Our present view of the site-specific endonucleases, which appear to be ubiquituous in the prokaryote kingdom, is probably heavily distorted by our search for tools for recombinant DNA technology. Only those enzymes having recognition sequences in the range of three to seven specific bases have been isolated. Of course the usefulness of these enzymes in the analysis of complex genomes, the rise of "reverse genetics", and the immediate breakthroughs in the area of gene expression in eukaryotes, particularly the understanding of tumour virus RNA processing and gene rearrangements in the expression of immunoglobulin genes has dominated the consciousness of the molecular and cell biologists during the last five years. There is a great diversity of staggering, symmetry, asymmetry and degeneration in the recognition sequences found. Taking into account also the genetic data on site-specific recombination and/or DNA degradation suggests that our present collection of endonucleases may only represent a narrow spectrum of specificities on an open-ended scale of complexity. The enzymes, themselves provide a rich pool to be exploited by the biophysicist and the biochemist to probe the subtleties of DNA-protein interaction.