MOUNT WILSON OBSERVATORY. 247 



in absorption but not in emission. The effect is less decided for lines of the 

 type of those in the iron spectrum. 



Production of Absorption Spectra. 



In the general work with the electric furnace, three methods for the pro- 

 duction of absorption spectra have been developed: (1) a diaphragm may be 

 placed in the furnace tube, and the spectrum of the vapor reversed by the 

 incandescent background; (2) light from a high-power tungsten lamp may 

 be passed through the furnace vapor heated to various temperatures; (3) 

 wire explosions, according to the method of Anderson, may be used as the 

 continuous source. By the first method, temperatures of the diaphragm 

 above 3200° C. are obtainable, the column of absorbing vapor being heated 

 in proportion. The second method admits of independent heating of the 

 source and the absorbing vapor, and when a quartz window is employed has 

 proved very effective in the region of short wave-length. The tungsten 

 filament heated to 2900° C. builds up a continuous spectrum extending well 

 into the ultra-violet, while the absorbing vapor may be left at a much lower 

 temperature and permits the selection of low-temperature lines in this region. 

 The use of wire explosions, when the material of the wire does not intro- 

 duce disturbing fines, supplements the second method, and on low-dispersion 

 spectrograms has given the shortest wave-lengths yet obtained for the furnace 

 absorption spectra. 



The study of the absorption spectrum of iron in the ultra-violet by the three 

 methods illustrates the useful features of each. By the first method, absorp- 

 tion fines of iron have been obtained as far as X2447, but necessarily only 

 for high temperatures of the absorbing vapor. The second method reaches 

 about the same limit, but the furnace lines which are absorbed at lower 

 temperatures may be studied. With the third method, the absorption of 

 iron vapor at a temperature as low as 1600° C. can be followed to X2298. 



The possibility of observing absorption lines in the ultra-violet for low 

 furnace temperatures is a distinct addition to our means of studying such 

 lines; for the emission spectrum of the furnace extends only as far as the 

 continuous spectrum of a black body at the same temperature, which is only 

 to about X3500 for the low-temperature stage usuaUy employed. If the con- 

 tinuous spectrum is produced by a very hot external source, lines absorbed 

 by a vapor in the low-temperature furnace can be observed at least 1200 a 

 beyond the point at which the furnace at that temperature can emit them. 

 Since many important series fines occur in this region, the ability to classify 

 them on the same basis as those of the visible spectrum is of much assistance. 



Ultra- Violet Spectrum of Iron. 



A list has been published which contains 892 iron lines between X2298 and 

 X3878, classified according to their intensities in the arc and at three differ- 

 ent furnace temperatures. The methods of producing absorption spectra, 

 described in the preceding section, have furnished many additional data for 

 this region, especially as regards the selection of low-temperature lines, and 

 have afiowed some extension of the ultra-violet limit of the investigation. 

 An additional list of 262 fines between X3884 and X4531 supplements the 

 results for this region published in 1913, and gives the classification of some 

 fines previously omitted, as well as some others which have been revised in 

 accordance with recent experimental results. 



