EBV was the first candidate human tumor virus. It is found in several human cancers, particularly lymphomas and carcinomas, and has potent transforming activity in vitro. Yet the virus persists benignly for the lifetime of more than 90% of the human population. Thus it seems that EBV has the potential to be highly pathogenic yet rarely manifests this potential. Studies over the last several years show this is because the virus actually persists in resting memory B cells and not proliferating cells. EBV needs its growth-promoting ability to gain access to the memory compartment but has evolved to minimize its oncogenic potential. These studies also reveal that the different EBV-associated tumors apparently arise from different and discrete stages in the life cycle of B cells latently infected with EBV. This raises the question of how actively EBV participates in the development of human tumors. Does the virus cause the disease, or is it simply a passenger? In the case of immunoblastic lymphoma in the immunosuppressed patient, the virus almost certainly plays a causative role, but in other cases, such as Burkitt's lymphoma, the contribution of EBV remains less clear.