About 60% of both Duchenne muscular dystrophy (DMD) and Becker muscular dystrophy (BMD) is due to deletions of the dystrophin gene. For cases with a deletion mutation, the "reading frame" hypothesis predicts that BMD patients produce a semifunctional, internally deleted dystrophin protein, whereas DMD patients produce a severely truncated protein that would be unstable. To test the validity of this theory, we analyzed 258 independent deletions at the DMD/BMD locus. The correlation between phenotype and type of deletion mutation is in agreement with the "reading frame" theory in 92% of cases and is of diagnostic and prognostic significance. The distribution and frequency of deletions spanning the entire locus suggests that many "in-frame" deletions of the dystrophin gene are not detected because the individuals bearing them are either asymptomatic or exhibit non-DMD/non-BMD clinical features.