the right in this same row, the comparative value estimates are 4.7, 4.6, and 5.8; and still further to the right, the chroma estimates are 6.4, 6.4, and 6.3. Comparisons of this character disclose no systematic differential influence of background on either hue or saturation, but it must be noted that such an influence could be reduced by the relativity of the judgments and the influence of color constancy (25-28, 37, 41, 85, p. 595). Briefly, that is because all the samples on a given chart are always seen in relation to each other and never on more than one background at a time. There is, however, a relatively marked and consistent effect in the case of value, and this effect is in the expected direction of simultaneous lightness