LYMAN. — THE FALSE SPECTRA FROM DIFFRACTION GRATINGS. 45 



by plane reflection gratings as well. The author has examined two 

 instruments ruled upon speculum metal by Rowland's engine. In the 

 experiment the light was collimated and brought to a focus by quartz 

 lenses. In order to reduce the phenomena to the simplest possible form, 

 a line in the visible blue spectrum of magnesium was separated out by a 

 prism spectroscope and thrown upon the slit of the collimator. The 

 camera was focussed on the image of this line in the first spectrum. 

 The first grating was ruled in 1883 and had 14,438 lines to the inch. 

 The image of the line used appeared in its proper place in the first 



o 



spectrum — = 4481 Angstrom units — and was not accompanied by any 

 ghosts of the common kind- That is to say, there was no doubling of 

 the line, nor were there any faint reproductions very near it. In the 

 region near wave-length 3000, however, four sharp lines occurred ; and 

 again near wave-length 1550 four more reproductions were present. 

 Thus this grating produces eight false spectra of a lower order than the 

 first, corresponding to the real line 4481. Besides these eight, numerous 

 faint reproductions may be detected, but they are of extremely feeble 

 intensity. 



The second grating examined was ruled in 1887 and had also 14,438 

 lines to the inch. The spectrum obtained with it, however, was very 

 different from that given by the first instrument. The first spectrum of 

 the line 4481 was, as before, sharp and without ghosts, but the eight 

 distinct false spectra were rejdaced by at least seventy reproductions 

 of very feeble, but nearly equal, intensity. These extended between 

 positions corresponding to wave-lengths 2000 and 900. 



The results obtained with these two plane gratings are exactly similar 

 in character to those obtained with the concave "ratings called No. I and 

 No. II and recorded in a previous paper. These gratings seem to belong 

 to two types, the one in which the false spectra are all of nearly equal 

 intensity and feeble, the other in which some few of the false spectra are 

 many times more intense than the others. In the one case the grating 

 gives a background of faint lines ; in the other sharp, strong false spectra 

 are present. 



The author wishes to call attention to the plate which accompanies 

 this article. It is from a concave grating of 6-foot radius and shows the 

 false lines whose positions have been discussed in this paper. The plate 

 is taken directly from a negative by photographic process. The two 

 groups at positions corresponding to wave-lengths 956 and 1834 are very 

 faint in this reproduction. Their positions are indicated, however. The 

 character and dispersions of the two stronger groups is well shown. All 

 the lines in this plate are false. 



