PEIRCE. — RESISTIVITY OF HARDENED CAST IRON. 187 



and the temperature coefficients of their magnetism are ahnost always 

 much smaller than those of forged steel magnets. Cast iron for per- 

 manent magnets must, however, be really hard, and, unfortunately, 

 mechanical tests of the hardness of this metal are often deceptive ; it 

 seems desirable, therefore, to inquire whether the electric resistivity of 

 a piece of chilled cast iron is a criterion of its temper. 



This paper gives briefly a few of the results of a large number of 

 observations made originally with the object of testing the relative 

 efficiencies of different methods of hardening cast iron for magnets, in 

 use in the Jefferson Laboratory. The details of this work have mainly 

 a local interest and are not enumerated here, but some general facts may 

 be useful to persons who have to make such magnets for themselves. 



Each of the test pieces was a rod about 30 cms. long and a little less 

 than 0.6 cm. in diameter. These were all milled down from stouter 

 pieces about 1.5 cms. in diameter which were usually cast in sets of a 

 dozen from a grid pattern to insure that they should be of the same 

 kind of iron. Different specimens from the same grid, however, often 

 showed different resistivities before they w^ere annealed and occasionally 

 one or two pieces from a grid would differ sensibly from the other 

 pieces after all had been softened with great care. These differences 

 are to be expected, as Karsten showed long ago, for the outer layers of 

 a mass of chilled cast iron sometimes contain a greater proportion of 

 combined carbon than the inner layers in which most of the-carbon may 

 be free, and an unequal chilling of a grid in the mould would -naturally 

 make the material slightly different in different parts. It is easy in 

 practice to avoid abnormal specimens. All the test pieces were pre- 

 pared, annealed, and hardened by Mr. George W. Thompson, the 

 mechanician of the Jefferson Laboratory, whose experience in treating 

 cast iron extends over many years. 



The measurements of the specific resistances of the rods (usually 

 three for each specimen) were mostly made with the help of a standard 

 Kelvin Double Bridge, but in a few cases the test piece was connected 

 in series with a standard manganin resistance bar and a constant stor- 

 age battery, and the small potential drop across a measured length of 

 the rod was compared with the corresponding drop across the standard. 

 Three commutators were used with this apparatus so that the effects 

 of disturbing electromotive forces at the contacts might be avoided. 

 The ultimate standard was Wolff No. 2718 furnished with the certifi- 

 cate of the Reichsanstalt. 



In the determinations of the temperature coefficients of resistivity 

 two large tanks of water were used. One of these was approximately 

 at room temperature. The water in the other, which was kept in 



