Throughout the pre-industrial period, the energy supplied to mechanical devices has been obtained mostly from muscle. Apart from the water mill, windmill, and some weapons, animal or human labour invariably drove machinery, and it was not until comparatively late that the thoughts of educated men turned to this problem. The Minutes of the Royal Society in 1670, for example, record that the most famous mechanic of the age, namely their Secretary, Robert Hooke, examined the problem, see Jennings (1995, 15). “[He] produced a contrivance of his to try, whether a mechanical muscle could be made by art, performing without labour the same office which a natural muscle doth in animals”. However, Hooke's normally fertile and inventive brain let him down on this occasion for his invention, based on the thermal expansion of solids, did not impress the members and he “was ordered to consider more fully of it”. It was not until the twentieth century that significant progress was made, and described by Hill (1935, 353–81), one of Hooke's successors at the Royal Society, who gave the Thomas Hawksley Lecture to the Institution of Mechanical Engineers in 1935. The performance of muscles has remained a pertinent consideration and one of great interest to engineers. In the field of bioengineering, for example, the engineer seeks to control and enhance muscular activity in injured or diseased patients, and in modern robotics, the computer-controlled hydraulic ram tirelessly fulfils the functions of the muscle.
The force, work and power that can be developed by animals and humans are often misunderstood. This has not been helped by the fact that engineers, for more than two hundred years, have measured the power output of machines in units of horsepower. This is defined as 33000 ft lb/minute (about 0.746 kW) and has, perhaps, encouraged a belief that the power developed by a horse is invariably one horsepower. In general, we can say that man or beast: