1894.] iipon the Dimensions and Volume of Iron Rings. 



FIG. l. 



95 



It need hardly be said that the experimental difficulties in the way 

 of determining to a ten-millionth part the changes which took place 

 in a length of less than l in. were very considerable. 



The annealed ring will, as before, be distinguished as Ring I, and 

 the hardened one as Ring IT. 



Table I. 



The changes observed in the widths of the two rings (transversely 

 to the magnetisation) are indicated in Table I and in the curves of 

 fig. 2. It will be seen that they are quite similar in the two cases, 



