502 



PROFESSOR KNOTT ON SOME RELATIONS BETWEEN 



To bring out the form of the dotted curve more clearly, it is plotted on a larger scale so 

 far as the field is concerned. Also, to prevent confusion with the lines of the other 

 curve, it is taken with a negative line current. These and other particulars are clearly 

 indicated in the figure. 



The two nickel cycles are represented by figs. 2 and 3, Plate III. The former is 

 for a field cycle, and is essentially of the same character as the first curve for the iron. 

 The latter is for a current cycle ; and the fundamental difference between the characters 

 of the field and current cycles is apparent at a glance. It should be mentioned that 

 these curves are but samples of a great number of experiments which were made with 

 various values both of currents and fields. So far as the purpose of the present paper is 

 concerned, however, the features are essentially the same in all cases. 



It will be noticed that the field corresponding to the maximum twist varies according 

 to the manner of the approach, being smaller for the descending than for the ascending 

 branches. Also the twist itself is greater at the ascending maximum, so that the 

 complete cyclic graph intersects itself in two points. For zero field a definite residual 

 twist is left, being positive or negative according to the sign of the recently acting field. 

 The reversed field which reduces the twist to zero is, for both iron and nickel, almost 

 identical in value with the field which reduces the residual longitudinal intensity to zero, 

 as studied in such full detail by Professor Ewing. # Such a correspondence is quite to be 

 expected, since residual strain effects will always be associated with residual magnetic 

 effects, and will probably vanish when these vanish. 



In the case of the iron the residual twist for zero field is greater than the twist for 

 the highest field. This phenomenon, when observed experimentally, is a very striking 

 one, the twist increasing when the twisting agent is apparently removed. The reason 

 of it is, of course, at once evident when the cyclic graph shown in fig. 1 is considered 

 in its full significance. For it matters not whether the field is reduced to zero suddenly 

 or gradually — in either case the twist must pass through its maximum, so that the 

 residual twist is really in great measure the after-effect of the maximum twist. The 



Line 

 Current. 



Field Cycle. 



Limiting Range 

 of Twist. 



Residual 

 Range. 



1-06 



8-5 



4-16 



3-36 





15-1 



4-28 



3-78 





231 



3-93 



3-88 





363 



3-23 



3-78 





57-8 



2-21 



3-59 





86-1 



1-68 



3-51 





106-0 



1-28 



3-40 





132-1 



111 



3-36 





243-2 



0-78 



3-33 



* "Experimental Researches in Magnetism," Phil. Trans., 1885, and "Magnetic Qualities of Nickel," Phil. 

 Trans., 1888. 



