produced in Soft Iron by Magnetism. 501 



Here must be noted a peculiarity which was observed in 

 every bar that I experimented with. 



If the current be gradually increased from zero up, at a 

 certain point we reach a maximum amount of expansion: now 

 a further increase in current will produce a decrease from 

 that maximum ; but if, on the other hand, we gradually 

 decrease the current, we are able to obtain a still greater 

 expansion. 



This increase in the case of soft iron is from "05 to *20 of 

 the total amount of expansion otherwise obtained. 



Mr. Bidwell has noted a peculiarity in the behaviour of a 

 steel bar which will come under this head. At lower points 

 in the expansion this peculiarity does not seem to exist. I 

 have plotted (fig. 3) from the above table columns I., 



Fisr. 3. 



III., and IV., which give a graphical illustration of this 

 phenomenon. 



I have adopted, as the method which gives the most uniform 

 and certain results, that of quickly but gradually increasing 

 my currents (not using the first contact), counting my fringes 

 from the zero point for my lighter currents and from the 

 turning, or maximum, point for my heavier currents. For 

 attaining this end I constructed a resistance out of fine 

 platinum wire placed in water to prevent fusing, with a 

 sliding metallic contact. By using this resistance directly, and 

 as a shunt for the smaller currents, I have been able to have 

 the current perfectly under control. 



The following tables give the detail results for the different 

 bars. Column I. gives the total, amount of expansion in 

 fringes. Column II. gives the corresponding currents. 

 Column III. gives the intensity of the field computed by the 



