38 CELESTIAL, ATMOSPHERIC, AND 



explanations, he introduces another idea into his train of 

 reasoning, saying that distant objects seem to be darker 

 because the visual rays fail to reach them, or only partially 

 do so, or because the rays are weakened by reflection. 

 From one or both of these causes, therefore, a bright or 

 white object may appear to be of some colour between 

 white and black, e.g., light red, greenish yellow, or dark 

 blue, which successively approach black. If the visual 

 rays are strong, white would be changed to light red; 

 if less strong, white would be changed to greenish yellow ; 

 and if weak, it would be changed to dark blue. Now the 

 greater the periphery from which the visual rays extend to 

 the Sun, the stronger and more concentrated the rays, but 

 the outer periphery of the primary rainbow is the greatest, 

 and therefore its colour is light red, which is nearest to 

 white. Reasoning in the same way, it follows that the 

 inner part of the primary rainbow is dark blue, and the 

 middle part greenish yellow.* 



Aristotle proceeds to deal with the secondary rainbow 

 and says that this also has three colours, formed by reflec 

 tion, the inner part of the secondary rainbow being light 

 red, the outer part dark blue, and the intermediate part 

 greenish yellow. His explanation of this phenomenon is 

 meagre and presents many difficulties, but the following 

 seems to represent his views. The secondary rainbow has 

 its colours duller than those of the primary and also in 

 inverse order, compared with those of the primary, for the 

 same reason, for the visual rays are weaker because the 

 reflections causing the secondary rainbow take place at a 

 greater distance than those causing the primary rainbow, 

 thus causing the colours to be dull. Again, more rays 

 extend to the Sun from the inner part of the secondary, 

 which inner part is nearest the observer, like the outer part 

 of the primary rainbow. The visual rays, therefore, being 

 more numerous and stronger at the inner periphery, its 

 colour will be light red, for reasons similar to those given 

 when explaining the order of the colours of the primary 

 rainbow, and the other colours proceeding radially outwards 

 will be greenish yellow and dark blue.! 



It will be evident that his explanations depend on some 

 ingenious assumptions, notably that relating to the pro 

 duction of colour-effects by the weakening of the visual 



* Meteorol. iii. c. 4. ss. 20-25. f Ibid. iii. c. 4, ss. 30-32, 



