Fig. 6. Jumping performance of model 1. (A) Model 1 did not permit rotations other than flexion—extension at the hindlimb joints. In a normal starting position (shown in Fig. 7A), jump distance was very short compared with the real frog (blue versus black recordings in D). Hence, to assess better its jumping potential, model 1 was placed in an unnatural starting position in which the plane of the hindlimbs and the long axis of the pelvis were oriented at 42° to the ground. The purple, orange and green arrows represent the ground reaction forces (GRFs) at the starting position that were produced by a unit extensor torque (1 N m) about the hip, knee and ankle joints, respectively. GRFs are in normalized units (i.e. N per N m of torque), so a torque value of 0.009 N m at the hip will produce 0.15 N of GRF (i.e. 0.009 N mx15 N N^-1 m^-1). At the starting position, a unit hip extensor torque produced the largest propulsive GRF. (B) The path of the center of mass (COM) of the frog during the ground-contact phase of the jump for 100 simulation runs in which the magnitudes of the extensor torques driving each relaxed DOF were randomly varied. The red path in B—D represents the simulation run in which the actual torques produced by the real frog were used to drive the model. The blue path represents a simulation run in which model 1 was placed at a more natural starting position in which the pelvis was oriented at 15° to the ground. (C) The vertical V_V and horizontal V_H velocity of the COM for the red and blue runs did not match the velocity of the real frog (black lines). (D) The predicted jump distances for the red and blue runs were shorter than those for the real frog. (E) The vertical and horizontal velocities were tightly correlated (r²=0.97, P<0.001) during simulations, signifying that take-off angles were the same for each run and equal to the angle of pelvis tilt. This occurs because the vectors of GRFs for a given torque are in the same direction for each joint (see A). (F) Accordingly, the magnitudes of vertical and horizontal velocities were tightly correlated to GRF (r²=0.90, P<0.01 for vertical and r²=0.81, P<0.01 for horizontal velocities).