The compass does not actually point to true north. This is because the magnetic north pole is not quite on the actual northernmost point of Earth – our slushy rock, and its magnetic field, are two separate things. The dynamics of spinning suggest that the magnetic field will be roughly aligned with the true north–south axis, but this is no guarantee.

The difference between magnetic north (where the compass points) and true north is known as magnetic declination, and sailors have known about this for a long time. They also knew the magnetic declination is not equal everywhere on Earth. A ship traveling from Bristol will experience a magnetic declination of nearly zero degrees. When it arrives at Lisboa, it might see a declination of two degrees. Navigating by compass, this is important information, so navigational charts have long had the local magnetic declination printed on them.

Back in the 1500s, the assumption was that the declination did not change. If e.g. Cabo Verde had a declination of eight degrees when it was measured, then eight degrees would be the declination also twenty or fifty years later when another ship passed by.

In the mid 1500s, Borough measured the declination from multiple locations in London and discovered small variation even on a city scale, but the differences were small enough to be averaged out and/or ignored. Then fifty years later, Gunter repeated similar measurements, and discovered that Borough’s measurements were consistently wrong in the same direction. Gunter published his numbers, acknowledged that they were different from Borough’s, and explained that it had to be because Borough was using complicated equipment and had misunderstood the readings, thus reporting incorrect figures.

A decade later, Gellibrand again made similar measurements and found numbers slightly different from both Borough and Gunter. Gellibrand knew Gunter as a very careful person and thus concluded that magnetic north had to be moving, rather than both Borough and Gunter being sloppy with their measurements. This discovery only gained broad acceptance in England at first, where Gellibrand had a good reputation. Outside of England it took longer for it to be accepted.

There are two things I think we should pay attention to in this story:

1. Although measurement error is always the most likely reason for differing outcomes using the same process, we should always be open to questioning the fundamental constants.
2. An idea is not necessarily accepted because it is correct; often acceptance of an idea rests on personal reputation as much as anything else. Not only did it take time for Gellibrand’s discovery to be accepted outside of his circle of reputation, but since Borough didn’t have a high reputation with Gunter, Gunter assumed that Borough sucked at taking measurements.

If we try to keep in mind how the people at the time felt about the magnetic declination and how surely the north pole must have a fixed location, we can start to imagine a world where the fundamental constants of the universe might, indeed, also be changing.