20 



tion, the mechanical strains to which elastic solids are subjected under 

 certain conditions with the coloured curves which those solids ex- 

 hibit in polarized light. In a paper published in the 'Quarterly 

 Journal of Mathematics/ he has treated in a very general manner 

 the passage of rays of light through optical combinations. 



The research specially mentioned in the award of the Rumford 

 Medal, was commenced by him many years ago, and as early as 1852 

 he had made an instrument for examining the mixture of the colours 

 of the spectrum ; but the first paper which he published on the sub- 

 ject was read before the Royal Society of Edinburgh in 1855. 

 Helmholtz had in the meantime published some important investi- 

 gations on the composition of the colours of the spectrum ; but as 

 the main object of his research was merely to make out under what 

 circumstances two colours of the spectrum could be combined so as 

 to give white light, his results could not be applied to test quanti- 

 tatively a mathematical theory of the composition of colours. 



Such a theory is virtually contained in the empirical construction 

 which Newton long ago gave for determining the effect of combining 

 in any proportions the colours of the spectrum. Newton's construc- 

 tion appears, however, to have been based on estimation only ; and 

 even the mathematical theory involved in it does not seem to have been 

 clearly pointed out ; though it has long been known, or suspected, 

 that colour depends in some way on three elements, whether the 

 triplicity exist objectively in some quality of the light itself, or sub- 

 jectively in our organization. In the paper already mentioned, Pro- 

 fessor Maxwell has not only exhibited in its essential simplicity the 

 theory involved in Newton's rule, but has put the theory to the test 

 of exact quantitative experiments. The instrument which he devised 

 for that purpose, and which he calls a ' colour-top,' enables the ob- 

 server to establish an exact match between two different sets of 

 coloured papers, the colours of which are mixed by rotation in pro- 

 portions which can be read off by the graduation. Such a match can 

 always be established between any four colours (including white as a 

 colour), by properly varying their proportions. If we suppose (n) ob- 

 servations to have been taken with the instrument, establishing each 

 a match between two groups made up of three standard colours, and 

 a fourth colour different for each observation, the results may be re- 

 garded as expressing the (n) fourth colours as linear functions of the 



