It is assumed that relaxation of vertebrate skeletal muscle is the result of steric blockage of myosin attachment by troponin-tropomyosin. The extent to which actin binding of nucleotide-free myosin subfragment [S] 1 and actin binding of subfragment 1 containing the ATP analog .beta.,.gamma.-imido-ATP (AMPPNP) was inhibited when troponin-tropomyosin-actin filaments were in the relaxed state was determined. Maintaining the relaxed state requires, in addition to Ca removal, a very low ratio of subfragment 1 to actin because saturation the actin filament with myosin reverses relaxation in the absence of Ca. Actin binding of the 2 subfragment species was inhibited to a different extent when the actin filaments were relaxed. These data are analyzed in terms of the steric model of relaxation under 2 different assumptions, model 1 and model 2. According to model 1, which represents the classical case of competitive inhibition, the S-1 binding site on actin is occupied either by troponin-tropomyosin or by S-1, but never by both simultaneously. The S-1-actin binding constants are unaltered by Ca removal. In terms of model 1 the binding sites for each of the 2 S-1 species must be different and nonoverlapping so that in the relaxed state troponin-tropomyosin blocks a larger fraction of the sites for nucleotide-free than for AMPPNP containing S-1. According to model 2, troponin-tropomyosin in the absence of Ca is bound to all S-1 binding sites at all times, but it occupies them only partially. Each binding site can contain both tropomyosin and S-1 at the same time. With only part of the binding site available to S-1 during relaxation, the number of accessible sites is the same as in the presence of Ca. The binding constant for nucleotide-free subfragment 1 was lowered 10-fold and that for S-1-AMPPNP only 2-3-fold. In terms of model 2 that means that tropomyosin blocks about 15% of the rigor binding site and about 12% of the S-1 AMPPNP binding site; i.e., both binding sites could largely overlap.