1890.] On a new Magnetometer. 2(19 



different metals used in the construction of the instrument are known 

 with sufficient accuracy. It is probably better not to include the 

 length of the bitilars in the constant of the instrument, except for 

 rough work, as both variation in temperature and in the hygrometric 

 state of the air will produce sensible alterations in length. 



An objection may be taken to the method when veiy accurate 

 determinations are desired on the ground that a knowledge of the 

 value of the acceleration of gravity at the place of observation is 

 requisite before absolute determinations can be calculated. The 

 writer believes, however, that a quartz fibre suspension for the magnet 

 would be preferable to the bifilar for magnetic survey work. It is 

 perhaps needless to say that provision would be made for clamping 

 magnet and needle during transport. To convert the readings of the 

 instrument into absolute measure, it will be necessary to determine H 

 at as nearly as possible the same place and at the same time by com- 

 paring the indications of the instrument with those of a large 

 standard instrument of the bifilar type previously described. 



In connexion with the erection of such a standard instrument, the 

 points to be borne in mind are, that all corrections arising from 

 (1) torsion of the silk fibres, (2) uncertainty in the position of the 

 poles of the large magnet, (3) couple exerted by needle on magnet, 

 shall be made as small as possible. To effect this, the bifilars should 

 be longer, the magnet should be made of thin band steel instead of 

 wire, and the diameter of this magnet should be increased in order 

 that r may be measured more accurately, and in order that the magnetic 

 moment of the large magnet may be increased without unduly 

 increasing the deflection of the needle. The writer is at present 

 engaged in erecting such an instrument in the physical laboratory of 

 the Yorkshire College, and hopes to be able to attain with it results 

 approximating in accuracy to 1 part in 10,000. 



In conclusion, the present method of determining H is believed to 

 be very much superior to Gauss's in the following respects : 



(1.) The necessity of making a determination of a time of vibration 

 (always a tedious operation) is avoided. 



(2.) The determination of a moment of inertia is avoided. 



(3.) The determination of the distance between the two poles of 

 the deflector is avoided. 



(4.) Variation of magnetic moment of the deflector during the 

 progress of an experiment produces no error ; neither does variation 

 in inductive action of the earth produce an error. 



(5.) The magnet needs only to be very feebly magnetised, as its 

 action on the needle is due to the sum of the actions of the two poles. 



(6.) The time occupied in a determination of H is only a few 

 minutes when once the constant of the instrument has been 

 determined. 



