Dip and Variation of the Magnetic Needle. 



107 



more than half the stations. When however several observations 

 have been made, the result clearly should not rest solely upon the 

 extreme observations. In such cases I have employed the method 

 exemplified on the fourth page of this article, (p. 96.) In selecting 

 the observations, I have preferred where practicable, to compare 

 those which were made by the same individual. Such observa- 

 tions may be presumed to be affected by a constant error, which 

 will have little influence upon the annual motion. Thus at Bur- 

 lington, Vt. we have (Vol. 34, page 302) a series of observations 

 by Mr. Johnson, from ISIS to 1834, quite consistent with each 

 other ,• but in 1837 another observation by Prof. Benedict, deci- 

 dedly inconsistent with the former. Without stopping to enquire 

 which observations represent most accurately the mean variation 

 at that place, it is obvious that the former are much the best 

 suited to indicate the annual motion. Like considerations have 

 guided my selection at other places. 



