We used two large procaryotic sequence data bases called LIB1 (bacteriophage) and LIB2 (E. coli and S. typhimurium) (Stormo et al., 1982a) for the sequences of ribosome binding sites. Twenty-five new sites were included: T4 gene 67, (Völker et al., 1982), T4 lysozyme, IPIII (Owen et al., 1983); E. coli genes: thrB, thrC (Cossart et al., 1981), rpsT (Mackie, 1981), rpsB, tsf (An et al., 1981), ndh (Young et al., 1981), aroH (Zurawski et al., 1981), alaS (Putney et al., (1981), rpoD (Burton et al., 1981), tufA (Yokota et al., 1980), unc1, unc6, uncC, uncB, uncdelta, uncA (Gay and Walker, 1981a,b; Kanazawa et al., 1981), tufB (An and Friesen, 1980), lexA (Horii et al., 1981; Miki et al., 1981; Markham et al., 1981), ampC (Jaurin and Grundström, 1981; Jaurin et al., 1981), EcoRI endonuclease, methylase (Greene et al., 1981; Newman et al., 1981), DHFR (Swift et al., 1981; Zolg and Haumlautnggi, 1981). Sequences other than ribosome binding sites were stored in a library called SITELI. The corresponding Delila instructions were stored as modules in a single file called SITEIN and the Module program was used to extract the instructions for each analysis. The sequences for carAB, argI and argR were from Cunin et al. (1983). The lacZ "pseudo"-operator sequence was from Kalnins et al., (1983). The remaining SITELI sequences described in Results were from the GenBank (TM) magnetic tape, release 14.0, (November 1983) which is available from Bolt Beranek and Newman Inc., Cambridge, Mass.