465 



On comparison of Rütti nauer's researches willi ours it appears 

 tliat we made use of comparatively narrow capillar}' tubes as circuit 

 of the current, the German investigator on the other hand of com- 

 paratively wide tubes. We derived, however, already before, that 

 two different formulae must be valid for these cases, and this as a 

 conserpience of the fact that in the case of wide tubes the laws of 

 PoisEUiLLE should be applied when taking the diffusion phenomena 

 into account, for narrow tubes those of Knudsen-Langmuir. Por in 



A Q^) 

 the first case the electric mass-transportation c^ may be put 



equal to : 



«.y(p,-p,)|/ J, 



in the second case to 



c, 



I p V M 



In the first case the theoretical formula for the pressure effect — 

 taking into account that Q=z — — for tubes with round cross-section 



— is ecpial to : 



AQ I ,— A I ,— 



p,-^p,^Lp=f,-^j-\/M=f\-.--VM. . . (7) 

 ap JJ* ap Q 



in the second case to : 



^ "^^ a JD' ''' a" D ^ ^ 



in which 



./i = ;= , /j =^ -=, a z=a .[ — ] and a' = a — 



c, 1/7' c, 1/7' \jrj :x 



When on grounds to be given later, the gradient g is taken 



inversely proportional to a, we may write equations 1 and 11 as 



follows: 



Lp = f^.^VM ...... (///) 



p Q 

 res p. 



I) In which Cj is a constant. Compare further Equation 9, p. 390, These Proc. 

 XXIII, No. 2 and 3. The factor Q has been introduced, because A now denotes 

 current density, in our former paper current intensity. 



3) Compare Equation 3, p. 382, These Proc. XXllI, No. 2 and 3, 1920. 



3) Compare Equation 1, p. 382, loc. cit. 



*) These equations have been obtained from the Equations 1 and 3, p. 382, 

 These Proc. XXI II, 2 and 3, after multiplication by \lp. This has been done on 

 the strength of what was said in footnote 3, p. 385 of our paper of 1920. 



30* 



