1. Efficient local excitation of nerve‐trunk or muscle elicits contractions quite as strong as the most vigorous voluntary effort; this is in opposition to the results of Mosso, FICK, and others.2. In prolonged quiescent ischæmia there is evidence of failure of conduction of impulses from nerve to muscle before the muscle itself has lost contractility.3. There is no proof of early failure of propagation of the impulse from nerve to muscle in the process of fatigue by serial induced shocks in man and animals with the circulation intact, and under ischæmic conditions in animals stimulation through the nerve‐trunk by serial induced shocks can exhaust almost completely the contractile power of the muscle, without any failure at the myo‐neural junction showing itself. But continued faradisation of the nerve‐trunk, with and without an active circulation, affords evidence (in animal experiments) of such failure prior to complete exhaustion of the muscle. Such failure in the field of conjunction does not explain the primarily “central”, nature of voluntary fatigue, since in the latter condition faradisation of the nerve‐trunk still causes muscular contraction, showing absence of “myo‐neural” block; this also applies to fatigue during ischæmia.4. The evidence available does not indicate rise in the threshold value of the stimulus for the peripheral neuro‐muscular mechanism as an important factor in voluntary fatigue.5. While voluntary fatigue is primarily central in its incidence, there may or may not be an overlapping effect in the way of failure in the contractility of the muscle itself; this varies according as relatively quick or slow series of voluntary efforts are used to induce fatigue. In addition, the diminution of contractility for a quick series is shown by a comparison (a) of the amount of work elicited by peripheral stimulation for a given period of time after voluntary fatigue from quick and slow series with (b) the amount for a similar period of time from the unfatigued muscle. With a slow series there is little evidence of reduction of contractile power.6. After the induction of voluntary fatigue, equal periods of rest and of peripheral stimulation show striking differences as regards recovery in the power and number of the voluntary contractions, before fatigue again ensues, the recovery after the latter being comparatively insignificant as compared with the recovery after the former. This points to the influence of local conditions arising in the muscle, having an inhibiting influence probably through afferent nervous impulses on the central nervous system—since the failure of response to voluntary effort is not adequately accounted for by changes in peripheral excitability or in muscular contractility.7. Recovery after voluntary fatigue is better if the load is removed from the muscle than if the muscle is left loaded or kept in static contraction.8. Ischæawmia per se does not markedly affect the contractile response to a voluntary effort for about 15 minutes. After this a pronounced weakening develops, and peripheral stimulation then causes contraction that is less weakened than the response to voluntary effort; the nerve‐trunk becomes less effective than the muscle excited directly—an indication of “myo‐neural” depression.9. In a limb artificially rendered ischæmic, voluntary muscular fatigue from serial contractions is also largely in the central nervous system, but is much earlier in occurrence than when the circulation is intact, e.g. 2 to 3 minutes as compared with 15 to 20 minutes (weight 2 kilos: rate 30 per minute). This earlier central failure is attributable to the influence of afferent inhibitory impulses from the working muscle.10. The fatigue of static voluntary contraction seems to be essentially central in origin. It differs markedly from fatigue induced by a series of voluntary efforts in regard to the effects of (a) rest, (b) peripheral stimulation, and (c) ischæmia, equal periods of rest and peripheral stimulation by a faradic current showing similar behaviour as regards recovery in static fatigue.11. The fatigue records of static voluntary contraction from muscle (a) with the circulation intact and (b) during ischæmia are apparently identical if the fatigue in both is complete in about 3 to 5 minutes. As the contractile powers of the muscle are not similar at this point when tested by serial peripheral stimulation, fatigue in short ischæmic static voluntary efforts is also central. In efforts lasting longer than about 5 minutes, the ischæmic static contraction approaches at this point the base‐line more rapidly than does the contraction with the circulation intact. The more rapid decline appears to be contingent on the presence of afferent impulses, since the contractility of the muscle is not exhausted.12. A “staircase” of contractions may be elicited under intact circulatory and under ischæmic conditions; the phenomena are probably not due to variations in the H ion concentration.13. In view of the evidence submitted as to the importance of afferent nervous impulses in the processes of voluntary fatigue for local muscular actions under normal and ischæmic conditions, it is probable that afferent nervous impulses of similar origin tend to limit the muscular response in more widespread voluntary efforts, in addition to factors involving (a) the contractility of the muscles themselves, (b) the central nervous system directly.I have to record my indebtedness to Professor MacWilliam for much valued advice in this investigation.Part of the expenses connected with the research was defrayed by a grant from the Carnegie Trust for the Scottish Universities.