Same maths, different names · Audit 10
A field carried by a flow as if painted on it: vortex lines in water, magnetic field lines in a plasma, potential vorticity in the atmosphere, the density of stars in a galaxy’s phase space. Each field has its own theorem and its own name for it. This audit records that they are one kinematic statement about how a geometric object is dragged by a flow, what dynamical condition each field needs before the statement applies, and where each breaks.
In plain terms
Stir a cup of coffee with a drop of cream in it and the cream is pulled into ribbons that follow the flow. It is carried; it does not move on its own. Some quantities in physics behave the same way but are less obvious than cream. In 1858 Hermann von Helmholtz showed that in a fluid without friction the lines around which the fluid swirls, the vortex lines, are carried by the flow as if they were dyed into it: a vortex line always consists of the same fluid particles, and the strength of a tube of them never changes. Lord Kelvin restated this a decade later as a rule about circulation, the total amount of swirl around any loop that moves with the fluid, which stays constant. In 1943 Hannes Alfvén found that in a fluid that conducts electricity perfectly, such as the plasma of the Sun, magnetic field lines are carried in exactly the same way. He said the matter was “fastened” to the field lines; everyone since has said the field is frozen in. In 1942 Hans Ertel found the atmospheric version, a quantity called potential vorticity that every parcel of air keeps as it moves, and weather forecasting still rests on it. In a galaxy, or a plasma too thin for collisions, the density of stars or particles in the space of positions and velocities is carried along unchanged. Each of these fields has its own theorem, its own founder and its own name.
The mathematics underneath is one statement, and it is about geometry rather than about any particular fluid. Some quantities are naturally measured along lines, some across surfaces, some per volume. A flow drags lines, surfaces and volumes along with it, and if a quantity is carried “as painted”, the amount measured on a moving line, surface or volume never changes. The rule that says so is called Lie transport, after Sophus Lie, and once a quantity is known to be transported that way, the consequences that follow depend on what kind of quantity it is. For a flux carried through surfaces, as vortex lines and magnetic field lines are, knots and links in the lines can never be untied while the flow stays smooth; for a quantity carried per parcel, as potential vorticity is, there is no such promise. In that language Helmholtz’s, Kelvin’s, Alfvén’s and Ertel’s theorems are the same sentence with different nouns, and the nouns matter.
Two things make the audit worth doing. The first is that the theorem is exact but its precondition is not free: the fluid must have no friction, the plasma must conduct perfectly, the air must not exchange heat, the particles must not collide. Where those conditions fail depends on the fluid and on the flow. Friction and resistance act most strongly where the field is finely structured, and in many flows, though not all, the carrying process itself creates that fine structure by stretching and folding; in others it was there at the start, or is imposed from outside. Whether a given flow creates it is a fact about the flow, not about the theorem; but where the condition fails, the interesting physics, from the origin of the Sun’s magnetic field to solar flares to the way turbulence dissipates, lives at the breakdown. The second is that “the field moves with the fluid” is a looser statement than it sounds. For a magnetic field there are many velocities that carry it equally well, and a field line has no identity of its own beyond what a chosen velocity lends it; that point, made by Newcomb in 1958, is where careless talk about “field lines moving” goes wrong.
The rest of the page is a ledger. It lists the fields, what each calls the theorem, when each first had it, what each assumes, and a verdict: exact when the theorem is exactly the transport statement within the row’s governing equations, limited when it holds only after a named approximation, analogy when only some features line up. Every claim carries a note on how far the source was actually read.
What is being audited
Take a flow with velocity field v(x, t) and a geometric object carried by it: a scalar, a density, a line-integrand (a 1-form), a flux (a 2-form) or a volume element. The object is frozen into the flow when its integral over any material curve, surface or volume, one that moves with the fluid, is constant in time. In the language of forms this is one equation:
vorticity, barotropic flow: ∂tω = ∇×(v×ω), the 2-form transported exactly; the velocity 1-form only up to an exact differential, (∂t + 𝓛v)v♭ = d(½|v|² − h − Φ), so that d/dt ∮C(t) v·dl = 0 on closed loops only (Kelvin)
Lagrangian solution (Cauchy, incompressible case): (ω/ρ)(X(a,t), t) = ∇aX · (ω0/ρ0)(a)
magnetic field: ∂tB = ∇×(w×B) with ∇·B = 0 ⇒ d/dt ∫S(t) B·dS = 0 for surfaces carried by w (Alfvén, with w the fluid velocity of ideal MHD)
baroclinic flow with an advected scalar θ and p = p(ρ, θ): (∂t + 𝓛v)Ωa = ρ−2 dρ∧dp ≠ 0, but Ωa∧dθ is transported, and dividing by the mass form gives D/Dt [(ωa·∇θ)/ρ] = 0 (Ertel), a transported scalar
a frozen scalar on phase space: ∂tf + v·∇xf + (F/m)·∇vf = 0 (Vlasov), with the phase-space volume itself invariant (Liouville)
The first line is the whole structure. A k-form ω is Lie-transported by v exactly when its integral over every material k-dimensional region is conserved; Cartan’s formula turns that into a differential equation. The remaining lines are the same equation for particular objects, and the differences between them are the content of the page. Barotropic vorticity and the ideal-MHD magnetic field are 2-forms (fluxes through surfaces): their transport equation is the curl of a cross product, and their Lagrangian solution carries the initial field forward by the deformation gradient, a formula Cauchy wrote for vorticity in 1815 and which reappears for B/ρ in every magnetohydrodynamics text, with the density in each case conserved by the same flow that does the transporting. The velocity 1-form is not transported in the strict sense: (∂t + 𝓛v) applied to it gives an exact differential, which integrates to zero around a closed material loop and not along an open one, so Kelvin’s theorem is a statement about circulation on closed loops, not about the 1-form itself. In a baroclinic fluid the absolute-vorticity 2-form is not transported either, since dρ∧dp need not vanish; what is transported is the 3-form Ωa∧dθ, because dρ∧dp∧dθ = 0 when p depends on ρ and θ alone, and dividing that by the transported mass form gives Ertel’s potential vorticity, which is therefore a transported scalar and not a density. On phase space the Hamiltonian flow preserves the volume form, so the particle density f is carried as a scalar too.
