Autosomal dominant facioscapulohumeral muscular dystrophy (FSHD1A) is associated with contractions of the polymorphic D4Z4 repeat array on chromosome 4qter. The disease has a high frequency of new mutations of mitotic origin. Pulsed-field gel electrophoresis-based studies show that mitotic mutations leading to somatic mosaicism occur equally frequently in patients and parents. Nevertheless, somatic mosaicism in FSHD is mainly reported in asymptomatic parents by applying standard Southern analysis after linear gel electrophoresis. Explaining this apparent discrepancy, we here demonstrate that somatic mosaicism in FSHD patients goes largely undetected using the standard diagnostic technique, indicating that linear electrophoresis is unsuitable to identify mosaic patients. As a consequence, the phenotype of mosaic patient's offspring will be underestimated, whereas the recurrence risk in the symptomatic mosaic individuals will be overestimated. Moreover, somatic mosaicism may partly explain the observation of anticipation in de novo kindreds. Therefore, clinicians should always consider pulsed-field gel electrophoresis analysis in de novo FSHD families, in particular when the patient's phenotype is much milder than expected based on D4Z4 length proper.