But we are told that modern science has proved that the law of the Conservation of Energy is an absolute law of the universe, and that though man can not construct a machine which will realize the dream of perpetual motion, the material universe is in reality such a machine. It becomes us to speak with some degree of diffidence in regard to a question which lies outside of our special department of study. Nevertheless we must confess that we have a growing suspicion of all so-called "absolute laws" in the domain of physical nature. And we are confirmed in this mistrust by the fact that physicists themselves are not agreed in regarding this law of conservation of energy as universally true. "That the amount of energy in the world is unchangeable, the sum of the actual or kinetic and potential energies being a constant quantity, has been by some writers overstrained. It may be taken as a postulate, and is probably true, but it is a proposition equally incapable of proof and of disproof."[260] "This principle," says Sir J. Herschel, "so far as it rests upon any scientific basis as a legitimate conclusion from dynamical laws, is no other than the well-known dynamical theorem of the conservation of _vis viva_ (or of 'energy,' as some prefer to call it), _supplemented to save the truth of its verbal enunciation_ by the introduction of what is called 'potential energy,' a phrase which I can not help regarding as unfortunate, inasmuch as it goes to substitute a truism for the announcement of a dynamical fact. No such conservation, in the sense of an identity of total amount of _vis viva_ at all times and in all circumstances, in fact, exists. So far as a system is maintained by the mutual actions and reactions of its constituent elements _at a distance_ (_i. e._, by force), _vis viva_ may temporarily disappear, and be subsequently reproduced between certain limits. Collision, indeed, between its ultimate particles or atoms, regarded as absolutely rigid, and therefore inelastic (_for that which can not change its figure can have no resilience_), can not take place without producing a permanent destruction of it, which there exists no means of repairing.... If, indeed, we could be assured _à priori_ that the system [of the universe] is one of simple or compound periodicity, in which a certain lapse of time will restore every molecule to identically the same relative situation with respect to all the rest, we should then be sure that in the nature of things there would take place, so to speak, a winding up from a lower to a higher state of potential energy, to be subsequently exchanged for newly created _vis viva_. But, as we can have no such _à priori_ assurance, can only assume such restoration to be possible, and can see no means of effecting it, if possible, otherwise than by foresight and prearrangement; the one equally with the other is an unknown function, variable within unknown limits, and susceptible of