A Rhesus D (RhD) red blood cell phenotype with a weak expression of the D antigen occurs in 0.2% to 1% of whites and is called weak D, formerly Du. Red blood cells of weak D phenotype have a much reduced number of presumably complete D antigens that were repeatedly reported to carry the amino acid sequence of the regular RhD protein. The molecular cause of weak D was unknown. To evaluate the molecular cause of weak D, we devised a method to sequence all 10 RHD exons. Among weak D samples, we found a total of 16 different molecular weak D types plus two alleles characteristic of partial D. The amino acid substitutions of weak D types were located in intracellular and transmembraneous protein segments and clustered in four regions of the protein (amino acid positions 2 to 13, around 149, 179 to 225, and 267 to 397). Based on sequencing, polymerase chain reaction-restriction fragment length polymorphism and polymerase chain reaction using sequence-specific priming, none of 161 weak D samples investigated showed a normal RHD exon sequence. We concluded, that in contrast to the current published dogma most, if not all, weak D phenotypes carry altered RhD proteins, suggesting a causal relationship. Our results showed means to specifically detect and to classify weak D. The genotyping of weak D may guide Rhesus negative transfusion policy for such molecular weak D types that were prone to develop anti-D.