220 SECTIONAL ADDRESSES 



Miles (^®) to be ' certainly the most sensitive indicator of colour weakness 

 that we possess,' and in the summary at the conclusion of the article, he 

 suggests that in testing men for mercantile establishments, only Plate V 

 need be used at the original interview. It is also regarded in the Report 

 on Colour Vision Requirements in the Royal Navy as one of the most search- 

 ing plates. It is true that 8i per cent, of the colour blind read the figures 

 as 21, and the others in different erroneous ways, but if we look at the 

 results from the N and the N — groups, we must modify our opinion. Only 

 56 per cent, of the N group passed, whereas 1 1 per cent, failed, that is, 

 read the 74 as 21. The remainder read the 7 as a 2 or as a 9 or as a i, so 

 that the figures read as 24 (7 per cent.), or 94 (2 per cent.), or 14 (i per 

 cent.) : or the 4 was read as a i, and the numbers read as 71 (20 per cent.), 

 or 91 (2 per cent.) or 11 (i per cent.). In the N— group, the percentage of 

 those passing is even smaller, 29 per cent., and the failures 31 per cent. 

 The other variations also occurred. It may be, of course, that this plate 

 offers a very delicate test of colour weakness, and therefore is very effective 

 in picking out colour defect of varying degree. But sometimes it was the 

 only error the individual made not only in this test, but in a group of 

 tests. 



The next four plates seem to be very significant. The normal group 

 shows a perfect pass in all four, and the colour blinds almost a complete 

 failure. 



The next two plates containing the hidden numbers differ very much 

 as regards efficacy for detection. The hidden 5 is certainly not visible 

 to the normal eye, and the fact that 4 per cent, with normal colour vision 

 saw it easily is a curious result. These 4 per cent, have perfect colour 

 vision on all the tests, and one would be almost inclined to rate them as 

 N+. Whether the supersensitive see the 5 or not requires further investi- 

 gation. The 2, on the other hand, is not satisfactory. It could be seen 

 fairly easily by all groups as will be evident from the percentages quoted. 

 Twenty-seven per cent, of the N group and 48 per cent, of the N— group 

 were able to read it. 



The last two plates are very satisfactory. 



The Ishihara test is a very reliable test of colour blindness and did not 

 allow any of the colour blinds to pass. It also seems to detect colour 

 weakness in a highly efficient manner.*" 



An attempt was made to investigate these plates with different-coloured 

 filters to discover if possible the regions of the spectrum the light from 

 which caused the figures invisible to certain varieties of colour blindness 

 to become invisible to the normal eye, and the figures read by colour 

 blinds to become visible to the normal eye.*^ 



In order to obtain illumination from different regions of the spectrum, 

 Ilford and Wratten filters were used. The light was that of an arc lamp 



'^ The Personnel J., 1931, IX. Also /. General Psychology, 1922, 2, 535-544. 



*" Dr. Rabkin, Kharkov, has forwarded to me a copy of a test he has just 

 produced. Polychromatic Plates for Colour Sense Examination. It contains 

 features similar to those of the Ishihara, though sometimes circles and triangles 

 are substituted for numbers. From a preliminary survey, it seems very serviceable. 



^^ I am indebted to Prof. Drever for this investigation with the filters. 



