486 Scientific Proceedings, Royal Dublin Society. 



Table II. If we plot as abscissae the values of the load on the wire, ami as 

 ordinates the corresponding values of the areas of the cyclic curves, the 

 points will be found to lie in a straight line and show that when the load is 

 increased six times the circular magnetism or area of the cyclic curve is 

 decreased about 38 per cent. 



These results confirm those obtained previously with a wire which was 

 tested under the same conditions of load and magnetic field. In the former 

 test the area of the cyclic curve when the wire was loaded with 10^ grammes 

 per sq. cm. was 30 sq. cm., which value wlien reduced from the scale and 

 distance used then to the scale and distance used now becomes 20'1 sq. cm. as 

 against 20'8 sq. cm. for the new wire, which was both a little thicker and 

 softer. 



The above results were obtained when the wire was the full length of 

 226 cm., and in a uniform longitudinal magnetic field throughout its entire 

 length ; tests were now made when the wire was 0-8 times and 0'5 times 

 this length respectively, so as to find out how the twist and circular 

 magnetism varied when the length of the wire was altered to 181 cm. and 

 113 cm. The top suspension for the wire was made longer to suit the 

 lengths required, the wire still remaining in a uniform longitudinal magnetic 

 field throughout its whole length, and each length was put through the same 

 series of tests as was done with the full-length wire. That is, the maximum 

 twist was observed for the given current density in the wire of 100 amperes 

 per sq. cm. ; when the wire was placed in longitudinal magnetic fields of 

 tivelve different strengths, and also when subjected to three different values 

 of longitudinal load, instead of the five loads used with the long wire. As 

 in tlie previous case when the wire was of the full length, the greatest twist 

 occurred when it was in a magnetic field of 2'5 units, and at this point 

 a complete cycle was taken in each case. 



The results obtained are shown in Table III., and one set of the 

 results for each of the three lengths when tlie load on the wire was 0"5 x 10' 

 grammes per sq. cm. are exhibited as curves in fig. 3 ; the results for 

 the other two loads when plotted sliow the same general contour in the 

 shape of the curves with the maximum gradually decreasing. 



