A topic in the Open Knowledge Graph — a free, open map of 15,290 topics and the order to learn them in.
Liouville's theorem states that the density of microstates in phase space is conserved under Hamiltonian evolution: ∂ρ/∂t + {ρ,H} = 0. This shows that the 'volume' occupied by an ensemble in phase space remains invariant as the system evolves, establishing a connection between deterministic dynamics and statistical ensembles.
From Hamiltonian mechanics, you know that the state of a system with N degrees of freedom is a single point in a 2N-dimensional phase space — N coordinates q₁, …, qN and N momenta p₁, …, pN. The equations of motion, q̇ᵢ = ∂H/∂pᵢ and ṗᵢ = −∂H/∂qᵢ, define a flow field in this space: each point moves along a trajectory determined by the Hamiltonian H. From ensemble theory, rather than tracking one system, we imagine a vast collection (ensemble) of identical systems in different microstates — a cloud of points in phase space. The phase space density ρ(q, p, t) describes how densely those points are distributed.
Liouville's theorem says this cloud flows like an incompressible fluid. As the ensemble evolves, the shape of the cloud distorts — it can stretch and twist — but its total volume never changes. The mathematical statement is the continuity equation ∂ρ/∂t + {ρ, H} = 0, where {ρ, H} is the Poisson bracket Σᵢ (∂ρ/∂qᵢ ∂H/∂pᵢ − ∂ρ/∂pᵢ ∂H/∂qᵢ). This is the phase-space analog of the incompressibility condition ∇·v = 0 for fluid flow. The proof uses Hamilton's equations directly: the divergence of the phase-space velocity field is identically zero, ∂q̇ᵢ/∂qᵢ + ∂ṗᵢ/∂pᵢ = ∂²H/∂qᵢ∂pᵢ − ∂²H/∂pᵢ∂qᵢ = 0, by symmetry of mixed partial derivatives.
The physical meaning is profound. An equivalent way to state the theorem is that ρ is constant along any trajectory: dρ/dt = ∂ρ/∂t + {ρ, H} = 0. If you ride along with a point in the ensemble, the density around you never changes. This means Hamiltonian evolution preserves information: no two trajectories can merge (that would compress the cloud and violate the theorem), and no information is ever lost about the initial microstate. This has deep implications. For a stationary (equilibrium) ensemble, we need ∂ρ/∂t = 0, which requires {ρ, H} = 0 — so ρ must be a function of H alone. This is why the canonical ensemble takes the form ρ ∝ exp(−H/kT): it is a function of H and therefore automatically satisfies the Liouville condition.
Liouville's theorem also provides the foundation for the ergodic hypothesis: if the theorem guarantees that a finite-volume region of phase space preserves its volume forever, then a single trajectory might (under ergodic conditions) eventually visit every part of the energy surface. The time average then equals the ensemble average — the key step that connects the trajectory of a single physical system to the averages computed from the statistical ensemble. Without Liouville's theorem, the whole framework of connecting deterministic mechanics to statistical mechanics would lack a rigorous underpinning.