76 



MR BROUN ON THE BALANCE MAGNETOMETER, 



The differences, when 1^-^%^, were too irregular and too few, on some days, 

 to give a good value of a. 



17. It is very rare that periods of such magnitude can be found free from 

 considerable irregularities. In general, however, it is conceived that smaller pe- 

 riods will give equally good results, and by a shorter method. 



If we consider the equations 



— y^ yv*r _ _ 



r a 



.ypH 





ra 



h*r 



' 'z'+r 



■ h-i' 



it is obvious, that if the temperatures rise and fall considerably throughout the 

 period selected, and no attention be paid to the sign of t^—tj,„ in the summations 

 of the differences, the coefficients of a will nearly destroy each other. 



18. In the following cases the sums for each day have been compared with the 

 sums of all the days after it in the period selected. By this means irregularities 

 in the force upon any day have their effect on the final result to a considerable 

 extent destroyed, as it is probable that the results will be as much too great in 

 some cases, as they are too small in others. 



The whole differences have been summed without regard to days, and the 

 signs oitf—tp^r have been disregarded. 



The equation is, therefore, simply 



l±M-„ 



Table V. 



Determinations of the Temperature Correction for the Balance Magnetometer, 

 from comparisons of the Daily Observations at different periods. 



From the above table it would appear that neither the removal of the needle 

 and readjustment, nor the alteration of the time of vibration, has affected the 

 temperature correction. 



