Magnetic Tractive Force. 161 



determined for these by means of a du Bois Magnetic Balance*", 

 which had been tested at the Reichsanstalt. These experi- 

 ments, as well as all the tractive-force measurements, were 

 made with increasing magnetization f. This integral formed 

 only a comparatively small correction, so that it was not 

 necessary to know it very accurately. 



Final Experiments. 



From the galvanometer-throws the values of B and H were 

 calculated, and their difference gave, on being divided by 47r, 

 the magnetization I. The greatest values of I obtained for 

 the soft iron and the hard steel were 1818 C.G.S. and 1556 

 C.G.S. respectively, and these occurred at a field of about 

 10,600 C.G.S. At stronger fields I appears to have rather 

 smaller values J ; a similar apparent diminution of I was also 

 observed by Ewing§. 



It must, however, be remarked that the isthmus method is 

 never very suitable for accurate measurements, and that the 

 arrangement described above is particularly unfavourable to 

 accurate measurements of I from the difference of B and H ; 

 firstly, because of the disturbing influence of the gap in the 

 isthmus ; and, secondly, because the bars are not continuous 

 with the pole-pieces, as in the original arrangement of Ewing 

 and Low. 



No valid conclusions can therefore be drawn from these 

 experiments as to the course of the magnetization-curve ; the 

 induction in the isthmus, however, which is the quantity of 

 chief importance in these experiments, was, in consequence 

 of its direct determination, known with sufficient accuracy. 



With each value of the current, the measurements were 

 repeated several times. For given values of the galvanometer- 

 throws, the minimum tractive force was very constant, so that 

 even at the highest values it could be determined to within 

 about 30 grammes weight. 



The greatest magnetic tractive force measured was 9430 

 grms. weight, the cross section of the iron bar being 0*1896 

 sq. centim. This gives a pull of 49*73 kilogs. weight per sq. 

 centim. (707*4 lb. wt. per sq. inch). 



* See du Bois, Zeitschr. fur Instr-Kunde, xii. p. 404 (1892). 



t The curves obtained with the balance will be published in another 

 connexion. 



\ In later experiments made with a continuous isthmus (see p. 160, 

 footnote) this diminution of I was not observed. 



§ Ewing, I. c. § 93; this was attributed by him to non-uniformity of 

 the field near the isthmus. 



