A WAVE-LENGTH COMPARATOR FOR STANDARDS OF LENGTH. 25 



means of the black-lacquered screws k, vertically over each other but to the left, 

 en echelon with B and C, the four images forming a rhombus. This arrangement 

 avoids all possibility of the spectra overlapping, as will be obvious from fig. 8, while 

 (/! corrects for the dispersion introduced by g a . The full aperture of the iris is just 

 sufficient, with wedge-discs of 35 minutes angle, to enable all four corners of the 

 rhombus to be included in the field. 



The image B should then be made to cover the image A to exactly the right extent, 

 for the production of interference bands which shall be precisely vertical in the field 

 of the micrometer eye-piece, and at the same time of the convenient width, equivalent 

 to somewhere between 100 and 200 drum-divisions of the micrometer. This is 

 achieved when the overlapping occurs to the extent and is of the character shown in 

 fig. 8, the displacement being only horizontal. The adjustment is now complete in 

 white light. 



The large total-reflection prism is then replaced by the constant-deviation prism, 

 already adjusted with vertical faces on its separate table, which only requires to be 

 dropped into position with its pegs in the holes of the circle plate, and locked there 

 by pulling round the little locking handle-pin projecting through a slot in the table. 

 The position of the circle on its transverse bed will require to be altered to the 

 necessary extent to bring the centre of the receiving face of the prism in the normal 

 line of the interference discs. The telescope may also have to be moved slightly 

 parallel to itself, by means of the right-and-left movement a at the lise of the pedestal, 

 to direct it to the proper part of the exit face of the prism. On setting the circle to 

 the reading which is known by preliminary work to approximately adjust the prism 

 for the production of bands of red hydrogen light, the four spectra should be seen in 

 the telescope, still using the common eye-piece, as shown by the horizontal lines in 

 fig. 8. 



The goniometer lamp is then removed and the hydrogen Geissler-tube fitting placed 

 in position. The common eye-piece is also removed and the micrometer eye-piece 

 substituted, but without the front lens-combination, leaving only the special single 

 lens in position between the micrometer and the iris diaphragm, in its draw-tube of 

 the correct length for focussing the silvered reference ring of the glass disc y 2 - On 

 actuating the Ruhmkorff coil, the four spectra will again be seen as in fig. 8, although 

 the magnification is not so great. But instead of consisting of continuous spectra, as 

 with white light, they will now show only a faint continuous spectrum, but at the 

 positions which the Ha, H/8, and Hy bright lines would occupy in a hydrogen 

 spectrum there are images of the rectangular signal-spot in red C-light, greenish-blue 

 F-light, and violet light (near G). Only one set of these images those in red hydrogen 

 light if the prism has been set for this will be in the field at once, the dispersion of 

 the constant-deviation prism separating the different sets very considerably, to the 

 extent of several field diameters. The iris diaphragm may now be closed to the 

 extent which is shown by the dotted semicircle in fig. 8, which is a faithful repre- 



VOL. CCX. A. E 



