While investigating the reflex elicited in the semi-tendinosus muscle of frogs by a single break induction shock applied to one of the digits of the same side, it was found that a very slight increase in the initial passive stretch of the muscle greatly enhanced the magnitude and duration of the response recorded isometrically. Thus, as shown in fig. 1, at 1 gm. initial passive stretch the plateau was reached at 6 gms., while at 3 gms. the plateau height was 12 gms., and at 4·5 the plateau was approximately 15 gms. Above the reflex curves in this figure are shown the responses to single shocks (maximal) applied directly to the muscle, which incidentally show that the reflex response was repetitive in nature. Previously Liddell and Sherrington (1) had made the same observation upon the responses of mammalian muscle, and found that the enhanced response with increasing initial stretch occurred both in reflexes and in the response to direct motor stimulation. The effect would appear, therefore, to be peripheral in its origin. The experiments about to be described were performed in an attempt to analyse this peripheral effect in the hope that it might contribute something to the theories of muscular contraction. Method. On account of the technical difficulty involved in stimulating the unexcised semitendinosus or sartorius muscles of the frog through their motor nerves, the majority of experiments were made with the gastrocnemius; a sufficient number, however, were performed upon these two muscles of the thigh to ensure that the general results were the same for all three. The preparations were usually decerebrate, though some were spinal. The decerebration was accomplished by quickly crushing the fore part of the brain with artery-forceps. By this method less blood is lost than by any of the other methods of decerebration, and the preparation remains longer in good condition, especially as regards the blood-pressure. In one experiment following this method the preparation reacted normally at the end of seven days, even though the nerve to the recording muscle was cut. Spinal frogs were unsatisfactory because the respiration ceased with the destruction of the brain, and, despite the respiration through the skin, their blood soon became venous, which, as will be shown presently, was immediately reflected in the muscular response. Artificial respiration was attempted in several cases with the spinal animal, and while in one experiment it proved successful, it is attended by many annoying difficulties. The frog was dissected before the shock due to decerebration had passed off. A small slit was made beside the urostyle on the side intended for use, and the sciatic trunk carefully withdrawn by means of a small glass hook and severed as high as possible in the lumbar region, especial care being taken to avoid severing any blood-vessels in the operation. As long a stretch of sciatic trunk as possible was now laid bare in the thigh for application of the electrodes. The Achilles tendon, together with a bit of the plantar aponeurosis, was now freed, drawn through a hole in the skin and securely tied with a loop of linen thread. Sometimes the foot was cut off at the ankle after previously ligating to prevent bleeding. The knee-joint was now transfixed by three strong pins, each inserted at a different angle, and the leg securely fixed in a vertical position to the cork floor of the moist chamber, immediately beneath the steel hook which attached the tendon to the isometric recording lever. The preparation was rendered immobile by pins appropriately placed, and by nerve section.