﻿Fixed- Arm Spectroscopes. 339 



the slit and the focal plane of the observing lens, of a re- 

 flecting surface, which has an angular motion equal to one- 

 half the change in angular motion of the ray refracted at 

 minimum deviation. In the instruments of the Littrow type 

 a reflector is also used, but for an essentially different 

 purpose (to return the ray through the prism) and in an 

 essentially different way from that here indicated. It hardly 

 seems possible that the use of a reflecting surface in this way 

 can be new, as it is something which would naturally suggest 

 itself to anyone who might consider the problem ; but no 

 reference has been found to its employment, perhaps because 

 the necessity for a " fixed-arm " spectroscope does not often 

 seemingly arise. I hope to show, however, that at least some 

 of the forms which have been developed during the last 

 three years are simpler, more convenient, and less expensive 

 than the ordinary form, and for spectrometric work quite as 

 accurate. 1 will briefly describe these forms in the order of 

 design and use, for this, perhaps, naturally is also the order 

 of increasing simplicity. 



Case I. — The case which first led me to a consideration of 

 the problem was one in which it was desired to examine the 

 radiations from a Geissler tube by means of the wave- 

 comparer*. Here the instrument which took the place of the 

 usual observing eyepiece was a massive apparatus weighing 

 about 500 pounds, which required great steadiness of mount- 

 ing, while the Geissler tube, which itself served as the slit, 

 was attached to a mercury-pump on one side and to a sodium- 

 amalgam generating apparatus, with its attached drying- 

 tubes, mercury-valves, &c, on the other, in such a way as to 

 make its movement impossible. At first, indeed, the tube 

 was mounted on a slit-arm of the spectroscope, sufficient flexi- 

 bility being secured to allow of different lines being brought 

 on the slit of the wave-comparer by the use of long lengths 

 of glass tubing, but experiments soon showed that it was 

 necessary to keep the passage from this Geissler tube to the 

 pump as short and large as possible. If any considerable 

 length of tubing intervened, unusual precautions had to be 

 taken to keep all parts of it perfectly dry and clean, for only 

 under these conditions, which, as the tubes were frequently 

 changed, were almost impossible to maintain, could the 

 McLeod gauge attached to the pump be relied upon to give, 

 even approximately, the true pressure in the tubes, particu- 

 larly when the electric discharge was passing. 



The arrangement adopted in this case was that shown in plan 



* See a paper by Professor A. A. Michelson, u Application of Inter- 

 ference Methods to Spectroscopic Measurements," Phil. Mag. Sept. 189:2. 



2 A 2 



