After exposure of mouse embryo cells to ts-P155, an early temperature-sensitive mutant of polyoma virus (Py) that transforms but replicates poorly at 39°, a transformed cell line (Cyp line) was obtained which was propagated at this temperature ( P. Bourgaux, L. Delbecchi, K. K. Y., Yu, E. Herring, and D. Bourgaux-Ramoisy, 1978 Virology 88, 348–360), Whereas viral DNA is not readily detectable in Cyp cells maintained at 39°, it is synthesized in large amounts by these cells from the 30th hour after temperature shift-down. We have thus examined the low molecular weight, covalently closed, cyclic DNA (CCC-DNA) accumulating at 33° in cultures of various clones of Cyp cells isolated in liquid medium, as well as in cells recloned in soft agar. This DNA was found to consist primarily of genome-size Py DNA (P155 DNA I), for all inducible clones analyzed except one. In cells from that clone (C12-22/a1), two discrete species of CCC-DNA molecules were consistently detected at 33°. The predominant species, referred to as Rm I, sedimented in sucrose solution, and migrated in agarose gels, as a molecule of about 4.4 × 10 6 daltons. Under these conditions, the minor species behaved as P155 DNA I. By comparison with other clones, cells from clone C12-22/a1 produced lower amounts of virus following temperature shift-down. When used to infect mouse embryo cells at 33°, the virus produced by C12-22/a1 cells directed the synthesis of DNA indistinguishable from P155 DNA I by restriction enzyme analysis. In contrast, Rm I molecules found in C12-22/a1 cells displayed a distinctive pattern after cleavage by restriction enzymes. Several enzymes were found to cleave P155 DNA and Rm I into the same number of fragments, only one fragment migrating differently in the two digests, as if the Rm I fragment was 1 × 10 6 daltons heavier than the corresponding P155 fragment. In all Rm I digests, the missing—or enlarged—viral fragment was the one overlapping Py DNA map positions 26.6 to 32.6. As compared with P155 DNA, Rm I contains one additional site for HindIII as well as one for BamHI, both located in the 1.0 × 10 6 dalton addition. By the use of the Southern transfer procedure, it was demonstrated that the sequences which are cleaved by either of these enzymes in Py DNA are not present in the addition. It thus appears that Rm I is a molecule comprising possibly the whole of the ts-P155 genome, as well as a piece of 1 × 10 6 daltons of DNA that possibly originates from the host cell.