The Do(a) antigen was discovered after I began my career in immunohematology and I have been fortunate to be involved in several fascinating discoveries in the Dombrock blood group system. The Do(a) antigen and its antithetical antigen, Do(b), have a prevalence that makes them useful as genetic markers. The paucity of reliable anti-Do(a) and anti-Do(b) has prevented this potential from being realized; however, our ability to type for DO alleles at the DNA level has made it possible to test cohorts from different populations. In 1992, the Dombrock blood group system was expanded to include three phenotypically related antigens, Gy(a), Hy, and Jo(a), when it was discovered that the Gy(a-) phenotype was the null of the Dombrock system. Based on the knowledge that the Dombrock glycoprotein is attached to the RBC membrane via a glycosylphosphatidylinositol linkage and subsequent to the assignment of the corresponding gene to the short arm of chromosome 12, expressed sequence tags from terminally differentiating human erythroid cells were analyzed in silico to identify the DO gene. This allowed determination of the molecular basis of the various Do phenotypes and the realization that DO is identical to the gene encoding a mono-ADP-ribosyltransferase, ART4. No enzymatic activity in RBCs has been demonstrated and the function of this glycoprotein, on the outside surface of RBCs, has yet to be determined. This review is a synthesis of our current knowledge of the Dombrock blood group system.