Wave propagation directions of post-midnight chorus in the near-equatorial region at L shells of 6 to 7 have been determined from OGO 5 triaxial search coil magnetometer data. Several methods were used to estimate the wave normal directions: minimum variance, imaginary part of the cross-spectral matrix, eigenvector of Hermitian cross-spectral matrix, and fit of dispersion relations for one-wave and two- wave models to cross-spectral matrix. Comparisons of the methods and results are discussed in detail. Wave propagation at all frequencies within chorus tones was found to occur most frequently along the magnetic field with median and average cone angles of 9.1¿ and 12.2AmD, respectively. All methods of analysis gave similar results. It is concluded that the wave growth is maximum for waves propagating parallel to B. Our results are consistent with emission in a narrow beam but not as a broad angular distyribution. The two wave propagation model typically resulted in negligible improvements in the residual beyond that obtained by assuming that only one wave was present. In only a few cases did substantial improvement result. These cases were unusual periods when it was obvious that simultaneous rising and falling tone emissions were present and crossing. It is concluded that in most cases only one chorus wave was present. In some cases the two-wave model gives misleading indications of waves with large energy fluxes propagating at very oblique angles near the resonance cone. This is because the wave group velocity goes to zero near the resonance cone, and if electric field data are not used, then a small amount of magnetic noise is modeled with a large wave energy flux. Some examples have been found of a lower frequency burst propagating almost exactly along the magnetic field followed by a higher frequency burst frequency burst travelling at large angles to the magnetic field. These paired bursts are quasi-periodic with a delay of about 6 s between pairs of bursts, the phenomenons is presumed to be associated with pulsating auroral patches. The physical implications are discussed.