Fig. 1. The distal end of the tibia was attached to the arm of a servo-motor via a stainless steel pin. In both preparations, the angle between the steel pin and tibia was 110° because the body–coxa joint is held at a constant 20° during locomotion (Kram et al., 1997). The hindlimb was chosen due to its vertically oriented joint axes, where vertical deflections of the leg are absorbed by either the body–coxa joint or passive deflection of the exoskeleton. The servo-motor input sinusoidal oscillations from 0.01 to 100 Hz and 0.1 to 1.0 mm and recorded the induced forces. (A) Ventral view of the joint axes of rotation in the meta-thoracic leg. (B) Sagittal view of the joint axes of rotation in the meta-thoracic leg. (C) In the fixed-coxa preparation, the leg was removed and affixed with epoxy resin to 0.95 cm-thick Plexiglas. The tarsus (gray broken line) was removed. (D) In the free-coxa preparation, the cockroach was tethered to a bronze rod via the metanotum. These two preparations were chosen to bound the effect muscle activation at the body–coxa joint could have on leg properties. The tarsus (gray broken line) was removed.