In chronic experiments concerning the activity-dependent plasticity of muscle properties, a simple and noninvasive method was used for monitoring changes of twitch speed in conscious adult cats. The animals had been provided with implanted electrodes for nerve stimulation, and a hand-held force transducer was pressed against the fully extended ankle joint while single test pulses were delivered to the common peroneal nerve. In the present report, this technique for the recording of ankle twitches is subjected to critical analysis and evaluation. The measurements were highly reproducible with respect to contraction time (time-to-peak) but less so for half-relaxation time and twitch amplitude; other methods should be used for the long-term monitoring of contractile force. The total force (torque) of the ankle twitch was mainly produced by tibialis anterior (about 45%), peroneus longus (PerL; 27%) and extensor digitorum longus (23%). The ankle twitch produced by PerL alone had about the same contraction time as that of all the muscles together. Among muscles that had become changed as a result of long-term electrical stimulation there was, in general, a good correspondence between the contraction times from simple external recordings of ankle twitches and those separately measured for PerL under general anesthesia (force transducer then directly connected to PerL tendon).