514 Prof. Riicker. On the Magnetic Permeability of 



bilities of which could be accurately determined, it may be safely 

 inferred that for weaker mixtures, at any rate, it will still be true, 

 and so a number of mixtures weaker than one-fifth were prepared, 

 and their susceptibilities calculated by proportion from those of the 

 stronger ones. Values so obtained will, however, be subject to the 

 error of mixing. 



Finally, a series of experiments, referred to above, was made to 

 determine whether the permeabilities of the mixtures varied with the 

 magnetic force. The susceptibilities obtained when 2, 3, and 6 

 Grove's cells were used in turn are given in the following table : 



The sums of the significant figures in the three columns are 495, 

 488, and 476 respectively, thus indicating a slight increase of per- 

 meability with the magnetic force. The differences between the in- 

 dividual observations are, however, too great to allow us to rely on 

 this result, and the table can only be considered as proving that no 

 serious error will occur if we assume that the permeabilities of the 

 mixtures are independent of the magnetic force. 



[Added Sept. 12, 1890. Experiments made afterwards confirmed 

 this view and extended the range over which its accuracy was tested. 

 The weakest field employed was about twice the earth's field in the 

 United Kingdom,] 



It would have been difficult to obtain accurate results with a 

 weaker field, but as the law of proportionality between the mag- 

 netic force and the induction appears to hold very approximately 

 for forces between 5'0 and 1'7 C.G.S. units, it is probable that 

 it is also valid for smaller values. It is true that Silow (' Wiede- 

 mann's Annalen,' vol. 11. 1880, p. 330) has stated that the 

 susceptibility of ferric chloride is a maximum when the inducing 

 force is about 0'4 C.G.S. unit, and that for that value it is between 

 two and three times as great as for fields at strengths snch as 

 those at which we have experimented, but the change in a range 

 of magnetic force much less thau 1*0 to 5'0 was veiy marked. Thus, 

 between forces of about 1'4 and 2'5, the susceptibility altered by 

 about 13 per cent. An effect such as this could not possibly have 

 escaped our notice, and there can be no doubt that for forces such as 

 those with which we have dealt the susceptibility of magnetite changes 



