We have studied elongation of SV40 DNA F1 by E. coli RNA polymerase looking specifically at the length of the transcript as a function of time. By running the transcription reactions at 18 degrees C with limited enzyme and adding heparin or rifampicin after elongation has started, we have achieved almost exclusive initiation from the SV40 DNA preferred promotor size [Zain, B. S., Weissmann, S. M., Lebowitz, P., & Lewis, A. M., Jr. (1973) J. Virol. 11, 682-693]. In the region within 1500 nucleotides of the initiation we observe nine prominent sites and a number of minor site where hesitation during elongation occurs. The positions of these hesitation points or pause sites are not effected by changes in the salt concentration, the simultaneous lowering of the concentrations of all the NTPs, or by increases in the RNA polymerase concentration, implying that the pause sites are a consequence of the RNA, DNA, and RNA polymerase ternary complex. The pause sites are not an artifact of the lowered temperature (18 degrees C) used in the experiments since they are also observed at 37 degrees C. The first four of these sites have been sequenced by using the 3'-O-methyl analogues of the ribonucleotide triphosphates. We have found no sequence homology between the pause sites. The kinetics of the pause reactions do not fit a first-order model but do correspond to a scheme were continuation through a pause site and termination at a pause site are both represented. For one of the pause sites, the relaxation time for continuation through the pause site was determined to be approximately 2.5 min and for the termination approximately 50 min at 18 degrees C. If the concentration of one of the NTPs is lowered to 10 muM, the strength of a pause site can be increased if that NTP is contained in the pause. Also, minor pause sites are observed at regions in the RNA sequence which are rich in the NTP that has the lowered concentration. When GTP is replaced by ITP during transcription, a new set of pause site quite different from the normal sites of hesitation are observed. The major new pause sites occur at or near sequences in the RNA which are rich in I-U residues preceded by a region rich in C residues. This indicates, as has been previously noted, that sequences where the DNA.RNA hybrid is quite stable followed by a region that is very unstable may cause termination. When BrUTP replaced UTP, very little effect was observed on the pause sites. The addition of p termination factor causes termintion to increase in all the pause sites with a length greater than 300 nucleotides. In the type of experiments performed here, those pause sites had continuation relaxation times greater than 45 s at 37 degrees C. This implies that regardless of the nature of a pause, p will cause at least some termination at all hesitation sites with a relaxation time greater than 45 s. All the results are discussed in terms of a kinetic model for the termination of elongation.