i the Sensation of Colors. 251 



the variation of the 

 i primary colors with a variation of postion in 

 the normal spectrum is uniform,* it follows that the change of * 

 color of the normal spectrum should be most rapid about A=582, 

 as it of course is. It is also obvious that if the total quantities 

 of the three colors sue nearly the same in different parts of the 

 spectrum (I here refer to these colors not as really objective, 

 hut as measured in the usual objective way) then the part about 

 A=582 must be the brightest, another familiar fact. 



I may observe that there is a modification of our formula for 

 a sensation of light, which probably better represents the rela- 

 tions of the sensations. Writing, in the first place, 



i=zli \=Jj k=K* 

 the formula is 



log x . i-f-log y . \ -hlog z . k. 

 This loses its validity when any of the logarithms become neg- 

 ative. If 2 is the smallest of the three quantities, we may sub- 

 stitute 



X=- Y=^ 



and the formula becomes 



log X.i+log Y.j-f-logs(i+j4 k). 

 When x or y is smallest there will be two other formulas. 

 Now, as the variation in the brilliancy of the light affects only 

 the last term of the last formula, and not the first two depend- 

 ing on X ami Y, it is more than probable that the eye is habit- 

 uated to separating the element of sensation which this last 

 term represents, and which is continually changing its values, 

 from the rest which remains constant. It is, therefore, likely 



- ■ ■ 

 wofef, the red, or the green, is contained in the smallest propor- 

 tion, is one which has a relation to the natural powers of 

 ' "'crimination. 



My observations have been made with 



which I am indebted to the liberality of the trustees of the 

 tfache Fund. I shall describe it on the occasion of }■ 

 some work of a more serious character. The colored disks 

 m ade use of were very kindly lent me bv Professor Eood. 

 * I will show this in a note in the next number of this Journal. 



