LYMAN. — ON THE PROLONGATION OP SPECTRAL LINES. 35 



is no longer symmetrical with respect to the slit, the arms have 

 broadened. 



The horizontal band which does not turn as the aperture is rotated 

 may be due to either of two causes, or to a combination of them. The 

 subordinate or secondary maxima which accompany every line in the 

 spectrum furnish the first reason, irregularities in the grating ruling 

 furnish the second. The author has shown that these irregularities 

 may not only furnish a background, but, under favorable circumstances, 

 may even produce sharp reproductions of real lines. The background 

 or nearly continuous baml can be noticed with almost every grating, and 

 can be best observed in that portion of the extreme ultra-violet where 

 no real lines are obtained. Its intensity varies greatly with different 

 gratings. In investigations where long exposures are necessary it often 

 proves very inconvenient, for faint real lines are much obscured by its 

 presence. 



In many cases the horizontal band due to diffraction through a rec- 

 tangular opening is much stronger and more troublesome than the band 



due to irregular ruling. In this case there is a 

 remedy at hand. The ends of the ruled space 

 may be covered with slanting pieces of blank 

 paper and the rectangular ruled space thus con- 

 verted into a parallelogram. The effect of this 

 Figure A. arrangement is to revolve the horizontal streamer 



due to the shape of the opening, the vertical 

 streamer remaining fixed, with the result that the background of the 

 spectrum is very materially cleared. The author can recommend this 

 device to all those who investigate faint spectral lines, and to whom a 

 clear field is a necessity. 



Jefferson Physical Laboratory, 

 Harvard University. 



