258 CARNEGIE INSTITUTION OF WASHINGTON. 



Supplementary photographs of the iron and manganese spectra have 

 been made, especially in the infra-red. In the visible region, long expo- 

 sures at low temperatures were used to determine the lines first radiated 

 by the vapor. For this work, current from the new 500-k. w. generator 

 was frequently used on account of the close adjustment of the voltage 

 which the control system of the generator permits. The photographs 

 of the barium furnace spectrum already supplied to Professor F. A. 

 Saunders for use in tracing the various series of this element has been 

 supplemented by photographs of selected regions taken with an iron 

 comparison spectrum. A further study of the absorption spectra of 

 calcium and iron was made by placing a plug at different points in the 

 furnace-tube and observing the effects of balancing emission and absorp- 

 tion, which often caused lines to become invisible in the photographs, 

 although they were strongly radiated by the vapor. 



An investigation was made of the possible effect of a potential differ- 

 ence at the ends of the furnace-tube filled with ionized vapor, the ques- 

 tion having been raised as to whether the sensitive high-temperature 

 lines (class III) require conditions equivalent to those of a low-tension 

 arc instead of simply a higher temperature. The method used was to 

 reduce or eliminate the potential difference while the high temperature 

 was maintained, and compare the spectrum with that emitted at the 

 same temperature with the usual operation of the furnace. A reduc- 

 tion of the potential to which the vapor was subjected was produced by 

 vaporizing iron in an insulated tube inside the tube carrying the heating 

 cmTent, and also by the use of direct current from the new generator, 

 by means of which the momentary high values of the potential occur- 

 ring with an alternating current were avoided. By the latter method 

 the test spectra of high-temperature Unes were produced with a po- 

 tential difference of only 0.5 volt per centimeter. Spectrograms 

 were also made with no cun-ent in the tube. The tube, protected 

 by a jacket, was heated by the current, which was then broken, the 

 spectrum being photographed as the tube slowly cooled. This method 

 with no current was supplemented by experiments in which an insu- 

 lated tube was heated by an external arc. In both cases the spec- 

 trum was the same as that of the regular resistance furnace at the same 

 temperature. While it is obvious that some current may be conducted 

 by the ionized vapor of the resistance furnace, the experiments have 

 shown that for the temperatures used in the classification of Unes the 

 low potentials required are ineffective in producing any change in the 

 spectrum. 



The production of the ''cyanogen" band at X3883 in absorption under 

 high dispersion was undertaken by Mr. King to supply data for the 

 work of Mr. St. John on the wave-lengths of the sokr lines belonging 

 to this band. Since unsynmietrical structure may influence the 

 measurements, it is desu-able that absorption lines from a laboratory 



