MAGNETISM AND TWIST IN IRON AND NICKEL. 



517 



current the lagging is positive ; for the intermediate current it is positive at the lower 

 end, but on the whole negative — indeed almost zero ; and for the strongest current it is 

 distinctly negative. Fig. 3 gives three out of the five experiments for the twist cycle 

 ±225°. All show negative lagging; although in the case of the smallest current there 

 is a slight indication of positive lagging near the extreme values. The graph, in fact, 

 has three loops ; and a slightly weaker current for this twist of ± 225° would probably 

 have given a much closer graph intersecting itself in several places. 



In fig. 4 (PL IV.) the manner in which the area of the graph varies with current and 

 twist is represented graphically for all the cases tabulated in Table IV. There are seven 

 curves in all, corresponding to the seven ranges of twist. The abscissae give the currents 

 along the wire ; and the ordinates are proportional to the areas of the cyclic graphs. 

 For all but one of these curves there is a maximum point. The lag-area begins positive, 

 attains a maximum for a particular value of current, then begins to decrease, and 

 ultimately, if the current is strong enough, becomes zero and changes sign. For the twist 

 cycles ±90° and ± 135° the current is not taken high enough to show the change of 

 sign ; but a close study of the graphs of fig. 1, or of the numbers in the table, will show 

 how these graphs begin to contract when the current is strong, as if preparing for a 

 change of sign for a current of about 6 or 7 amperes. Except for the two lowest twist 

 cycles, all the curves which have a maximum point cut the zero line. The curve for 

 ±225° has apparently no maximum point; but it may possibly have one for a very 

 small value of current, as indicated by the dotted line. From fig. 4 we may construct 

 the following table showing the maximum positive area for each twisting, the current 

 corresponding to that area, and the current corresponding to the vanishing point of the 



area. 



Twist. 



Max. A 



7T. 



Current 

 corresponding. 



Current for Zero 

 Area. 



90° 

 135° 

 160° 

 170° 

 180° 

 190° 

 225° 



68 

 105 

 105 

 60 

 54 

 16 

 1 



11 



•6 ? 



•42 



•40 



•38 



•17 



•04? 



7 ? 

 5 ? 

 3-4 

 15 

 1-05 



•48 



•08? 



The general truth here indicated is that increasing range of twist and increasing 

 current alike have the tendency to bring about a negative lagging, the first effects of 

 untwisting from either limit being greater than the immediately preceding effects of 

 twisting up to that limit. 



In fig. 5 (PL IV.) the ranges of intensity for three of the cyclic twistings are plotted 

 against the currents. From these it is evident that the range increases somewhat rapidly 

 at first, increasing more and more slowly as the currents and twistings are taken higher 

 and higher. And this holds for all cases, the range of polarity increasing with the twist 



and with the current. 



VOL. XXXVI. PART II. NO. 18). 



4 H 



