While most Tâ«3Ű⏠single-stranded RNA (ssRNA) viruses package in vivo about 3,000 nucleotides (nt), in vitro experiments have demonstrated that a broad range of RNA lengths can be packaged. Under the right solution conditions, for example, cowpea chlorotic mottle virus (CCMV) capsid protein (CP) has been shown to package RNA molecules whose lengths range from 100 to 10,000 nt. Furthermore, in each case it can package the RNA completely, as long as the mass ratio of CP to nucleic acid in the assembly mixture is 6:1 or higher. Yet the packaging efficiencies of the RNAs can differ widely, as we demonstrate by measurements in which two RNAs compete head-to-head for a limited amount of CP. We show that the relative efficiency depends nonmonotonically on the RNA length, with 3,200 nt being optimum for packaging by the Tâ«3Ű⏠capsids preferred by CCMV CP. When two RNAs of the same length-and hence the same charge-compete for CP, differences in packaging efficiency are necessarily due to differences in their secondary structures and/or three-dimensional (3D) sizes. For example, the heterologous RNA1 of brome mosaic virus (BMV) is packaged three times more efficiently by CCMV CP than is RNA1 of CCMV, even though the two RNAs have virtually identical lengths. Finally, we show that in an assembly mixture at neutral pH, CP binds reversibly to the RNA and there is a reversible equilibrium between all the various RNA/CP complexes. At acidic pH, excess protein unbinds from RNA/CP complexes and nucleocapsids form irreversibly.T he remarkable capacity of the capsid protein (CP) of cowpea chlorotic mottle virus (CCMV) to self-assemble in vitro into nanometer-size capsids around a broad range of anionic materials-its own single-stranded RNA (ssRNA) genome (2, 6), heterologous ssRNAs (3,4,9,24), organic polymers such as polystyrene sulfonate (11,16,23), metal oxide particles (19, 28), functionalized gold nanoparticles (5), and negatively charged nanoemulsion droplets (13)-as well as into a wide range of structures in the absence of any polyanions (1,8,26) has spurred interest in the mechanism of assembly. But a fundamental question has rarely or only inadequately been addressed: how efficient are the assembly processes?To answer this question, it is first necessary to formulate a precise definition of efficiency of assembly, along with a procedure for determining it. In some cases, efficiency has been quantified as the fraction of filled capsids in the mixture of filled and empty capsids observed in electron micrographs of the assembly mixture (14), but this depends on the formation of empty capsids under the conditions of filled-capsid assembly. In other cases, the relative efficiency of packaging an RNA molecule has been defined as the ratio of the number of in vivo capsids containing it to the number formed with the genomic RNA (5), when both are present in the same CP-expressing cell; however, this requires additional knowledge about the relative levels of replication of both RNAs.Several assembly experiments, both in vivo and in...