270 On a new Magnetometer. [June 19,' 



(7.) The instrument, exclusive of telescope and scale, can be made 

 ;at a very small cost. 



[Note added June 17. In comparing the accuracy of the pro- 

 posed instrument with the Kew, it is necessary to distinguish between 

 the determination of N in absolute measure, say, for laboratory pur- 

 poses, such as the measurement of the strength of electrical currents, 

 and determinations where the requirements are the estimation of 

 differences in the value of H at different stations, say, for the purposes 

 of magnetic surveys. For an absolute determination a great deal 

 can be said for an instrument in which the only measurements are 

 certain angles, certain distances, and a certain weight, and which 

 does not require a determination of the influence of the earth's 

 inductive action on the magnet, nor of the variation of the magnetic 

 moment of the magnet with temperature, nor of the position of the 

 poles of the magnet. With reference to magnetic survey work, a 

 comparison may be instituted between the Kew instrument and that 

 under discussion by assuming the constants of each to be known ; then 

 in each case angular deflections are being measured, and other things 

 being equal, the accuracy obtainable will be approximately propor- 

 tional to the magnitude of the angles observed. Now in the Kew 

 instrument, the deflection produced by the deflector in its near posi- 

 tion is about 24, and in its far position about 12. Something like 

 the difference between these deflections, or 12, will represent the 

 order of angle to be estimated as accurately as possible. It is not 

 possible in the Kew instrument to materially increase these angles 

 without unduly increasing the influence of the distribution of magne- 

 tism in the deflector. In the present instrument there seems to be no 

 reason why deflections of 45 or thereabouts should not be obtained. 

 This reasoning would seem to show that the present instrument could 

 be made considerably more sensitive than the Kew. 



It need, perhaps, scarcely be mentioned that the writer is not advo- 

 cating the use of telescope and scale for measuring the angular 

 deHections in preference to an azimuth circle. The former method 

 (quite unsuitable for measuring large angles) was only adopted in 

 the first instance to roughly test the capabilities of the instrument in 

 the absence of any graduated circle. In a final instrument all the 

 deflections will be referred to a graduated azimuth circle, as in the 

 Kew instrument at present. 



A few points of detail may be just mentioned in conclusion : (1.) To 

 eliminate any error arising in the bifilar suspension from the distance 

 between the centres of the silk fibres being slightly different from 

 the breadth of the metal hook, the writer proposes to control the 

 distance between the upper ends of the bifilars from the outside by 

 an aperture in a piece of metal, and to control the distance between 



