of Electrons, and the Spectrum of Canal Rays, 685 



latest paper. It must be remembered that the intensities 

 given for the various lines rest on eye estimations o£ the 

 blackening of the negative, and therefore great accuracy is 

 not claimed for them. Stark appears to base his hypothesis 

 mainly on the following facts : — 



(1) As regards intensity of the undisplaced line, there 

 is parallelism between it and that of the band spectrum ; 

 both are attributed to the residual gas, the undisplaced 

 line to slowly moving positive ions produced by collisions 

 between canal-ray ions and atoms, the band spectrum to 

 the atom at recombination (§ 11). 



(2) As regards intensity of the displaced line, there is 

 parallelism between the velocity of the canal-ray ions 

 and the temperature of a radiating gas (§ 12). 



(3) Between the undisplaced and displaced lines there 

 is a dark space, attributed by Stark to the supposed 

 small emission at small velocities of the canal-ray ion, 

 and not to a dearth of slowly moving canal-ray ions 

 (§14). 



In § 13 Stark assumes that the intensity E of the displaced 

 line is given by an equation of the form 



E = nJ(u 2 , X) 



where n is the number and v the velocity of canal-ray ions 



1 



i 





S i X 





^ * 







s 







(n) , ' 



/ v < 













It X 





" 



/ / v 







/ / J" 



s ' 





/ / 



\ 







\ 





/ s 





(J) j* 



^s£' 











emitting the line of wave-length X. He supposes (1) n to 

 be given in terms of v by the curve (n) in the diagram ; (2) 

 J to be given by the curve J, from the parallelism between 

 velocity and temperature. Then we get the curve (E), 

 which is of the form required by experiment. 



The first supposition requires that canal-ray ions exist 

 in appreciable numbers for all velocities between the small 

 ones corresponding to the undisplaced line up to velocities 

 of order 10 7 cm./sec. belonging to the displaced line. We 

 must enquire how far this supposition is legitimate. Eor 

 this purpose it is necessary to consider the generation of the 

 canal-ray ions in the vacuum-tube. 



