Measurements of the yield behaviour of a cohesive bulk solid in the Jenike shear tester are normally done in two steps. First, pre-shearing to a steady state yielding, and secondly, shearing under a lower normal stress to establish the shear stress at incipient failure. Due to the limited strain of the sample in this apparatus, the material must be consolidated in an earlier stage. This is done by the twisting action of a loaded cover, at which the vertical load and the amount of twists are chosen such that a suitable stress—strain curve will be obtained in the pre-shearing stage. The ratio n between the vertical load during consolidation and pre-shearing, respectively, can have a strong impact on the results. In this article, results of measurements with various n-values are given, leading to a variation in shear stress levels for both steady state and incipient failure. This can be explained by variations in the real normal stresses on the shear plane, which can substantially deviate from the externally applied normal stresses, due to contact between ring and base. To establish the real normal stress on the plane of shear, a somewhat modified shear tester was used. Furthermore, measurements were performed where the contact between ring and base was avoided by a removeable pair of half-rings. The results of those measurements and the implications for the derived quantities are amply discussed, resulting in some recommendations for a more reliable testing procedure.