Many clays and stones contain particles of magnetic oxides of iron. These particles, if heated above their Curie points, which range up to 670° C., lose whatever magnetism they have; and when they cool back through their Curie points, they acquire a new ‘thermoremanent’ magnetization under the influence of the surrounding magnetic field, which generally is the magnetic field of the earth. That field is changing continuously, both in direction and intensity, and the course of its secular change is not yet understood; the change is compound, one factor being the main field, which may be fairly stationary over long periods, and the other being the numerous minor regional fields, which move and alter relatively quickly and largely determine the local variations in the magnetic field. So it is dangerous to extrapolate values for local variations either for more than a century or two in time or for more than five to ten degrees in space. At present the best hope for discovering past changes in the earth's field is from the thermoremanent magnetization of burnt clays and stones, where the date of the burning is reasonably closely fixed from other evidence. Such knowledge is obviously of interest to geophysicists, but for periods and places where the past course of the earth's field has been ascertained, archaeomagnetism—that is the study of the thermoremanent magnetization of archaeological remains—can help archaeologists too. It should be evident on reflection that if an archaeomagnetic specimen is to be useful certain requirements are necessary. First, the locality where it was magnetized must be known. Secondly, for the study of direction, the sample's orientation at the time when it was magnetized must be recorded, so that the inclination [or dip] and declination [or compass bearing] of its own thermoremanent magnetism can be related to the horizontal and to true North respectively.