SF. \SATION8 IN TERMS OF LUMINOSITY. 



2G1 



sensation of white, the amount being shown by the portions between the horizontal 

 parallel lines. I, II, and III, &c., show colours which are due to the mixtures of 

 two sensations. A and D are the two most interesting colours. If we take away 

 the green sensation from A we have the mixture of red and violet which a green- 

 hlind person would match with white. Similarly, if we take away the red sensation 

 from D we have a mixture of green and violet which the red-blind person would 

 match with white. 



Fig. 1. 

 A B C D f I 2T Iff IT V 



d a b d b d 



a v a a v a G v- ft a * 



/f G ft G 



G V /f V 



The position which D occupies in the spectrum can readily be found by the normal 

 eye, by finding that colour which, with red alone added, matches the white employed, 

 in other words by finding the complementary colour to the red. The position of A in 

 the spectrum is much less readily determined by the normal eye, since it requires 

 the addition of both red and violet to make the white, a condition which is also 

 necessary with B and C. The position can of course be determined by the aid of the 

 green-blind eye, but a preliminary measurement of colour sensations involving the 

 assumption of its position enables it to be fixed with the required accuracy. Before 

 the measurements herein recorded were made such a preliminary set of observations 

 was gone through, and the position found which was subsequently confirmed by a 

 green-blind person. (See XXII.) 



(IV.) Precautions to be taken. 



There were several considerations that had to be taken into account in making 



these measures. In the first place, the white light used in the observations had to be 



of the same quality, that is, the relative luminosities of the different rays of the 



spectrum had to be constant, for it must not be supposed that the positions in the 



spectrum of A and D are fixed points except for the same quality of white light. 



They may be separated from one another by nearly the same interval in the 



spectrum when different qualities of white light are measured, but the larger the 



proportion of blue contained in the white the more they will approach the more 



refrangible end of the spectrum. For instance, the positions will be nearer the red 



with the white light emitted by the crater of the positive pole of the electric light 



than they would be with light of the sun on a June day near noon. Again, the 



final equations, for white light, given in terms of the three sensations will vary 



according to the white light employed, and they will also vary according to the 



extent of the area of the retina on which the colour images fall. This last variation 



is caused by the absorption by the macula lutea, and may differ in different eyes. If, 



