Alternatives to the species concept for virus taxonomy

The species concept seems to be inescapably based on the occurrence, in a population, of regular genetic interchange which leads to the formation of common gene pools and on the presence of barriers to interchange which distinguish one gene pool from another. Present knowledge indicates that some types of viruses (e.g. Orthomyxoviridae, Reoviridae) may exchange genes and have a gene-pool population structure; application of the species concept here is legitimate and likely to be feasible in practice. However, it seems that other types of viruses (e.g., tobamoviruses, tombusviruses) do not indulge in regular gene exchange and that common gene pools, distinct from each other, do not occur. Rather, there is clonal multiplication, accumulation of variants, and a fanwise radiation of types with little or no genetic exchange across the rays of the fan. Here, the species concept cannot usefully be applied at either the theoretical or the practical level. If viruses within some major groups are not amenable to being classified and named as species, it follows that attempts to apply the species concept to all viruses should be abandoned. An alternative system that can embrace all cases is already used by plant virologists and merits careful examination by others. Adoption of a 'nonspecies' general approach need not exclude the proposition that in some instances virus species do exist and can be identified and named, as special subsets of the general case. If we cease to aim for the universal pigeonholing of viruses into genera and species, binomial latinized names lose their chief justification.(ABSTRACT TRUNCATED AT 250 WORDS)