600 MAGNETIC FIELD. 



1170. INTENSITY. The method of oscillations (1137) has often 

 been applied to compare the horizontal component of the terrestrial 

 field. In this case, the correction for the induced magnetism may 

 generally be neglected, but the variations in the magnetic moment of 

 the magnet with the needle should be allowed for. The observation 

 is made at the temperature /; let M be the magnetic moment at 

 zero, and a the coefficient of variation, we have 



or 



N 2 



7T 2 K - = M H. 

 i -at 



Hence, for each observation the square of the number of oscil- 

 lations N 2 , should be divided by the binomial for the corresponding 

 temperature. 



The value of a is always less than a thousandth, so that it is un- 

 necessary to know very exactly the temperature of observation. 



1171. This coefficient should be determined directly for each 

 needle. The numbers of oscillations N and N' for the two different 

 temperatures / and /' give the value of a by the equation 



C. L, CORY. 



N 2 N' 2 



i - at i - at' 



The experiment is, however, rather delicate, for if we do not 

 allow for the changes of the surrounding temperature, it is difficult 

 to know the temperature of a needle which oscillates freely in a 

 heated space. 



It is better to arrange the experiment so as to observe directly 

 variations in the magnetic field of the bar considered as to its action 

 on another needle which serves as declinometer. 



A sort of calorimeter is formed of two tubes of Bohemian glass, 

 fitted in each other, which leave a stationary layer of air between 

 them. The bar is contained in a glass or metal tube and placed in 

 the centre of the inner vessel, which is filled with water and closed 

 by a wooden lid. Hot water at a temperature of about 60 is first 

 introduced; the liquid is then agitated by blowing air in at each 

 observation. In these conditions the cooling is very slow, and 

 the liquid takes more than six hours to revert to the surrounding 

 temperature. 



