It may be possible to estimate the age of a feature that has been fired in place, even though that feature is underground and has not been excavated. This is because the analysis of a magnetic map may reveal the direction of magnetization within that feature. Some failed attempts at this type of dating are described here. The failures have been caused by non-uniform magnetization and perhaps by the rotation of a feature.
Three examples are illustrated. The first is a study of the hearths of prehistoric furnaces in Wales; this example shows the difficulty that can be caused by overlapping anomalies. If the directions of magnetization are not all the same (for example, if one feature has a natural origin), it will probably be impossible to separate the directions. If the directions of magnetization are the same, then the overlap of the anomalies may cause little difficulty.
A second example is the blocks of slag that were formed below early iron furnaces in Denmark. It was found that about half of these blocks have a complex magnetization, apparently caused by lumps of metallic iron near their upper surface; these blocks may contain two or more directions of magnetization. The other half of the blocks have a simple magnetization with a single direction; these blocks appear to be suitable for dating.
The third example is a medieval kiln in Crimea. An analysis of this kiln reveals a direction of magnetization that is over 30° from its correct value; it is possible that this kiln has rotated, for it is on the side of a steep slope.
This dating procedure is at an early stage of development. Its ultimate accuracy is unknown; however, errors in dates of 50–200 years or more will probably remain.