In mdx mice, a model for Duchenne muscular dystrophy, the timing between the replication of myoblasts and their incorporation into myotubes was determined autoradiographically. Thirty-eight mdx mice aged 23 d were injected with tritiated thymidine to label myoblasts replicating early in the dystrophic process. At intervals from 8 h to 30 d after injection the tibialis anterior muscles were removed, processed for autoradiography and analysed for labelled central myonuclei (derived from the progeny of myoblasts which had been labelled at 23 d). At 8 h after injection there were no labelled central myonuclei, showing that the labelled myoblasts had not fused within this time. At 1 d, 2% of central myonuclei were labelled, at 2 d, up to 32% were labelled, at 3 d ∼60% were labelled, and at 4 d the labelling peaked at 74%. In the 27 mice sampled from 5–30 d after injection, the levels of central myonuclear labelling varied enormously: from 1–63%. However, there was a consistent decrease in the numbers of labelled central myonuclei with time. This may have been due to dilution of the relative numbers of labelled myonuclei due to other, nonlabelled, myoblasts replicating after the availability of tritiated thymidine, and fusing. It was also possible that labelled myofibres underwent subsequent necrosis and were eliminated from the muscle. The proposal that a regenerated myofibre can undergo a subsequent cycle of necrosis and regeneration was supported by evidence of some necrotic myofibres with labelled and unlabelled central nuclei. These results have implications for understanding the cellular biology and pathology of dystrophic muscle, particularly in relation to myoblast transfer therapy as a potential treatment of Duchenne muscular dystrophy.