The influence of resistivity and Hall current on the Rayleigh-Taylor problem involving two superposed fluids of finite density in the presence of gravitational and magnetic fields normal to the fluid interface is examined. Unlike the related problem in which the magnetic field is parallel to the interface, it appears that the dispersion relation does not exhibit singular behaviour in the zero resistivity limit. The ‘potentially stable’ situation is considered throughout. The results are compared with earlier ideal and resistive theories, and an apparent anomaly regarding the existence of normal modes in such systems is resolved.

One correction to column 2, table 2, page 455, has already been given (J. Plasma Physics, vol. 4, 1970, p. 643). In addition, the computed downstream convective velocities in columns 2, 3 and 4 of table 2 are given in a co-ordinate system moving at 29·8 km s-1 in the direction opposite planetary motion (-T direction), not in a heliocentric system as stated. The measured T-component and magnitude of downstream velocity in column 1 of table 2 only must be read as -59 ± 14 and 169 ± 20 km s-1, respectively, for a valid comparison with these computed results. However, the calculations have also assumed that the shock is moving at 29·8 km s-1 in the -T direction with respect to Earth (i.e. shock velocity is zero in the heliocentric system). This describes a non-physical situation, which could only apply to one of a series of crossings of a bow shock with an oscillating location. The conclusions of the paper are unchanged if the shock is assumed stationary, and if measured, computed and biased results are all expressed in a heliocentric system.