1. A study has been made of the relation between force and speed of shortening, or lengthening, in isolated twitch and slow muscle fibres, dissected from the iliofibularis muscle of Xenopus laevis. Both after-loaded and quick-release contractions were studied. Twitch fibres were stimulated electrically to give tetanic contractions (5-20 degrees C); slow fibres were activated by a rapid change to solutions with high K concentration (30-75 mM; experiments at 21-24 degrees C).2. The velocity of slow fibres was constant during shortening over 10% length change in after-loaded contractions, except at forces exceeding about 0.8 of isometric tension, P(0). In quick-release experiments, shortening velocity was found to depend not only on the relative load, P/P(0), but also on the instant when the release was made. With increasing time after onset of contraction the initial rate of shortening decreased; also, a progressive fall in speed during shortening became more marked.3. The fall in initial shortening speed with time of release from the onset of a contracture was more pronounced at high [K](o) than at low.4. The relation between the relative force, P/P(0), and shortening velocity, V, in after-loaded contractions (75 mM-K) and quick-release contractions (45 mM-K, early releases) in slow fibres could be represented by a hyperbola with the constants a = 0.10P(0), b = 0.11 lengths/sec; extrapolated V(max.) was 1.10 lengths/sec.5. Isometric tension and maximum shortening velocity in slow fibres were very nearly constant between 32 and 75 mM-K. a/P(0), however, was clearly reduced at 32 mM-K, representing a more curved P-V relation.6. Force-velocity data for twitch fibres (quick-release contractions, 20 degrees C) were reasonably well fitted by a hyperbola (a = 0.38P(0), b = 1.97 lengths/sec, V(max.) = 5.20 lengths/sec), but a systematic deviation was observed for forces exceeding 0.6P(0).7. a/P(0) for twitch fibres was found to be independent of temperature in the range 5-20 degrees C. Q(10) for b was 2.24 (10-20 degrees C), and 2.86 (5-10 degrees C).8. V(max.) for twitch fibres was calculated to be 6.34 lengths/sec at 22.5 degrees C, the average temperature in the slow fibre experiments. The maximum shortening velocity in twitch fibres is thus 6 times higher than in slow fibres.9. When loads in the range 1.1-1.4P(0) were quickly applied to an actively contracting slow fibre, lengthening of the fibre occurred in two phases, an initial rapid phase, followed by a phase of extremely slow lengthening. In corresponding experiments on twitch fibres lengthening was rapid at first and then gradually became slower.10. Factors affecting the shape of the force-velocity curve are discussed. Calculations based on A. F. Huxley's (1957) model for muscle contraction indicated that cross-bridge turnover rate is about 15 times lower in slow than in twitch fibres.