In navigating long distances, the compass was as important an instrument for indicating direction as the sand-glass was for marking time. The compass was developed in China around AD 1100, and independently in northern Europe shortly thereafter. By the end of the twelfth century the compass was used for navigation by mariners in the Mediterranean.

As long as voyages in the Atlantic were confined to routes along the west coasts of Africa and Europe, the compass served mariners well in guiding their course. But when they began to venture west across the ocean, it no longer seemed to read correctly. Until then, everyone believed that magnetic north coincided with true north at the pole. They soon found out this was not so. Not only did the two norths not coincide, but the difference between them increased the farther west and north they sailed. In northern waters east of Ireland, the change was in the opposite direction. This discrepancy between magnetic north and geographic north is called magnetic variation, and it varies in different parts of the world according to regional influences.

Until the mid-sixteenth century, mariners were unaware of this variation; consequently, as they sailed west they found their position did not correspond with their location according to the charts. If a navigator departed England and followed his compass due west, his course would take him on a gradual curve south; instead of arriving in Newfoundland as intended, the landfall would be somewhere along the mid-Atlantic coast of the United States.

Having learned that true, or geographic north, differed from magnetic north, instrument makers in some of the northern countries produced compasses in which the compass card (the flye) was mounted on the magnetic needle in alignment with the amount of magnetic variation. Thus, while the needle pointed to magnetic north, the fleur-de-lis on the compass card indicated true north. These were called a "varied compass." By this arrangement the compass had a built-in correction value for the amount of magnetic variation. This was fine, as long as voyages were limited to regions where the amount of variation did not appreciably change. On other compasses, the position of the card on the needle as it pointed to magnetic north could be changed according to the amount of local variation. These were called "true compasses."

On this Admiralty chart published in 1854, magnetic variation is shown on the compass rose in dotted lines, with the fleur-de-lis pointing toward magnetic north. Here, the amount of variation for Portland Harbor is 11°55'. By the year 2000, however, this variation will have changed to 16°52' West. Solid lines on the rose are oriented to geographic, or true north. With this arrangement, unless the navigator used extreme caution it would be easy to confuse the two, and plot an erroneous course. Present-day charts reduce this problem by using two concentric rings; the outer ring for plotting by geographic coordinates, and an inner ring for plotting with the magnetic compass.

Since maps and charts are drawn with meridian lines of true north and south, and parallels of true east and west, the navigator needs to know the amount of magnetic variation at his position. Only then can he correctly plot the ship's course by continually changing the magnetic course heading relative to true north. To determine the amount of variation, the navigator uses a meridianal compass which has a special attachment casting a shadow to true north when the sun reaches its meridian at noon. This arrangement allows him to read the amount of variation directly off the compass card.