// '/.* <md Regional Magnetic Disturbance*. 



513 



If now we take the means of the results of Ihese four series of 

 experiments as giving the values of the susceptibilities, and divide 

 each number by the strength of the mixture, the ratio is found to be 

 nearly constant. It must be remembered that errors which cause 

 deviations from the mean value are in part, and probably in large 

 part, due to uncertainty as to the exact composition of each mixture. 

 As the magnetic oxide settles quickly in water, tlie amounts added to 

 the glycerine were probably only approximately proportional to the 

 volumes used ; but, as the susceptibility of each mixture is absolutely 

 determined without reference to its supposed composition, this will 

 not affect the accuracy of the results. 



The numbers in the last column are consistent with the view that 

 the susceptibility of any mixture varies directly as the percentage of 

 magnetic oxide which it contains. The matter may be further 

 tested by means of the sets of experiments described immediately 

 below, in which the susceptibilities of three of the liquids (including 

 the weakest and the strongest mixtures) were again measured. The 

 mean results are as follows : 



In this case the strongest mixture gives the smallest result. On 

 the whole, then, and for the purposes of this investigation, the values 

 of K/8 must be considered as independent of the strength of the 

 mixture. Any difference which exists 'could only be certainly de- 

 tected by a very careful determination of the quantity of magnetic 

 oxide present in the unit of volume in each case. Since this law 

 holds good within the limits of the series of liquids the suscepti- 



VOL. ILVIII. 2 N 



