Although DNA sequence homology is believed to be a prerequisite for homologous recombination events in procaryotes and eucaryotes, no systematic study has been done on the minimum amount of homology required for homologous recombination in mammalian cells. We have used simian virus 40-pBR322 hybrid plasmids constructed in vitro as substrates to Tn3 (9, 24), also catalyze recombination events between specific sites containing small regions of homology. The integration and excision of lambdoid phages (15) and the recombination events seen with the phage P1 lox-cre system (10) are also mediated through short homologous sequences. In eucaryotes, the FLP protein of the yeast 2,u plasmid catalyzes site-specific recombination between inverted repeats present on the plasmid (4). Other demonstrations of the importance of short homologies include nonhomologous recombinations lacking site specificity. Nonhomologous recombination between phage A DNA and pBR322 DNA in Escherichia coli may involve 10 to 13 base pairs (bp) of homology (18), whereas deletions in E. coli may be formed by recombination between homologous regions of 6 to 17 bp (1).Because sequence homology is such an important feature of recombination, several investigators have studied the effects of decreasing the length of homology available for specific recombination events. First, Singer et al. (30) determined that a minimum of 50 bp of homology is required by the major recombination pathway of bacteriophage T4. More recently, Gonda and Radding (8) found that E. coli RecA protein efficiently paired molecules that shared 151 bp of homology but failed to pair molecules in which homology was limited to 30 bp. In the present work, we have used previously described pBR322-simian virus 40 (SV40) hybrid plasmids (34) to investigate the minimum amount of homology required for recombination in monkey cells. After transfection of these plasmids into permissive cells, homologous recombination events leading to the formation of infectious SV40 virions can be easily scored by a plaque assay (see below). By progressively decreasing the length of homology from 5,243 to 0 bp, we have measured recombination frequency as a function of the length of homology. Our results show that the steepest drop in recombination frequency occurred between 163 and 214 bp, with lower levels of recombination occurring when there was only 14 bp of homology.