Let us consider an analogy for the energy distribution of the particles of a system. In a party, the people have various degrees of activity. Some may be for moments quite still, just sitting; while others may be dancing luriously. The majority will be drinking, talking, smoking, walking and so on. People change their degree of activity to become either more or less active, but the total effect is much the same. We talk of the party "warming up"- meaning that the average degree of activity has increased. Often, this is connected with the consumption of alcohol-that is, an internal conversion in the party from potential to thermal energy! The ideal party, however, is one in which the "system" remains near to equilibrium all the time, so that there is a common emotional atmosphere in which all people, in time, pass through the range of activities. We can look at the changing degrees of activity of people-or changing energy states of particles- in terms of exchanges or interactions between them. Boltzmann, in fact, did investigate the energy distribution of the particles of a gas in terms of collisions, and showed that there would be a movement towards the equilibrium state-starting from any arbitrary distribution. But what of the relation between entropy (5) and the number of arrangments (W) of particles with regard to energy states, i.e. S = log W? Every one of the arrangements of Wmax corresponds to the same energy distribution. We have then a link between the notion of entropy and the notions of and . An increase in entropy means an increase in microscopic complexity, but also an increase in macroscopic uniformity. In the case of a gas, we deal with a system that is constituted by an enclosure - e.g. a metal cylinder. The actual enclosure is a barrier restricting the range of behaviour of the particles within it; they cannot move through it and, in certain cases, they cannot gain or lose energy through it. A closed system is one effectively surrounded, contained, or restricted by an actual barrier. But there are also open systems where the barrier is one of energy - an open beaker of water is a simple example. When we have open system and potential barriers, can the same "principle of maximum entropy" apply as with closed systems? Perhaps there is some way in which complexity can come under control- only in that way is ordered change possible for macroscopic structures.