76 Intelligence and Miscellaneous Articles. 



of any pieces of steel may be eliminated by repeated weighings 

 with reversed direction of the magnets ; but the suspicion that tem- 

 porary magnetization, by the influence of the magnets used, came 

 into play could not be altogether set at rest. I hope shortly to 

 repeat these measurements with a balance quite free from iron, 

 and arranged for reading by reflection. I3ut even preliminary 

 experiments show, that in this way values can be obtained which 

 are constant, and accurate to within \ per cent., in which we are 

 quite free from disturbances due to the incessant changes in the 

 terrestrial magnetism. This latter I consider as the most essen- 

 tial advantage of the method. 



A control whether magnets of suitable size undergo any change 

 of their moment, a question which comes into play in many 

 magnetic measurements, can be effected by means of this method 

 with any sufficiently fine balance, even though it is not quite free 

 from iron. Three nearly equal magnetic bars are required for 

 complete determination. Those which I used were hollow tubes 

 10 centim. in length and 1*3 centim. in diameter. This form of 

 tube is convenient, for they can be easily suspended, and retain com- 

 paratively powerful magnetization. In making the measurements, 

 two of the bars are placed in suitable supports and suspended by 

 hooks from the upper bend of the hook which supports the pans. 

 One of the magnets, V, hangs vertically, and the other, H, horizon- 

 tally, and so that the prolongation of its axis cuts the centre of V. 

 When H turns its north pole to V, whose north pole is upper- 

 most, the north pole of H repels that of V upwards, while it 

 draws up the south pole of V. The south pole of H is conversely 

 driven down, and with it the second pan of the balance. For the 

 same load in both pans, V rises therefore and H sinks. The 

 converse is the case when one of the magnets is inverted. The 

 difference of the load which must be applied, the position of H 

 being unchanged, in order to keep the balance in equilibrium, 

 when the north pole of V is first turned upwards and then down- 

 wards, is four times the vertical force with which the magnets act 

 on each other. By inverting H and repeating the weighing, we 

 can eliminate the actions of the permanent magnetization of the 

 steel parts of the balance. We may finally suspend the vertical 

 magnets horizontally and conversely : the same result is then ob- 

 tained if the mean polar distance of the two magnets does not 

 appreciably vary. 



If m and m } are the moments of the two magnets, a the distance 

 of their centres, which is equal to the distance of the terminal 

 knife-edges (30 centim. in the balance which I used), the difference 

 of weight Gr reduced to absolute units of force, when one of the 

 magnets is inverted, is given by the expression 



The value of G for the strongest magnet I have hitherto used was 



