Let us take a very simple, indeed the simplest possible, example. If two hydrogen atoms having just the correct total energy for the formation of a hydrogen molecule have approached one another and combined, the law describing what has happened must indicate that at a definite moment the combination was complete and the process at an end. This is an example of an irreversible process, since the molecule does not _spontaneously_ break up again. Moreover, the mathematical formulation of this process must include the definite age of the system at which the process was complete, this age being measured from some selected initial moment. This process provides an interesting limitation to a principle put forward by Maxwell as the basis of physical science. He suggested that the laws of physics must be considered to be eternal and unchanging and that therefore they must be expressed in a form which does not contain the time explicitly. This means that for physical laws there can be no difference between to-day and to-morrow. The laws are concerned with small changes which systems undergo in small time intervals, and need not express any fundamental distinction between one moment and another. Such laws cannot express the fact that anything sudden ever occurs which makes an essential change in the system as when two systems become one, or when one system breaks up into two. The laws of organic growth or the evolution of individual systems must display the fact that at a certain age of the system special things happen, such as the combination of two hydrogen atoms, or the attainment of maturity by an organism. Maxwell’s principle puts a limitation on the form of physical laws which precisely eliminates the laws that would be appropriate for organisms. But there is no reason why a broader physics should not try to frame this new type of law that would be applicable to the history and development of individual systems, and it is probable that if this could be done the reversible laws of Newton, Maxwell, and Einstein would appear as approximations which were valid when nothing of special interest was happening, i.e. when only spatial movements were involved without synthesis, disintegration or the emission of light.