Entropy is the logarithm of a count
We met entropy earlier as a measure of how spread-out a system is. Statistical mechanics gives it a startlingly literal meaning. Count the number of microstates — the detailed molecule-by-molecule arrangements — that all look like the same macrostate. Take the logarithm of that count. Multiply by the Boltzmann constant. That is the entropy, full stop. This is statistical entropy, and the recipe is carved on Boltzmann's gravestone as S = k log W. The whole mystique of the word melts away: statistical entropy is just a tidied-up way of saying 'how many ways can this happen?'
Why a logarithm, of all things? Because entropy must add up when you join two systems, but the *counts* multiply. Put two dice together and the number of combined outcomes is 6 times 6 = 36, not 6 plus 6. The logarithm is the one operation that turns that multiplying into adding: log of (A times B) is log A plus log B. So the Boltzmann entropy formula is precisely the bookkeeping that makes entropy behave like the extensive, additive quantity classical thermodynamics always insisted it was.
When the count refuses to reach one
Sometimes a substance cannot tidy itself into a single arrangement even at the lowest temperatures, because some molecules can point in more than one direction with the same energy and get frozen before they can agree on one. Carbon monoxide is the classic case: each little molecule can lie head-first or tail-first, and as it freezes there is no energy reason to prefer one. The crystal locks in a jumble, leaving a leftover count greater than one — and so a leftover entropy that survives all the way down. That stubborn remainder is called residual entropy, and it is a vivid reminder that entropy really is about counting arrangements, not about temperature alone.
The equipartition theorem: everyone gets the same slice
Now to the second great shortcut. A molecule can stash energy in several distinct ways: sliding left-right, up-down, and front-back; spinning about its axes; and so on. Each independent way of holding energy is called a degree of freedom. The equipartition theorem makes a remarkably democratic claim: at an ordinary temperature, every such degree of freedom holds, on average, the very same amount of energy — a slice equal to one-half of the thermal energy kT. No way of moving is favoured; the energy is shared out evenly, like a pie cut into equal pieces.
Once you accept this, you can count up the energy of a substance almost on your fingers. A single atom of a noble gas can only slide in three directions, so it holds three half-slices of kT and nothing more. That immediately predicts how much its internal energy rises as you warm it — and therefore its heat capacity. The equipartition theorem turns 'how much heat does this gas soak up?' into a matter of *counting the ways it can move*.
Where the fair-share rule breaks (and why that's wonderful)
Equipartition is a generous rule, but it comes with fine print: a degree of freedom only collects its half-slice of kT once it is *easy to wake up*. If a way of moving has its first rung set far higher than kT — as a molecule's stiff bond vibrations often do at room temperature — there is not enough thermal cash to climb even one step, and that mode stays frozen out, contributing nothing. So the fair-share rule applies only to the modes that are switched on; the stiff, high-rung ones sleep until you heat things enough to wake them.