LIGHT. 131 



To take a more refined instance : nitrogen is perfectly 

 colourless, oxygen is perfectly colourless, but chemically uni- 

 ted in certain proportions they form nitrous acid, a gas which 

 has a deep orange brown colour. I know not how the col- 

 our of this gas, or of such gases as chlorine or vapour of 

 iodine, can be accounted for by the ethereal hypothesis, with- 

 out calling in aid molecular affections of the matter of these 

 gases. 



Colour in many instances depends upon the thickness of 

 the plate or film of transparent matter upon which light is in- 

 cident ; as in all those cases which are termed the colours of 

 thin plates, of which the soap bubble affords a beautiful in- 

 stance. 



When we arrive at the more recent discoveries of double 

 refraction and polarisation, the effects of light are found to 

 trace out as it were the structure of the matter affected, and 

 the crystalline form of a body can be determined by the 

 effects which a minute portion of it exercises on a ray of 

 light. 



Let a piece of good glass be placed in what is called a 

 polariscope, or instrument in which light that has undergone 

 polarisation is transmitted through the substance to be exam- 

 ined, and the emergent light is afterwards submitted to anoth- 

 er substance capable of polarising light, or, as it is termed, an 

 analyser ; no change in effect will be observed. Remove the 

 glass, heat it and suddenly or quickly cool it as to render it 

 unannealed, in which state its molecules are in a state of 

 tension or strain, and the glass highly brittle, on replacing it 

 in the polariscope, a beautiful series of colours is perceptible. 

 Instead of subjecting the glass to heat and sudden cooling, 

 let it be bent or strained by mechanical pressure, and the col- 

 ours will be equally visible, modified, according to the direc- 

 tion of the flexure, and indicating by their course the curves 

 where the molecular state has been changed by pressure. So 

 if tough glue be elongated and allowed to cool in a stretched 



