87 
1921-22.] New Method of investigating Colour Blindness. 
26. This observer thought he was colour blind, as he had a difficulty in naming 
shades of pink and peacock-blue. Had been rejected for the R.A.F. by 
wool test as partially colour blind. In the bead test he left pink and 
yellowish green in the box. Almost normal. 57 patches. 
27. A pronounced green anomaly, requiring only *06 times as much red as the 
average. In bead test put peacock-blue, greenish yellow, gold, yellowish 
white, and amber in green, and pink in blue. 18 patches. 
28. Bead test correct with a slight effort. When shown a spectrum, called green 
with a sliglit trace of yellow “yellow.” When examined by Rayleigh test, 
was found to be a green anomaly, only requiring '309 times as much red as 
average. 34 patches. 
The spectrum colours lie approximately along the red-green and blue- 
green sides of the triangle. The total number of steps on these sides 
should therefore correspond roughly to the number of monochromatic 
regions obtained by the Edridge-Green colour-perception spectrometer ; 
only roughly, however, for the latter instrument works from end to end 
of the spectrum, whereas the new instrument goes from a red with some 
orange in it to ultramarine blue. Five of the observers whose diagrams 
are given in this paper were tested by the perception spectrometer four 
years ago. The following table gives the results by the two methods : — 
No. of observer’s diagram in present paper 5 4 6 1 1610 
Number of steps on two sides of triangle 30 22 1 16| 16 8-| 6 
Number of steps by perception spectrometer 26 17 15 18 8 5 
The agreement is as good as can be expected. The colour perception 
spectrometer has the advantage that it works with spectral colours ; the 
new instrument has the advantages that the readings are self-checking, 
that all possible variations of colour are tested, and that the results are 
exhibited in their relation to the colour triangle of Helmholtz and Maxwell. 
Of the 23 colour-blind observers, 14 were taken from a list of the 
20 worst cases, compiled from the surveys of the past few years. The 
other 6 had left the University. The remaining 9, who are quite as 
bad as the 14, offered themselves in the course of the investigation. So 
that the diagrams give a fair idea of the worst cases we should meet 
with in a random collection of 1000 men. 
It is maintained that the colour blind fall naturally into two classes. 
I am unable to separate the diagrams into two classes, or even to recognise 
two pronounced types. According to the original view of Helmholtz, 
described under his name in so many textbooks, these two classes should 
be the red-blind and the green-blind. The green usually employed was an 
