■i-C) On Gases under the Influence of the Electric Discharge. 



the other connected by a discharge-tube and galvanometer 

 with the earth. 



The deflection amounted to 145 millims., with a pressure of 

 about 10 millims. in the discharge-tube; so that a deflection 

 of about 1*5 millim. would correspond to a discharge. The 

 quantity of electricity given by each discharge would therefore 

 be sufficient to convert the band spectrum into the line spec- 

 trum in a tube of 1 to 2 millims. width (p. 409). But if we 

 have even a very powerful excitation of the coil, and replace 

 the metallic contacts by a mercury break, we shall hardly 

 succeed in bringing to bear on unit surface in a tube 10 times 

 as wide, and consequently of 100 times the section, so much 

 electricity that the same change would take place. This only 

 takes place when we employ Leyden jars; and then the con- 

 ditions become so complicated that their discussion has hitherto 

 been impossible. 



Wullner*, from a series of careful experiments, has urged 

 objections against the above view of the origin of spectra, and 

 has employed the spectrum of nitrogen as his starting-point. 

 As he had not these separate measurements of the influence 

 of the thickness of the layer, of the intensity of the discharges, 

 of the pressure, and of the quantity of energy given off, it is 

 not possible to follow the discussion of his experiments into 

 details. The appearance of single nitrogen-lines together with 

 the bands may be explained by the reasons given above. An 

 objection formerly raised by me against Wiillner's theory has 

 been set aside by recent experiments of Dewar and Scottf . 

 But that with hydrogen a change of line spectrum into band 

 spectrum cannot be brought about by change of the thickness 

 of the radiating layer may be seen from the remark of Lockyer, 

 that a layer of incandescent hydrogen only a fraction of a 

 millimetre thick in a Geissler's tube shows the same bright 

 lines as the layer in the sun's photosphere thousands of 

 kilometres in thickness. Moreover the gases would have to 

 possess an enormous absorption-coefficient for the luminous 

 rays corresponding to the bands, whereas exactly the opposite 

 follows from the experiments of GouyJ. Moreover Wiillner 

 has always taken the change of temperature into account. 



My next problem will be to examine further the region 

 sketched out in this research — namely, to determine the rela- 

 tion between the intensity of the separate lines of the spectrum, 

 the quantities of electricity, and the quantities of energy given 



* Wullner, Wied. Ann. viii. p. 590 (1879). 

 t Dewar and Scott, Beibl. iv. p. 309 (1880). 



t Gouy, Beibl. ii. pp. 349, 411 (1878), iii. p. 611 (1879), iv. p. 476 

 (1830). * 



