100 



SCIENCE. 



[N. S. Vol XII. No. 290. 



will take its place. Such is, and such 

 ought to be, the life history of all scientific 

 theory. 



The more promising a new theory ap- 

 pears, the more is it deserving of a careful 

 and critical scrutiny, both from its adher- 

 ents and from its opponents. The hypothe- 

 sis of electrified corpuscles, which is in- 

 volved in the modified Crookes theory, has 

 proved its right to a hearing. It now has 

 a right to demand the severest of friendly 

 criticism. An elaborate critical discussion 

 of the theory would be out of place in an 

 address of this kind, even if sufficient time 

 for the purpose were available. I wish, 

 however, to call attention briefly to some 

 points in connection with the subject which 

 I think have not previously received the 

 attention that they deserve. 



Let us compare, for example, the values 

 of e/m determined by different observers. 

 The discrepancies between the values ob- 

 tained by Wiechert and by J. J. Thomson 

 is not surprising, since they were the first 

 determinations of this kind that had been 

 made. As a preliminary test of the theory, 

 the fact that results obtained by such 

 widely different methods were of the same 

 order of magnitude is eminently satisfac- 

 tory. A number of new determinations 

 have been made, however, during the past 

 two years. Since the more recent determi- 

 nations were undertaken with a full under- 

 standing of the necessary experimental 

 precautions, we should expect a close agree- 

 ment among their results. But discrepan- 

 cies of considerable magnitude still remain. 

 It appears to me that the variation in the 

 values of e/m obtained by different observ- 

 ers is greater than can be accounted for by 

 experimental errors. To bring out this 

 point, and in the hope of getting some idea of 

 where the cause of the discrepancy is to be 

 sought, I have prepared the following table, 

 which contains practically all the values of 

 e/m that have been obtained by experi- 



ments upon the kathode rays. Some of the 

 values obtained by other methods have also 

 been added for comparison. 



The values of ejm are arranged in groups 

 according to the method by which they 

 were determined. The results of the most 

 recent experiments, and presumably, there- 

 fore, the most accurate ones, are in each 

 case placed last. 



Leaving aside the results of Schuster and 

 Wien and the first results of Wiechert, all 

 of which were obtained by experiments of 

 a purely preliminary character, we see that 

 the results obtained by different observers 

 show a satisfactory agreement, provided 

 that the same method was used. Compare, 

 for example, the two results of Kaufmann, 

 obtained by dififerent modifications of the 

 same method, with that obtained by Simon. 

 A more satisfactory agreement could scarce- 

 ly be desired. Similarly, the values ob- 

 tained by Lenard agree quite well with 

 those that were obtained by J. J. Thomson 

 when using the same method. But the 

 smallest value obtained by the first method 

 is twice as great as the largest value ob- 

 tained by the last method. The results ob- 

 tained by the second and third methods 

 agree fairly well with each other, and are 

 intermediate between the two extremes just 

 mentioned. Wiechert's later determina- 

 tions, however (Method III.), are subject 

 to a possible constant error, so that these 

 results must be regarded as uncertain.* The 

 third method is liable to experimental error 

 for several reasons, notably because its re- 

 sults are especially likely to be influenced 

 by the conductivity of the residual gas. 

 The eiiect of this source of error, as pointed 

 out by Thomson, would be to make the re- 

 sults larger than they should be. Objec- ' 

 tions might also be raised to the assump- 

 tions on which the method is based. On 

 the whole, it appears to me that the results 

 of the first and fourth methods are to be 



* WUd. Ann., 69, p. 739, 1899. 



