Intrinsic muscle properties, such as dependence of force on length and velocity, have been hypothesized to help organisms to respond to destabilizing conditions predictably and rapidly, but relatively few experiments have tested this hypothesis in a quantitative way. These stabilizing (or destabilizing) effects are particularly challenging to understand during rhythmic muscle activity, and most behavior involves rhythmic motion. Our study uses a modified work loop protocol to measure how muscle in the silver lamprey, Ichthyomyzon unicuspis, responds to perturbations during swimming. We use these data to develop a model to predict the muscle function. A section of axial musculature was dissected and used to perform standard in vitro work loops. A baseline sinusoidal length change was imposed on the muscle and force was measured. To examine the effects of muscle activation, the muscle was stimulated at different phases during the cycle. Next, we added a pseudorandom stimulus composed of sums of small sinusoidal perturbations at multiple frequencies to the baseline oscillation. Using a new system identification technique based on harmonic transfer functions, we can characterize how the muscle responds to the perturbations and how the response depends on the phase of the baseline oscillation. The model makes testable predictions about how the muscle would respond to any small perturbation at any phase. Preliminary results indicate that both the effective stiffness and damping properties of the muscle change within a cycle. Moreover, these properties differ for muscle depending on the phase during the length-shortening cycle when the muscle is activated.