Rock* and Regional Magnetic Disturbances. 



515 



very slowly, and that the variation is not sufficient to affect seriously 

 the argument of this paper. 



We now turn to the method of comparing the susceptibilities of the 

 rock specimens with those of the liquid by means of Professor 

 Hughes' induction balance. 



In the first place it wns necessary to mate certain thai the effects 

 observed were due only to the permeabilities, and not to the conduc- 

 tivities, of the bodies under investigation. The weaker liquids were 

 practically non-conductors, but the stronger ones conducted feebly. 

 When, however, a solution of salt in water, of rather greater conduc- 

 tivity than the strongest mixture, was introduced into the balance, 

 which had previously been adjusted, no 1 sound whatever could be 

 detected, thereby proving that the very different effects obtained with 

 the magnetic oxide were not, in any way, due to Foucault currents in 

 the mixture. Two of the rocks which produced the greatest effect in 

 the balance were also chipped out into the form of horse-shoes, 

 and by dipping the ends into two mercury cups or into two cups 

 containing acid and water, they were used to com/plete circuits, in 

 which a mirror galvanometer was included. They appeared, as thus 

 tested, to be non-conductors. We are, therefore, confident chat the 

 experiments are not vitiated by Foucault currents set up within 

 either the liquids or the rocks. 



The first test applied to the method was to measure by the aid of 

 the balance the susceptibilities of the different mixmres relatively to 

 each other. Thus in the case of three liquids a, b, and c, say, the sus- 

 ceptibility of b was found by using the values of the susceptibilities of 

 a and c which had been obtained by the absolute method. In the 

 following table the numbers thus obtained are compared with those 

 given directly by the absolute method : 



The agreement between the last two columns is sufficient to justify 

 the induction balance method. The strengths of the fields in the 

 balance were different from those employed in the absolute method, 

 and Foucault currents might affect the results. It is clear, however, 

 that neither of these possible causes of difference produces any 



2 N 2 



