We have described studies on the biological fate of a minicircular DNA molecule that is a specific, complex deletion mutant of SV40. When the minicircular DNA alone was used to infect monkey cells, its replication was not detected. However, after infection with the minicircles and SV40 DNA together, incorporation of (3H)thymidine into both species of viral DNA was demonstrated. This finding suggests that circular, duplex viral DNA segments, much smaller than SV40 DNA, are able to be replicated in vivo. Furthermore, 26% of the (3H)thymidine-labeled, superhelical DNA sedimented more rapidly than SV40 DNA I (21S) in neutral sucrose gradients (22S-32S). A similar amount of this rapidly sedimenting DNA was also detected when intact DAR DNA containing the triplication mutant was tested. Cleavage of the purified, rapidly sedimenting DNA with R.EcoRI produced 10.4S segments (one-third the size of unit-length SV40) in addition to full-length linears (14.5S) and a new cleavage product (16.7S). Cleavage of the 21S DNA I molecules also produced 10.4S DNA. These results indicate that the minicircular molecules are amplified in vivo, yielding not only the original triplication mutant but also a heterogeneous population of oligomers in which the 10.4S segment has been reiterated as many as 6 to 9 times. Our studies support the model proposed by Khoury et al. (1974) for the generation of the original DAR triplication mutant. In our experiments, cells were infected with a minicircular DNA molecule formed in vitro, which then served as a precursor in vivo in the formation of trimers and higher oligomers, as predicted by the proposed model. The DAR triplication mutant first appeared after the third passage in primary monkey kidney cells and rapidly became the predominant species in later passages (Fareed et al. 1974)...