516 Prof. Riicker. On the Magnetic Permeability of 



appreciable error. Even if the magnetic force and the induction are 

 not strictly proportional, the ratio of the permeabilities of the liquids 

 is, to the degree of accuracy attained, the same in the balance as in 

 the solenoid. 



Finally, the method was applied to a specimen of basalt which 

 Professors Thorpe and Riicker had brought from the Island of 

 Canna, in the West of Scotland. A piece of this had been cut into the 

 form of a rectangular bar, and its magnetic properties had been in-, 

 vestigated by Dr. Hoffert. His experiments are described in the 

 published account of the magnetic survey ( ; Phil. Trans.,' A, 1890). 

 The permanent magnetisation was determined by three methods, and 

 the susceptibility was found from the times of vibration when 

 the bar was suspended in a known magnetic field with the directions 

 of permanent and induced magnetisation alternately coincident and 

 opposed. Dr. Hoffert found that the value of *: was about 0'0015. 

 Unfortunately, the particular bar used by him has been mislaid, but 

 we have measured the permeability of another fragment of the same 

 specimen, and find 



K = 0-00132. 



Observations to be described below prove that differences such as this 

 exist between specimens of the same rock. We do not therefore 

 regard these numbers as giving any test of the accuracy either of 

 Dr. Hoffert's or our own observations, but the agreement between 

 them is sufficient to prove that there can be no doubt as to the order 

 of the magnitudes of the quantities under discussion. 



The range over which the instrument could be employed was 

 between susceptibilities about five times greater and ten less than 

 that of Canna basalt, i.e., from about 0'00792 to O'OOOIS. The 

 higher of these values could only be obtained by extrapolation, as 

 liquids of such great permeability could only be formed by using so 

 small a quantity of glycerine that the magnetic oxide settled too 

 quickly to enable us to obtain reliable results. In the tables given 

 below, the statement that the magnetic susceptibility of a substance 

 is zero means only that it is distinctly less than that of the weakest 

 liquid employed, i.e., than about G'00013. The details of a single 

 experiment are given by way of example. The readings of the com- 

 pensator scale are in centimetres. 



Experiment. 



Equal volumes of the liquid of strength 2/9 were placed in test- 

 tubes in the two arms of the balance. 



rlO-1 



Zero readings < 1Q . Mean = 10' 11 = z } . 



LlO-15 



