Isolated whole frog gastrocnemius muscles were electrically stimulated to peak twitch tension while held isometrically in a bath at 4 degrees C. A quartz hydrophone detected vibrations of the muscle by measuring the pressure fluctuations caused by muscle movement. A small steel collar was slipped over the belly of the muscle. Transient forces including plucks and steady sinusoidal driving were applied to the collar by causing currents to flow in a coil held near the collar. The instantaneous resonant frequencies measured by the pluck and driving techniques were the same at various times during a twitch contraction cycle. The strain produced by the plucking technique in the outermost fibers was less than 1.6 x 10(-4%), a strain three orders of magnitude less than that required to drop the tension to zero in quick-length-change experiments. Because the pressure transients recorded by the hydrophone during plucks and naturally occurring sounds were of comparable amplitude, strains in the muscle due to naturally occurring sound must also be of the order 10(-3%). A simple model assuming that the muscle is an elastic bar under tension was used to calculate the instantaneous elastic modulus E as a function of time during a twitch, given the tension and resonant frequency. The result for Emax, the peak value of E during a twitch, was typically 2.8 x 10(6) N/m2. The methods used here for measuring muscle stiffness are unusual in that the apparatus used for measuring stiffness is separate from the apparatus controlling and measuring force and length. © 1990, The Biophysical Society. All rights reserved.