This review surveys the different experimental approaches which describe the binding of tRNA to mRNA-programmed ribosomes and the control of tRNA selection. This selection is best described by the two-step model proposed by Hopfield and demonstrated by Thompson and his collaborators. The model involves a first control at the initial reversible binding of tRNA to the ribosome and a second control, the proofreading control, which promotes rejection of the incorrect tRNA from a high-energy intermediate during the transition from the initial to the final binding state. Streptomycin, neomycin, and ribosomal fidelity mutations appear to affect both control steps. Their effect can be related to the location of the mutated ribosomal proteins and to the conformational changes induced in the ribosome by the misreading agents. An alteration of the first control probably results from a distortion of the codon-anticodon interaction, while an alteration of the second control may be caused by a change in the association between ribosomal subunits.