TEKRESTEIAL PHENOMENA. 41 



noticed that, in Fig. 1, the rainbows are shown greatly mag 

 nified and in positions suitable for showing the lifting up of 

 a rainbow at Athens at midday, as the year advances. 



Several other interesting statements about rainbows are 

 made by Aristotle. He says, for instance, that, in conse 

 quence of the juxtaposition of the light red and greenish 

 yellow, an orange colour is seen in some cases,* and that the 

 colours of the rainbow, light red, greenish yellow, and dark 

 blue, are almost the only ones which painters cannot 

 produce by mixing other colours.! 



It is well known that various colours can be produced 

 by mixing red, yellow, and blue pigments in suitable pro 

 portions. Aristotle considered the extreme colours of the 

 rainbow to be some shade of red and of blue respectively, 

 but it is not clear what was the intermediate colour ; some 

 passages suggest that it was green, others, like the one cited 

 above from Meteorol. iii. c. 2, s. 5, suggest that it was yellow. 

 He calls it irpdaivov, which usually meant leek-green, but it is 

 unlikely that he misunderstood what was known by painters 

 about the mixing of colours, and, mainly for this reason, it 

 has been assumed in this chapter that the intermediate 

 colour was some shade of yellow. The colours and their 

 arrangement, according to this view, are shown in Fig. 2. 

 The colour |v0o v, which Aristotle considered to be due to the 

 juxtaposition of coloured lights, viz., greenish yellow and 

 light red, is assumed to be orange, a colour which ZavQov 

 sometimes denoted. It will be noticed that Aristotle men 

 tions four colours of the rainbow and yet, in some passages, 

 says that each rainbow has three colours only. The 

 inconsistency is apparent only, for he makes it clear that 

 the three colours are those which he considered to be 

 due to reflection. The number of colours of the rainbow 

 which can be distinguished varies with different observers, 

 being usually five, six, seven, or even more. Aristotle 

 discusses the possibility of the formation of more than two 

 rainbows, but concludes, incorrectly, in Meteorol. iii. c. 4, 

 s. 33, that three or more are not produced. With respect 

 to lunar rainbows, he says correctly that these are only 

 rarely seen, and adds that, during a period of more than 

 fifty years, he had known of only two instances.! The 

 colours of a lunar rainbow can be seen, but are by no means 



* Meteorol. iii. c. 4. s. 26. f Ibid. iii. c. 2, s. 5. 



J Ibid. iii. c, 2, s. 9. 



