254 Mr. S. Bidwell's Experiments illustrating 



the deflections to the right or left of the zero-point of the 

 scale, caused by exciting the magnet, being generally as much 

 as 15 or 20 centim.* 



A number of experiments were now made with pieces of 

 iron cut into various shapes, and occasionally with fine copper 

 wires arranged in different forms: and the conviction gradually 

 forced itself upon me that the effect in question was not due 

 to any such action of the magnet upon the current as that 

 which Mr. Hall supposed. 



The possibility of a thermoelectric effect suggested itself at 

 an early stage, the heat being derived from the current flow- 

 ing through the strip, and the changes in direction being 

 possibly produced by some effect of magnetization upon the 

 thermoelectric properties of the metal. Such effects, I 

 thought, could only occur at the points where the wires were 

 attached which went to the galvanometer. Disks of bismuth 

 and antimony were therefore inserted at the contacts of the 

 wires and the strip, and their relative disposition varied in 

 order that the thermoelectric effect, if any existed, might be 

 exaggerated. But it was found that, whether the disks were 

 of bismuth or of antimony, or of the two metals in various 

 combinations, the extent of the galvanometer-deflections was 

 unaffected, except in such a manner as could be perfectly 

 accounted for by ordinary thermoelectric action. The idea of 

 mechanical strain then occurred to me ; and I may mention 

 that this occurred to Mr. Hall, but only to be rejected by him 

 as utterly inadequate to account for the phenomenon. I 

 attached a string by means of sealing-wax across the middle 

 of a thin sheet of iron, which was cemented to glass and con- 

 nected with the battery and the galvanometer as usual. On 

 pulling the string in a transverse direction I found unmis- 

 takable indications of a galvanometer-deflection in the same 

 direction as if the sheet of iron had been acted upon, not by a 

 mechanical pull, but by the electromagnetic force. The effect 

 was small, but I have since greatly increased it by the follow- 

 ing device : — A strip .of thin iron was cemented between two 

 thin slips of deal about 20 centim. long, forming a kind of 

 sandwich. The sandwich was attached to a board by means 

 of four screws, the distance between the middle screws being 

 about 7 centim. The two ends of the iron were connected with 

 a battery, and the middle points of its opposite edges with 



* That the magnetizing current used by Hall was not absolutely con- 

 stant is evidenced by a statement in his paper in the ' Philosophical 

 Magazine ' for September 1881, that the readings of a tangent-galvano- 

 meter, placed in the circuit for a special purpose, "slowly decreased with 

 the running-down of the current," 



