Two basically different hypotheses have been advanced to explain the behavior of molluscan muscles in cases in which relaxation of the muscle is extraordinarily prolonged. In one hypothesis, tetanic activation due to prolonged activity in an intrinsic ganglion network is postulated; in the other, changes in the mechanical properties of the muscle capable of maintaining tension generated by the contractile system are proposed. Experiments reported here were designed to test these hypotheses. Recordings were made of electrical activity in a number of circumstances in which the muscle relaxes slowly, and this activity was absent in some cases and in others was not found to correlate well with rate of relaxation. Quick release of the muscle during and after a stimulus which induced slow relaxation showed disappearance of the active state long before decay of tension. Contractile tension decreases with length below rest length whereas passive tension due to stretch following D. C. stimuli remains approximately independent of length. The latter has the same mechanical basis as prolonged relaxation following D. C. stimuli. Thus initial contractile tension and the tension remainder during prolonged relaxation appear to originate through different mechanisms. These results lead us to favor the second hypothesis above. A means by which this could be achieved in vivo is discussed.