an Explanation of Hall's Phenomenon. 253 



influences. In addition to the possible sources of error men- 

 tioned by Mr. Hall, I was much troubled by one which he does 

 not seem to have noticed, and w r hich may probably account for 

 the uncertain nature of some of his results. It was found 

 that, though the galvanometer-deflection indicating the Hall 

 effect was invariably reversed w r hen the direction of the 

 magnetizing current was reversed, it sometimes appeared to 

 be altogether independent of the direction of the current 

 passing through the strip. It occurred to me to excite the 

 magnet without passing any current at all through the strip ; 

 and I found that when the key K was depressed a steady de- 

 flection of the galvanometer occurred, which was reversed on 

 reversing the magnetizing current. That this effect was not 

 due to the direct action of the magnet on the galvanometer- 

 needle was proved by the fact that, when the key was raised, 

 the deflection immediately ceased. It appeared at first sight 

 as if the magnet had the power of generating a current in the 

 strip of metal. But this, of course, was impossible if the mag- 

 netic force remained constant. I concluded therefore that it 

 was an effect of induction caused by weakening of the magnet 

 in consequence of the running-down of the battery, or perhaps 

 of the resistance of the wires being increased by the heating- 

 effect of the current; but it was difficult to see how this could 

 induce a current in the strip which was symmetrically situated 

 with regard to the lines of force. Finally, I found that the 

 effect was produced, not in the strip at all, but in the wire 

 which passed from the strip between the arms of the electro- 

 magnet to the galvanometer. When this wire was led out on 

 one side of the yoke, the deflection produced was opposite to 

 that which occurred when the w r ire was led out on the other 

 side ; and if the wire was turned Tialf a dozen times round the 

 magnet, the resultant galvanometer-deflection w r as so great 

 that the spot of light disappeared from the scale. This was 

 ultimately remedied by forming the w r ire in question into a 

 loop embracing the yoke of the magnet, as showm at L 1/ 

 (fig. 1). From this loop an independent wire W', attached 

 to a sliding binding-screw D, was led to the galvanometer. 

 It was found that the position of D could be so adjusted that, 

 K being steadily depressed, the magnetism of the electro- 

 magnet might be interrupted, or even reversed, without causing 

 any material deflection of the spot of light even at the instant 

 of the change. When so adjusted it was, of course, impos- 

 sible that the comparatively small changes in the magnetic 

 force, caused either by polarization of the battery or by the 

 heating of the wires, could have any effect upon the galvano- 

 meter. The Hall effect now came out with great distinctness, 



