in the Ultra-red Region of the Spectrum of the Sun. 241 



the numbers a, b, and c directly to the position of that normal 

 in the three different orientations. The delineation (fig. 2) 

 refers to the third position, in which the normal to the grating 

 is defined by the number c, while to the incident ray E, which 

 in the first position coincided with the normal, the number a 

 still belongs. Hence the angle of incidence is 



B = a — c, 



if, as was the case with our arrangement, the mirror has to 

 be rotated from the first position to the third in the direction 

 of diminishing numbers. 



Since in the second position the direct reflection has the 

 same orientation as the ray A in the figure, the angle $i be- 

 tween A and E is equal to the angle between the direct reflec- 

 tion in the second position and the incident ray E. Now the 

 angle between the mirror-normal in the second position and 

 the ray E is equal to b — a, since the mirror must be turned 

 from the first to the second orientation in the direction in 

 which the numbers increase ; therefore the angle 6\ between 

 the direct reflection and E is 



Sj = 2(&-a). 



From the observation of the quantities a, b, and c we can 

 therefore calculate B and B 1} and consequently X also. 



III. Measuring. 



In order that the numbers a, b, c, read off" from the orien- 

 tation-circle may actually be connected with the angles 8 and 

 S x in the way above supposed, it is necessary that the plane of 

 incidence of the rays remain the same at all positions of the 

 grating, and parallel to the plane of the orientation-circle. 

 To effect this, first the spectrometer was set up, with the aid 

 of the adjusting-screws attached to its feet, so that the plane 

 of the orientation-circle was horizontal, which was rendered 

 visible by a spirit-level placed on it. Then the stand was 

 placed, by means of its special screws, so that its orientation 

 was horizontal, and kept that position while being rotated 

 360°. Thereupon the grating was mounted on the stand, and 

 at first in the position in which the incident rays were per- 

 pendicular to the plane of the grating. 



In order to be able to effect this arrangement with greater 

 exactness, instead of the pasteboard tube P x (fig. 1) a paste- 

 board disk with a rectangular aperture of 16 millim. height 

 and 3 millim. breadth was put before the aperture A in the 

 window-shutter. By this the incident pencil was made much 

 narrower, and hence a more accurate adjustment was rendered 

 possible. Moreover, in the slit s a cross of fine wire was 



