Paroxysmal nocturnal hemoglobinuria (PNH) is an acquired hemolytic blood disease. The dramatic symptoms that gave the disease its name and the unique nature of the underlying cellular abnormality made the disease for many years a curiosity amongst human blood disorders. Clinical manifestations of PNH are chronic hemolytic anemia with acute exacerbations, bone marrow failure, an increased tendency to thrombosis, and episodes of severe abdominal pain. PNH has a close although not yet fully understood relationship to aplastic anemia (AA), and probably also to acute myeloid leukemia (AML). Blood cells in patients with PNH have a defect in the biosynthesis of a complex glycolipid structure which is a glycosyl phosphatidylinositol (GPI) molecule and serves as an anchor for many surface proteins. Red cells, granulocytes, monocytes, lymphocytes and platelets are therefore deficient in all proteins which are anchored to the cell membrane by such a molecule. Biochemical analysis pinpointed the metabolic block in PNH cells to an early step in the anchor biosynthesis. The block in the biosynthetic pathway is due to the deficiency of a protein called PIG-A. The PIG-A gene maps to the X-chromosome. Cloning of the gene and analysis of the gene in PNH patients lead to the identification of a number of somatic mutations which occur on the active X-chromosome in an early hematopoietic stem cell. All mutations identified thus far inactivate or impair the function of the PIG-A protein and therefore fully explain the deficiency of the missing surface proteins, which in turn explain some of the clinical features of the disease. However, the characterization of the molecular lesion does not explain how the PNH clone can expand to the extent of contributing a substantial proportion of the patient's hematopoiesis. Thus a second factor is needed to explain the pathogenesis of PNH. We hypothesize that this is most likely the coexistence of bone marrow failure that produces paradoxically a growth or survival advantage for the PNH clone. The coexistence of more than one PNH clone in many patients supports this hypothesis and suggests that bone marrow failure is the primary event. The occurrence of the PIG-A mutation causing the absence of GPI-linked proteins on blood cells allows the PNH clone to flourish and to maintain hematopoiesis; thus it seems that the PIG-A mutation is nature's own gene therapy and the price that these patients have to pay is PNH.