By entropy we understand the quantity of heat of a body divided by its absolute temperature. If a quantity of heat, of Q calories, of a body at a temperature of 100° (absolute temperature, 373°) passes over to another body of 0° (absolute temperature, 273°), the total entropy of the two will have been decreased by Q/373, and increased by Q/273. As the latter quantity is the greater, the entropy of the whole will have increased. By itself, we know, heat always passes, either by radiation or by conduction, from bodies of higher temperature to bodies of lower temperature. That evidently implies an increase in entropy, and it is in agreement with the law of Clausius that entropy tends to increase. The most simple case of heat equilibrium is that in which we place a number of bodies of unequal temperatures in an enclosure which neither receives heat from outside nor communicates heat to the outside. In some way or other, usually by conduction or radiation, the heat will pass from the warmer to the colder bodies, until at last equilibrium ensues and all the bodies have the same temperature. According to Clausius, the universe tends to that thermal equilibrium. If it be ever attained, all sources of motion, and hence of light, will have been exhausted. The so-called “heat-death” (Wärmetod) will have come. If Clausius were right, however, this heat-death, we may object, should already have occurred in the infinitely long space of time that the universe has been in existence. Or we might argue that the world has not yet been in existence sufficiently long, but that, anyhow, it had a beginning. That would contradict the first part of the law of Clausius, that the energy of the universe is constant; for in that case all the energy would have originated in the moment of creation. That is quite inconceivable, and we must hence look for conditions for which the entropy law of Clausius does not hold.