1. The cases where there appears to exist a "quick" and "slow" contractile system in the same crustacean muscle are reviewed. 2. Blaschko, Cattell and Kahn (1931) showed that in the claws of Maia a stimulus of very low frequency produced a scarcely perceptible response. But the interjection of a single stimulus during such low-frequency excitation brought about a sudden rapid contraction which was subsequently maintained by the low-frequency stimulus. This effect is to be found in other muscles, including the flexor of the dactylopodite of the walking leg of Carcinus. 3. The effect is not due to the presence of a second excitable element in the neuromuscular system. The flexor in Carcinus leg can be shown to behave as though there were only one excitable element present. The effect is due to neuromuscular facilitation. The low-frequency excitation is unable to reach the majority of the muscle fibres; but it leaves their neuromuscular junctions in a condition in which the transmission of an interjected impulse by momentarily increasing the frequency, is so greatly facilitated that it reaches almost all the muscle fibres. Once the conduction path has been established between nerve and muscle in this way, the low-frequency excitation is sufficient to maintain it. 4. The most certain evidence for the existence of a double excitable mechanism in any crustacean limb muscle is that of Keith Lucas on the claws of lobster and of Astacus. These are highly specialised, and further are differentiated into cutter and crusher claws in Carcinus and many decapods. The muscles of the cutter claw show the same behaviour as the muscles of the walking leg. The muscle behaves as a single excitable unit when the nerve is excited near the base of the limb. 5. The adductor of the crusher claw and this muscle alone exhibits a double excitable system. Two distinct types of response are obtained at different thresholds. Two intensity-duration curves can be traced. The systems show very different rates of contraction. The whole system behaves as though there were two distinct muscles. The "slow" system has a lower threshold than the "quick" system. 6. The "quick" system differs sharply from the "slow" in the frequency of stimulation required to bring it into action. The frequency range of stimulation required to activate the "slow" system of the crusher claw is identical with the range for the whole adductor of the cutter claw and the flexor and extensor muscles of the dactylopodite of the walking leg. The "quick" system, on the other hand, is brought into action at a much lower frequency range. Apart from this, the behaviour of both systems is the same as that of all the other muscles investigated. Over their own characteristic ranges, the higher the frequency the more rapid the contraction. 7. The crusher adductor is equivalent to two units physiologically differentiated. The unit corresponding to the slow system is used in the normal movements of the crab, while the unit corresponding to the quick system is used when great tensions are suddenly required for the crushing of hard objects. The differentiation thus has a decided functional significance.