52 Lord Kayleigh's Investigations in Optics. 



_ Experience has not yet decided the question as to the rela- 

 tive advantages of large and small prisms. As generally used, 

 large prisms have the disadvantages of requiring a greater quan- 

 tity of material for a given thickness and of involving cumbrous 

 and more expensive telescopes. The first might be avoided 

 partially, and the second wholly, by the use of higher refracting 

 angles, or (perhaps preferably) by the addition of half-prisms 

 to the ends of the train (fig. 19). That prisms may act 



Fig. 19. 



as cylindrical telescopes was observed many years ago by 

 Brewster* ; and recently Christie has constructed half-prism 

 spectroscopes in which this property is taken advantage of. 

 In these instruments, however, the total thickness of glass is 

 too small for high resolving-power. 



In the arrangement of fig. 19 the rays from the collimator 

 are received on diminishing half-prisms, by which the width 

 of the beam is increased up to the point suitable for the big 

 prisms. Afterwards the rays are concentrated by magnifying 

 half-prisms until the width of the beam is the same as at first. 

 On this plan the larger prisms need be no higher than small 

 ones ; and the quantity of glass corresponding to a given total 

 thickness varies as the first power, instead of as the square, of 

 the linear dimension. 



I have experimented with a pair of 60° prisms, 3 inches 

 thick and only j inch high. The glass is of rather low den- 

 sity; so that when the position of minimum deviation is 

 adopted, the emergent beam is inconveniently wide. With 

 this material a larger angle would have been preferable ; 

 but much the same result may be arrived at by turning the 

 prisms a little, so as to increase the angle of incidence on 

 the first surface and the angle of emergence from the last 

 surface f. In this way the incident and emergent beams are 

 so far narrowed that they can be received on small telescopes; 

 and the combination is very economical in comparison with 

 one in which the position of minimum deviation is adopted, 



* < Optics' (London, 1853), p. 513. 



t A pair of prisms thus arranged is called by Thollon a couple. . . 



