Retroviruses have fundamental properties that other virus families do not have. They are RNA viruses which normally convert their nucleic acid to DNA (reverse transcription). They also normally insert this DNA into the genome of the host cell. This gives them two modes of transmission. They can be both infectious agents which spread horisontally, within a generation, like influenza or rotaviruses, but also normal genetic constituents which spread in Mendelian fashion, vertically, between generations. In fact, the reverse transcriptase may have shaped up to 70% of the human genome. However, retroviral genes (human endogenous retroviral sequences, HERVs) only constitute 1% of it. Although most HERVs are damaged by mutations many are expressed as RNA, a few also as proteins and viral particles. The latter are not known to be infectious. Obviously, human ancestors encountered many different exogenous retroviruses, some of which may still be extant. In fact, an exogenous retrovirus related to ERVs was recently discovered. It is the fifth human exogenous retrovirus, human retrovirus 5 (HRV-5). A common pattern is that ERVs first act like "selfish genes" after endogenisation (germ line integration). At this stage, the likelihood that they give rise to disease is greatest. As time passes, HERVs get more or less well assimilated. There are several examples of HERV genes which acquired normal physiological functions. Diseases now investigated for relation to HERVs are multiple sclerosis, schizophrenia and seminoma. It may be significant that viruses which have shown an association to these diseases belong to phylogenetically young sequences which still have some intact viral functions. Although attempts to associate HERVs with disease will be rather complicated, there probably will be high rewards. I think it is worth the effort.