As a preliminary step towards determination of the source of the natural remanence in highly-oxidised basalt lava flows, electron probe microanalysis has been carried out on the two main phases in each of two types of highly-oxidised iron-titanium oxide. The discovery of the source of NRM in these basalts is important because the correlations obtained recently by several workers between high oxidation and reversed polarity in basalts appear to support the possibility of self-reversal, even though some of the rocks studied give independent indications that the reversed magnetizations are due to field reversal.

In one type of grain analysed the probe data are consistent with a titanohematite phase and a pseudobrookite phase. This agrees with previous petrographic data. In the other type of grain the probe data are consistent with one phase of titanohematite but contradict petrographic data by showing that the other phase cannot be titanomagnetite. It is concluded that significant contributions to the NRM could possibly arise from the titanohematite (whose magnetic properties in basalts are little known) or the unindentified phase (unless subsequent work rules this phase out), but not in the pseudobrookite phase. However, the NRM may not lie in the analysed grains at all; and other workers are at present investigating this possibility.