A Quantum Theory of Colour Visiori. 



231 



However, it must possess a basis of fact, for it affords an explanation of the 

 relative luminosities of the colour sensations. For red, the energies are 

 equally divided ; for green, two go to the colour sensation and one to the 

 luminous ; for blue, the allocation is three and one. Hence the last is the 

 least luminous of the colour sensations. 



(22) Another matter finds explanation in these considerations. The 

 colourless interval between the general threshold and the colour threshold 

 is for long wave-lengths (beyond 670 /a/x) almost completely absent. " In 

 fact, even with very good dark adaptation, such a red light excites the 

 sensation of red."* The small energy. values requisite to excite both colour 

 and luminous sensations in the case of red sensation explains this observa- 

 tion. Very feeble electrons served to excite both. With less luminous 

 colours it is otherwise. In these cases, a relatively small number of the 

 electrons liberated from the sensitiser carry into the cone the requisite 

 energy to build up the stimulus to the value of the colour sensation. For the 

 blue sensation, the colourless interval is therefore the greatest. 



(23) On these views white is a sensation arising in the loss of form attending 

 the activation of all nine fibres. This loss of form causes it to resemble the 

 luminous sensation. It is neither a true luminous nor yet a true colour 

 sensation, but something sharing the properties of both. It is a nine-fibre 

 sensation, sui generis. 



(24) One of the most remarkable facts of vision is its marvelloTis range : 

 from the feeblest twilight to full mid-day sunshine. In photopic vision a 

 very large number of electronic stimuli probably do not and cannot take 

 effect, even if we assume a very brief refractory period. The figures are 

 instructive. 



I assume at least nine nerve fibres, discharging stimuli, in each cone, and 

 that the normal refractory period is 10~* sec. A total of 9 x 10* stimuli can 

 be accepted per second per cone. If we estimate the number of cones in the 

 fovea as 2 X 10*, this area might transmit 18 x 10^ stimuli. The number of 

 electrons or quanta involved is one-third of this where white light is con- 

 cerned ; or, say, 6x10* quanta per second. 



A standard candle gives 5 = 10^ ergs per second (Rayleigh), and from this 

 4 ergs per second reach 1 sq. cm. at a distance of 1 m. Assume a ten-candle- 

 power light, and that the pupil has closed down to the area of 3 sq. mm. 

 Further, suppose the image of the source of light just covers the fovea. 

 Under these conditions 0"22 ergs reach the fovea in one second. 



Now the value of a "green" quantum is about 5 x 10"" erg. Hence the 

 energy reaching the fovea will give rise to about 4x'10^'^ quanta or electrons 



* Parsons, loc. cit., p. 61. 



