The discovery of IS1 as the cause of 'spontaneous' mutations in the bacterium E. colt' marked the (renewed) beginning of the field of transposon research ~. After the discovery of bacterial transposons, the importance of older observations of transposition, such as those made by McClintock on maize, was widely recognized. Many early workers in the field of transposition have gone on to higher things, but some have stayed with IS1, and are making important new discoveries. Recently, two groups 2.3 have discovered a very elegant .and subtle regulation system that takes care of one of the main concerns of a transposable element: how to regulate jumping activity to avoid killing the host. The IS 1 element is 768 bp long, has short inverted repeats (20 bp) and creates a duplication of 9 bp of target DNA upon integration ~. It contains two relevant open reading frames: insA and insB. Mutational analysis has shown that both reading frames are essential for transposition. The InsA protein has been detected, and DNase footprint analysis has shown that it specifically binds to the ends of IS1. Until recently the function of InsB was unclear and the protein had not been detected in vit,o. Another observation added to the mystery: the level of IS 1 transposition is not affected when it is placed downstream of a strong promoter, even thotlgh lnsA levels are elevated. How can insB be essential if its gene product is undetectable, and why does the level of transcription of the transposase operon not affect transposition frequencies? The first indication of a solution to this conundrum came from ',\ork by Sekine and Ohtsubo 2, who synthesized mutant derivatives of IS1. Their hypothesis for the regulation of IS 1 transposition has been confirmed and extended in a recent article by Escoubas et al. from the groups of Galos and of Chandler-~. The following picture is the result of the work of these two groups (Fig. 1). IS/ encodes two proteins: InsA, encOded by the insA gene, and InsAB', encoded by both insA and insB. The latter protein results from a -1 frameshift that occurs in a specific region just upstream of the stop codon of insA. It is not precisely true to say that the protein is encoded by insA and insB, as the fusion protein is encoded by almost the complete insA gene (up to the frameshift site) and insB', which proceeds from the frameshift site in frame with insB (insB was the name assigned to the region of insB' starting at the most upstream ATG sequence; now we know that lnsB probably does not exist per se, we can forget the insB gene). The frameshift event is probably the result of a -1 slip in a slippery region, A<,C (Fig. 1). A potential s tem-loop structure just downstream of this sequence may facilitate the frameshift by introducing a 'pause' in translation in this region. As a consequence, the ratio between the two IS l-encoded gene products, lnsA and InsAB', depends on the frequency of this frameshift. Both proteins bind the ends of IS 1 w:ith their 'InsA' domains: InsAB' catalyses transposition, and lnsA prevents it by competing with the lnsAB' transposase for binding to IS/ DNA ends. The frameshift frequency determines the ratio between InsA and lnsAB', and therefore eventually the transposition frequency as well. In the crucial experiment supporting this model, Sekine and Ohtsubo showed that insertion of one extra A residue into the 'slippe W' region leads to an ahnost 100 times higher level of transposition. Their explanation of this resull is that in this mutant no frameshifting is needed for the synthesis ot InsAB', so that the level of transposase is enhanced (and no repressor is made). The region of translational slipping shows a remarkable