Continuing concerns among virologists are what range of isolates is covered by one virus name and whether such names relate to categories broadly equivalent to biological species of higher organisms. In the potyviruses, tobamoviruses, and probably other groups of plant viruses with monopartite RNA genomes, exchange of genetic information among isolates is not known to occur; our ability to delineate separate viruses apparently depends on the extent to which biological fitness involves several correlated changes in the genomes of variant clones. Although some such correlated changes are found where viruses infect different plant families, intergrading forms occur in many instances where the host ranges of variants overlap. In plant viruses with multipartite RNA or DNA genomes, the extent of gene pools can be assessed from the ability of isolates to form pseudo-recombinants by reassortment of their genome parts. In the nepoviruses, clusters of virus strains sharing a gene pool resemble, but seem more sharply delimited than, the clusters based on nucleotide sequence homology or serological specificity. In the tobraviruses gene pools do not coincide with serological groupings, and in the geminiviruses biologically very distinct entities have much genome homology and are closely related serologically. The biological species concept seems inappropriate or impractical for many plant viruses, and a more flexible and pragmatic approach to assigning virus isolates to nameable categories is advocated.