512 



The Eev. T. R. Robinson's Expmmmtal Researches on the 



Here M is the resultant of {z) and (c — 2) ; M' of {z + 6) and (c — 2 — 6). 

 In the first five pairs z and 2+6 are constant, so that the variation of M and 

 M' is due to that of c ; but even so, it is impossible to get their relation. The 

 variable u cannot be obtained from 2? (a + w), unless the nature of the function 

 R be known. Still, some information may be derived from them. 



The ratio of M to M\ and the values of the quantities on which they 

 depend, are given for each pair in — 



Table XXIII. 



The ratio of if to M', notwithstanding some irregularity in the second and 

 third, decreases slowly in the first five ; so slowly as to show that the compo- 

 nents of each which depend on the constant terms predominate much over the 

 others. In the sixth, where z and 2 + 6 are considerably greater, while the 

 others are lessened, the ratio is increased, though the absolute magnitudes are 

 diminished. As c - 2 and c — 2 — 6, having a common difference, tend to 

 equality by these increases, I should expect that the ratio would ultimately be 

 that due to ( 1 ) and ( 7 ), unless something beside the mere distance interferes 

 with the induction. The last column, which gives the ratio of the first M to 

 those of the other circuits, shows a more rapid decrease, but though in this 

 case there is only (c-2) variable, the law seems unattainable. 



It occurred to me as possible that iron in a state of high excitement may 

 not have the same power of transmitting induction as when less magnetic. If 

 so, when each pole has its own helix, as the M then magnetizes the keeper, 

 the M' of the other pole would be transferred with less power, and so also for 

 the remainder of the circuit. In this case L is the resultant of 2M+ 2M', and 

 this would account for the resultant being less than the sum. As ^|r would 



