M. E. Goldstein on Spectra of Gases. 337 



similar observations that the discharge is at the same time 

 every where the same throughout the whole circuit, and that 

 therefore it cannot be continuous in one part and discontinuous 

 in another. 



I now turn to the spectra we observe in these different kinds 

 of discharges. 



I must first mention an observation regarding a change in 

 the discharge without the aid of any of the means mentioned 

 above. 



The tube filled with carbonic oxide, which has already been 

 mentioned, showed in the beginning stratifications with a dark 

 space and feeble light. The discharge was widened into a 

 continuous band in the rotating mirror. 



The spectrum of carbonic oxide consists of a well-known 

 group of bands. After a short time the phenomena character- 

 istic of the continuous discharge disappeared, and the rotating 

 mirror showed distinct images of the tube indicating a disrup- 

 tive discharge. Nevertheless the spectrum consisted of the pre- 

 viously mentioned bands without the addition of lines. 



Tubes filled with nitrogen show generally the band-spectrum. 

 When the discharge was made disruptive by means of the car- 

 bonic-oxide tube, the spectrum of nitrogen remained the same, 

 only showing a variation of intensity. 



A tube filled with nitrogen and another filled with hydrogen 

 were introduced into the same circuit. The spectrum of nitro- 

 gen consisted of bands ; the spectrum of hydrogen showed Ha, 

 H/S, H7. If Wullner's hypothesis were correct, such a result 

 would be impossible ; both tubes, according to Wullner, should 

 give spectra of the same order. The images of both tubes in 

 the rotating mirror consisted of a continuous field of light. 



If one of the electrodes of a vacuum-tube has only a small 

 surface, or if it is contained in a narrow tube, the spectrum may 

 be different, according to the direction of the current. If the 

 large electrode is positive, the discharge consists of a greater or 

 smaller number of single sparks ; if the current goes in the other 

 direction, the discharge is continuous. 



If we introduce such a tube into the circuit together with an- 

 other tube, the discharge, as has been said, is continuous or dis- 

 ruptive according to the direction of the current, yet the spec- 

 trum of the other tube is unaltered. It often happens that in 

 one and the same tube both spectra are seen at the same time, 

 the narrow part of the tube showing the line- spectrum, while the 

 wider parts of the tube show the band-spectrum ; yet the dis- 

 charge has every where the same character. 



If the gas in a tube is exhausted to very small fractions of a 

 millimetre pressure, the discharge consists of a number of sparks 



