The site-specific DNA inversion system Cin encoded by the bacteriophage P1 consists of a recombinase, two inverted crossing-over sites and a recombinational enhancer. The latter approximately 75 bp long genetic element is bifunctional due to its location within the 5' part of the cin gene encoding the recombinase. In order to determine the essential nucleotides for each of its two biological functions we randomly mutated the recombinational enhancer sequence sis(P1) and analysed both functions of the mutants obtained. Three distinct regions of this sequence were found to be important for the enhancer activity. One of them occupies the middle third of the enhancer sequence and it can suffer a number of functionally neutral base substitutions, while others are detrimental. The other two regions occupy the two flanking thirds of the enhancer. They coincide with binding sites of the host-coded protein FIS (Factor for Inversion Stimulation) needed for efficient DNA inversion in vitro. These sequences appear to be highly evolved allowing only a few mutations without affecting either of the biological functions. Taking the effect of mutations within these FIS binding sites into account a consensus sequence for the interaction with FIS was compiled. This FIS consensus implies a palindromic structure for the recombinational enhancer. This is in line with the orientation independence of enhancer action with respect to the crossing-over sites.