Repeated first-order triangulation measurements made from 15 monuments in central and southern California are used to explore the structure of the crust and upper mantle beneath southern California, as well as to estimate the distribution of slip during the great 1857 earthquake. The monuments are divided into two subnetworks, with repeated observations covering the respective time intervals from 1884 to 1923 and from 1898 to 1923. The angle changes occurring within these times are modeled as a sum of two contributions: postseismic displacements following the 1857 earthquake and plate tectonic loading. The postseismic displacements have been calculated using a elastic / viscoelastic coupling model that, for any fault geometry, yields the exact displacements on a spherically stratified Earth. The southern California crust and upper mantle are modeled as an upper elastic plate underlain by a viscoelastic asthenosphere. For several different combinations of structural parameters, inversions are performed for the slip distribution of the 1857 earthquake. In general, the best fits to the data are achieved with a triangular slip distribution, an elastic plate thickness of 16 km, a crustal thickness of 16 to 33 km, and a mantle viscosity of 0.4 to 0.8 · 1019 Pa-s. Although these inversions identify the best rheological models, they lead to observed residuals that are inconsistent with a normally distributed data set. A preferred model of slip is constructed that accounts for the budget of total slip during the earthquake cycle and satisfies a normally distributed data set. The preferred slip distribution is similar in shape to the mapped surface slip distribution, but is greater in amplitude by a factor of 1.5 to 2. In addition, the San Jacinto fault, which has no known surface slip associated with the 1857 earthquake, is inferred to have accommodated 1 to 3 m of slip during the 1857 earthquake and appears to be the favored site for tectonic strain accumulation since 1857. Inferred fault slip exceeding known surface slip implies either unrecognized secondary faulting or a slip deficit at the surface relative to the slip at depth.