102 Barrett, Brown & Hadfield — On the Electrical Conductivity and 



resistances R by means of the terminal L. The resistance coils, marked 1 to 6 

 in the above fig. 7, had values of 41, 10, 5, 3'8, 0'94, and 0-54: ohms, respectively, 

 and were connected with a series of mercury cups. When L was in the position 

 shown in the diagram, the whole of the resistances were in circuit; then, by 

 inserting the second terminal L' into the next mercury cup and removing L, the 

 magnetising current was suddenly increased without any interruption of the 

 circuit. Again, by putting L into the third mercury cup, and removing L', the 

 next step up in the magnetising current was made, and so on, till all the coils 

 were cut out and the maximum magnetising current reached ; the process was 

 then repeated backwards, stepping down till all the resistances were again in, 

 when the circuit was broken by removing L, The reversing key E, K was then 

 thrown over, and a second series of steps up and of steps down were made; the 

 current was then once more reversed, and a third series of like steps made; this 

 completed the cycle for that one rod. Between each of the steps one observer 

 read the deflections on the magnetometer M, and another read the strength of the 

 magnetising current as measured by a pair of direct reading Weston ammeters, 

 and a Kelvin magneto-static amperemeter, one of which is shown diagram matically 

 at G. One of the Weston instruments was graduated to read 0*001 ampere per 

 division, and the other to read 0"05 ampere per division, while the magneto-static 

 read 0"2 ampfere per division. These three instruments thus gave us a continuous 

 scale measuring from 0-001 amjjfere up to 18 ampferes, the Weston ammeters 

 being switched out whenever the current became too large for their respective 

 ranges. Before use, these instruments were carefully tested by means of a Kelvin 

 standard ampfere balance. 



In order to reduce the rod under test to a non-magnetic state before putting it 

 through the cycle, the reversing key was rapidly oscillated by one observer, 

 whilst another (by shifting the terminals L L' up and down tlie mercury cups) 

 threw a gradually increasing or decreasing magnetising current round the 

 solenoid ; by this means each test-rod could, in a very short time, be reduced to a 

 neutral state, as shown by the magnetometer needle returning exactly to its zero 

 position. Each rod that was tested was taken through a complete magnetic 

 cycle, with a maximum magnetising force of 45 C.G.S. units. This involved 

 a total of 36 steps in each experiment. 



The effect of the various series of steps will be better seen by referring to 

 fig. 8, where the axes OH and OB represent, respectively, the magnetising force 

 and the induction produced. The first series of steps up gives the initial curve 1, 

 starting from the origin ; the next series down gives the part 2 from the point 

 of greatest induction to where the curve meets the axis O B ; the circuit is now 

 broken and the current reversed ; the next series of steps up gives that part of 

 the curve marked 3, 3, cutting the axis H, and extending down to the other 



