'...the combination of a clear night sky and a good compass together, many seamen believed, could also tell a ship's longitude. For if a navigator could read the compass and see the stars, he could get his longitude by splitting the distance between the two north poles - the magnetic and the true.

'The compass needle points to the magnetic north pole. The North Star, however, hovers above the actual pole - or close to it. As a ship sails east or west along any given parallel in the northern hemisphere, the navigator can note how the distance between the magnetic and the true pole changes: At certain meridians in the mid-Atlantic the intervening distance looks large, while from certain Pacific vantage points the two poles seem to overlap. (To make a model of this phenomenon, stick a whole clove into a navel orange, about an inch from the navel, and then rotate the orange slowly at eye level.) A chart could be drawn - and many were - linking longitude to the observable distance between magnetic north and true north.

'This so-called magnetic variation method had one distinct advantage over all the astronomical approaches: It did not depend on knowing the the time at two places at once or knowing when a predicted event would occur. No time differences had to be established or subtracted from one another or multiplied by any number of degrees. The relative positions of the magnetic pole and the Pole Star sufficed to give a longitude reading in degrees east and west.'

from The Illustrated Longitude
Dava Sobel
William J. H. Andrewes