The sensitivity of electrical stimulation tests for detecting low frequency fatigue (LFF) and its duration were examined in the adductor pollicis muscle of ten normal subjects. Supramaximal stimulation was applied to the ulnar nerve at the wrist and values for test stimulation force ratios of 10:100 Hz, 15:100 Hz and 20:50 Hz (2 sec at each frequency with 5 min rest between each ratio) were obtained for fresh muscle. Fatigue was then induced by voluntary isometric contractions at 50% maximal voluntary force (MVC) repeated until only 30% MVC could be achieved. Contractions lasted 10 sec with 5 sec rest between each. The three test ratios were then repeated to monitor recovery at intervals up to 72 h after activity. High frequency forces returned to fresh values by 24 h but low frequency forces were all still significantly reduced. Forces at 10 and 15 Hz were still significantly reduced at 48 and 72 h (10 Hz greater than fatigue than 15 Hz). The low/high frequency ratios, calculated once 50 and 100 Hz forces had recovered, also demonstrated differences in recovery rates. Repeatability tests indicated that 10 Hz force was more variable than other frequencies and forces at all stimulation frequencies were repeatable on different days with a coefficient of variation of less than 15%. Values for the 15:100 Hz ratio from fresh muscle in 22 normal subjects were 0.48 +/- 8. The 15:100 Hz ratio is suggested as the most appropriate test ratio for detecting LFF since 15 Hz force is more sensitive than 20 Hz and more stable than 10 Hz.