A Method for Determining the Individual Accuracy of Astrometric Catalogues

We present a method for determining empirically the actual accuracy (external errors) of proper motions and positions given in astrometric catalogues. The method is based on a comparison of a sufficient number of independent proper motions and positions. The method can be applied, for example, to a comparison of the basic FK5 and the future Hipparcos catalogue with the aim to determine the external errors of the proper motions in both catalogues separately. A few variations of the same basic method are presented: The t_min_ method, the μ_0_ method, and the K method. The expected statistical uncertainty of the derived external errors is discussed, especially as a function of the number of stars used. Numerically simulated catalogues confirm the validity of the results produced by the proposed method. The most disturbing effect may be undetected astrometric binaries.