Travel times between sources and stations in the Australian shield are used to develop P- and S-wave velocity models for the Precambrian shield of western and central Australia. Errors in travel times are minimized by using sources for which locations and origin times are constrained by local data or fault scarps. Apparent velocities between 1500 and 2200 km for P (8.8 km/sec) and S (4.8 km/sec) require a high-velocity lid, and the maximum range of observation of this branch (2200 km) requires that the lid extend to a depth of 205 ± 20 km. Because of the distribution of sources and receivers, there are no observations from the shield at ranges greater than 2300 km, so we use times from stations in Phanerozoic eastern Australia and published global summary times as a guide for interpretation at greater ranges. Allowing for known differences in teleseismic travel-time residuals between the Precambrian shield and eastern Australian, the P times are best fit by a model that has a small velocity inversion at the base of the lid, and the S times are best fit by a model with a more substantial reduction in velocities. However, both P- and S-wave travel times can be satisfied with models that have velocities below 210 km that are concordant with global averages such as IASP91. Thus, the velocities inferred beneath the lid are low relative to those of the lid but not necessarily low relative to global averages.