142 Mr. A. Schuster. Experiments with [June 5, 



It will be seen from this table that the proportion of red to green 

 which I require to make an artificial yellow is 0'86. This number is 

 not comparable with those given by Lord Bayleigh for his own and 

 Mr. Balfour's vision, as everything depends on the particular green 

 which he uses to make the match. I was aiming in setting up the 

 instrument to choose such a red and green as would without the Nicol 

 make a yellow, and perhaps took too much of a yellowish-green in 

 consequence.* Table II exhibits the results obtained for various 

 observers. The means only of all observations taken at different 

 times, and with both eyes, are given. 



Table II. 



Number of Mean of Mean difference 



Observer. readings. readings. of readings. 



1 601-0 



2 17 -1-0 2-3 



3 10 +1-5 5-3 



4 25 19 



6 12 -4-0 4-4 



7 5 +3-0 3-4 



8 5 -1-0 2-8 



9 20 -4-0 1-8 



10 5 -7-0 2-2 



11 5 +2-0 3-6 



12 7 +3-3 2-0 



13 5 1-0 



14 15 -73 2-3 



15 15 +3-0 2-7 



16 6 -11-5 0-9 



18 5 -5-0 3-6 



19 5 -5-0 4-0 



20 5 -12-0 2-0 



21 20 -7-2 3-1 



22 5 -4-0 2-8 



23 12 -1-5 4-3 



24 5 -11-0 1-6 



25 10 +3-5 2-9 



26 20 -5-3 1-8 



27 20 1-3 

 29 18 +1-2 2-0 



* [Note, added May 23. Since writing the above, I have determined the wave- 

 lengths of the colours selected, and was surprised to find how much more yellow 

 the green was than I thought or intended ; the wave-length of the green which 

 was mixed with the red was about 5'620 x- 10 ~ 5 . The red was a full red near C, 

 the yellow about the sodium yellow.} 