Three conditions travel with the identity, and each row of the ledger is read against them.
First, the theorem is kinematic and the precondition is dynamical. Lie transport says what happens to an object once it is known to obey ∂tω + 𝓛vω = 0; it does not say which objects do. That comes from the equation of motion, and each row buys it with an ideal assumption: inviscid barotropic flow with conservative forces, for vorticity; infinite conductivity, for the magnetic field; adiabatic motion with a thermodynamic closure p = p(ρ, θ), for potential vorticity; no collisions, for the phase-space density. The verdict judges the transport statement within the row’s ideal equation. Whether the ideal equation is a good model is the “model” line, and where it fails is the “breaks” line.
Second, the geometric type of the object is not optional. A scalar, a 1-form, a 2-form and a density transform differently under the flow map, and “frozen in” means something different for each: a scalar keeps its value, a 1-form keeps its line integrals, a 2-form keeps its fluxes, a density keeps its volume integrals. The vorticity vector is frozen as a 2-form, so ω/ρ, not ω, follows material line elements; Kelvin’s circulation is conserved on closed loops because the 1-form is transported only up to an exact term; Ertel’s potential vorticity is a scalar because it is a transported 3-form divided by the mass form. Getting the type right is what makes the dictionary column checkable, and getting it wrong is the commonest error in accounts of this subject, including the first draft of this page.
Third, “carried by the flow” presupposes a flow, and the mathematics does not single one out. For a divergence-free B, any sufficiently regular w with ∇×(E + w×B) = 0 preserves its flux; for vorticity, v + λω gives the same cross product as v and so the same transport, wherever the resulting flow is regular. What picks out one velocity is the physical model: in ideal MHD the bulk velocity is the material velocity that the fluid equations and the ideal Ohm’s law define, not a convention, and in Hall MHD the induction equation names the electron velocity instead. Other flux-preserving velocities exist without being material velocities. Newcomb’s 1958 paper distinguishes flux preservation from the weaker preservation of field-line connections, and the general lesson is that the identity of a field line through time is a property of the pair (field, transporting velocity) together with the equation that selected the velocity, not of the field alone. The Cauchy formula for B/ρ also presupposes a density conserved by that same velocity; it does not transfer to an arbitrary w with the fluid’s density.
Two columns in the ledger do the real work. The dictionary column says which object is frozen and of what geometric type, so that a verdict can be checked rather than believed. The direction column says which way the implication runs: from an ideal equation of motion to the transport statement, or from the transport statement to a consequence such as a topological invariant. The verdict judges the correspondence within the row’s stated governing equation; the adequacy of that equation is answered on the “model” line.
The ledger
Each verdict applies to the correspondence named in that row’s dictionary column, not to the field as a whole. The physical adequacy of the starting equation is stated on the “model” line and does not enter the verdict.
