68 Bar us — Displacement Interferometer. 



was tried without apparent disadvantage, but a light paper 

 cylinder and air damping is of course preferable, being far 

 more sensitive. The mica B should be somewhat adjustable 

 so that the beam^> may be given the proper direction, roughly. 

 The methods for doing this will be described below. 



The whole apparatus, covered at least as far as C, was then 

 mounted on one arm of the interferometer by the aid of a large 

 clamp of the ordinary pattern, the clamp having been screwed 

 into the base A at F. As the arm is of gas pipe, F is an ordi- 

 nary nipple of suitable length. It is advisable to further sup- 

 port the tin case with wood uprights from the table to avoid 

 the effect of tremors due to the vicinity of an active labora- 

 tory ; but no great difficulty was here encountered. In fact, 

 the ellipses of the interferometer were obtained in all cases 

 after a preliminary rough adjustment of the mirrors with sur- 

 prising ease, and they were quite as stable and available for 

 measurement as if the mirrors had been rigidly fixed. Natu- 

 rally this was quite contrary to our anticipations, but it is 

 shown by the large number of experiments made, each of 

 which required a new adjustment. 



The measurement consists merely in a compensation of the 

 displacement of the movable mirror on the electrometer, by 

 the micrometric displacement Ai\ 7 of the mirror belonging to 

 the other component beam. The center of the displaced 

 ellipses in the field of the telescope is thus brought back to the 

 fiducial sodium line. The compensation admits of an accu- 

 racy of about i^N — - 0001 centimeter without resorting to the 

 interference rings. Other remarks on the optical method are 

 given in the next paragraph. 



3. Cylindrical electrometer; movable cylinder with in. Inter- 

 ferometry. — The frame-work and the mounting here is in 

 general the same as in the preceding case and the movable 

 cylinder kk is the same, but differently suspended. Details 

 must be omitted here and other forms of the instrument (disk 

 pattern, closed field pattern) will be described below. 



The general treatment of the displacement interferometer 

 has been given in my earlier papers. The field of the telescope 

 after the appropriate two (of the four) undeviated images of 

 the slit, from front and rear of the grating plate, have been 

 put into coincidence horizontally and vertically (the images 

 being usually bluish and yellowish), contains but three super- 

 posed spectra, although there are four white images, two from 

 the front and two from the rear of the plate. For the compo- 

 nent beam reflected from the rear face of the grating is not 

 diffracted. It is thus advantageous to eliminate the third spec- 

 trum also, and to keep only the two which are superimposed 

 for interference in the field. This may be easily done as fol- 

 lows : the two component beams reflected from the mirror 



