208 Prof. J. S. Townsend 



on 



resemble the X rays in not being deflected by a magnetic 

 field. 



This hypothesis seems to be in accordance with most of the 

 experiments, but as there is a good deal of uncertainty as to 

 the way in which gases are ionized by X-rays, it can only be 

 ■considered as being adopted as a temporary explanation. 



[Professor Millikan in quoting these experiments has inad- 

 vertently stated that ions with double charges are produced 

 when the secondary rays come from a freshly cleaned 

 surface.] 



In Mr. Haselfoot's experiments which were made on the 

 same lines as the above, the ions being produced by a radio- 

 active substance, rather more experimental difficulties were 

 encountered, and a value of N x <?=1*26 x 10 10 was found 

 for positive ions when a force of 2 volts per centimetre was 

 used, and 1*37 x 10 10 with a force of 4 volts per centimetre, 

 thus giving a 9 per cent, difference, which according to 

 Professor Millikan shows that the method is inconsistent. 

 A reference to the curve for deducing the values of N x e 

 from the experiments shows that an error of 1 per cent, in 

 the determination of the ratio of two charges gives rise to an 

 error of about 2 per cent, in Nxe when a force of 4 volts 

 per centimetre is used, and 1 per cent, when a force of 2 volts 

 per centimetre is used, also in the previous case a larger 

 experimental error is liable to come in. The probable error 

 with 4 volts per centimetre is about 4 times as great as that 

 with 2 volts per centimetre. Mr. Haselfoot calls attention to 

 the probable inaccuracy of the number 1*37, and points out 

 that 1*26 is the most reliable number. 



The objection therefore that has been raised to the method 

 is one that could be raised to any form of investigation, — 

 namely that under some conditions the accuracy is not high. 



What is of importance is that the accuracy is as high as 

 can be expected from the theory as shown by the determi- 

 nations with negative ions, with a suitable range of forces; 

 and it ought to be possible to obtain the values of X X e to 

 an accuracy of 2 or 3 per cent. With the negative ions 

 generated from molecules by Rontgen rays it has been shown 

 that in all cases they have the same charge, which is con- 

 firmed by the experiments on the effects produced by collisions 

 from which it is obvious that all negative ions must be pre- 

 cisely the same. It is extremely difficult to believe that the 

 method fails when positive ions are investigated. 



It may be pointed out that these experiments also brought 

 to light a large effect due to moisture, which was subsequently 

 confirmed when the velocities under an electromotive force 



