﻿to maintain a Current in a Gas at Low Pressures, 219 



will absolutely coincide, giving two solutions for each value 

 of the ordinate. The quantity (3 is known to be small in 

 comparison with a from other work ; consequently for each 

 large value of ad the corresponding value of fid can be read off 

 directly. Such a curve is given in fig. G. 



Fier. 6. 



0-30 

 t 



0-20 

 0-10 















\ 





























\___ 



l £ 3 4 J^ Sot 



It should be noted, however, that a small error in the 

 determination of the sparking potential will make a very 

 large variation in the value of {3, since the part of the curve 



where - is read off is so much steeper than the corresponding 



part for -. So this is not a good way to determine /3. 



Notwithstanding this fact, the results obtained seem to fall in 

 extremely well with the theory. In fig. 7 (p. 220) the heavy line 



gives the curves showing the relation of - and — for hydrogen, 



as determined by Professor Townsend. The two marked 

 points are the values determined for the two curves under 

 discussion, B and C, fig. I ; and the dotted line shows how 

 closely they coincide with a continuation of previous ob- 

 servation. The calculation is given in Table I. But it 

 should be noted that in the case of the larger values for /3, and 

 probably at the higher pressure too, it is very doubtful if the 

 equation from which these values have been derived can be 

 expected to hold very accurately. The assumption upon which 

 it is based is that, there is a large number of collisions 

 between the ions and the neutral molecules of the o- a s. But at 



