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



as coloured. This peculiar behaviour of r^fS^ taken in connection with the 

 fact that other colours are in minute sizes seen only as uncoloured spots, 

 throws a side light upon component colour theories such as the Young- 

 Helmholtz specimen. We would naturally expect according to this 

 theory that, as the coloured surface became small its influence would 

 be to stimulate only isolated minute areas in the retina, presumably 

 areas so minute as to contain only the one class of nervous elements 

 which respond to one of the original colour components red, green or 

 violet (blue). But we find that this expectation is fulfilled in the case 

 of red, and not in the case of green and violet, which would seem to 

 contradict the theory. Especially is this convincing when we reflect 

 that green and violet can be seen as colourless light of which they are 

 supposed to be constituent elements, at smaller sizes than they can be 

 seen aiS green and violet. 



Our characteristic thresholds on the black ground displayed a degree 

 of regularity and mutual agreement in tendency among the difl'erent 

 observers' results, which was gratifying. They all coincided in having 

 for the lowest space thresholds red, yellow and blue (to blue-green), and 

 for the regions of highest thresholds those about (orange), orange-yellow, 

 yellow-green and violet. The marked correspondence we have graphi- 

 cally presented in curves which will be given with the larger exposition 

 of the problem and results subsequently. 



On a grey ground wherein the light contrast influence was to all 

 intents eliminated, we obtained results which, considered relatively, 

 among the various colours with regard to their characteristic space 

 thresholds, were well in agreement with the results obtained upon the 

 absolutely black ground. That is to say, the colours of the lowest 

 characteristic space thresholds were again red, the region around 

 yellow, and that around blue and blue-green, while the regions of highest 

 characteristic thresholds were as previously, the orange, green and violet. 



It is not permitted us, in the scope of this brief paper, to pause to 

 discuss these results. For this we refer to the paper which will appear 

 later. Meanwhile, all we can do is to call attention to the fact which is 

 noticeable, namely, that the colours which have the highest thresholds 

 on these colourless grounds, are respectively just about the comple- 

 mentaries of those wherein the lowest thresholds obtain. Whether this 

 is significant or insignificant, we cannot here debate, but at any rate, it 

 is a balanced regularity of behaviour which adds a certain aesthetic 

 interest to the findings of our experiments. There is no a priori reason 

 which would lead us to expect that, of themselves, complementary colours, 



