68 



Variation of the Magnetic Needle. 



so that it was necessary to subtract this quantity from all the 

 observations to obtain the true variation. The mean varia- 

 tion for each hour of the day, and for each month of the 

 year, as deduced from these observations, and corrected for 

 the error 3' 25", are given in the following tables. 



The whole number of observations was 5125, and the 

 mean of all made the variation 6° 22' 35" W, which may be 

 assumed as the mean variation at Salem in the year 1810. 



These observations were made about two miles south of 

 the place where the late President Willard observed the va- 

 riation in August 1781 to be 7° 2' W, as may be seen by ex- 

 amining his paper on the subject, in the first volume of the 

 Memoirs of the Academy. The difference of the variation at 

 the two places, at the same time, was probably not more than 

 2'; so that from 1781 to 1810, a- period of twenty nine 

 years, it had decreased about 38 minutes, or T 19" in a 

 year, which is at nearly the usual rate. From which I am in- 

 clined to believe, that the variation has not experienced any 

 change in its direction, in this part of the country, and that 

 the needle continues to approach the true meridian with 

 nearly the same velocity as at the time of the earliest obser- 

 vations on record. 



