The termination/release phase of transcription must involve at least three major steps: cessation of elongation; release of the transcript; and release of the RNA polymerase. We have devised a novel method for measuring the rate of Escherichia coli RNA polymerase release during transcription termination. The method is based on a kinetic analysis of the rate of RNA synthesis during steady-state transcription. Using this method with defined transcription units, we have found that RNA polymerase release occurs rapidly from several rho-independent terminators. Enzyme release from the T7 early terminator occurs within 13(+/- 3) seconds of the cessation of elongation. Neither nusA protein nor supercoiling of the DNA template affects the rate of enzyme release. However, addition of excess sigma factor significantly increases the rate of enzyme recycling during the steady state. Since added sigma factor does not alter the rates of initiation and elongation by E. coli RNA polymerase holoenzyme, it appears that sigma factor stimulates one or more steps in the termination/release process and reduces the rate of enzyme release to a few seconds. We present evidence that suggests sigma may be directly involved in catalyzing release of the core RNA polymerase from the DNA template during transcription termination. The rapid rates of enzyme release we measure make it difficult to be certain of the exact pathway of events that occur in the termination/release phase of transcription. The most plausible pathway involves initial release of the RNA transcript followed by release of core RNA polymerase from the DNA. Studies on the properties of core polymerase-RNA complexes indicate that core polymerase and the RNA transcript probably do not dissociate as a complex from the terminator. Furthermore, these core-RNA complexes are too stable to represent significant intermediates in the termination/release pathway, at least in the early steps of the reaction.