Spectroscope of one kind of Glass. 2G9 



to the refracting anole, i.e. D = A, then 



, A 



yU,= 2cOS^. 



The glass being selected, and the ray o£ light which is to be 

 directly transmitted being chosen, the refracting angle must 

 be constructed in accordance with this simple formula. Since 

 the refracting angle is thus to be equal to the angle of 

 minimum deviation, it is clear that the light from the colli- 

 mator must be directed upon the prism at right angles to the 

 side from which the light is to emerge, and when the light 

 emerges the ray selected will be at right angles to the side 

 upon which incidence took place. If then the emergent 

 lio-ht is received upon an obtuse-angled isosceles prism so 

 placed that its base is parallel to the selected ray, the whole 

 of the light will pass through, suffering reflexion once, and 

 emerging in the following way: — The selected ray will be 

 parallel to its course before incidence on the isosceles prism, 

 and therefore still at right angles to the face on which the 

 first incidence took place. 



The rest of the rays will have changed sides as regards the 

 selected ray, and therefore be in a condition to receive an 

 extra dispersion by a prism of the same refracting angle as the 

 original one, but having its point turned so that the selected 

 ray is brought back to parallelism with its original course. 



Now the first prism and the isosceles may be and are in 

 fact merged into one right-angled prism, as shown in the 

 accompanying plan, where also the hypothetical condition of 

 the separation of these two is also shown, for explanation 

 merely, in the right-hand top corner. In the plan the course 

 of the ray selected is shown hy ab c d e f g h i k, and then into 

 the telescope (p. 270). 



The collimator and the two first prisms form one rigid 

 system, the telescope and the two last prisms another, and 

 these move relatively round the symmetric point F. 



It is clear that the telescope might have been fixed so as to 

 receive the light emerging from e. In that case parallelism 

 but not coUinearity would have been secured. With the four 

 prisms both conditions are reached. 



When a ray which is not the selected one is brought into 

 the centre of the field, it is clear that though between the 

 second and third prisms the ray is not })arallel to the colli- 

 mator, it is still passing symmetrically through the four 

 prisms as a whole, nnd therefore with minimum deviation. 



