Plasmid R1 replication was studied in shifts between two steady states of copy number. The copy number was varied in two ways. First, we utilized the fact that it decreases with increasing growth rate. To minimize the metabolic effects of changes in the growth rate, the downshifts were obtained by adding alpha-methylglucoside to cultures growing in glucose-minimal medium, and the upshifts were obtained by adding glucose to cultures growing in the presence of glucose plus alpha-methylglucoside. Second, we used a temperature-dependent copy mutant of plasmid R1 (pKN301). Plasmid pPK301 shows a threefold higher copy number at 40 than at 30 degrees C. In both types of shift, plasmid replication immediately adjusted to the postshift differential rate. The copy number asymptotically adjusted to the new steady state. Hence, the system that controls plasmid R1 replication sets the frequency of replication without measuring the actual copy number. It has been suggested that plasmid R1 replication is under negative control by an R1-mediated repressor protein. Among the replication control models that involve negative control, the Pritchard inhibitor dilution model, the Sompayrac-Maaløe autorepressor model, and the plasmid lambdadv system all predict gene dose-independent copy number control.