Equipotential Lines of an Electric Current. 141 



The first three of these comparisons were carefully made ; 

 the others less carefully. In the second the electro-magnet 

 was worked by a battery current. In all the others the cur- 

 rent came from a Gramme dynamo. The coil was always 

 withdrawn from the field very briskly on the end of a lever 

 which carried it three or four feet away from the magnet. 

 The reversal was effected by means of a mercury commutator, 

 which was worked by pulling a string and probably turned in 

 a quarter of a second or less. The time of a single swing of the 

 ballistic needle from one elongation to the next was small, not 

 much more than three seconds. The discrepancy between the 

 results of the two methods when comparatively weak fields are 

 used is very great, about 10 per cent. The difference dimin- 

 ishes as stronger and stronger fields are used, and in fields above 

 9000 appears to be not greater than 2 per cent. Direct com- 

 parisons of the two methods were not made in fields much 

 stronger than those given in the table. Indirect comparisons 

 made in a field about 12,000 and in another about 16,000 indi- 

 cate that the difference continues to decrease as the field in- 

 creases in strength. 



It has been already stated, that the loops on the iron crosses 

 5 and 7 gave their induction currents by remaining between 

 the poles when the magnet current was reversed. The indica- 

 tions which they gave were therefore subject to a correction in 

 order to make them comparable with those given by the test- 

 coil upon sudden withdrawal from the magnetic field. In the 

 results to be given this correction will be applied in accordance 

 with the indications of the table above. It will be assumed 

 that the proper correction in the strongest fields used is 1 per 

 cent. 



The main object of the whole series of experiments being 

 to test the relation between the magnetic induction through 

 the crosses and the transverse effect in them, it was desirable 

 to get the induction current from the loop as nearly as possi- 

 ble simultaneously with the measurement of the transverse 

 current from the arms of the cross. This latter current was 

 measured by a series of readings of an astatic galvanometer, 

 the magnet current being reversed after each reading. At 

 every such reversal, the swing of the ballistic needle due to the 

 induction current from the loop of wire was noted, .so that 

 the two series of observations, one upon the transverse effect, 

 the other upon the induction, were carried on at the same 

 time. To avoid error due to direct action* of the electromag- 

 net upon the ballistic and the astatic galvanometers, or to in- 

 duction in parts of the ballistic galvanometer circuit which 



* When the magnet was strongest its direct action upon the astatic galvanom- 

 eter about 140 feet distant was very important. 



