Determination of the Magnetic Susceptibility of Rocks. 461 



lias to be moved to produce silence after the introduction of a small 

 piece of a rock into one of the tubes is noted. This is repeated with 

 liquids of different strengths. Then if a, b are the differences 

 observed between the rock of susceptibility x and two liquids of 

 susceptibility Jc l9 /^ respectively, we have 



a : x _ 



b K 2 x b a 



due regard being given to the signs of the differences. 



The liquid employed was pure glycerine in which finely divided 

 magnetic oxide of iron was suspended, for no solutions of metallic 

 salts possess a susceptibility sufficiently high for our purpose. These 

 liquids were very carefully prepared, and in the light of the expe- 

 rience now gained it is desirable to add to the statements made in 

 the earlier paper on this point. Natural magnetite cannot be ground 

 'fine enough to remain long in suspension, though experiments in 

 which a coarsely powdered mixture of magnetite and manganese di- 

 oxide (minerals of about the same specific gravity) was shaken in 

 water and allowed to settle rapidly into a solid mass were fairly 

 successful. 



Finally, the magnetic oxide was prepared artificially by adding 

 ammonia to a boiling solution of sulphate of iron of which one-half 

 had been oxidised by nitric acid. The black precipitate, dehydrated 

 by alcohol and dried at 100, gives a friable dark brown mass about 

 one and a half times as magnetic as a good specimen of the powdered 

 mineral. 



By grinding with glycerine between glass plates the oxide is 

 reduced to minute yellow scales. Glycerine mixtures made with this 

 substance give no indications of settling for forty-eight hours. 



Numerous experiments with various suspensions of natural magne- 

 tite called attention to the fact that in a magnetic field the particles 

 rotated slowly to set their axes along the lines of force, thus giving a 

 fictitious susceptibility largely dependent on the permanent magneti- 

 sation of the grains of magnetite. With the artificial oxide this effect 

 is extremely small, but, to avoid it altogether, we prepared gelatinised 

 glycerine mixtures which at ordinary temperatures become solid. 

 In this state we repeatedly determined, firstly, their absolute sus- 

 ceptibility with the magnetometer apparatus, and, secondly, the 

 value in sonometer scale divisions of each balance tube filled with 

 the jelly. With these particular tubes one division = susceptibility 

 O0001075. The balance tubes being thus standardised, the absolute 

 magnetic value of a liquid is determined by the induction balance 

 under exactly the same conditions that hold during the tests of 

 rocks. 



2 M 2 



