﻿Ionization by Alpha Rays. 

 Table XXIII. 



477 



I 



1 



Aluminium. 



1 

 Gold. 



y- 



P. 



n i 



»i 



Error. 



P. 



n x 



»i 



Error. 



1021 



•1265 



(oalc). 

 •0822 



(corr.). 

 •0863 



-2-3 % 



•1338 



(calc). 



■0878 



(corr.). 

 •0975 



-2-8 % 



71-54 ... 



•1361 



•0799 



•0839 



-3 9 



•1475 



•0909 



•0990 



-1-6 



511 ... | 



■1504 

 •1638 



•0870 

 •0896 



•0930 

 •0896 



+2-5 

 +0-2 



•1748 



•0956 



■0956 



-3-3 



41-48 ... 



•1645 



•0851 



•0900 



4-0-5 











: 40-8S ... 





... 







•1849 



•0970 



•1064 



4-2-8 



31-11 [ 



•2161 

 •1993 



•0867 



•0875 



•0930 

 •0871 



-fl-8 

 -11 



•2273 



•0929 



•0960 



-2-3 



i 2074 ... 









•3726 



•1123 



•1165 



+41 



1037 ... 



i 



1-588 '0750 



1 



■0796 



-06 



1-878 



•1610 



•1167 



4-2-7 



M 



[ean '0892 



Mean -1013 



n x (co 

 ri 



rr.) obs 



erved 



•0937 1 

 •0742 j 



ratio 

 126 





n x (con 



n' „ 



■•) obsei 



ved -1040 1 ratio 

 •0819 J 1-27 



be noted that the apparently large divergence for the smallest 

 value of y has really no special significance. At the foot of 

 the table are given the values of n x and of n' determined 

 experimentally, reduced to the same date as the values of n 1 

 in the table. It will be observed that in both cases the 

 calculated value of n x is slightly less than the observed, and 

 that in both cases the ratio of the observed value of n x to 

 that of n is the same. 



Discussion of the Results. 



15. (1) From the identity for gold and for aluminium of 

 the ratio just mentioned, the conclusion may be drawn once 

 more that the number of delta rays excited by a given stream 

 of a rays is the same whether they fall on gold or aluminium, 



(2) The calculations have been carried out in such a 

 manner as to throw all errors, whether of experiment or of 

 theory, on the quantity n x . Accordingly, the agreement of 

 the calculated values n x among themselves and the agreement 

 of the mean with the observed mean is sufficiently satisfactory 

 to afford very convincing proof of the substantial truth of 

 the theory and of the accuracy of the experiments. The 

 difference between the various values of n x is almost certainly 

 due to experimental error, for if the values of P from Table 

 XXI. which are calculated with the observed and not the 

 calculated values of jjl are taken, this difference is very much 



