The relationship between semitendinosus muscle force and knee joint kinematics during isometric torque production was examined in the frog (Rana pipiens) hindlimb. Passive muscle sarcomere length was monitored by laser diffraction during knee rotation, and joint center of rotation was determined later using principles of rigid body kinematics. Contractile force at the distal tibia, resulting from semitendinosus contraction, was also measured, and, using the kinematic data, a torque vs. joint angle curve constructed. Muscle sarcomere length varied from 3.6 micron at full knee extension to 2.0 micron at full knee flexion. Effective lever arm varied almost as a sine function, with optimal lever arm at 90 degrees of flexion. Joint torque increased linearly from 0 to 140 degrees of flexion and then sharply decreased to 160 degrees of flexion. Thus the optimal joint angle occurred at an angle (140 degrees) that was neither the angle at which muscle force was maximum (160 degrees) nor the angle at which the effective lever arm was maximum (90 degrees). These data indicate that knee torque production in the frog results from the interaction between muscular and joint properties and not either property alone.