STUDIES IN SPECIAL SENSE PHYSIOLOGY 31)1 



not, indeed, a complete super-position, because of the bend in 

 the spectral diagram between green and blue, which means that 

 mixtures of green and violet are less saturated than spectral 

 cyanide blue ; we can, however, generalise a little. 



If spectral greenish-blue be mixed with red in certain propor- 

 tions, it matches a mixture of green and violet 



a.GBl. + b.R. = c.Gr. + d.V. 



Hence 



a.GBl. = c.Gr. + d.V.-b.R. 



or we have the unmixable colour in terms of our three chosen 

 colours. 



It is to be remarked that colour equations, as they are termed , 

 of the form 



a.R. + b.Gr. Cc.V. = d.R. 



seem to be a justifiable way of expressing the facts. Addition is 

 uniform, the same result being always obtained when the same 

 quantities are summed ; it is commutative, the order of opera- 

 tions does not affect the result ; it is associative and homogeneous. 

 Analogies may be discerned in the process of colour mixing. 

 Green + (mixed with) Red + Violet = Red -f Green + Violet = (Red 

 -f Violet) + Green. 



If we define subtraction in terms of arithmetical quantity, as 

 uniform, non-commutative, and non-associative, similar analogies 

 can be observed. This, however, is of little importance ; the 

 justification of using the symbol of addition in an arithmetical 

 sense is sufficient for our purposes. 



We can now consider some experimental points. Since practi- 

 cally all chromatic stimuli may be expressed in terms of three, 

 researches can be planned in the following way. A definite 

 spectrum, e.g. the prismatic spectrum of an Auer gas lamp, and 

 three lights of arbitrary but constant intensity, e.g. a red, a green, 

 and a violet, are selected ; each part of the spectrum is then 

 matched by a mixture of all the three or of two, or by only one 

 of them. In this way, Koenig and Dieterici investigated normal 

 colour vision ( 10 ). Details of the results need not detain us ; all I 

 wish to emphasise here is the general conclusion that the experi- 

 mental facts of colour stimulation can be graphically represented 

 by a plane figure and the possible forms of stimuli reduced to 



