We have constructed the PRM promoter of phage lambda and eight variants, which represents intermediates in the conversion of this promoter to one that has complete homology to the consensus sequences in the -10 and -35 regions. The in vivo activity of these promoters was determined from the beta-galactosidase or galactokinase activities in cells harboring plasmids, in which the cloned promoters were driving the expression of these genes. Additionally, the kinetics of the interaction of Escherichia coli RNA polymerase with the same series of promoters was measured as a function of RNA polymerase concentration. This allowed the overall rate of functional or open complex formation to be dissected into the equilibrium constant for binding of the polymerase to form a closed promoter complex and the rate of subsequent isomerization to yield the open complex. The following conclusions can be drawn from the data presented: (1) The consensus sequence is optimal for promoter function both in vivo and in vitro. (2) Alterations of the -10 and -35 regions have similar effects on the kinetics of RNA polymerase binding in vitro; with one exception, the same holds for promoter activity in vivo. (3) The in vitro rate of RNA polymerase binding to a promoter is solely determined by the number of positions at which its -10 and -35 regions match the consensus promoter sequence. The functional importance of a match does not appear to be determined by the sequence conservation at the particular position. (4) The extent to which a particular base change affects the kinetic parameters depends on the sequence of the promoter into which it is introduced.