Infection of EBV-negative human B-lymphoma cells of the lines BJAB and Ramos with EBV from P3HR-1 or B95-8 cells resulted in gradual conversion of these cells to EBNA synthesis. Whereas B 95-8 virus-infected cells exhibited a uniform brilliant EBNA fluorescence, two distinct fluorescence patterns were observed in P3HR-1 virus-converted BJAB and Ramos cells, a faint granular and a brilliant fluorescence, with predominance of the faint granular pattern. Cloning of P3HR-1 virus-converted BJAB cells resulted in 20 clones, 11 of them showing the heterogenous parental pattern, six revealing exclusively faint granular EBNA staining, and three with brilliantly stained nuclei, containing also a varying percentage of EBNA-negative cells. Further subcloning of one of the latter clones resulted in 26 subclones with brilliant EBNA expression, always segregating a significant percentage of EBNA-negative cells and one entirely EBNA-negative subclone. Reassociation kinetics did not reveal striking differences in the genome content of clones showing exclusively the faint granular or the brilliant type of EBNA expression. The EBNA-negative clone did not contain detectable amounts of EBV-DNA. Upon superinfection of the converted clones by the parental P3HR-1 virus, a significant increase in EA induction was noted when compared to non-converted BJAB and Ramos cells. This accounted in particular for cells with faint granular EBNA expression. These data support previous interpretations (Fresen and zur Hausen, 1976), suggesting the existence of at least two populations of EBV molecules within P3HR-1 cells. The reason for the apparently labile association of P3HR-1 EBV genomes inducing the brilliant EBNA flourescence in BJAB cells still remains obscure. The possible existence of a "helper" effect, exerted by the faint granular EBNA-inducing virus in stabilizing the persistence of the former, is discussed.