Extensor digitorum longus muscles of normal mice (C57BL/10ScSn hereafter called C57) were orthotopically transplanted into dystrophin-deficient mice (mdx) and reciprocally, mdx Extensor digitorum longus muscles were transplanted into C57 mice. After an initial phase of degeneration, transplanted muscles regenerate nearly completely, as evaluated from the maximum isometric force of muscles isolated 60 days after the surgery. In other similar experiments, instead of isolating the grafted muscles, we excised the antero-external muscles of the leg, including the grafted muscle. Cryostat cross-sections at three levels along the muscles were immunostained with an anti-dystrophin antibody. No muscle cells of dystrophin-deficient muscles grafted into normal mice took the antibody except a few 'revertant' fibres, while all the muscle cells of the normal host were immunostained. Reciprocally, all the muscles cells of normal grafts were stained, whilst no antibody stained the cells of the surrounding muscles of the dystrophin-deficient host. These experiments show that very few if any of the myoblasts or muscle precursor cells, active during the regeneration of grafted muscle, migrate into the adjacent muscles. These results could be explained by the absence, in our work, of injuries of the grafted and adjacent host muscles epimysium and the absence of extensive inflammatory reactions. This lack of myoblast mobility suggest that when myoblast transfer is applied to muscle therapy, it will be necessary to inject myoblasts within each muscle to obtain an efficient treatment.