with the Aid of a Grating. 



413 



As the compensation for colour cannot here be perfect, the 

 two slit images obtained are very narrow, practically linear 



spectra, but they are strong enough to produce interferences 

 like the normal images of the slits with which they nearly 

 agree in position. Other very faint slit images also occur, 

 but they may be disregarded. The doubly-diffracted slit 

 images are often useful in the adjustments for interference. 



Among the normal slit images there are two, respectively 

 white and yellowish, which are remote from secondary 

 images. If these be placed in coincidence both horizontally 

 and vertically along EG, fig. 1, the observation along D(jt 

 through the telescope will show a magnificent display of 

 black, apparently confocal ellipses, with their axes respec- 

 tively horizontal and vertical, extending through the whole 

 width of the spectrum, from red to violet, with the Fraun- 

 hofer lines simultaneously in focus. The vertical axes are 

 not primarily dependent on diffraction, and therefore of 

 about the same angular length throughout ; the horizontal 

 axes, however, increase with the magnitude of the diffraction, 

 and hence these axes increase from violet to red, from the 

 first to the second and higher orders of spectra, and in 

 general as the grating space is smaller. It is not unusual 

 to obtain circles in some parts of the spectrum, since ellipses 

 which have long axes vertically in one extreme case, have 

 long axes horizontally in the other extreme case. The inter- 

 ference figure occurs simultaneously in all orders of spectra, 

 and it is interesting to note that even in the chromatic slit 

 images shown in fig. 2, needle-shaped vertical ellipses are 

 quite apparent. 



It is surprising that all these interferences may be obtained 

 with replica or film-gratings, though not, of course, so sharply 



