1. We compared the rates of work and heat production during ramp shortening with those during staircase shortening (sequence of step releases of the same amplitude, separated by regular time intervals). Ramp or staircase shortening was applied to isolated muscle fibres (sarcomere length, 2.2 microns; temperature, approximately 1 degree C) at the plateau of an isometric tetanus. The total amount of shortening was no greater than 6% of the fibre length. 2. During ramp shortening the power output showed a maximum at about 0.8 fibre lengths per second (Lo s-1), which corresponds to 1/3 the maximum shortening velocity (Vo). For the same average shortening velocity during staircase shortening (step size, approximately 0.5% Lo) the power output was 40-60% lower. The rate of heat production for the same average shortening velocity was approximately 45% higher during staircase shortening than during ramp shortening. 3. The relation between rate of total energy output and shortening velocity was well described by a second order regression line in the range of velocities used (0.1-2.3 Lo s-1). For any shortening velocity the rate of total energy output (power plus heat rate) was not statistically different for staircase (step size, approximately 0.5% Lo) and ramp shortening. 4. The mechanical efficiency (the ratio of the power over the total energy rate) during ramp shortening had a maximum value of 0.36 at 1/5 Vo; during staircase shortening, for any given shortening velocity, the mechanical efficiency was reduced compared with ramp shortening: with a staircase step of about 0.5% Lo at 1/5 Vo the efficiency was approximately 0.2. 5. The results indicate that a cross-bridge is able to convert different quantities of energy into work depending on the different shortening protocol used. The fraction of energy dissipated as heat is larger during staircase shortening than during ramp shortening.