According to the Kolodner and Tewari model [Kolodner, R.D. and Tewari, K.K. (1975) Nature, 256, 708.], plastid DNA replication involves displacement-loop and rolling-circle modes of replication, which are initiated on a pair of origins of replication (ori). In accordance with the model, such a pair of oris -oriA and oriB- was described in Nicotiana tabacum [Kunnimalaiyaan, M. and Nielsen B.L. (1997b) Nucl. Acids Res. 25, 3681.]. However, as reported previously, both copies of oriA can be deleted without abolishing replication. Deletion of both oriBs was not found [Mühlbauer, S.K. et al. (2002) Plant J. 32, 175.]. Here we describe new ori inactivation lines, in which one oriB is deleted and the other copy is strongly mutated. In addition, lines oriA and oriB were deleted from the same inverted repeat. In contrast to the expectations of the model, neither oriA nor oriB is essential. Some of the deletions led to reduced growth of plants and reduced plastid DNA copy number in later stages of leaf development. The gross structure of plastid DNA was unchanged; however, the location of the ends of branched plastid DNA complexes was different in the inactivation mutants. Taken together, the results indicate that there are additional mechanisms of plastid DNA replication and/or additional origins of replication. These mechanisms seem to be different from those found in eubacteria, which, according to the endosymbiont theory, are the progenitors of plastids.