170 Dr. E. H. Hall on the " Rotational Coefficient'' 



should be undertaken as soon as practicable. It will require 

 very great intensities of the magnetic field and a very large 

 battery* to carry these metals through a range of magnetiza- 

 tion corresponding to that through which nickel has been 

 examined. It seems doubtful whether the magneto-electric 

 machine can be here employed, as the current which it pro- 

 duces may not be sufficiently uniform to be used with ad- 

 vantage. 



The examination of the non -magnetic metals also should be 

 continued as fast as circumstances will permit, with the object 

 of determining the sign and, when practicable, approximately 

 the magnitude of the rotational coefficient in every case. 



In my article of last November I stated that, in accordance 



* In the experiments here detailed, the largest battery used has con- 

 sisted of 48 large Bunsen cells arranged 8 in series. The resistance of 

 each cell was probably something more than an ohm ; the resistance of 

 the electromagnet is, I think, rather less than an ohm. The resistance of 

 the connexions was considerable, however ; and the battery probably gave 

 about its best effect. 



After this powerful battery had been applied to the electromagnet, a 

 rather singular effect was observed on returning to the use of weak cur- 

 rents. In making observations in the usual way to determine the strength 

 of the field produced by these weak currents, it was found that the im- 

 pulses given to the galvanometer-needle were very capricious. These 

 observations may be arranged under two heads, + and — , according to 

 the direction of the current through the electromagnet, this current being 

 usually reversed after each withdrawal of the little coil from between the 

 poles. * The observations being arranged in this way, it would be found 

 that there were occasional sudden changes of many per cent, in the read- 

 ings in the same column. Of course the most obvious explanation of the 

 phenomenon was that some connexion was loose, either in the circuit of 

 the galvanometer and the test-coil or in that of the electromagnet. That 

 the fault was not in the former circuit was made probable by the fact that 

 by means of the earth-inductor, which was in the same circuit, quite uni- 

 form deflections of the galvanometer-needle were produced. To test for 

 a fault in the magnet circuit, a tangent-galvanometer was introduced into 

 it and its deflections observed during the series of observations on the 

 strength of the field. 



The readings of the tangent-galvanometer decreased slowly with the 

 running-down of the current ; but the changes were quite regular, and not 

 at all of a character to account for the irregularity of the other observa- 

 tions. The most plausible explanation I could finally propose was, that 

 this irregularity in the strength of the magnetic field was due to a sort of 

 uncertain struggle between the action of the present weak magnetizing 

 current, and the magnetization previously induced by the strong currents 

 in the poles of the electromagnet, which are not, I believe, of very soft 

 iron, and are probably capable of considerable permanent magnetization. 

 I do not by any means feel able to assert, from my rather hasty observa- 

 tions, that there can be no other explanation. I have, however, simply 

 thought the matter of sufficient importance to justify me in recording what 

 seemed to be the fact. 



