42 On the Vision of Colours, 
the analogy of solar light to that resulting from com- 
bustion; and that the former is modified by the 
transparent blue atmosphere, as the latter is by the 
transparent blue liquid. Now the effect of a trans- 
parent coloured medium, as Mr. Delaval has prov- 
ed, is to transmit more, and consequently imbibe 
fewer of the rays of its own colour, than of those of 
other colours. Reflecting upon these facts, I was 
led to conjecture that one of the humours of my 
eye must be a transparent, but coloured, medium, 
so constituted as to absorb red and green rays prin- 
cipally, because I obtain no proper ideas of these in 
the solar spectrum; and to transmit blue and other 
colours more perfectly. What seemed to make 
against this opinion however was, that I thought red 
bodies, such as vermilion, should appear black to 
me, which was contrary to fact. How this difficulty 
was obviated will be understood from what follows. 
Newton has sufficiently ascertained, that opake 
bodies are of a particular colour from their reflect- 
ing the rays of light of that colour more copiously 
than those of the other colours; the unreflected rays 
being absorbed by the bodies. Adopting this fact, 
we are insensibly led to conclude, that the more rays 
of any one colour a body reflects, and the fewer of 
every other colour, the more perfect will be the 
-colour. This conclusion, however, is certainly er- 
roneous. Splendid coloured bodies reflect light of 
every colour copiously ; but that of their own most 
— 
