Effects of stimulation frequency history on length-force characteristics were determined for rat medial gastrocnemius muscle (GM). The peripheral nerve was stimulated supramaximally according to two stimulation protocols. First, a complete set of length-force data were obtained by stimulating the nerve with a decreasing stimulation frequency (DSF) staircase composed of five successive pulse trains (200 ms each) of 100-, 50-, 40-, 30- and 15-Hz stimulation. Then length-force data were obtained using constant stimulation frequency (CSF; during the isometric contraction inter-stimulus interval was constant). The acquisition order of a complete set of length-force data was: first 15-Hz, then 30-, 40-, 50- and finally 100-Hz stimulation. For all DSF conditions, both optimum muscle length as well as active slack length were shifted significantly (P < 0.05) to lower muscle length with respect to CSF. Muscle length range between active slack and optimum length for all DSF conditions was increased significantly with respect to CSF. DSF thus caused a marked shift of the length-force relationship to lower muscle length compared to the CSF-dependent length-force relationship. As a result of this shift, muscle force enhancement (potentiation) was non-linearly related to muscle length; force enhancement decreased exponentially with increasing muscle length. In addition, DSF-dependent length-force characteristics are not scaled and shifted versions of those for CSF. Possible factors affecting these length-force characteristics are higher intracellular calcium concentration, myosin light chain phosphorylation, fatigue during sustained contractions, interaction between aponeurosis and fibre length, distribution of fibre mean sarcomere length with respect to muscle length, and muscle length changes during unfused tetanic contractions. It is concluded that length-force characteristics of rat GM are dependent on both short-term stimulation frequency history as well as stimulation frequency per se.