It is shown both theoretically and by examination of the Leiden experimental data on Ce${\mathrm{F}}_3$ and tysonite that in rare earth compounds the Verdet constant $V$ should exhibit the same dependence upon temperature as does the magnetic susceptibility $\chi$, even when Curie's law is not obeyed. In consequence of the similarity in the mode of variation with temperature for $V$ and $\chi$, existing data on the temperature variation of $\chi$ can be considerably enhanced by drawing upon the Faraday measurements of Becquerel and de Haas. In particular, information concerning the $\chi \text{'}\mathrm{s}$ for individual crystallographic axes thus becomes available in tysonite and the tests of calculations of the susceptibility with crystalline potentials become correspondingly severe. The triclinic field proposed by Kramers for tysonite does not appear to account for the behavior of individual principal susceptibilities, although adequate for the mean characteristic of a crystal powder. It is shown that somewhat better agreement with experiment can be achieved by assuming tentatively that the local field has rhombic symmetry, with three possible orientations for the rhombic axes differing from each other by 120°, so that all told the crystal has the desired hexagonal symmetry. Such staggered rhombic fields are compatible with Oftedal's x-ray analysis of the crystallographic structure of tysonite, except that the deviations of the fields from axial character seem rather large.

• Received 12 May 1934

DOI:https://doi.org/10.1103/PhysRev.46.17