ON THE RELATION OF OPTICS TO PAINTING. 97 



sun's disk. The moon however is a gray body, whose 

 mean brightness is only about one fifth of that of the 

 purest white. 



And when the moon- irradiates a body of the purest 

 white on the earth, its brightness is only the hundred- 

 thousandth part of the brightness of the moon itself ; 

 hence the sun's disk is 80,000 million times brighter 

 than a white which is irradiated by the full moon. 



Now pictures which hang in a room are not lighted 

 by the direct light of the sun, but by that which is re- 

 flected from the sky and clouds. I do not know of any 

 direct measurements of the ordinary brightness of the 

 light in a, picture gallery, but estimates may be made 

 from known data. With strong upper light and bright 

 light from the clouds, the brightest white on a picture 

 has probably l-20th of the brightness of white directly 

 lighted by the sun ; it will generally be only l-40th, or 

 even less. 



Hence the painter of the desert, even if he gives 

 up the representation of the sun's disk, which is always 

 very imperfect, will have to represent the glaringly 

 lighted garments of his Bedouins with a white which, 

 in the most favourable case, shows only the l-20th part 

 of the brightness which corresponds to actual fact. If 

 he could bring it, with its lighting unchanged, into the 

 desert near the white there, it would seem like a dark 

 grey. I found in fact, by an experiment, that lamp- 



II. H 



