506 Scientific Proceedings, Royal Dublin Society. 



abscissae the values of the longitudinal magnetic field, and as ordinates the 

 corresponding values of the twist, we obtain the curves shown in fig. 2 (p. 505), 

 from which we see that the longitudinal magnetic field in which the maximum 

 twist occurs is independent of the temper of the wire. The top curve is 

 that given by the softest wire of temper Hg, and the lowest curve that given 

 by the wire of temper Hi; and the intermediate curves repi-esent the 

 intermediate tempers or degrees of magnetic softness. 



Again, from the same table, if we plot on the axis of abseissse the values 

 of the rigidity of the wires at different tempers, and as ordinates the corre- 

 sponding values of the maximum twist for the five wires when they are 

 subjected to the three different longitudinal loads, it will be found that the 

 points in each of the three eases lie very approximately in a straight line. 

 When these three straight lines are produced, they cut the axis of abscissae 

 at practically the same point which represents a rigidity of 774 x 10'^ grammes 

 per sq. em. ; and therefore a wire of this rigidity should give little or no 

 twist when subjected to simultaneous longitudinal and circular magnetism. 

 An attempt was made to test this experimentally ; but the nickel could not 

 be hardened beyond a certain limit by means of hanging a weight on the 

 end of it and raising it to a bright cherry-red heat. Weights equivalent to 

 1'5 X 10' and 2 x 10^ grammes per sq. cm. were hung on the wire when it was 

 being heated, and the highest rigidity obtained was 757 x 10'* grammes per 

 sq. cm., — that is, a little above the wire of temper Hi — and the twists for the 

 various longitudinal magnetic fields (with a load on the wire of 1'5 x IC* 

 grammes per sq. cm.) when plotted gave a flat curve situated below the lowest 

 curve in fig. 2, and having a maximum twist of approximately 25 millimetres 

 on the scale, which value of maximum twist falls practically on the straight line 

 corresponding to the given load. This, however, shows that the graphs 

 relating the maximum twist and hardness or temper of the wires, for the 

 different longitudinal loads, are not really straight lines, but are flat 

 curves, which for the higher tempers very gradually approach the axis of 



In order now to test the effect of varying the length of the wire, a fresh 

 No. 16 nickel wire was taken and prepared for experiment in the usual way, 

 without getting it to any definite temper — which was, however, between 

 the tempers Hj and H|, judging from the values of the twists obtained. 

 Three different lengths were tested, each under three different loads, and in 

 sixteen different longitudinal magnetic fields. 



All the results are given in Table III ; and one set of them is shown as 

 curves in fig. 3 — that is, the results obtained when the longitudinal load 

 on the wire was 1'5 x 10^ grammes per sq. cm. 



