358 



NA TURE 



[September 9, 1922 



fundamental facts which recommend the resonance 

 theory of hearing are the smallness and rapidity of 

 those motions which constitute the external stimulus 

 of audition. But how much stronger is the argument 

 based on the corresponding facts for vision ! For in 

 vision we are concerned with the frequencies of light, 

 many millions of millions per second, and with dis- 

 placements correspondingly minute. Further, we now 

 know that in all receptions for wireless telegraphy 

 there must be the tuning of a sympathetic vibrator 

 before the detection of the excessively minute and 

 rapidly alternating disturbances which constitute the 

 signal. 



Thus the question naturally arises, Can a syntonic 

 theory of colour vision prove tenable ? To answer this 

 we must review the main facts of the case, put forward 

 a specific hypothesis based upon them, and test it 

 against the facts. 



Chief Facts of Colour Vision. — These may be 

 classed under five heads : the spectrum, spurious colour 

 mixtures, true colour mixtures, time phenomena, and 

 colour blindness. 



(i.) The Spectrum of white light extends over nearly 

 an octave (to borrow the musical term) from red with 

 a wave-length of nearlv four-fifths /x to violet of wave- 

 length two-fifths fa. The spectrum has no gaps in it, 

 thus showing that we have continuous vision over the 

 range in question. The colours change gradually all 

 the way along the spectrum, and we have six or seven 

 common names for the chief colours occurring. But 

 Dr. Edridge Green finds that the spectrum can be 

 divided into bands, each seeming monochromatic and 

 each different from its neighbours. The number of 

 these bands varies with the observer, but he finds it 

 may reach from 18 to 27 for those with very sensitive 

 colour vision. 



(ii.) Spurious Colour Mixtures. — By mixing pigments 

 (in water colour or oils) and by superposing coloured 

 glasses or films we really execute a double subtraction 

 of colour and not a true addition at all. Thus blue 

 and green pigments mixed or blue and green glasses 

 superposed almost always give green. This is because 

 each pigment or glass subtracted (or absorbed) certain 

 colours from the whole spectrum and left certain 

 colours, and that of the colours thus left, green was 

 the only one common to both residues. By the same 

 methods red and green will give a low colour approach- 

 ing grey, or it may give an absolute black. This last 

 effect is easily obtained by superposing two good 

 films, each of which transmits only a limited portion 

 of the spectrum, with no colour common to both films. 



(iii.) True Colour Mixtures. — These true additions of 

 colour may be obtained by converging two or more 

 coloured beams of light, by the colour top, or by 

 stippling, weaving, or the various colour processes now 

 used in book illustrations. These true mixtures give 

 results quite different from those of the spurious 

 mixtures and thus throw valuable light on colour 

 vision. Thus blue and yellow do not make green but 

 white or a light pinkish tinge ; red and green make 

 yellow; red, green, and violet make white, while red 

 and violet give a colour not found in the spectrum. 



(iv.) Time Phenomena. — It is known that the full 

 acquisition of a visual sensation is not attained under 

 something of the order of a tenth of a second, and that 



NO. 2758, VOL. I IO] 



the vision also persists for a like period after the 

 stimulus is withdrawn, before dying out completely. 

 Again, there are effects of fatigue, so that after gazing 

 for twenty seconds at a red object and then at a white 

 surface, a green image appears of the size and shape 

 of the previous red object which had fatigued the eye 

 for red and thus caused the white to appear deficient 

 in red and so look green. 



All the facts of colour vision hitherto enumerated 

 apply to those with normal sight. 



(v.) Colour Blindness. — Finally there are the facts 

 of colour blindness. Some patients are blind to red, 

 others are blind to green or to two of the three colour 

 sensations, some are blind to all three. 



Syntonic Hypothesis. — In forming a syntonic hypo- 

 thesis of vision, in which the initial response of the eye 

 to the stimulus of light is supposed due to the sym- 

 pathetic vibration of something, we must indicate the 

 number of different vibrators imagined to be present 

 at each element of the retina, also the frequencies and 

 dampings natural to them. Since the tri-colour theory 

 of vision has been so successful in many ways it is 

 natural to try first if three vibrational responders 

 could form the basis of a syntonic theory. Obviously 

 if three responders are to suffice their " resonance " 

 or response curves must replace the sensation curves 

 usually drawn to indicate the degrees to which the 

 three colour sensations are excited by the various 

 spectral colours. These curves are much spread and 

 overlap, so that no portion of the spectrum is left 

 without power to excite one or more of the colour 

 sensations. The response curves of the vibrators now 

 postulated can be equally spread by rightly choosing 

 the damping (or dying away of their free vibrations) 

 natural to the vibrator. For, the greater the damping 

 the flatter is the response curve, the less the damping 

 the more sharply tuned is the response. Thus, by the 

 hypothesis of three strongly damped responding 

 vibrators we can account for the visibility of the 

 continuous spectrum just as easily as by the vaguer 

 hypothesis of three colour sensations. What natural 

 frequencies must be assigned to these vibrators ? 

 Probably such as to respond to light of wave-lengths 

 rather less than 0-76/^, about 0-55/*, and rather more 

 than 0-4/x. Or we can think of the matter in musical 

 terms as follows, rather sharper than Cjf, about Fj, 

 rather flatter than B, all three to be in the same 

 octave. 



The Hypothesis Tested. — We have now to test this 

 postulated set of vibrators against the facts of true 

 addition of colours. This test may be carried out 

 mathematically or by experiments with a simple 

 arrangement or model in which the vibrational re- 

 sponders are crudely imitated by pendulums RGV with 

 paper cones as bobs (Fig. 1). (Any form or type of 

 vibrators will serve equally well, as is shown by mathe- 

 matical theory.) They hang from a stretched hori- 

 zontal cord AB which is set in motion by the swings of 

 a heavy pendulum CD, as shown in the accompanying 

 figure. To represent a second light stimulus, simul- 

 taneously imparted, a second cord, AE, and pendulum 

 FH are provided ; the two cords are connected by a 

 bridge of wood near v when both drivers are in action. 



By the use of this model it is easy to make the 

 desired tests of the hypothesis. In speaking of it we 



