It has been reported that immunofluorescent labelling of dystrophin in muscle from patients with Becker muscular dystrophy (BMD) is invariably patchy or discontinuous. This observation has led to the suggestion that BMD dystrophin molecules, which are usually smaller than normal due to the presence of "in frame" gene deletions, cannot be assembled into a complete lattice network under the plasma membrane and instead form isolated patches. Our experience with immunoperoxidase labelling of BMD muscle indicates that complete gaps in the reaction around fibres are uncommon. We have therefore compared immunofluorescence and immunoperoxidase labelling patterns on sets of serial sections from 6 BMD patients using a monoclonal antibody to dystrophin. No difference was detected between the two types of label used: the incidence of discontinuous labelling was rare in both cases. We suggest that significantly different patterns of dystrophin labelling may be obtained using different primary antibodies, and that caution needs to be exercised in extrapolating models of structure/function relationships from observations of antibody binding patterns.