108 Mr. H. E. Ives on Hue Difference 



as determined by Steindler and others. Nutting * has con- 

 structed an equal hue difference scale of the spectrum by 

 this means, and the black dots placed on the spectrum line 

 represent his division. These dots show equal hue distances 

 in the colour triangle, for the immediate neighbourhood and 

 in the direction that the spectrum line there runs. 



A discussion of the law by which hue changes occur in 

 the colour triangle will not be undertaken here. Probably 

 equal hue steps are represented by equal logarithmic incre- 

 ments starting from the vertices and sides of some triangle, 

 apparently neither the equal sensation nor the equal lumin- 

 osity one f- For the present purpose it is sufficient to note 

 first, that these steps are smaller near the sides and vertices, 

 and that the steps are approximately three or four times as 

 large in going in the blue direction as in going at right 

 angles thereto. 



On the basis of this purely experimental knowledge it is 

 possible to apply approximate corrections to the distances 

 plotted by the dotted lines in fig. 2. Thus if we take as our 

 unit in the spectrum-white hue step calculations the hue 

 step where this is shortest, namely, in going from the 

 spectrum parallel to the red-green side, near *5 /jl, then the 



quotient ^- applying to the mixture with the yellow and 



green parts of the spectrum will be subject to a reduction 

 to as little in places as one-fourth. Such corrections, which 

 of course can only be approximate, are shown by the dashed 

 lines of fig. 2. They are of dominating influence in the case 

 of the white and blue comparison lights, and of less effect 

 with the red and yellow-white. In the case of the yellow- 

 white the equal luminosity mixture line falls on the yellow 

 side so much nearer the triangle side as to profit by the 

 resultant decrease of size of the equal hue distances, with 

 the result of keeping the distances in hue steps nearly pro- 

 portional to the triangle distances. 



In regard to the final step, that of finding the critical 

 speeds from these calculations, these are to be read off 

 directly, in arbitrary units, from the plotted curve of fig. 1, 

 provided the value of 8 S for some definite colour triangle 



*"A Method of Constructing the Natural Scale of Pure Colour," 

 Nutting, Bulletin Bureau of Standards, vi. p. 89 (1909). 



t In the first paper on nicker photometer theory it was assumed for 

 purposes of calculation that equal steps along a mixture line are equal 

 hue steps. No important change is made in the result there obtained by 

 the actual apparently varying size of these steps. 



