Mr. N. A. Kent on the Zeeman Effect. 279 



investigated show values of —jtt which are the same as those 



investigated in each set — an assumption which is certainty 

 not unjustifiable. 



3rd. That the average value of -^ —obtained from 



Preston's, Reese's, and my determinations — for the third set 

 in n=3, or J (12 + 10*7 + ll'l) = li'27, is so related to the 

 average value for n = 4= (as given in the same table, III.) in 

 the homologous sets of both subordinate series, or J (15*1 

 4- 14-9) = 15-00, that 



(^L 3 : G?fiL4 ::3:i; aS i (ll-*7)=8-7», while 

 £(15*00) = 3*75. Thus, if we assume — „. proportional to — 



w 



here " e " is the amount of electricity carried by the particle 

 of mass " m," we may say that the ratio of the charge to the 

 mass of the particle varies directly with " n" for the third set 

 of lines in the second subordinate series where " n" has either 

 the value 3 or 4. 



(The first subordinate series shows no lines for n = 3. The 

 value 15*1 was averaged with 14*9 because of the probability 

 of obtaining thus a more correct mean.) 



4th. That there appears to be no relation between sets 1 and 

 2, ?i = 4, and 1 and 2, n = 3, or, taking total mean values, 

 between 



16-9, 15-0 for n=d, and 



5-0, 8'6 for n=4. 



However, inasmuch as the wave-lengths of the three lines 

 forming the spectroscopic triplet are so related to each other 

 that, given the wave-length of one of them, those or! the other 

 two can be calculated — thus forming a connexion between the 

 members of the triplet — we would not expect that here again 

 a relation would appear between the first, second, and third 

 lines of the triplet in any one series unless, indeed, it were a 

 relation equivalent to that just mentioned, namely, the 

 possibility of the calculation of the wave-lengths of any two 

 lines of the triplet given that of the third. 



U2 



