We study here the effect on phage T7 RNA polymerase activity of lac repressor bound downstream of the T7 promoter. When repressor binds in vitro at an operator centered at +13 or +15 with respect to transcription start, it does not prevent initiation, though the transcript yield is reduced. However, the processivity of the polymerase is depressed and transcript extension is blocked at positions +4 and +6, respectively. These results indicate that repressor and polymerase do not simply exclude each other from the promoter. Rather, they would come into steric conflict and compete for establishment or retention of interactions with the same segment of DNA, without this leading to the immediate displacement of either polymerase or repressor. The resulting destabilization of the transcription complex would depress both initiation rate and enzyme processivity. In contrast to the above results, little reduction in runoff transcription is observed when operator is centered at +47. The decreased sensitivity of polymerase to repressor bound at +47 versus +13 or +15 is likely to be due to the higher stability of the elongation complex during the transcription of downstream regions in comparison with the first transcribed nucleotides. We also show that under conditions of leaky repression and with operator centered at +13, a mutant T7 RNA polymerase showing normal promoter affinity but a slower elongation rate is more sensitive to repression than the wild-type enzyme, both in vitro and in vivo. In vitro, this higher sensitivity is largely due to a reduced ability of the mutant to overcome the elongation block at position +4. The parallel between the in vitro and in vivo data suggests that in vivo the repressor also does not prevent polymerase from binding to promoter, but interferes with subsequent steps in initiation and transcript extension, in this case presumably largely extension beyond +4.