Synopsis This paper covers laboratory testing, both static and cyclic (reversed loading), of reinforced concrete beams aimed at establishing the repair capability and consequences of synthetic resin injection. Beams designed to fail both in tension and in shear were tested to failure, repaired and re-tested. Unidirectional and reversed loading sequencies were applied. The purpose of cyclic loading was to establish the effect of an interposed resin layer within cracked concrete, resulting from ‘wedging’ and possible resin fatigue. The work has established that, if cracks are accessible to the resin (i.e.not less than 0·1 mm wide), failed beams can be reinstated to load-deflection and ultimate load behaviour at least as good as those of the unfailed beam. If; however, bond between reinforcement and concrete has been destroyed, or the reinforcement extended beyond its elastic limit, total reinstatement is doubtful. Extensive shear cracking can be reinstated totally and beams failed in this way are both stiffer and stronger than originally. Within the limits of this test programme, resin within a crack has no deleterious effect upon repaired beam’s behaviour under reversed loading.