The theoretical background for the two most common indirect rock tensile strength measurements, namely the point load test and the Brazilian test, is discussed. The exact theoretical stress field variation within the loaded specimens has not been evaluated earlier in the literature, strangely enough. It is shown that in the point load test this variation is much larger than earlier studies have shown, and that these earlier studies actually have taken the minimum tensile stress along the loaded diameter as a measure for t rock tensile strength, a rather unusual way of interpreting data. In the case of the Brazilian test, the three-dimensional ‘correction’ to the two-dimensional theoretical solution is shown to be considerable, even for rather thin samples. These theoretical remarks show that the interpretation of indirect tensile strength test data for rock materials is still more difficult than previously believed.