The CDAN2 gene, responsible for congenital dyserythropoietic anaemia, type II (CDA II), was recently mapped to 20q11.2. We report data on an additional member of a previously studied CDA II family. This member had always been regarded as haematologically normal. Unexpectedly, she had the same microsatellite assortments around the CDAN2 alleles as her three sisters with CDA II. In particular, she was a homozygote for microsatellites D20S863 and D20S841. This prompted an analysis of all facets of her phenotype. The Ham test was negative. The bone marrow smears contained a normal proportion of binucleate erythroblasts. Electron microscopy revealed the absence of extensive stretches of cisternae beneath and parallel to the inner surface of the erythroblast plasma membrane. Proteins of the endoplasmic reticulum, which contaminate the reticulocyte plasma membrane in CDA II patients, were missing. Only the shape of the band 3 peak appeared slightly altered. This case exemplifies that homozygosity (or compound heterozygosity) for a deleterious gene may be silenced, or almost completely silenced. In recessively inherited diseases, suppressed phenotypes tend to be overlooked in siblings where both patients and unaffected individuals are expected.