Ox THE MAGNETIC EFFECT OF ELECTRIC CONVECTION 129 



a commutator was placed, so that the potential of the latter could be 

 made plus or minus at will. All parts of the apparatus were of non- 

 magnetic material. 



Over the surface of the disc was suspended, from a bracket in the 

 wall, an extremely delicate astatic needle, protected from electric 

 action and currents of air by a brass tube. The two needles were 1-5 

 cm. long and their centres 17-98 cm. distant from each other. The 

 readings were by a telescope and scale. The opening in the tube for 

 observing the mirror was protected from electrical action by a metallic 

 cone, the mirror being at its vertex. So perfectly was this accom- 

 plished that no effect of electrical action was apparent either on charg- 

 ing the battery or reversing the electrification of the disc. The needles 

 were so far apart that any action of the disc would be many fold greater 

 on the lower needle than the upper. The direction of the needles was 

 that of the motion of the disc directly below them, that is, perpendicular 

 to the radius drawn from the axis to the needle. As the support of 

 the needle was the wall of the laboratory and the revolving disc was on a 

 table beneath it, the needle was reasonably free from vibration. 



In the first experiments with this apparatus no effect was observed 

 other than a constant deflection which was reversed with the direction 

 of the motion. This was finally traced to the magnetism of rotation 

 of the axis and was afterward greatly reduced by turning down the 

 axis to -9 cm. diameter. On now rendering the needle more sensitive 

 and taking several other precautions a distinct effect was observed of 

 several millimetres on reversing the electrification and it was separated 

 from the effect of magnetism of rotation by keeping the motion con- 

 stant and reversing the electrification. As the effect of the magnetism 

 of rotation was several times that of the moving electricity, and the 

 needle was so extremely sensitive, numerical results were extremely 

 hard to be obtained, and it is only after weeks of trial that reasonably 

 accurate results have been obtained. But the qualitative effect, after 

 once being obtained, never failed. In hundreds of observations extend- 

 ing over many weeks, the needle always answered to a change of electri- 

 fication of the disc. Also on raising the potential above zero the action 

 was the reverse of that when it was lowered below. The swing of the 

 needle on reversing the electrification was about 10- or 15- millimetres 

 and therefore the point of equilibrium was altered 5 or 7| millimetres. 

 This quantity varied with the electrification, the velocity of motion, 

 the sensitiveness of the needle, etc. 

 9 



