Hypothesis of Colour Vision with Mechanical Illustrations. 343 



values corresponding to wave-lengths 0*4//,, 0*55 //-, and 

 0*76 /L6 respectively, thus representing violet, green, and red ^ 

 the value of the log. dec. per half period being retained 

 throughout 8 = 0'2. 



It will be seen that these resonance curves in their 

 general features are very similar to those current as repre- 

 senting the various degrees of excitation of the colour 

 sensations. The following contrasts are. however, to be 

 noted, (a) these resonance curves do not fall completely to 

 zero, (b) there is no special convexity of the curve for the 

 red "responder" for very short wave-lengths. 



From the curves shown it is thus seen that the damping- 

 has to be considerable even for the half period. Accordingly 

 the absolute time to reduce the natural vibrations of one of 

 these responders to a negligible quantity would be extremely 

 short. But vision persists for about a tenth of a second and 

 also the full normal perception is not reached under about 

 the tenth of a second. Hence in addition to the syntonic 

 vibration as a primary response to the stimulus of incident 

 light we must also postulate a secondary effect executed by 

 another mechanism whose operation requires time of the 

 order of a tenth of a second for its completion. 



We may therefore suppose that some atomic or electronic 

 vibratory mechanism first responds in syntonic fashion to 

 the stimulus of the light received, and that this response 

 initiates some physiological or chemical change, the com- 

 pletion of which occupies an appreciable fraction of a second. 

 The nerves are then supposed to be excited by the results of 

 this change. Further, it may be supposed that, after the 

 light is cut off, the vision persists until the ordinary physio- 

 logical operations reinstate the normal condition of things. 

 And this process may be imagined as occupying about the 

 tenth of a second. 



Referring again to the " resonance " curves plotted in 

 fig. 1, their drooping ends might be continued to zero, on 

 the supposition that when the vibratory response fell below 

 a certain limit the physiological change initiated by it fell 

 still more abruptly and soon became imperceptible. 



Such hypothetical continuations are shown on fig. 1 by 

 broken lines. Thus the full line curves may be taken to 

 represent both the syntonic responses and the physiological 

 changes supposed to be directly proportional to them. The 

 broken lines, on the other hand, represent only the values of 

 the physiological changes supposed to be much smaller than 

 would result from simple proportionality to the vibrations 

 that initiate them. The whole of the curved lines, full and 



