In Nickel and Cobalt. 171 



in 



with Prof. Rowland's suggestion, I had tested the Kerr effect 

 with one specimen of nickel, and found it to be of the same 

 sign as the effect which Kerr had observed with iron. In 

 order to prevent mistakes, the experiment was repeated with 

 iron, or rather, I suppose, with steel, the result being the same 

 which Kerr had obtained. The surface of nickel first used in 

 this way was the coating upon one of the plates of Prof. Row- 

 land's absolute electrometer, the metal beneath being brass. 

 Two other specimens of nickel have since been tried. One 

 was a coating deposited electrolytically directly upon the iron 

 pole of the electromagnet, the other was a nickel film fastened 

 with soft cement to a plate of glass. Probably none of 

 these specimens was pure ; but (and this is a matter of more 

 importance) the third was of precisely the same character and 

 origin as the specimen in which the transverse effect was 

 studied. The Kerr effect is of the same sign in all three 

 plates of nickel, i. e. of the same sign as the effect in iron. 



One specimen of cobalt has also been tested for this effect. 

 A block of cast cobalt, quite similar to that from which was 

 cut the cross mentioned in the first part of this article, was 

 sawn in two, and one of the fresh surfaces was made quite 

 smooth with a file and then polished with emery. It is not 

 difficult to get a sufficiently good surface. An hour's work 

 might prepare it. 



With sunlight and a tolerably strong magnetic field, say 

 4000 (cm.-grm.-sec), the rotation produced by cobalt was de- 

 tected, and found to be of the same sign as that observed with 

 nickel and iron. 



The fact that nickel behaves like the other magnetic metals 

 in optical effect, but differently from them in the transverse 

 electrical effect, is on its face undoubtedly an argument against 

 the theory which refers the two effects to the same cause. In 

 order if possible to examine the optical effect in a somewhat 

 different manner, an attempt has been made to detect an action 

 of magnetized nickel upon polarized light transmitted directly 

 through it. For this purpose a thin piece of glass was coated 

 over a part of its surface with nickel by Wright's process*, 

 the action being stopped before the nickel film became thick 

 enough to be opaque. It was found, however, that the glass 

 alone, although only about £ millim, thick, perceptibly rotated 

 the plane of polarization of the light sent through it when 

 subjected to the very strong magnetic field employed. The 

 action produced by the nickel and glass together was of the 

 same sign as that produced by the glass alone ; and as the 

 magnitude of the effect could not in either case be measured 



* Amer. Journ. of Science, Jan. 1877, p. 49, and Sept. 1877, p. 169. 



