Average source parameters of the 1968 Meckering, Australia earthquake are obtained by the inversion of body waves. The objectives of the inversion are the elements of the moment tensor and the source-time history. An optimum source depth of 3 km is determined, but because of source complexity the point source assumption fails and the moment tensor obtained at that depth has a large nondouble-couple term, compensated linear vector dipole = 34 per cent. The source parameters of the major double-couple are: strike = 341°; dip = 37°; rake = 61°; and seismic moment = 8.2 × 1025 dyne-cm. The source-time function is of approximately 4 sec duration, with a long rise time and a sharp fall-off. The fault length is constrained on the surface by the observed surface break, and results from vertical displacement modeling suggest a width of approximately 10 km in the middle, assuming a dip of 37°. That restricts the entire faulted area to lie above 6 km depth. Two finite fault models for the earthquake are presented, with rupture initiating at a point (1) near the top of the fault and (2) at the bottom of the fault. Both models produce similar long-period synthetics, but based on the short-period waveforms, model 1 is favored. It is argued that such a rupture process is the most reasonable in this cold shield region.