The Helmholtz coil system of the MOLSPIN anisotropy delineator is shown to be partly sensitive to the specimen shape, particularly if the sample departs from the optimum shape. The effect is found to be most marked when measuring weakly anisotropic rocks which have an anisotropy below about 13 × 10−12 m3 and susceptibility less than 50 × 10−10 m3. However, comparison of MOLSPIN data to measurements undertaken on low and high field torque magnetometers, suggest the results are not unduly affected by sample shape if the rock anisotropy signal is sufficient to dominate this spurious effect. Some suggestions are also made for minimising noise factors and thus improving instrument sensitivity.