A critical step in the life-cycle of retroviruses is the integration of the double-stranded DNA copy of their RNA genome into the genome of the host cell (1 ). Although the provirus DNA sequences flanking the site of integration are precisely determined and characteristic of each virus (2 ), the vast majority of the integrations are the product of nonhomologous recomination events, resulting in the pseudorandom integration of the provirus into the host-cell genome (3 ,4 ). Retroviruses can therefore act as agents of insertional mutagenesis. The insertion of the provirus into the genome could, in principle, result in either gene activation or gene inactivation. The insertional inactivation would be the result of provirus integration within the coding or regulatory sequences of the gene of interest, thus disrupting the expression of a functional gene product (5 –13 ). Insertional activation could be the product of the integration of viral-promoter enhancer elements, contained within the long terminal repeat (LTR) sequences in the vicinity of a silent gene, resulting in the increased transcription of that gene (14 –26 ). In addition to these direct as-acting effects, there can be indirect trans -regulatory effects resulting from the presence of the viral genome within the cell, but irrespective of its position of integration. This could be the product of genes or other regulatory elements encoded by the virus. The position-independent effect(s) of retrovirus integration would be easy to identify; they would be present in all cells or in a vastly larger number of cells than would be compatible with low-frequency integrations into specific genomic domains.