378 Mr. J. S. Ames on the Concave Grating 



until the flame and image come in line. The reflecting-prism 

 is now pnt in place so as to illuminate the entire grating, and 

 the slit opened. The spectrum formed at the camera-box is 

 observed by the eye, or thrown on a piece of paper ; and the 

 back-screw of the grating-holder is turned until it falls at the 

 right height. The camera-box is moved along its way, and 

 in general the spectrum rises or falls ; and this is corrected by 

 the side-screw of the grating-holder. These two adjustments 

 are repeated many times until the spectrum stays in place 

 however the camera-box is moved. Then the slit is narrowed, 

 and revolved until the best definition is secured. The instru- 

 ment now should be in perfect adjustment ; and to test this 

 an exposed photographic plate, of which the emulsion has 

 been partly scraped, giving it a lattice-work appearance, is 

 put in the camera-box, emulsion side toward the grating. 

 The spectrum formed on the plate and the emulsion itself 

 ought now to be in focus at the same time in all orders of 

 spectra ; that is, if the plate is observed with an eyepiece, 

 there should be no parallax between the two. In general, 

 further adjustment is found necessary. It was to this end 

 that the theory of errors, as above given, was deduced. Let 

 the camera-box be placed in focus when it is near the slit ; 

 and then, as it is moved away from it, suppose the parallax 

 increases proportionally to the distance along the way. This 

 w^ould lead one to think that the two beams were not exactly 

 at right angles. Similarly for the other displacements. It is 

 found in practice that it does most good to turn the grating- 

 holder slightly around its vertical axis. 



If, in setting up the instrument, a micrometer-eyepiece is 

 used instead of a camera, practically the same adjustments are 

 found necessary. 



Use of Instrument. 



Gratings in Practice. — Special gratings should be selected 

 for special purposes. Every grating has spectra of different 

 brightness on the two sides ; and one should be used which is 

 bright in the particular spectra desired. But more than this, 

 even if the red of any one spectrum is bright, the violet may 

 not be. This fact must be especially noted in working beyond 

 the visible spectrum. Further the various parts of the grating, 

 especially if it is concave, may give spectra of varying 

 brightness. For instance, the second spectrum may be uni- 

 formly bright for all parts of the grating, while one end of the 

 grating may give a bright third spectrum and the other end 

 a faint one. This fact may be brought out by viewing the 

 grating directly with the e}'e. It is only when extreme 



