The physicochemical properties of the interactions of RNA polymerase (RPase) with promoter and nonspecific DNA sequences have been investigated. These show that nonspecific binding is principally an ionic interaction and that promoter binding is more complex, involving nonionic interactions. Nonspecific binding has been shown to be very important in the promoter search, and one-dimensional diffusion can account for the rate at which RPase finds the promoter. Significant differences have been reported in the binding process for various promoters and in the effects of regulatory proteins. Further investigation of these differences will lead to a better understanding of the selectivity and regulation of the initiation process. The pathways of the initiation process have been outlined, by recent studies and considerable progress has been made in determining the rates of interconversion of the intermediate states. A number of questions remain about the detail of initiation and the effects of various parameters on the reactions. Of particular importance is the identification of the point at which the enzyme becomes truly processive. In addition, the step which is rate limiting has not been identified in either the productive or nonproductive process. The mechanistic features of the steps after bond formation are just beginning to yield to investigation. Use of substrate analogs with RPase has led to a picture of the polymerization site according to the ability of the enzyme to incorporate analogs. Base specificity appears to be determined primarily by interaction with the template rather than the enzyme, but the ribose moiety must interact with the site quite specifically. The orientation of the phosphate residues has been determined by NMR, which has also proved to be a valuable probe of the initiation site. At this site base specificity is resident in the enzyme and expressed through the interaction of the base and intrinsic metal, as shown by studies with the Cobalt substituted enzyme. In both initiation and polymerization, the reaction has been shown to proceed by inversion of configuration. Techniques similar to those used for initiation will probably be applied to the polymerization reaction as well, which has not recently received as much attention with respect to mechanism. Functional phenomena such as pausing make the polymerization process particularly promising for producing insight into RPase reactions.