The experimental evidence available for animal and plant RNA viruses, as well as other RNA genetic elements (viroids, satellites, retroelements, etc.), reinforces the view that many different types of genetic alterations may occur during RNA genome replication. This is fundamentally because of infidelity of genome replication and large population sizes. Homologous and heterologous recombination, as well as gene reassortments occur frequently during replication of retroviruses and most riboviruses, especially those that use enzymes with limited processivity. Following the generation of variant genomes, selection, which is dependent on environmental parameters in ways that are poorly understood, sorts out those genome fits enough to generate viable quasispecies. Chance events can also be destabilizing, as illustrated by recent results on fitness loss and other phenotypic changes accompanying bottleneck transmission. Variation, selection, and random sampling of genomes occur continuously and unavoidably during virus evolution. Evolution of RNA viruses is largely unpredictable because of the stochastic nature of mutation and recombination events, as well as the subtle effects of chance transmission events and host/environmental factors. Among environmental factors, alterations resulting from human intervention (deforestation, agricultural activities, global climatic changes, etc.) may alter dispersal patterns and provide new adaptive possibilities to viral quasispecies. Current understanding of RNA virus evolution suggests several strategies to control and diagnose viral diseases. The new generation of chemically defined vaccines and diagnostic reagents (monoclonal antibodies, peptide antigens, oligonucleotides for polymerase chain reaction amplification, etc.) may be adequate to prevent disease and detect some or even most of the circulating quasispecies of any given RNA pathogen. However, the dynamics of viral quasispecies mandate careful consideration of those reagents to be incorporated into diagnostic kits. Broadening diagnosis without jeopardizing specificity of detection will be challenging. There is a finite probability (impossible to quantify at present) that a defined vaccine may promote selection of escape mutants or a particular diagnostic kit may fail to detect a viral pathogen. Of particular concern are the potential long-term effects of weak selective pressures that may initially go unnoticed. Variant viruses resulting from evolutionary pressure imposed by vaccines or drugs may insidiously and gradually replace previous quasispecies. The great potential for variation and phenotypic diversity of some important RNA virus pathogens (human immunodeficiency virus, the hepatitis viruses, the newly recognized human hantaviruses, etc.) has become clear. Prevention and therapy should rely on multicomponent vaccines and antiviral agents to address the complexity of RNA quasispecies mutant spectra.(ABSTRACT TRUNCATED AT 400 WORDS)