550 CAPTAIN AHNKY AND MA.IOK-GKNKRAL FEHT1NG 



$ XXIX. Apparatus juiU ;i .l<//>teil. 



After many experiments, it was determined to fix a double-image prism behind the 

 collimating lens. This double-image prism of Iceland spar was made by Mr. A. HILGER, 

 with his usual ability, and was so adjusted that when the central half of the collimator 

 slit was used the two spectra, while of the same length, were separated by one-eighth 

 of an inch on the focussing screen, the ordinary camera lens being employed. The 

 reflecting apparatus was also slightly altered by substituting for the fixed reflector 

 a second right-angled prism attached to the card so as to reflect the light through 

 the second lens L, v . There was a great advantage in this, for with the fixed reflector 

 the colour patch travelled across that formed by the direct beam, and thus the same 

 parts of the image of the prism's face were not always superposed. The plan of 

 attaching the reflector to the slit card got over this difficulty, and rendered the 



measurements more accurate. 



Fig. 15. 



XXX. Adjustment of the Instrument. 



The adjustment of the instrument, when using the double-image prism, required 

 care, and the following plan was adopted. The whole slit of the collimator was 

 illuminated by light from the arc in which lithium and sodium were vaporised. The 

 two spectra now overlapped, since the separation of one-eighth of an inch was only 

 obtained when the slit was one-fourth of an inch in height. The bright lines of the 

 lithium in the two spectra were then made to coincide by turning the double-image 

 prism ; the central portion of the slit in the collimator was then used, and the slit in 

 the card passed through the two spectra. If the collimator slit was properly adjusted 

 m the vertical and a bright line in one spectrum traversed the centre of, say, the top 

 part of the aperture in the card, the same bright line in the other spectrum ought to 

 traverse the centre of the bottom part of the aperture. If this were not so the colli- 

 mator was readjusted, and the same operation gone through. To make doubly certain 

 that the adjustment was correct, the direct and reflected rays from different parts of 

 the continuous spectrum of the positive pole were made to form superposed patches 

 on white card, and shadows of a rod were cast by each so as to touch. The rotating 

 sector was placed in front of the brightest, and the illumination of the two equalised. 

 If the same aperture of sector equalised the illumination throughout the spectrum 

 the adjustment was considered as complete, if not, a new adjustment was made till 

 such was the case. It was found in practice that a very good adjustment could be 



