Longitudinal Stress 07x the Magnetization of Nickel. 125 



remark. Here, then, we see how the already diminished loop 

 for negative twisting, as shown in fig. 9, has vanished alto- 

 gether in fig. 10. At the same time, the curve begins to 

 cross the lines of zero magnetization into the negative region. 

 Fig. 11 is a further development in the same direction. In 



both of these curves (10 and 11), the quantity -^, the rate of 



change of magnetization with twist, still changes sign at par- 

 ticular twists. The negative tail — as it might be called — so 

 evident in fig. 10, has disappeared in fig. 11 ; while at the 

 same time the negative magnetization has numerically in- 

 creased. But now passing to the higher load (see fig. 12), 

 we see that the various changes discussed above have their 

 final end in a simple single-looped curve, with a large portion 

 in the region of negative magnetization, and with no true 

 minimum-points for ^. The twist for which, in the first three 



cases, the value of -7- changes sign, works towards the left 



as the load is increased. Thus : — 



For W=400, d^/dT=0 at + 8°. 

 W = 527, „ =0 at -20°. 

 W = 655, „ =0 at -60°. 



For W = 1910, we may regard this critical twist as being too 

 great to be included in the range of greatest twist applied. 



It was remarked, while discussing the experiment made in 

 the earth''s vertical field, that the ratio of the two opposite 

 magnetizations gradually tends to unity as the load is increased; 

 but this does not seem to be generally the case. There is a 

 certain limit beyond which there is an opposite tendency. 

 The following calculations show that this must be the case: — 

 Let -t-3 and — 3 be the greatest magnetization during positive 

 and negative twists respectively, then in the field = 4' 94 units, 

 and 



W= 655; +3 = 188-6, -'^= 49-9, +3/_3 = 3-78 ; 



W= 910; +3 = 152-1, -:^ = 151-3, +3/-:^ = l-00; 



W = 1270; +3 = 119-3, -3=113*1, +3/-3 = 105; 



W=1910; +3= 86-3, -3= 77-4, +3/_3 = l-ll. 



The following experiments in field =6-71, show how this 

 ratio depends on the strength of the field. The changes which 

 the magnetization-curve undergoes while it is changing its 

 sign are quite analogous to the preceding two cases, as a glance 

 at figs. 13, 14, and 15 will show. From figs. 15, 1 6 (PI. III.), 

 17 (PI. IV.), however, we see how the opposite magnetization 

 becomes smaller as we increase the field . The ratio + 3/( —3) 

 is always a very large quantity and has a minimum value for 



