Because of the small size and genetic simplicity of small DNA viruses, parvoviruses would appear to be excellent models for studying viral evolution and adaptation. In an earlier publication we hypothesized the evolution of sequences of cellular "junk" DNA into protective interfering transposons. These transposons would interfere with invading pathogenic viruses by competing with the pathogen DNA for replicative enzymes. We speculated that a small, defective parvovirus, the adeno-associated virus (AAV), which usually requires the presence of a pathogenic helper virus to replicate, may have evolved from such a piece of cellular "'junk" DNA. Our theory predicted that AAVs, as a consequence of their defective nature, developed under pressures favoring maintenance of their transposon like qualities. In contrast, disease-causing, autonomous, non-defective parvoviruses such as the BI9 agent of humans and the canine parvovirus, even though their origins may have been in cellular DNA, would appear to have developed under totally different evolutionary pressures. In this paper we will present evidence for a common ancestry for the defective and autonomous parvoviruses and discuss the divergent paths this evolution may have taken in establishing the two genera.