A TEST FOR ELECTRON POLARIZATION 



479 



with the possible exception of the value for the 103 volt beam, all are 

 equal sensibly to zero. 



TABLE III 



The large values found for both (p +- a^ cos 0:2) and ai in the case 

 of the 103 volt beam are due, we believe, to some departure from the 

 usual conditions of the experiment which occurred while observations 

 on this beam were being made. Fewer measurements were made on 

 this beam than on any of the others, and the discordant values found 

 for its constants are traceable to an exceptionally low value — based on 

 eight readings only — which was obtained for D2{d = 90 degrees). The 

 discordance of this value with those obtained for the other deflections 

 is evident from the figures set down in the last column of Table IV. 

 These are the differences between the various deflections and the 

 mean of the deflections Di, D3 and D^. It will be noted that the 

 departure of the value obtained for Do from this mean is three times as 

 great as that for any of the others. The fact that this single unusual 

 departure is responsible for the exceptionally large value not only of 

 (p + 0-2 COS 0:2) but of ai as well, is reason, we think, for regarding it as 

 accidental. We believe, therefore, that we are justified in disregarding 

 the result obtained in this case, and in concluding from the values found 

 for (p +- a2 cos a2) for the other beams that the amplitude of the polar- 

 ization efi"ect is zero within the limits of uncertainty of our measure- 

 ments — that is, within about one half of one per cent. 



TABLE VI 

 103 Volt Beam 



Experiments designed to test for the polarization of electrons by 



