642 MR JAMES RUSSELL ON 



acting, when the magnetising force at right angles is superposed ; and the heavy dash 

 curves, the component of the induction in the direction of the force first acting after 

 the superposed force has been repeatedly reversed. On the other hand, the faint full 

 lines measure the component of the induction in the direction of the magnetising 

 force last acting, and which has been superposed ; and the faint dash curves measure 

 the component of the induction in the direction of the force last acting and which 

 has been repeatedly reversed. Reference to the tables (VII. and VIII.) shows that 

 the heavy full line curves, CL (or LC), when taken along with the faint full line 

 curves, CL (or LO), co-ordinate the circular and longitudinal components of the resultant 

 induction under the CL (or LC) conditions. Also that the heavy dash curves, 

 CLL (or LCC), taken along with the faint dash curves, CLL (or LCC), co-ordinate 

 the circular and lougitudinal components of the resultant induction under the CLL (or 

 LCC) conditions. 



§ 18. It will be observed that both the faint line curves, CL and LC, are near 

 their origin above and afterwards decidedly below the normal curves (dotted lines). 

 As the first effect is small, it was thought advisable to repeat the measurements of the 

 circular component of the induction under the LC conditions with shield A, limiting the 

 values of the circular field between H c = and H c = 22 C.G.S. units. Table IX. gives 

 these results obtained in a slightly different manner from those given in Tables VII. and 

 VIII. Four sets of readings were taken, increasing the longitudinal force by increments 

 from zero, and at each increment superposing the circular field, kept at a constant value 

 for each set. The second column gives the measurements of the longitudinal force, 

 which, as will be observed from the third column, produces no throw of the galvanometer 

 connected so as to measure B c . The circular field, at the particular value given in the 

 first column for each set of readings, being now superposed, the circular components 

 of the resultant inductions are obtained in scale divisions, fourth column, and in C.G.S. 

 units, last column. In fig. xiv. the circular component B c is plotted against the 

 superposed force H,., for four different values of the longitudinal field, viz. : — H t = 0, 

 6 '5, 16 '4, and 41 C.G.S. units. By taking readings whenHj = 0, the normal B-H curve 

 (dotted line) is obtained with a minimum of error, due to altered conditions when com- 

 pared with the full line curves. Fig. xiv. thus shows on an enlarged scale, and for 

 another quality of iron (shield A), the earlier positions of the B c curves under the LC 

 conditions. It is at once evident that the initial effect mentioned at the beginning of 

 this section is confirmed. 



Figs, xi., xii. and xiii. show that the effects of field superposition are essentially 

 the same in the B c as compared with the B t curves, although more pronounced in the 

 former than in the latter case. This difference is fully accounted for by the fact that in 

 the Bj curves the longitudinal induction completes its circuit through a large air re- 

 luctance. 



§ 19. The distinction between "circular" and "longitudinal" may thus be regarded 

 as arbitrary. The conclusions now to be drawn can be equally well or better stated in 



