ON LIGHT. 265 



laying on of successive washes of different transparent- 

 colours, the tendency is to produce, first, a tint very 

 remote from that expected to result from their union ; and 

 secondly, becoming more and more muddy and sombre, 

 the greater the number of such heterogeneous layers of 

 colour. Hence the maxim in water-colour painting, to 

 secure brilliancy by using only a single wash of colour, 

 if possible, to produce the required effect. The painter 

 should never forget that his notion of colour (as com- 

 pared with that of the photologist) is a negative one. 

 He operates solely by the destrnciion of light, and his 

 aim should always be to destroy as little as possible. 

 His direct action (unknown to himself) is upon the tint 

 complementary to that which he aims at producing. 



(49.) Each particular coloured medium has its own 

 peculiar and specific scale of absorptive action, differing 

 inter se in the most singular and capricious manner. In 

 many, indeed in most cases, the spectrum viewed through 

 such a thickness as to give a strong colour to common 

 daylight, in place of being seen as a continuous band of 

 graduating colour, is broken up into distinct coloured 

 spaces, more or less intense, and more or less well- 

 defined, separated by dark intervals. This is particularly 

 the case with coloured gases or vapours. Thus the red 

 vapour of nitrous gas, especially when its absorptive 

 action is intensified by heat, breaks up the spectrum 

 into a succession of narrow spaces, alternately dark and 

 bright, from one end to the other. 



(50.) When coloured flames are examined with such a 

 "spectroscope" as above described, the phcenomena are 



