with DMD, and that subject I.2 had transmitted three types of gametes: (1) one bearing the deletion (II.3); (2) one non-deleted and non-recombined (11.2, II.5, II.6); (3) one recombined but non-deleted (11.4). These results show that II.4 (in spite of the intragenic recombination), 11.2, 11.5, and II.6 are in fact not carriers and therefore the prenatal diagnosis performed in 1987 for 11.6 misdiagnosed a normal male fetus (III.3). This is the first case report of erroneous prenatal diagnosis owing to germinal mosaicism overlooked by genomic probes: carrier status had been evaluated with 5' informative markers, whereas the DMD mutation was located towards the 3' end of the gene. This stresses the importance of systematically searching for a DMD associated deletion. In these families all potential carriers should be re-evaluated regardless of previous results, using informative markers or densitometric scanning to investigate the deleted region. Our results also stress that germinal mosaicism must be looked for not only in families with an ascertained de novo mutation but also in those presenting as straightforward familial cases.