By studying the oldest-known rocks, we can also learn about the strength and direction of the ancient magnetic field. When these rocks were formed and solidified, not only did they freeze the products of radioactive decay so that the rocks could be dated, but they also froze-in the strength and orientation of the magnetic field at the time of solidification. The problem is that the farther back you go in time, the less secure will be your idea of how the rock was oriented in space. Continental drift has occurred so that for the most ancient rocks that are several billion years old, it is hard, if not impossible, to recover what their orientation was on the surface of the Earth. Without this knowledge, you cannot interpret the orientation of the trapped magnetic field in terms of a global picture of what the Earth's magnetic field looked like...if it was shaped like a bar magnet, or where the poles were geographically located.
From samples that go back several hundred million years, geophysicists have learned that the surface field has undergone many dramatic changes. Periodically every 300,000 years, the polarity of this field flips over, and like the solar magnetic field, there is a magnetic cycle to this change. Also, the poles wander considerably over the surface of the Earth, and even now you can measure this 'secular change' of the magnetic pole locations. Although the strength of the main field waxes and wanes in strength during these cycles, and even reaches near-zero for perhaps a few decades or centuries, the maximum strength doesn't seem to have exceeded 1 gauss. Currently it is at about 0.6 gauss, so there are some internal mechanisms that seem to regulate this maximum field strength from cycle to cycle. The oldest magnetic traces we have to work with are about 100 million years old.