Theory of Colour Vision. 407 



28'6, 41, and 1*6 at Maxwell's three fundamental wave- 

 lengths. The dispersion of Abney's apparatus was pre- 

 sumably about the same as that of Maxwell's, and Maxwell 

 used the solar spectrum. I have therefore multiplied the 

 three columns of Maxwell's tables VI. and X. by 28'G, 41, 

 and 1*6 respectively, added tlie three figures in each row, 

 divided the sum into each of the three figures separately, and 

 expressed the results as percentages. I thus obtained the 

 columns headed K and J in Table I. 



Unfortunately Konig and Dieterici also do not express 

 their results in units of equal luminosity ; they choose their 

 units so as to make the quantity of each fundamental colour 

 present in sunlight numerically the same. 1 have not been 

 able to find what factors they multiplied up by. So I have 

 reduced the results on p. 819 to units of equal luminosity in 

 the following manner : — I have taken \ rg and X gv (p. 820) 

 as 573 and 496 /jl/jl for Konig, and as 570 and 493 yu,//, for 

 Dieterici ; at these wave-lengths the red and green, and 

 green and violet present in the spectrum are, in their units, 

 numerically the same. I have next ascertained from each 

 of Abney's papers what the ratio of the luminosities should 

 be at these wave-lengths, and taken the mean for each 

 wave-length ; at Konig's \ rg the red should be 2*095 times 

 as bright as the green, and at Konig's \ av the violet should 

 be "116 times as bright as the green, while at Dieterici's \ rg 

 the red should be 1*94 times as bright as the green, and at 

 Dieterici's \ gv the violet should be '144 times as bright as 

 the green. I therefore obtained the values corresponding 

 to MpxwelPs 16 wave-lengths from the table on p. 819 by 

 graphical interpolation, multiplied the R and V columns 

 by the factors cited above, added the three figures in each 

 row, reduced each three columns to percentages, and finally 

 evaluated Konig and Dieterici's R, G, V in terms in 

 Maxwell's R, G, B. The results are given in the columns 

 headed Kn and D in Table I. 1 had to extrapolate both 

 the R columns on p. 819 to wave-lengths 475*9 and 465*9 fifi 

 in order to obtain smooth curves. 



Abney's results were taken in the case of the 1900 paper 

 from columns IV., V., and VI. on pp. 278-9, and in the 

 case of the 1906 paper from columns IV., V., and VI. on 

 p. 344, and plotted as functions of X. The values for 

 Maxwell's 16 wave-lengths were then obtained by graphical 

 interpolation ; Abney's R, G, V were then evaluated in 

 terms of Maxwell's R, G, B. The results are shown in the 

 columns headed A '00 and A'06 in Table I. 



