We describe the characterization of a transposable element from an archaebacterium. The bacteriorhodopsin genes from the wild-type and two mutant Halobacterium halobium strains have been cloned as BamHI fragments in pBR322. The cloned DNA fragments from the two mutants both contain a 1.1-kilobase-pair insertion sequence (ISH1) near the NH2 terminus of the bacteriorhodopsin coding sequence. ISH1 is present in the two mutants in an identical palindromic site but in opposite orientations. The complete sequence of ISH1 has been determined; it is 1,118 nucleotides long, it has 8-base-pair interrupted inverted repeats at the ends, and it duplicates an 8-base-pair (A-G-T-T-A-T-T-G) target sequence upon insertion. As for most eukaryotic and some prokaryotic transposable elements, the sequence of the ISH1 begins with T-G and ends in C-A. ISH1 contains an open reading frame 810 nucleotides long and codes for an RNA approximately 900 nucleotides long. The copy number of ISH1 ranges from one to five or more in different H. halobium strains. In at least one of the strains, one copy of ISH1 is present also on a plasmid DNA.