34 LECTURES TO SCIENCE TEACHERS. 



round in passing through, a prism, the red rays being bent 

 round the least, and the violet r&ys the most, while there are 

 kinds of light of all shades of refrangibility between the two 

 extremes. If I were to form a coloured image or spectrum 

 in this simple way it would not be what is called a pure 

 spectrum. Suppose, for simplicity of explanation, we had only 

 two kinds of light, blue and red, differing from one another 

 in refrangibility, then the incident light would be decomposed 

 into those two beams which would each diverge separately 

 from the source of light, or rather from the virtual images of 

 that source, and would be bent round to a very different 

 extent in passing through the prism; consequently if we 

 received them on a screen we should get two circular 

 patches of light, one blue and one red. Now actually, as 

 I have said, you have all intermediate shades of refrangi- 

 bility, and therefore this compound fan-shaped beam, which 

 passes through the prism, must be regarded as made up of 

 a vast number of cones of light differing from each other in 

 refrangibility, which increases from the red end to the blue 

 end. Consequently if you were to receive the whole on a 

 screen, any one point of the screen would not be illuminated 

 solely by one kind of light, but by all the kinds the refrangi- 

 bility of which lay within certain limits ; in fact there would 

 be a spectrum made up in this way, each circle that we draw 

 representing the section of one of those cones, and each over- 

 lapping the neighbouring circles. How then shall we arrange 

 to procure a pure spectrum, and that without loss of light ? I 

 say without loss of light, because a very simple mode, in 

 theory, of obtaining a pure spectrum would be to limit a beam 

 of light by one hole, and then by another at a distance ; the 

 diverging beam, which passes through the first hole, would be 

 limited by the second, so as to transmit only a very narrow 

 pencil of light, which you might regard as a mere ray, and if 

 you allowed that to fall upon a prism, it would be bent round 

 differently for the different kinds of light of which it consists, 

 so that you would get in that way a pure spectrum, but at an 

 enormous sacrifice of light. How then are we to obtain such a 

 pure spectrum without loss of light 1 This diagram represents 

 (Fig. A) a beam of sunlight diverging through a small hole, 

 forming a pencil of light. If that were received on a convex 

 lens at a sufficient distance (Fig. B), it would be brought to a 

 focus again on the other side, and would diverge from that 



