The binding curve of myosin subfragment-1 (S1) to F-actin is not a simple hyperbola: at high concentrations of S1 the binding curve can be transformed into a linear plot ('normal' binding), but at small concentrations of S1 the binding complications deform the binding curve and produce nonlinear transforms ('anomalous' binding) [Andreev, O. A., and Borejdo, J. (1992) J. Muscle Res. Cell Motil. 13, 523-533]. This anomalous behavior may result either from the heterogeneity of S1 in regard to light chain isoforms or from the cooperativity between S1's. To distinguish between these possibilities we measured the affinity and the orientation of S1(A1) and S1(A2) with respect to F-actin. Affinity was measured in vitro by ultracentrifugation in the presence of F-actin, and orientation was measured in vivo by a combination of polarization of fluorescence and linear dichroism. We found that both the affinity and the orientation depended on the relative concentration of S1 isomer and actin: when S1 was in excess or was equimolar with actin (filament saturated with S1), each isomer bound F-actin with an affinity of 2 x 106 M-1 and was oriented approximately perpendicularly to the muscle axis. When actin was in excess (filament unsaturated with S1), each isomer bound F- actin with an affinity of 1.2 x 107 M-1 and was oriented more parallel to the muscle axis. S1(A1) and S1(A2) labeled on the light chain had different polarizations when bound to unsaturated filaments but had the same polarizations when bound to saturated filaments. These results excluded heterogeneity as a reason for anomalous binding and suggested that binding occurred with negative cooperativity. We think that the negative cooperativity occurs when saturation of actin filaments with heads leads to the lack of vacant adjacent sites on a filament and a consequent prevention of S1 binding to two actin protomers.