The purpose of this study was to estimate the relative contributions of central and peripheral factors to the development of human muscle fatigue. Nine healthy subjects [five male, four female; age = 30 (2) years, mean (SE)] sustained a maximum voluntary isometric contraction (MVC) of the ankle dorsiflexor muscles for 4 min. Fatigue was quantitated as the fall in MVC. Three measures of central activation and one measure of peripheral activation (compound muscle action potential, CMAP) were made using electromyography (EMG) and electrical stimulation. Measures of intramuscular metabolism were made using magnetic resonance spectroscopy. After exercise, MVC and electrically stimulated tetanic contraction (50 Hz, 500 ms) forces were 22.2 (3.7)% and 37.3 (7.1)% of pre-exercise values, respectively. The measures of central activation suggested some central fatigue during exercise: (1) the central activation ratio [MVC/(MVC + superimposed tetanic force)] fell from 0.94 (0.03) to 0.78 (0.09), (2) the MVC/tetanic force ratio fell from 2.3 (0.7) to 1.3 (0.7), and (3) the integral of the EMG (iEMG) signal decreased to 72.6 (9.1)% of the initial value, while the CMAP amplitude was unchanged. Intramuscular pH was associated by regression with the decline in MVC force (and therefore fatigue) and iEMG. The results indicate that central factors, which were not associated with altered peripheral excitability, contributed approximately 20% to the muscle fatigue developed, with the remainder being attributable to intramuscular (i.e., metabolic) factors. The association between pH and iEMG is consistent with proton concentration as a feedback mechanism for central motor drive during maximal effort.