444 ACCOUNT OF THE ARCTIC REGIONS. 



ror of observation, at an elevation of 3 feet, would 

 not be a quarter of a mile ; whereas, at the height 

 of 15 feet, the usual situation of the observer in 

 merchant ships, the eiror of observation would 

 amount to 5'. And if the visible horizon were on- 

 ly half a mile distant, the error of observation, at 

 the height of 3 feet, would be still only 2' ; but, at 

 15 feet, 13'. Again, if, by estimating the distance 

 of the visible horizon, we attempt to make allow- 

 ance for the increased dip, according to tables cal- 

 culated for the purpose, we shall find that a small 

 mistake in the estimated distance of the horizon, 

 will be productive of a considerable error in the la- 

 titude, as deduced from an observation from the 

 deck ; but that the error at the elevation of 3 feet 

 will be trifling. Suppose the estimated distance of 

 the horizon to be one-fifth of a mile, but its true 

 distance one-tenth, which is an error that might 

 easily be committed ; the dip of the sea correspond- 

 ing to these two distances, at the height of 3 feet, is 

 respectively, about 8'.4, and 16'.8, the difference 

 8'.4 being the error which would be produced in the 

 latitude ; but, at 15 feet elevation, the dip of the sea, 

 at the two distances of one-fifth and one-tenth of a 

 mile, is 42^, and 85, the difference 42 J miles being 

 the error of the latitude ; and, at 20 feet height of 

 the eye, the error would be 56 miles. Hence, when 

 the sea is so smooth that a boat can be lowered to 

 the water's-edge, and the eye placed within 3 feet of 



