DESCRIPTION OF APPARATUS AND METHODS. I9 



from the mirror also forms an image of the sHt upon the sHt. Since 

 the slit lies in the axis of the collimating mirror, the prism and its 

 mirror are truly vertical and parallel only when they reflect the image 

 of the slit back upon the slit. 



The angle m (fig. 2) between the mirror and the prism face was deter- 

 mined by reflecting the beam from the prism face back upon its path, 

 thus forming an image upon the slit (or, better still, upon a fine line on 

 a piece of paper just above it) and then noting the angular rotation of 

 the prism table necessary to reflect the slit image upon the mirror back 



Fig. 2. 



upon the slit. By this method the variation in the settings was only 

 from 5" to 6" from the mean. 



To determine the " zero setting " of the instrument it is necessary to 

 determine the angle p (fig. 2), at which the incident rays from the colli- 

 mating mirror must fall upon the prism mirror, in order that the rays 

 reflected from it will enter the prism and emerge from it at minimum 

 deviation. If we know this angle for the sodium lines, after fixing the 

 spectrometer arms and rotating the prism table, then all the succeeding 

 wave-lengths will pass through at minimum deviation. 



