After two decades of investigation the proposed aetiological link between herpes simplex virus type 2 (HSV2) and human carcinoma of the cervix remains unproved. The initial seroepidemiological studies, which were of a retrospective nature, were in favour of the aetiological relationship, whereas the more recent prospective studies suggest that HSV infection is a co-variable. The detection of HSV genes and their products in tumour tissues has been plagued by problems of cross reactivity, but the most recent results indicate that while HSV2 DNA can be detected in a small proportion of cancer tissues, HSV proteins or tumour antigens probably do not persist in the tumours. The same situation applies with respect to in vitro transformation. Although HSV, and isolated DNA fragments from the viral genome, can transform tissue culture cells (mainly of rodent origin) and under some conditions can cause cervical abnormalities (including cancer) in experimental animals, in no instance need the viral genes be maintained in the transformed cells. A number of mechanisms, including viral mutagenesis, have been proposed to explain this unusual feature, but the relevance of this work to the induction of human cervical carcinoma by HSV2 in vivo may be difficult to prove.