Since its introduction in the 1970s, magnetic rubber testing (MRT) has been used successfully to inspect critical high-strength steel aerospace components for surface-breaking fatigue cracks. In the residual-field variant of MRT, inspections are performed following the application and subsequent removal of a magnetic field, relying on the remanent magnetisation of the component to produce crack indications. For certain geometries, residual-field MRT is capable of reliable detection of cracks as small as 0.43 mm (0.017 inch) in surface length. However, for other conditions of specimen geometry and magnet configuration, it was observed that crack indications could be shorter or even absent when using residual-field MRT compared with active-field inspections. This report presents the results of an experimental and theoretical study into the formation of residual-field MRT indications and the rationale for verification of adequate field strength. It is proposed that a combination of inadequate magnetisation, shallow crack depth and the presence of a reversed tangential magnetic field at the surface of the specimen contribute to these observed anomalies in residual-field MRT compared with active-field MRT. The results of a related investigation into active-field MRT are presented in a companion report.