But suppose that instead I had taken a film of a cup of tea as it was cooling. One end of the film would show the steam above the cup and the spoon changing in length as it changed in temperature. Passing along the film these effects would grow less marked until the successive photos showed no variation when the temperature of the tea was nearly that of the surrounding air. It would be obvious which way to wind this film, without using any subjective criterion supplied from memory of the individual process which had been photographed. This process is irreversible, but physics has hitherto assumed that all such processes are merely the statistical result of a chaos of molecular motions each of them perfectly reversible. The assumption of reversibility seems to some physicists so fundamental that they think there could be no science without it. But that is a mere prejudice arising from the fact that Newton conceived one particular way of giving mathematical formulation to the measurable features of physical processes. By suggesting that all the laws of nature might take a form similar to his law of gravitation, he made the implicit assumption that all elementary processes were reversible. Gravitational motions are so, at any rate within the accuracy of Newton’s law, and as a consequence of the confirmation of his law and the fact that it has been taken as a model for the whole system of modern physical conceptions, the latter are only appropriate for reversible processes. Apparent irreversibility, such as the cooling of a cup of tea, is attributed to statistical effects, and the second law of thermodynamics, which asserts that temperatures tend to uniformity, is treated as merely a statement of what is highly probable. This is probably quite legitimate, but even where no statistical effect can enter and the process is clearly irreversible physics usually adopts any measure rather than assume that a fundamental elementary process is irreversible.[2] We cannot be surprised at this, since if physics once admitted that any elementary process was irreversible it would have to give up the whole system of Newtonian conceptions. Matter, force, energy, action, and wave properties are all unsuitable for the treatment of irreversible effects since they all ultimately depend on Newton’s reversible law. An entirely new set of ideas is necessary for describing processes which necessarily proceed in one direction, so that one particular state of the system must precede another state. It appears conceivable that an alternative set of conceptions to replace the Newtonian might be established by demanding the irreversibility of all natural laws, as well as the demands hitherto made by physics, i.e. the permanence of matter and the conservation of energy.