238 TRANSACTIONS OF THE CANADIAN INSTITUTE. [VOL. V. 



the others in the facility with which it makes itself known. But the 

 facts are not as we would expect. They are on the contrary arrayed 

 with striking uniformity in flat contradiction of these natural expecta- 

 tions, as is displayed nicely in the curves which we have prepared, but 

 are unable to embody in this paper. 



If we consider in connection with this our characteristic threshold 

 results on a blue ground, we will find more material for wonder. On 

 the blue ground we find the lowest threshold marks are not in the 

 orange as we might have expected from the complementary relations of 

 blue and orange, but they are uniformly at the red. The next lowest 

 points are respectively purple and yellow-green ; whereas the highest 

 points are the orange and the blue-green, and the green. It may be 

 noticed that so far from the orange being seen earliest, as we would 

 expect, and at smallest space threshold, it was generally not distinguished 

 as orange at all, even at the fullest opening of the diaphragm apparatus 

 at a visual angle of something in the neighbourhood of 2°. It is from 

 the first called red, and remains so generally throughout the subsequent 

 transitions in surface size, quite in violation of our natural supposition, 

 that the orange should be seen easiest on a blue ground. We must 

 note also that not only did the entire range of observations made by 

 many different observers agree in the fact that on the red ground blue, 

 and on the blue ground red had the lowest thresholds, but the agree- 

 ment was striking and uniform to such a degree that the various curves 

 representing the several observers' results came almost to one point (the 

 lowest of all), at the blue and red in the case respectively of the red and 

 blue grounds. 



In short, it will be noticed, if we consider the blue ground results and 

 red ground results together, that red and blue act in both cases as we 

 would expect true complementaries should act, viz.: the blue ground 

 seems to facilitate the minute distinguishability of red more than the 

 remaining colour elements, and the red ground acts similiarly upon the 

 blue. What the significance of this singular phenomenon may be it is 

 perhaps difficult to say, but it seems to indicate, at least, that there is a 

 disturbance of the ordinary complementary relations of the various 

 members of the colour system when the colour surfaces are reduced to 

 small sizes. 



Whether such is the case or not we wish to call attention to a very 

 striking coincidence between the facts here disclosed and a similar com- 

 plementary behaviour of blue and red in the case of a very remarkable 

 colour blind (dichromate) investigated by Dr. Kirschmann and recorded 



