

1905.] On Colour- Vision by Very Weak Inght. 201 
it is easier to distinguish differences of colour between two parts of a 
spectrum that is not continuous. 
The method, which I finally adopted, of reducing the intensity of the 
illumination to the requisite degree, was to vary the distance between the 
spark-gap, or vacuum tube, and a piece of white paper, towards which the 
slit of the spectroscope was directed, the whole apparatus being in the dark 
room, and enclosed in a box blackened inside and covered with a black cloth, 
through which the eyepiece of the spectroscope projected. 
For bright-line spectra, it is better to have the slit rather wide. It is 
difficult, after staying some hours in total darkness, to tell exactly how the 
eye is focussed when straining to catch sight of an almost invisible object, 
and the intensity of the stimulus due to a small object is much diminished 
when the image of it, being out of focus, is spread over a larger surface of 
the retina. As soon as the object is caught sight of, the difficulty vanishes 
and the slit may be narrowed. 
To locate the position of a given line, I used a broad pointer with a square 
end, fixed horizontally in the focal plane of the 
eye-piece (Fig. 1). . This causes an easily visible Fre. 1. 
gap to appear and disappear as it is moved to 
and fro over a faint line. . 
With this arrangement I found that lines as 
far in the red as that of lithium always appeared 
red, however faint, if the eye was sufficiently 
rested, and that the sodium lines were green, but 
that, with more eomplex spectra, the middle 
portions were so much brighter that it was diffi- 
cult to see the ends. I therefore prepared a light- 
filter, composed of gelatine films stained to the requisite depth with aniline 
and other pink and purple dyes, painted on with a brush until the faintest 
visible spectrum of lamplight, seen through them, appeared no brighter at 
the middle than at the ends. 
Experiment 3.—Using the above described light-filter, I placed in the focal 
plane of the eyepiece a stop with four slits so spaced as to leave visible a 
strip of red near B, of yellowish-green near D, of blue close to F, and of 
violet as far beyond G as the light employed would admit. I found after- 
wards that a small strip of paper covered with Balmain’s Luminous Paint, 
aid partly across the white reflecting surface, would supply additional violet 
if necessary. , 
The illumination was made so faint that even after two hours in tota 
VOL. LXXVI.—B. P 
