46 



PROCEEDINGS OP THE AMERICAN ACADEMY. 



The pairs of sharp lines iu the immediate neighborhood of the charac- 

 teristic groups to which the treatment has been confined are reproductions 

 of the real lines at 2936.8 and 2928.9. They form the 47th, 46th, 24th, 

 and 23d spectra of the system iu which the real lines are the 7()th 

 spectra. The single line near the middle of the plate is the 23d false 

 spectrum of the real line at 4481.3. The lines not far from it are the 

 23d false spectrum corresponding to the real lines 3838.5, 3832.5, and 

 3829.5. The 24tli spectra of all these lines are too feeble to reproduce. 

 Their 46th and 47th spectra occur among real lines in the region between 

 3050 and 2500. The apparent wave-lengths of these false lines are 

 marked upon the plate ; the decimal place is omitted. The following 

 table gives the measured apparent wave-lengths of all the lines appear- 

 ing upon the plate, their calculated wave-lengths, their order as false 

 spectra, and the wave-lengths of the real lines of which they are the 

 reproductions. 



It is well to remember that this grating is not unique, but that most 

 reflection gratings produce false spectra. It is in the extreme ultra- 

 violet that these false lines are most easily seen, and it is in this region 

 that they may be most readily taken for real lines. The lines may 

 serve to show by their strength and sharpness the danger which they 

 offer in spectroscopic work. 



In conclusion, it may be well to repeat the chief result of this paper. 

 False spectra differ fundamentally in character from the commonly ob- 



