376 "WIIALE-I'ISHEIIV. 



may have been seen immediately before setting out 

 on the passage liomeward. As this error is inva-r 

 riably in one direction, the canse must be regu- 

 lar. Ships which have made the Norway land, have 

 been frequently supposed to be steering direct for 

 Shetland. This error has usually been attributed 

 to an easterly current ; if such a cause, however, do 

 operate at all, most certainly it is not the sole cause. 

 There is little dou])t but it is chiefly owing to the 

 increase of the westerly variation of the compass, 

 in consequence of the " local attraction" of the ship 

 on the magnetic needle, when steering on a course 

 to the westward of the magnetic meridian ; and 

 partly to an error in the departure of the Green- 

 land ships, arising from the situation of the south- 

 ern parts of Spitzbergen being laid down 3 or 4 

 degrees too far to the westward in the charts. 



That there is a kind of focus or centre of attrac- 

 tion in every vessel, towards which, in north lati- 

 tude, the north point of the compass is invariably 

 drawn, is a fact which is now sufficiently establish- 

 ed. When the compass, therefore, is placed near 

 the stern of the vessel, which is its usual situation, 

 its north point in the northern hemisphere, is at- 

 tracted towards the centre of the ship. Hence, if 

 the ship's head is directed towards the west, the 

 north point of the compass being deflected forward, 

 produces an error of a quarter to half a point to- 

 wards the north, in very high latitudes, and au 



