Glycans have the potential to carry more variation than either proteins or nucleic acids. Terminal glycan variation exists both between as well as within species as distant as bacteria and humans. The reasons for this extensive variation are still elusive. This includes the most well known example of polymorphic terminal glycosylation, the ABO histo-blood group family of antigens. A number of associations with infectious diseases have been described, which have become focused on differential adherence by different pathogens to ABO antigens at mucosal surfaces. Histo-blood group antigens can, however, also be carried by virus, as determined by the host cell. When the virus enters a new host, it is likely that it encounters natural antibodies with specificity for the histo-blood group antigens that it carries. We believe that this leads, not only to increased direct neutralisation of the virus, but also to an increased specific immune response to the virus. When taking both of these interactions with pathogens into account in a modelling study, we have shown that the two selective forces together can explain the ABO type frequencies typically seen in human populations and could thus explain how and why these types of terminal glycan polymorphisms have evolved.