Systematic errors in seismic travel-time data from mis-located shot and receiver positions and/or systematic timing delays can distort considerably the results of tomographic inversions. Standard crosshole tomographic inversion algorithms do not account for systematic errors. Given sufficient data redundancy, the coupled inverse technique allows velocities, source and receiver coordinates and systematic delays to be reliably determined. Efficiency of a coupled inverse technique in crosshole applications is demonstrated through inversions of synthetic data sets and calculations of the formal resolution matrix. Tests revealed that shot delays may trade off with shot coordinates, but neither coordinates nor delays show strong dependencies or correlations with velocity parameters. ¿ American Geophysical Union 1996