Brown — Mechanical Stress and Magnetisation of Nickel. 501 



clutch, and allowed to haug loosely under its own weiglit only. It was then 

 raised to a cherry-red heat — by means of the Bunsen burner — by heating 

 it from the top downwards ; this was done three times to make sure that the 

 wire was as soft as it could be made; and the hardness tlius obtained we here 

 call Ho. 



The highest degree of hardness was obtained by tlie same process, the 

 wire in this case being heated when a weight or load was on the lower end 

 of it equivalent to 10* grammes per sq. cm. ; this degree of hardness we call 

 Hi ; and the intermediate degrees were obtained in the same way when the 

 loads on the end of the wires were \ x 10°, \ x 10^ and | x 10° grammes 

 per sq. cm. respectively, and these degrees of hardness are indicated b) 

 Hj, Ha, and Hj. 



Fresh No. 16 nickel wires were taken in each case to obtain the various 

 degrees of hardness ; and the largest load of 10° grammes per sq. cm. did not 

 appreciably change the diameter of the wire, which was taken as 0'163 cm. in 

 all the tempers. It was found that the nickel wire was much easier to harden 

 by this means than an iron wire of the same size ; thus, an iron wire when 

 heated with a load on its lower end of 10° grammes per sq. cm. had its 

 rigidity increased about 1"8 per cent., whereas a nickel wire treated in the 

 same way had its rigidity increased by about 6 per cent. 



The simple rigidity and electrical conductivity of each wire were measured, 

 the former by means of a slightly modified form of Searle's horizontal 

 torsion apparatus, and the latter by comparison with a wire of pure copper. 

 The results are collected in Table I, where the rigidity is expressed in 

 grammes per square centimetre, and the electrical conductivity as a percentage 

 of pure copper, tlie cross-sectional area of each wire being very approximately 

 20'8 X 10"" sq. cms. 



Table I. 



If these results be plotted by taking on the axis of abscissse an arbitrary 



4 8 2 



