We examined the tension change induced by a rapid temperature jump (T-jump) in shortening and lengthening active muscle fibres. Experiments were done on segments of permeabilized single fibres (length (L0) ∼2 mm, sarcomere length 2.5 μm) from rabbit psoas muscle; [MgATP] was 4.6 mm, pH 7.1, ionic strength 200 mm and temperature ∼9°C. A fibre was maximally Ca2+-activated in the isometric state and a ∼3°C, rapid (< 0.2 ms), laser T-jump applied when the tension was approximately steady in the isometric state, or during ramp shortening or ramp lengthening at a limited range of velocities (0–0.2 L0 s−1). The tension increased to 2- to 3 ×P0 (isometric force) during ramp lengthening at velocities > 0.05 L0 s−1, whereas the tension decreased to about < 0.5 ×P0 during shortening at 0.1–0.2 L0 s−1; the unloaded shortening velocity was ∼1 L0 s−1 and the curvature of the force–shortening velocity relation was high (a/P0 ratio from Hill's equation of ∼0.05). In isometric state, a T-jump induced a tension rise of 15–20% to a new steady state; by curve fitting, the tension rise could be resolved into a fast (phase 2b, 40–50 s−1) and a slow (phase 3, 5–10 s−1) exponential component (as previously reported). During steady lengthening, a T-jump induced a small instantaneous drop in tension, followed by recovery, so that the final tension recorded with and without a T-jump was not significantly different; thus, a T-jump did not lead to a net increase of tension. During steady shortening, the T-jump induced a pronounced tension rise and both its amplitude and the rate (from a single exponential fit) increased with shortening velocity; at 0.1–0.2 L0 s−1, the extent of fibre shortening during the T-jump tension rise was estimated to be ∼1.2%L0 and it was shorter at lower velocities. At a given shortening velocity and over the temperature range of 8–30°C, the rate of T-jump tension rise increased with warming (Q10≈ 2.7), similar to phase 2b (endothermic force generation) in isometric muscle. Results are discussed in relation to the previous findings in isometric muscle fibres which showed that a T-jump promotes an early step in the crossbridge–ATPase cycle that generates force. In general, the finding that the T-jump effect on active muscle tension is pronounced during shortening, but is depressed/inhibited during lengthening, is consistent with the expectations from the Fenn effect that energy liberation (and acto-myosin ATPase rate) in muscle are increased during shortening and depressed/inhibited during lengthening.