METHODS The animals (kittens and cats) were anaesthetized with pentobarbital sodium. The muscles under investigation were isolated by dissection, care being taken to preserve their blood supply. The motor nerves werecut centrally and mounted on stimulating electrodes. Full precautions have been taken to ensure thatthe muscle contractions approximated as closely as possible to isometric contractions: for the leg muscles steel drills transfixed the distal end of the femur and the distal ends of the tibia and fibula; for the thigh muscles the drills were inserted through the greater trochanter and the distal end of the femur. The drills were clamped in chucks which in turn were fixed to a massive steel frame to which the recording strain gauge (Statham G1-8-350 or Gl-80-35) was also fixed by an assembly giving great flexibility of adjustment. The muscle tendon (or the patella in the case of crureus) was tied firmly toa short steel hook directly linked to the strain gauge; care was taken to align the strain gauge with the direction of muscle pull, which ofcourse was kept as close as possible to its naturalline of pull. The direction of muscle pull was about 600 from the vertical so that it was possible to have the muscle immersed in a paraffin pool with its motor nerve readily accessible. Controlled heating and stirring of the pool ensured that its temperature was almostalways in the range 37-38 C. This temperature control was im-portant because the speed of muscle contraction has a fairly hightemperature coefficient; TIo= 1-53 according to Gordon & Phillips (1953). Despite the oblique angle of the muscle pull there was a negligible sagging of the muscle, because its weight was greatly reduced by its immersion in paraffin and it had a fairly high initial tension-usually more than ten times the weight of the muscle.

Considerable effort has been madeto ensure that the various muscles were contracting under comparable conditions of initial tension. The relationship of initial tension to duration of the isometric twitch was described in detail by Fulton (1925, 1926), but unfortunately the twitches were distorted by friction of the myograph bearing, as revealed by the angle between the summit and the falling phase (Cooper & Eccles, 1929). A re-investigation was therefore necessary, and Fig. 1 shows the effect of length on the initial tension and on the time course and size of the twitch. Measurements from the series partly shown in Fig. 1 AE are plotted in Fig. 1 F with the muscle lengths as abscissae; and as ordinates the initial tensions, the twitch tensions, the twitch contraction times (onset to summit), and the half-decay times (summitto half-decay). It will be seen that around the optimum condition for active tension development (cf. Banus & Zetlin, 1938) progressive lengthening of the muscle causes lengthening both of the contraction time of the twitch and of the time from summit to half-relaxation. It was therefore necessary to adopt some criterion in order that the contractions of different muscles could be compared under standard conditions. The standard adopted was the muscle length giving the maximum twitch tension, as indicated by the vertical line in Fig. 1 F. In all figures zero on the tension scales has indicated zero muscle tension.