108 ON THE RELATION OF OPTICS TO PAINTING. 



which Fechner's law holds; and therewith objects are 

 to be represented whose brightness is beyond the 

 limits of this law. 



We find that the older masters, and pre-eminently 

 Rembrandt, employ the same deviation, which corre- 

 sponds to that actually seen in moonlight landscapes ; 

 and this in cases in which it is by no means wished to 

 produce the impression of moonshine, or of a similar 

 feeble light. The brightest parts of the objects are 

 given in these pictures in bright, luminous yellowish 

 colours ; but the shades towards the black are made 

 very marked, so that the darker objects are almost lost 

 in an impermeable darkness. But this darkness is 

 covered with the yellowish haze of powerfully lighted 

 aerial masses, so that, notwithstanding their darkness, 

 these pictures give the impression of sunlight, and the 

 very marked gradation of the shadows, the contours of" 

 the faces and figures, are made extremely prominent. 

 The deviation from strict truth to nature is very re- 

 markable in this shading, and yet these pictures give 

 particularly bright and vivid aspects of the objects. 

 Hence they are of particular interest for understand- 

 ing the principles of pictorial illumination. 



In order to explain these actions we must, I think,, 

 consider that while Fechner's law is approximately cor- 

 rect for those mean lights which are agreeable to the eye, 

 the deviations which are so marked, for too high or too 1 



