Brown — Mechanical Stress and Magnetisation of Iron. 117 



is further increased by about 16 per cent., the twist is slightly increased, or 

 about 2 per cent. It is difficult to see why this increase should take place. 

 It might, however, be due to a sligiit error in reading the value of the current 

 through the wire. 



It will also be seen that the area of the cyclic curve which measures the 

 circular magnetisation is decreased by about 9 per cent, for an increase of 

 70 per cent, in the load, and reaches a minimum at a load of 2-1 x 10° grammes 

 per sq. cm. ; and then there is a slight increase in the cyclic area for a further 

 increase of 16 per cent, on the load. 



The same wire, No. 16, was now annealed by passing tJie flame of a 

 bunsen burner slowly over it ; and was again tested for two different loads in 

 magnetic fields of different values. That is, it was tested after annealing with 

 a load on of 10^ grammes per sq. cm., when placed in magnetic fields of 

 strengths from 4 to 1 3'6 c, g. s. units. It was then re-annealed, and again 

 tested for a load of 243 x 10° grammes per sq. em. when placed in the same 

 fields. The results thus obtained are given in Table YII. and fig. 8. 



Table VII. 



If we plot the values of the magnetic fields as abscissse and the two last 

 columns in Table YII., as ordinates, we obtain the curves shown in fig. 8, 

 where the first curve at the top is that of maximum twist and magnetic field 

 for a load on the wire of of 10° grammes per sq. cm., and the second curve is 

 that for a load of 2'43 x 10* grammes per sq. cm. 



SCIENT. PECO. E.D.S., VOL. XII., NO. XII. X 



