This review of the considerable evidence linking Epstein-Barr virus (EBV) infection to risk and disease progression in multiple sclerosis (MS) builds on the background to the virus and its interactions with the human host available in the online supplement (see supplement, available online only). The evidence for a similarity in the geographic patterns of occurrence of MS and EBV infection (with infectious mononucleosis or EBV specific serology used as surrogate markers), when reviewed critically, is very limited. There is strong evidence however that people with MS are more likely to report a past history of infectious mononucleosis (thought to represent initial EBV infection at an older age), and higher titres of EBV specific antibodies are associated with an increased risk of developing MS. Elevated levels of the latter are apparent many years before MS onset (compared with non-MS controls) and there is a dose-response relationship between MS risk and antibody titre, with antibodies to the EBV nuclear antigen-1 particularly important. The evidence in relation to EBV DNA load in blood or CSF is conflicting, as is that in relation to T cell responses to EBV. Several hypotheses that have been proposed to explain the links between EBV and MS risk are reviewed and gaps requiring further research are identified.