Much of our current understanding of the replication of hepatitis delta virus (HDV) has been based on transfection studies using plasmid-DNA-based expression of HDV RNA. Although this approach is simple and has provided much useful information, there has always been the possibility that this highly artificial system can generate artifacts unrelated to true HDV RNA replication, as there are no DNA intermediates in natural HDV infections. For example, it is not always certain whether any feature observed in these systems is representative of true HDV RNA replication or reflective of artificial DNA-templated HDV RNA transcription. Moreover, HDV cDNA itself has been shown to contain endogenous promoters that may lead to synthesis of additional (unnatural) transcripts (1 ). A more reliable approach is by direct transfection of HDV RNA. However, this procedure requires that an independent source of small hepatitis delta antigen (S-HDAg) also be provided, presumably to transport the transfected HDV RNA to the nucleus (2 ), the site of HDV RNA replication, and/or to initiate RNA replication. For this reason, HDV RNA transfection was first done in cell lines that stably expressed S-HDAg from an integrated cDNA copy (3 ,4 ), which, however, still involves some HDV cDNA-templated RNA transcription. The next attempt involved transfection of HDV RNA that had been incubated with recombinant S-HDAg to allow the formation of high-molecular-weight ribonucleoprotein complexes (5 ,6 ). Although this approach most closely mimics natural HDV infections, it is very inefficient, requires highly purified S-HDAg, and can be used only for the transfection of genomic-sense HDV RNA (6 ).