Lifting Powei' of the Electro- Mac/net. 509 



in the note, were laid on the keeper in the same position as they had on the 

 magnet, its extremities would neither attract a small key nor hold a horse-shoe 

 keeper. 



I have already pointed out the gradual decrease of the magnet's power 

 during a series of experiments, and the fact that this decrease is prevented by 

 continually reversing the current. On this plan most of the experiments whose 

 results follow were conducted. It is made more effectual by exciting the 

 magnet, and, without disturbing the keeper, suddenly reversing. It seems that 

 the abrupt change of magnetic tension keeps the molecules of the iron in a state 

 of neutrality, which prevents them from assuming permanent polarity.* To 

 perform this easily, a commutator is attached to the magnet, and each set con- 

 sists of six, half direct and half reverse. By this method the results have 

 become far more consistent. The rest of the appai'atus is unchanged except 

 the rheometer, which is now on the construction discovered by M. Gaugain, 

 and demonstrated by M. Bravais.I Six rings, the largest 19' in diameter, 

 are placed parallel on a frustum of a cone, whose base is four times its axis : 

 the centre of the needle is in the vertex of the cone, the needle is three inches 

 long. This arrangement has the advantage of giving the proportionality of the 

 force to the tangent of deflection — a far wider range than in the ordinary con- 

 struction, and enabling to measure much higher currents. For such the largest 

 rincf alone is used : for small currents the whole six. A set of fifteen obser- 

 vations with the voltameter gives for its constants in the first case, 



i^=tan<^x log->(0-58298) |1 +log-> (6-7607) x (sin^^-f sin*0)|, 



which will serve for any instrument of the same dimension. 



In comparing the efficiency of different arrangements of spires or magnets, 

 the most obvious method is to excite them till the Ufts of the magnet are the 

 same, when the mean efficiency of each spire, = ^, must be inversely as t^, the 

 product of the force and number of spires. It would, however, involve an 

 immense waste of time to ascertain this equality, and therefore it is better to 

 refer them to a common standard. That which I have chosen is the action of 



* This has been so effectual that the residual magnetism is insensible, though the number of 

 times that it has been excited exceeds 1200. 

 t Comptes rendus, Jan., 1853. 



3 u 2 



