A transient perturbation applied to a limb held in a given posture can induce oscillations. To restore the initial posture, the neuromuscular system must provide damping, which is the dissipation of the mechanical energy imparted by such a perturbation. Despite their importance, damping properties of the neuromuscular system have been poorly characterized. Accordingly, this paper describes the damping characteristics of the neuromuscular system interacting with inertial loads. To quantitatively examine damping, we coupled simulated inertial loads to surgically isolated, reflexively active soleus muscles in decerebrate cats. A simulated force impulse was applied to the load, causing a muscle stretch, which elicited a reflex response. The resulting deviation from the initial position gave rise to oscillations, which decayed progressively. Damping provided by the neuromuscular system was then calculated from the load kinetics. To help interpret our experimental results, we compared our kinetic measurements with those of an analogous linear viscoelastic system and found that the experimental damping properties differed in two respects. First, the amount of damping was greater for large oscillation amplitudes than for small (damping is independent of amplitude in a linear system). Second, plots of force against length during the induced movements showed that damping was greater for shortening than lengthening movements, reflecting greater effective viscosity during shortening. This again is different from the behavior of a linear system, in which damping effects would be symmetrical. This asymmetric and nonlinear damping behavior appears to be related to both the intrinsic nonlinear mechanical properties of the soleus muscle and to stretch reflex properties. The muscle nonlinearities include a change in muscle force-generating capacity induced by forced lengthening, akin to muscle yield, and the nonlinear force-velocity property of muscle, which is different for lengthening versus shortening. Stretch reflex responses are also known to be asymmetric and amplitude dependent. The finding that damping is greater for larger amplitude motion represents a form of automatic gain adjustment to a larger perturbation. In contrast, because of reduced damping at small amplitudes, smaller oscillations would tend to persist, perhaps contributing to normal or “physiological” tremor. This lack of damping for small amplitudes may represent an acceptable compromise for postural regulation in that there is substantial damping for larger movements, where energy dissipation is more critical. Finally, the directional asymmetry in energy dissipation provided by muscle and reflex properties must be reflected in the neural mechanisms for a stable posture.