81 
1921-22.] New Method of investigating Colour Blindness. 
specially good colour vision take longest, as they naturally make more 
steps than the others. I conducted the test myself in each case; the 
observer was allowed to repeat the first step, if he desired, but not to 
repeat any other ; and every single observation is entered in the diagrams, 
as it was taken, without any smoothing or adjustment whatever. 
The instrument gives a feeling of what might be called sesthetic 
satisfaction to the person using it. All the readings check themselves. 
After producing the least perceptible difference of tint, the observer has 
next to set to equality. No matter how colour blind he is, he always 
does this in a confident manner. From my position when turning the 
screws I could alwa^^s see what proportion of the area of the aperture was 
covered by the different colours ; in no single case did the observer get 
the filter into an unreasonable position. It 
was possible, in producing the difference of tint, 
to make it appear in the right-hand image 
instead of the left-hand one, or to slip in a 
piece of wire gauze while pretending to turn 
the screw, and so substitute a change of 
intensity for a change of tint. But the 
observers always detected the trick at once. 
The two images were separated by a dark 
space. Had they bounded one another sharply, 
the observer would no doubt have obtained 
smaller steps. But this would be a disadvantage ; the tests would then have 
taken longer, and it will be observed from the diagrams that the number 
of steps obtained is quite sufficient to bring out all the characteristic 
features. The total number of patches is not the absolute number of 
colours which the observer can see ; it is only the number he sees under 
the conditions given by the apparatus. 
The results of the tests are given in the diagrams on pp. 82, 83, and 84. 
Here follow some notes on the different diagrams : — 
1. Normal. 57 patches. 
2. Normal. 65 patches. 
3. Normal. 98 patches. 
4. Normal. My own. No. 4 of perception spectrometer test. Differs from others 
in two respects ; I had some practice in using the apparatus before making 
the diagram, and I turned the screws and read the scales myself. In all 
other cases there was no preliminary practice, and I worked the screws and 
took the readings. 102 patches. 
5. Exceptioually good colour vision. This observer got most patches, namely 26, 
in the perception spectrometer test, and, like Newton, sees indigo as a 
VOL. XLII. 6 
