

1905. | On Colour-Vision by Very Weak Light. 215 
determined the increase in the sensitiveness of the rested eye to green. It 
appears to be less than the increase for blue. 
There is no colour about which so much latitude of expression is used as 
about white. Being composed, according to Newton’s view, of all the colours, 
it is more affected by variations in the composition of the light than, for 
instance, red, which may appear brighter or darker, but cannot change its 
hue. We habitually discount this effect in the case of white, without doing 
so for the other colours. Thus purple flowers often look red by lamp-light, 
and are spoken of as red; but white paper, though it looks yellow, is called 
white. Aud we do so whether using candles, arc lights, or incandescent gas, 
though the actual hue of the paper is different with each, and is strikingly so 
when they are compared side by side with daylight. With regard to the 
colour or colourlessness of very feebly illuminated objects, no experiment 
should be considered valid in which the colours are not contrasted. None 
but spectral colours should be used, and even these, owing to the overlapping 
of the colour sensations, must of necessity appear pale. Few persons have 
the faculty, which dyers by long practice develop, of being able to match 
colours by memory. Yet this is what has to be done in order to say, after 
two hours in absolute darkness, whether a colour so faint as to be barely 
visible is pale yellow, pale green, or pale blue. 
To sum up:—In my paper on Artificial Colour-Blindness, I described 
_ experiments showing that Hering’s argument in favour of a black-white 
sensation is invalid, in so far as it rests on the statement that by intense 
light all colours tend towards white. For the apparent whiteness—in 
the green region, for instance—is only a transitory stage in the production 
of green blindness, and is reached when the green sensation is reduced 
to the strength of the underlying blue and red, the mixture of the three 
being equivalent to white by candlelight, and, therefore, by courtesy, white. | 
And if we continue the experiment the whiteness gives place, as the eye 
becomes completely green-blind, to rich red and blue. 
I submit that the experiments described in the present paper indicate 
that Hering’s theory of the black-white sensation is also invalid in so far 
as it rests on the statement that by very faint light all colours appear 
white. 
For in this case also my experiments, many times repeated and extending 
over a number of years, show that the apparent whiteness is only a 
transitory stage in the recovery from the after-effects of light, and is due 
largely to positive as well as negative after-effects. And when the “ dazzle- 
tints,” as I have called them, have completely subsided, there is no interval 
