Report of the Committee on Colour- Vision. 383 



that the correct remain, which is often only the sample-skein. 

 He is shown what mistake he has made. Names are used to 

 remind him that one class of green may be yellow-green ; and 

 another, blue-green ; and, to induce him to avoid them, he is 

 advised only to select skeins of the same shade as the specimen, 

 although they be lighter or darker, and have neither more yellow 

 nor blue than that. If his first error arose only from a miscon- 

 ception or want of practice in handling colours, he begins 

 generally to understand what he has to do, and to do properly 

 what is required of him, 



" 2. Or else he selects and rejects immediately the skein of 

 the sample itself. This proves that he sees the difference of 

 colour. He is then shown the skein as the only correct one, and 

 asked to repeat the trial in a more correct manner. He is again 

 put on the right track as just before ; and the trial proceeds 

 rightly, unless the error arose from a defect in the chromatic 

 sense. Many seem, however, to experience a natural difficulty 

 in distinguishing between yellow-green and blue-green, or the 

 dull shades of green and blue. This difficulty is, however, more 

 apparent than real, and is corrected usually by direct comparison. 

 If the method requiring the name of the colour to be given is 

 used, a number of mistakes may be the result. If a skein of 

 light green and light blue alone are presented to him, asking 

 him to name them, he will often call blue green, and green blue. 

 But if, in the first case, a blue skein is immediately shown him, he 

 corrects his mistake by saying ' this is blue,' and ' that green.' 

 In the last case it happens so mutatis mutandis. This is not the 

 place for an explanation. It must suffice to say that the error 

 is corrected by a direct comparison between the two colours. 



" There is, according to the theory, one class of the colour- 

 blind — violet-blind — who, in consequence of the nature of their 

 chromatic sense, and, therefore, notwithstanding the comparison, 

 cannot distinguish blue and green. But our method has nothing 

 to do with this class of the colour-blind, because such are not 

 dangerous on railwaj^s. 



"(b) Another Process. — If the one examined place by the side 

 of the sample a shade, for instance, of yellow green, the examiner 

 places near this another shade, in which there is more yellow, or 

 even a pure yellow, remarking, at the same time, that, if the 

 first suit, the last must also. The other usually dissents from 

 this. He is then shown, by selecting and classing the inter- 

 mediate shades, that there is a gradation, which will diverge 

 widely if logically carried out as he has begun. The same 

 course is followed with colours of the blue shades, if the blue- 

 green were first selected. He sees the successive gradations, 

 and goes through with this test perfectly if his chromatic sense 

 is correct. 



" To ascertain further whether he notices these additions, or 

 the tints of yellow and blue in the green, we can ourselves take 

 the yellow-green and blue-green to ask him if he finds this to be 



