240 RECEXT PROGRESS OF THE THEORY OF VISIOX. 



painters make green by mixing blue and yellow pig- 

 ments, the union of blue and yellow rays of light pro- 

 duces white. The simplest way of mixing coloured light 

 is shown in Fig. 33. P is a small flat piece of glass ; b and 

 g are two coloured wafers. The 

 observer looks at b through the 

 glass plate, while g is seen re- 

 flected in the same ; and if g is 

 put in a proper position, its image 



\x i exactly coincides with that of b. 



It then appears as if there was 

 a single wafer at 6, with a colour 



produced by the mixture of the two real ones. In this 

 experiment the light from 6, which traverses the glass 

 pane, actually unites with that from #, which is reflected 

 from it, and the two combined pass on to the retina at 

 o. In general, then, light, which consists of undu- 

 lations of different wave-lengths, produces different im- 

 pressions upon our eye, namely, those of different colours. 

 But the number of hues which we can recognise is 

 much smaller than that of the various possible com- 

 binations of rays with different wave-lengths which ex- 

 ternal objects can convey to our eyes. The retina 

 cannot distinguish between the white which is pro- 

 duced by the union of scarlet and bluish-green light, 

 and that which is composed of yellowish-green and 

 violet, or of yellow and ultramarine blue, or of red, 

 green, and violet, or of all the colours of the spectrum 

 united. All these combinations appear identically as 

 white ; and yet, from a physical point of view, they are 

 very different. In fact, the only resemblance between 

 the several combinations just mentioned is, that they are 

 indistinguishable to the human eye. For instance, a sur- 

 face illuminated with red and bluish-green light would 

 come out black in a photograph ; while another lighted 



