INTENSITY. 60 1 



The bar being placed in a plane perpendicular to the magnetic 

 meridian, the deflections 8, 8' produced at different tempera- 

 tures are determined; we have then 



tan 8 tan 8' 



i -at i - at' 



As the experiment lasts some time, it is well to eliminate 

 variations of declination. The calorimeter is placed on a platform 

 movable about a vertical axis which passes almost through the 

 centre of the magnet. When the frame is turned through 180, the 

 action of the magnet on the declinometer changes its sign, and 

 the difference of the readings corresponds to twice the deflection 

 8. It is simply assumed that the horizontal component does not 

 appreciably vary during the experiments; if that' is the case, the 

 necessary corrections are made from the indications of a variation 

 apparatus. 



The coefficient a may have very different values according to the 

 nature and temper of the steel. Lamont* had conceived the idea of 

 a magnetic system formed of two opposite magnets M and M', the 

 coefficients of which a and a' were very unequal. The magnetic 

 moment M (i at) M' Q (i a't) of the system may be written 

 M M' - (M #- M00')// it is independent of the temperature if 

 the coefficients a and a are inversely as the corresponding moments. 

 We may thus form a resultant magnet compensated for temperature, 

 but the compensation seems difficult to realise. 



1172. On the other hand, the magnetism of a needle becomes 

 weaker in time : at the outset somewhat rapidly in the first few weeks 

 after magnetisation, and then much more slowly. 



This variation is very unequal according to the nature of 

 the magnets, the degree of magnetisation, and the time that 

 has elapsed since this operation was performed. Thus, for seven 

 needles observed during the voyage of the Recherche,^ the mag- 

 netic enfeeblement in a year varied from 0*0025 to 0-0486. The 

 corresponding correction is made for the intermediate observations, 

 assuming that this diminution has been proportional to the time. 



1173. The magnetism of a needle may, lastly, be suddenly 

 modified by a blow, the neighbourhood of a magnet or a piece of 

 iron, or by any other unknown cause. These accidents are avoided, 



* LAMONT. Handbuch of Magnetismus, p. 402. 1867. 



t Voyage de la Recherche en Islande et au Greenland. Magnetisme, Vol. II., 

 p. 326. 



