278 A. TANAKA.DATE 



sional]y as with other mstruments of a simihir description. 



There is one inconvenience however in this system, namely, that 

 the four magnets placed round the current have all their axes per- 

 pendicular to the direction of the current and consequently they are 

 subject to the effect of induction. The induction will be positive for 

 th3 two approaching mignets (that is their moments will be increased), 

 and negative for the receding two. The total etfect on the system 

 will be nearly nil for small currents; but with large currents, the 

 magnets may be permanently affected and may even be reversed in 

 maornetization. AYe can o-et rid of this inconvenience by fixing the 



O O JO 



four magnets to an axis Avhich is perpendicular to each axis of the 

 magnets as in Fig. 3. It will ^^■ o 



be seen that the system is 

 exactly the same as the previ- 

 ous one, if the four magnets 

 be of equal moments or if they 

 satisfy the condition ^M = 

 with respect to the fixed axis. 

 The former condition we can 



not hope to attain in practice, but the latter can be approximated to 

 thus : — 



Make some 10 or 20 magnets of the same steel wire and leave 

 them for some months or even for years, until the time-rate of the 

 fall of moments becomes insensible. Then measure the moment of 

 each magnet by any of the ordinary magnetometric methods, such as 

 the deflection of a mirror magnetometer at a constant distance from 

 each magnet, a knowledge of the relative strengths only being- 

 essential. The magnets I used in one arrangement had the following 

 relative moments. 



(a) [h) (c) {d) 



313.3 311.7 312.U 313.0 



