1888. ] and on a Photo-Voltaic Theory of Vision. 309 
changed, as well as those of the right, though usually to a smaller 
extent. 
A diagram of colour sensations is usually given as in fig. 1; 
each sensation being affected by the whole range of the visible 
Fig. 1. 
ee ee, 
spectrum. So far as these observations go, however, they seem 
to indicate that, with the exception of violet, rays of higher 
refrangibility than the greenish-blue do not affect the “red 4 
colour-sensation. The “green” sensation seems to be affected 
through a range extending from orange to blue, the lower limit 
for the “violet” sensation being somewhere in the green or 
yellowish-green. 
It also seems that the “green” sensation participates only 
to a very small extent in the perception of blue; this colour 
being distinguished from violet by not exciting the “red” colour 
sensation. 
It may here be noticed that a pure and uniformly illu- 
minated diffraction spectrum would appear throughout of equal 
brightness, and its tint would appear to change at a uniform 
rate if the colour-sensations followed the law indicated in fig. 2; 
the “green” sensation, however, being twice as strongly affected 
by white light as either of the others. If the straight lines in 
fig. 2 be replaced by more continuous curves as in fig. 3, there 
will be a variation of apparent intensity in different parts of the 
spectrum, and a still greater variation in rate-of-change-of-tint, 
which now becomes zero at the middle and ends of the spectrum, 
and attains a maximum at two intermediate points. 
It will be clear that the arrangement indicated in fig. (2) 
or in fig. (3), where each simple colour affects only two sensations, 
Fig. 2. 
Fig. 3. 
will give greater purity of tint than would be possible with the 
arrangement of fig. (1), where all three colour-sensations are 
always affected; for in the latter case, a certain portion of the 
