Muscle pathology, dystrophin expression and X-inactivation patterns were studied in the muscle of five asymptomatic females heterozygous for deletions in the dystrophin gene (non-manifesting carriers) and five symptomatic carriers (manifesting carriers). Muscle from the non-manifesting carriers showed an increase in the population of centrally nucleated fibres (9.0 +/- 2.8%; controls, 1.4 +/- 0.3%), frequent fibers with abnormally interrupted dystrophin staining (38 +/- 5%), and, in sections from three individuals, small numbers of dystrophin-negative fibers (1-4%). The amount of dystrophin measured by immunoblotting was reduced to 64 +/- 5% (P < 0.001 n = 5) of normal. The pattern of X-inactivation in muscle DNA was non-biased (50: 50-60: 40) in all cases. In the manifesting carriers both highly biased (90: 10) and non-biased patterns of X-inactivation were found, but no consistent relationship was apparent between the patterns of X-inactivation and the proportions of dystrophin-negative fibers. We conclude from studies of the non-manifesting carriers that the proportion of residual dystrophin is similar to the relative activation in muscle of the X-chromosome carrying the wild-type allele. Extreme bias of X-inactivation can be associated with early clinical symptoms and severe pathology. However, as non-manifesting and some manifesting adult carriers had identical patterns of X-inactivation, abnormalities in the distribution of dystrophin, as well as overall levels of expression, may be important for the development of myopathic pathology.