WEBER OS THE NATURE OF DIAM AGNETISM. 



483 



again inserted, the commutator again changed, and the bar 

 of bismuth withdrawn, &c. During this experiment, con- 

 tinued for about 1 minute, the state of the galvanometer was read 

 off at intervals of about 10 seconds. 



A second series of experiments was now made, but with this 

 difference, that the commutator assumed that position on with- 

 drawing the bar of bismuth which it had occupied in the first 

 series on inserting the bismuth, and vice versa. 



The third series was an accurate repetition of the first, and so 

 forth. iJBifckj"/") •i9flhut on &->i!;;po't .•luisrnofiii/jisg ovil 



Previous to commencing each series the state of the galvano- 

 meter was observed, without however waiting until the needle 

 had attained a perfect state of rest. Each series was begun by 

 withdrawing the bismuth. 



In the following table the whole of the readino;s made on the 

 galvanometer are an'anged together. The different series are 

 distinguished by Roman numbers ; the two states of the com- 

 mutator which occurred in the different series on the withdrawal 

 of the bar of bismuth are distinguished in the heading by A and 

 B. The state of the galvanometer before commencing each 

 series is also noticed in the headins:. 



Now if we compare the states of the galvanometer in the odd 

 alternate series, where the commutator occupied the position A 

 on withdrawing the bismuth from the annular conductor, with 

 the mean value in the bottom-line, it is seen that the latter is 

 always somewhat greater. For instance, the mean values ar^— . 



1. Series 514-92 = 51 2-3 + 2'62 



3. „ 518'72 = 515-9 + 2-82 

 •Si „ 522-00 = 51 7-0 + 5-00 

 "'t*'"'',, 528-87 = 524-7 + 4-17 



2 K 2 



