Watson, Color Vision in Monkeys. 21 



intensity scale of each of the above named monochromatic light 

 stimnli, we might next lav off a series of steps, the reaction D. L's. 

 (diiference limen) along the scale of each of the four bands. In 

 order to illustrate the foregoing method, let us suppose that we are 

 testing the ability of an animal to make a discrimination between 

 red and green. We have found previously a reaction threshold 

 value (R. L.) for green, say X, and a similar threshold for red, 

 say, 20 X (that is, red is lower in its stimulating effect than the 

 green). The next step in the experiment would be to confront the 

 animal in the usual way with these two stimuli at the intensity of 

 20 X for red and X for green. If the discrimination arose, we 

 would have to assume that the animal was discriminating between 

 the two colors by reason of the difference in their wave length. 

 Absence of discrimination at this level of intensity, however,, would 

 bo no proof of the lack of 'color vision' {stating the situation in 

 conscious terms for fJie sake of convenience) for the reason that the 

 values X and 20 X might represent the 'brightness thresholds' 

 at these places in the spectrum and not the 'color thresholds.' In 

 the absence of color discrimination at the level of the thresholds 

 (R. L.), we should have to carry our experiment further and test 

 the possibility of the discrimination arising when the intensities 

 of the two stimuli are raised respectively to the level of their pre- 

 viously determined first D. L, (with red, e. g., at the intensity 20 X 

 4" c, where c represents whatever constant the Weber-Fechner Law 

 requires, and green at X + c^. If discrimination failed also at this 

 ])oint, at intermediate points and at points high up on the intensity 

 scale, we would have just grounds for denying color vision in the 

 animal; or if discrimination were possible, for affirming it.^'^ That 

 many difficulties are in the way of the successful carrying out of this 

 experiment, the writer is painfull}'' aware. The chief difficulty in 

 the way of such an investigation will be found to lie in our present 

 crude methods of color photometry. With conditions as they are, 

 therefore, I felt that the safest method to use was the one adopted 

 in the present paper, namely, to alter the relative intensity of the two 

 bands by enormous steps, hoping that when tests on stimulus and 



"Provided, as is not the case, tliere were no otlier factors to consider. 



