86 COLOURS OF CLOUDS. CHAP 2. . 16. 



at a time when they either do not appear at all 

 over head, or do not there refract any colours. 

 If it were only from one place that these clouds 

 were seen near the Western horizon, we might 

 suppose that they were local, but as all over 

 large tracts of country the same appearances 

 would be seen probably at the same time, we 

 must conclude that the modification of cloud 

 is existing every where about, but that a parti- 

 cular angle with respect to the sun is necessary 

 to its being visible, or appearing as a coloured 

 cloud. 



We observe that clouds of the same variety, 

 having the same local or angular position with 

 respect to the sun, sometimes appear richly 

 coloured, and at other times scarcely coloured at 

 all ; a circumstance which renders it questionable, 

 whether the colour is from the cloud itself, or 

 whether the cloud only reflects the light which 

 is coloured by refraction in passing through the 

 haze of the atmosphere in the evening ?* The 



* Another very curious phenomenon, which I cannot help 

 ascribing to some peculiar atmospheric refraction is the alter- 

 nation of colours observable in certain of the fixed Stars, 

 particularly Antares, Betalgeus, and other red Stars. I first 

 observed this phenomenon in 1809, in September; in the 

 twinkling of the Star the colour changed, so that one twinkle 



