1. An experimental protocol designed to assess fatigability in motor units (Burke, Levine, Tsairis & Zajac, 1973) has been applied to the whole muscles of anaesthetized adult rats, and the association between the electromyogram (EMG) and force was monitored over the course of the test. 2. Both test muscles (soleus and extensor digitorum longus) exhibited a wide range of fatigability, which was defined as the decline in isometric peak force at 6 min, such that the data could be separated into five levels of fatigability. Fatigue indices for each test muscle were distributed across three levels. 3. The EMG was quantified with four measures of amplitude, four of duration, and one interaction term (area). Correlation analyses indicated that the EMG was adequately represented by one measure of amplitude (absolute amplitude), one of duration (peak-to-peak duration) and area. The best single measure was area. 4. The EMG-force associations for soleus varied markedly among its three fatigability groups. In contrast, over the course of the test, all three extensor digitorum longus groups displayed qualitatively similar EMG-force associations. 5. Multiple regression analyses indicated that the EMG parameters were able to predict peak force better for extensor digitorum longus than for soleus. Furthermore, for both test muscle, the prediction was best for the most fatigable group. 6. The associations between EMG and force exhibited three patterns for the two test muscles and three levels of fatigability. These differences suggested variation in the mechanisms, related to both fibre-type composition and susceptibility to fatigue, that dictate the performance elicited by this particular stimulus regimen. The mechanisms seem to include both intracellular and transmission processes.