130 Mr, B. Brudenell Carter [May 9, 



right or wrong each time. He has always an even chance of being 

 right ; and it would be an insoluble problem to discover how many 

 correct answers might be due to accident, or how many incorrect ones 

 might be attributed to nervousness or to confusion of names. 



We must remember that what is required is to detect a colour- 

 blind person against his will ; and to ascertain, not whether he 

 describes a given signal rightly or wrongly on a particular occasion, 

 but whether he can safely be trusted to distinguish correctly between 

 signals on all occasions. We want, in short, to ascertain the state of 

 his colour-vision generally; and hence to infer his fitness or unfitness 

 to discharge the duties of a particular occupation. 



For the accomplishment of this object, we do not in the least want 

 to know what the examinee calls colours, but only how he sees them, 

 wdiat colours appear to him to be alike and what appear to be unlike ; 

 and the only way of attaining this knowledge with certainty is to 

 cause him to make matches between coloured objects, to put those 

 together which appear to him to be essentially the same, and to 

 separate those which appear to him to be essentially different. This 

 principle of testing was first laid down by Seebeck, who required from 

 examinees a complete arrangement of a large number of coloured 

 objects; but it has been greatly simplified and improved by Prof. 

 Holmstren, who pointed out that such a complete arrangement was 

 superfluous, and that the only thing necessary was to cause the 

 examinee to make matches to certain test colours, and, for this pur- 

 pose, to select from materials which contained not only such matches, 

 but also the colours which the colour-blind were liable to confuse 

 with them. 



After many trials, Holmojren finally selected skeins of Berlin wool 

 as the material best suited for this purpose ; and his set of wools com- 

 prises about 150 skeins [showiij. The advantages of his method over 

 every other are that the wool is very cheap, very portable, and always 

 to be obtained in every conceivable colour and shade. The skeins 

 are not lustrous, so that light reflected from the surfaces does not 

 interfere with the accuracy of the observation ; and they are very 

 easily picked up and manipulated, much more easily than coloured 

 paper or coloured glass. The person to be tested is placed before a 

 table in good daylight, the table is covered by a white cloth, and the 

 skeins are thrown upon it in a loosely arranged heap. The examiner 

 then selects a skein of pale green much diluted with white, and throws 

 it down by itself to the left of the heap [shown]. The examinee is 

 directed to look at this pattern skein and at the heap, and to pick 

 out from the latter, and to place beside the pattern, as many skeins 

 as he can find which are of the same colour. He is not to be par- 

 ticular about lighter or darker shades, and is not to compare narrowly, 

 or to rummage much amongst tlie heap, but to select by his eyes, and 

 to use his hands chiefly to change the position of the selected 

 material. 



In such circumstances, a person with normal colour-sight will 



