Dr. Bowditch, President of the American Academy. xvii 
made, no notice was taken of the daily variation of the needle, which 
sometimes exceeds any of the changes that had been observed in 
New York; besides which, the variation found at the same place 
with different instruments will frequently vary half a degree; and 
by changing the place of an instrument, only a few feet, the same 
effect will sometimes be produced. 
As this subject was of importance in a country like ours, where 
the boundary lines of towns and tracts of land are determined by the 
compass, and was of still greater consequence in our navigation, 
Dr. Bowditch instituted a course of observations in the four differ- 
ent years above mentioned. These observations were conducted 
with great care, and a minute attention to all those accidental 
causes, which might produce irregularities in the needle. And, 
to show the necessity of the utmost caution in such a case, it will 
be sufficient to state, that on one day, when he was observing with 
two excellent theodolites standing in the same place, the variation 
differed above fifty minutes; which was greater than any of the 
changes observed in New York, that had given rise to the dis- 
cussions of the subject. 
Dr. Bowditch came to the conclusion, that the differences noticed 
in New York arose in a great degree from the shortness of the 
needles used by the observers, and, perhaps, in part from the imper- 
fection of their instruments. To obviate these difficulties, he pro- 
cured a needle of twenty-four inches in length, suspended on an 
agate, and mounted with great care in every particular. This in- 
strument was generally used in one of the rooms of his house ; but, 
in order to ascertain whether the building affected the needle, he 
fixed a true meridian line on a table in the garden adjoining his 
house and at the distance of thirty feet from any building; and he 
found, that, even in so short a distance between the two places of 
observation, the variation in the garden was less by 3’ 25” than in 
3 
