In this paper, we describe a production procedure of the one-to-one double helical complex of poly(dG)-poly(dC), characterized by a well-defined length (up to 10 kb) and narrow size distribution of molecules. Direct evidence of strands slippage during poly(dG)-poly(dC) synthesis by Klenow exo(-) fragment of polymerase I is obtained by fluorescence resonance energy transfer (FRET). We show that the polymer extension results in an increase in the separation distance between fluorescent dyes attached to 5' ends of the strands in time and, as a result, losing communication between the dyes via FRET. Analysis of the products of the early steps of the synthesis by high-performance liquid chromatography and mass spectroscopy suggest that only one nucleotide is added to each of the strand composing poly(dG)-poly(dC) in the elementary step of the polymer extension. We show that proper pairing of a base at the 3' end of the primer strand with a base in sequence of the template strand is required for initiation of the synthesis. If the 3' end nucleotide in either poly(dG) or poly(dC) strand is substituted for A, the polymer does not grow. Introduction of the T-nucleotide into the complementary strand to permit pairing with A-nucleotide results in the restoration of the synthesis. The data reported here correspond with a slippage model of replication, which includes the formation of loops on the 3' ends of both strands composing poly(dG)-poly(dC) and their migration over long-molecular distances (microm) to 5' ends of the strands.