8 DETERMINATION OF THE TEMPERATURE CO-EFFICIENT 



For January, 1843, . 

 For February, 1843, 

 For March, 1843, . 



803". 1 at 59.2 

 798".9at51.9 

 815". 1 at 48. 1 



On the 15th of April, 1843, the instrument was carefully examined and found in 

 adjustment. 



At 6 1 ' 50 m on May 4, 1843, the bifilar was disturbed, but readjusted on May 5, 

 before the regular observation at 2 h 21 P. M. A correction of 16 divisions 

 during the interval is to be applied to the readings. After this date the instrument 

 remained undisturbed. 



We have, therefore, for discussion the following continuous series of monthly 

 means of the readings of the bifilar magnetometer with its corresponding mean 

 temperature. The series extends over five years and one month. To obtain a 

 better view of the series, the correction of 900 divisions for the first twelve 

 months has been applied, it gives a negative value to the June mean of 1840. 



Under the supposition of a uniform progression in the change of the mean 

 monthly readings (due to change in the horizontal force and loss of magnetism of 

 the bar) the bifilar readings for a given period may be represented by the form : 



B = B m + Ae x + A* y 



where B m a mean bifilar reading for the period. 

 x the change during a period. 

 y the change in the reading due to a change of 1 Fahr. 



