Measuring Colour-sensations by Intermittent Light. 529 
Intermittent Stimulation—The very remarkable results obtaied by 
‘Shelford Bidwell with intermittent light seemed capable of interpretation as 
‘due to temporary colour-blindness induced by a succession of stimul. I 
itherefore repeated his experiments, using the pure colours of the spectrum.* 
‘It proved possible to sce the boundaries of the colour-sensations, by using 
spectral and monochromatic light instead of pigment colours for the first or 
‘for the second stimulus, but there was too much uncertainty as to which 
stimulus would prove effective to make the method reliable for general 
testing. 
_ In May of the following year, Shelford Bidwell announced his discovery 
of the “ External and Border Phenomena” in connection with negative after- 
images.t This gave the clue I wanted, showing that a very complete colour- 
blindness could be induced in one portion of the retina by light falling upon 
another part, and in my paper on “ Areal Induction” { I described an 
apparatus by which this could be done. Since then my aim has been to work 
out the details of a practical apparatus for colour-testing on these lines. 
The Apparatus. 
The Spectrum.—A series of experiments with Thorp’s celloidin gratings 
showed that though a sufficiently strong illumination could be obtained with 
them, the unavoidable small admixture of white light confused the results 
in the more refrangible parts of the spectrum. But for this objection the 
use of these gratings would have greatly simplified the construction of the 
apparatus. . 
With prisms, the red is so much compressed and the violet so much drawn 
out, that it is difficult to get a monochromatic field in the red without 
employing a dispersion so wide that the violet 1s too weak to produce 
fatigue. I have, however, found a way out of the difficulty. 
Any pair of similar prisms symmetrically situated with regard to a beam 
ot parallel rays may be treated as a glass prism whose angle is that between 
the outer faces, containing an air-prism of the angle between the adjacent 
faces. The pair of them together fulfil, therefore, the conditions of minimum 
‘deviation. 
Ii, then, they are simultaneously rotated in opposite directions upon the 
prism table, the combination continues in the position of minimum deviation, 
but behaves as though it were a single prism of variable angle. 
Accordingly, the spectrum traverses the field of view as in a constant 
* ‘Roy. Soc. Proc.,’ vol. 66, p. 213. 
+ ‘Roy. Soc. Proc.,’ vol. 68, pp. 276, 277, etc. 
techkOy. coc. roe, vol. 695 pp. 127, 128: 
