Trowbridge — Spectra of Gases. 425 



trum was seen in the capillary ; moreover, this continuous 

 spectrum was crossed by a dark line which resulted from the 

 absorption of heat in the colder layers of gas in the larger 

 portions of the tube. 



The broadening of the spectra of the vapors of metals which 

 I have observed in capillary tubes has thus its analogy in the 

 case of gaseous spectra. 



Having obtained reversals of the spectra of metallic vapors 

 under new conditions, I was naturally interested in the experi- 

 ment of Liveing and Dewar, especially since a controversy had 

 arisen between M. Cantor and E. Pringsheim in regard to the 

 possibility of the reversal of gaseous lines in Geissler tubes. 

 M. Cantor* concluded from his experiments that such reversals 

 do not occur in the phenomena of luminescence, such as one 

 obtains by the discharges of electricity in Geissler tubes. 

 Pringsheim objected to these conclusions on the ground that 

 Cantor did not observe a sufficiently narrow portion of the 

 spectrum of the gas and did not use sufficient dispersion. 

 Pringsheimf quotes the results of Liveing and Dewar in sup- 

 port of his position. 



In repeating Liveing and Dewar' s experiment, it occurred to 

 me that objection might be brought against it on the ground 

 that it was a spark discharge and not a clearly marked glow or 

 luminiscent discharge such as Cantor evidently had in mind. 

 I, therefore, placed a second spark gap (fig. 3, S) just outside 

 the inner capillary of the large Geissler tube provided with an 



s 



X 



r 



\/ 



inner capillary, as I have previously described in speaking of 

 the temperature inside a capillary and in the space just outside. 

 The discharge passed first through the capillary and then by 

 means of an outside connection through the second spark gap ; 

 thus the light from the capillary passed through the light from 

 the second spark gap. In both cases the light was a glow or 

 luminescence and not a white spark discharge, the pressure in 

 the tube being from one to two centimeters. 



A Rowland grating was employed and an eye-piece was 

 fixed on the C line of hydrogen. The second spark gap gave 

 a fine bright line of the apparent length of the slit, the capil- 



* Ann. der Phys., 3, 462, 1900. 

 fAnn. der Phys., 5, 1900. 



