New Analysis of Solar Light. 413 



true colour. I believe, therefore, that in the spectrum observed 

 through olive oil, the carmine-red complementary to the bright 

 green which lies contiguous to the blue is shed over the latter, 

 which is thus rendered violet. 



There are also other methods which I can recommend in this 

 and similar cases of testing. Let the absorbing substance be 

 placed before the greatest portion of the slit, and before the 

 remaining portion white paper, thick or thin, oiled or not oiled, 

 which is so chosen that the place to be investigated in the absorp- 

 tion spectrum shall be equally bright with the cori'esponding 

 place of that formed by the light which has passed through the 

 paper. When the absorption is by oil, it will be seen that in the 

 paper spectrum also the blue between the lines F and G appears 

 violet. For the experiment to succeed, it is necessary that the 

 breadth of the absorption spectrum shall be much greater than 

 that of the paper spectrum. 



By these facts it is plainly proved that subjective changes of 

 colour can take place in the spectrum, not only in the same 

 degree as when ordinary colours are brought together, but per- 

 haps more striking and illusive, on account of the greater vivid- 

 ness of the simple colours. In other cases these changes cannot 

 be refeiTcd to the induction of complementary colours. An 

 example of this, to which Brewster refers, is the band in reddish- 

 orange, which extends about from C to D in the smalt-glass 

 spectrum. It is much darker than the red and yellow portions 

 which lie next it, and seems when looked at between these with 

 an ordinary brightness of spectrum, to possess exactly the same 

 red tone as the extreme red. In a more brightly illuminated 

 spectrum it is observed to pass into orange. Brewster first called 

 the band orange-red*, but aftei^wardsf affirms that Sir John 

 Herschel found it to be pure red|, and thought he had observed 

 a change wrought in it by absorption. In this case also the 

 separation of the band from the remaining portion of the spec- 

 trum proves that its colour is not in the least degree changed. 

 The same remark applies to the green-blue tones of colour on 

 the green side of the line F, which, as Brewster remarks, on 

 being looked at through a deep blue glass (probably the smalt- 

 glass) become green. When they are isolated and examined there 

 is no alteration of colour observed. 



Finally, in some of Brewster's experiments another physiolo- 

 gical circumstance comes into play ; the same homogeneous light 

 at different degrees of intensity does not excite the same impres- 

 sion of colour. When dazzlingly bright, all colours appear white. 



* Edinburgh Transactions, vol. L\. part 2. p. 439. 



t In liis reply to Airy. 



X Treatise on Light, art. 49G and 50G. 



