388 STUDIES IN SPECIAL SENSE PHYSIOLOGY 



These statements owe their present form to Grassmann, and can 

 be summarised in the following way : 



1. If in a mixture one component be continuously varied, the 

 appearance of the mixture will likewise vary ; unequal lights mixed 

 with equal lights produce unequal mixtures. 



2. On the other hand, lights which appear equal give, when 

 mixed, equal mixtures. A corollary is that proportional increase 

 of the intensity of each component does not destroy a match. 



Passing to the actual observations, we at once note that the 

 effect of mixing spectral extremes is the production of a colour, 

 purple, not present in the spectrum at all. It thus follows that 

 any graphical representation of our results must take the form of a 

 closed curve, since, passing from red to violet, we can either travel 

 over the range of spectral colours or by way of purple. 



If we mix lights not belonging to the extreme ends of the range 

 our results are quite different. The simplest cases are those of 

 mixing colours of wave length not less than 540 /J./JL. For instance, 

 a red (670 /x/x) mixed with a yellow (580 /x/i) gives a pure colour 

 of intermediate wave length ; the greater the proportion of the 

 long wave-lengthened component in the mixture, the nearer will the 

 position of the mixed colour be to the red end, and conversely. 1 

 The mixing relations for this part of the spectrum are therefore 

 straightforward ; but the results obtained i.e. that two simple 

 lights when mixed give a colour matching that of a simple 

 light the wave length of which is intermediate between those of 

 its components are only valid for a small part of the range. If 

 we mix a blue-green (510 /x/x) with a blue (460 /x/x), the mixture, 

 although resembling, perhaps closely, a pure intermediate, e.g. 

 490 /x/x, does not match it perfectly. The mixed colour is paler, 

 or as we say, " less saturated," than any of the spectral colours. 

 This is still more evident when we choose our components in such 

 a fashion that the wave length of one is greater and that of the 

 other less than 517 /x/x. If one constituent is taken a little nearer 

 the red than 560 /x/x and the other diminished, in wave length, 

 in each experiment, then, with suitable proportions, the mixtures 

 pass from greenish yellow becoming paler and paler until we reach 



1 It is very important to remember that when we say that two colours are 

 " equal" or "match," we only mean that viewed under apparently the same 

 conditions they look alike, i.e. are followed by the same results in consciousness. 

 Neglect of this truism has been a fertile source of misunderstanding. 