| Field | Name used · earliest source located | Dictionary | Direction of implication | Assumptions added or dropped · model · where it breaks | Verdict |
|---|---|---|---|---|---|
| Fluid dynamics | Helmholtz’s vortex theorems; Kelvin’s circulation theorem; Cauchy’s vorticity formulaCauchy 1815 (printed 1827; incompressible) · Helmholtz 1858 · Thomson (Kelvin) 1869, read 1867 · Batchelor 1967, Sec. 5.3 | Vorticity ω = ∇×v ↔ a transported 2-form (flux through material surfaces) · the velocity 1-form ↔ transported up to an exact differential, so circulation ∮v·dl is conserved on closed material loops · vortex lines ↔ material lines · ω/ρ ↔ a material line element, carried forward by the deformation gradient in Cauchy’s formula · strength of a vortex tube ↔ a conserved flux | Euler equation → transport statement, exactly, under the ideal conditions; transport statement → Helmholtz’s three theorems and the conservation of linkages (Audit 12). | adds no viscosity; barotropic pressure, p = p(ρ); conservative body forces · model the Euler equationsbreaks with viscosity, which diffuses vorticity across material surfaces; with baroclinic torque ∇ρ×∇p ≠ 0, which creates circulation; with non-conservative forces; and in practice wherever stretching has made the vorticity fine enough for viscosity to act, which in turbulence is everywhere at the small scales | exact |
| Geophysical fluid dynamics | Ertel’s potential vorticity theorem; PV conservation; Rossby’s shallow-water PVRossby 1940 · Ertel 1942 · Hoskins, McIntyre & Robertson 1985 (review) | Absolute-vorticity 2-form Ωa ↔ not transported in general: (∂t + 𝓛v)Ωa = ρ−2dρ∧dp · potential temperature θ ↔ a transported scalar · Ωa∧dθ ↔ a transported 3-form, because dρ∧dp∧dθ = 0 when p = p(ρ, θ) · q = (ωa·∇θ)/ρ ↔ that 3-form divided by the transported mass form: a transported scalar, with ρq d³x the corresponding density · isentropic surfaces ↔ material surfaces | Rotating, stratified Euler equations with the thermodynamic closure → material conservation of q, exactly; the baroclinic source in the vorticity equation cancels in the combination. Hoskins et al. derive it from the vorticity equation and the Bjerknes circulation theorem; McIntyre’s encyclopaedia article shows the cancellation explicitly. | adds adiabatic motion (θ materially conserved); inviscid; a rotating frame; an equation of state p = p(ρ, θ), which a multicomponent fluid with a second advected scalar does not satisfy in this form · model inviscid, adiabatic, rotating compressible Euler flow with the closure p = p(ρ, θ); the hydrostatic primitive-equation versions used in forecasting carry their own approximations and are stated separately (McIntyre, p. 377, after eq. 4)breaks with diabatic heating and friction, which are sources and sinks of q; the invertibility principle that makes q useful is a separate result with its own assumptions (balance), not part of the transport theorem | exact |
| Magnetohydrodynamics | Alfvén’s theorem; frozen-in flux; frozen-in field lines; flux freezingAlfvén 1943 (“fastened”) · Batchelor 1950 (the analogy with vorticity) · Newcomb 1958 (line identity) · Moffatt 1978, Ch. 3 · Landau & Lifshitz, Electrodynamics of Continuous Media, Sec. 65 | B ↔ a transported 2-form (flux through surfaces carried by the material velocity) · B/ρ ↔ a material line element by Cauchy’s formula, with ρ conserved by the same velocity · flux preservation ↔ ∇·B = 0 and ∇×(E + w×B) = 0 for a regular w; line-connection preservation ↔ a weaker condition (Newcomb) · magnetic flux through a material loop ↔ circulation of the vector potential | Ideal induction equation ∂tB = ∇×(v×B) → flux conservation through surfaces carried by v, exactly; the induction equation follows from Ohm’s law with infinite conductivity, E + v×B = 0, and it is that law which makes the material velocity the transporting one. | adds infinite conductivity; a single material velocity · model ideal magnetohydrodynamicsbreaks with finite resistivity, which lets flux diffuse across the fluid while the velocity stays perfectly smooth, and which in resistive MHD permits reconnection where gradients are steep (Priest & Forbes 2000); with the Hall term, which breaks freezing into the bulk velocity while preserving it into the electron velocity (given barotropic electron pressure, negligible electron inertia and a regular electron flow); with electron-inertia and pressure-tensor terms, which break transport by the electron flow as well and support collisionless reconnection (Kuznetsova, Hesse & Winske 2001); and other flux-preserving velocities exist besides the material one, so “the motion of a field line” is well defined only once the velocity and the selecting equation are named | exact |
| Two-fluid and Hall plasmas | Canonical (generalised) vorticity; Hall MHD; frozen into the electron flowD’Avignon, Morrison & Lingam 2016 · Mahajan & Yoshida 1998 | Canonical vorticity of species s, ∇×(msvs + qsA) ↔ a transported 2-form carried by that species’ own velocity, given a barotropic pressure for that species · in Hall MHD (SI units), B ↔ transported by the electron velocity, and B + (mi/e)∇×vi ↔ transported by the ion velocity · the single-fluid theorem ↔ the further limit in which the Hall contribution is negligible at the scales considered | Ideal two-fluid equations with barotropic species pressures → one transported 2-form per species, exactly. Hall MHD neglects electron inertia and keeps the Hall term; ordinary ideal MHD keeps the bulk inertia, carried mostly by the ions, and additionally neglects the Hall term. | adds collisionless, unresistive species equations with barotropic scalar pressures; each species treated as a fluid; D’Avignon, Morrison and Lingam identify barotropic electron pressure as what supports the transported magnetic flux and barotropic ion pressure as what supports the second invariant · model ideal two-fluid or Hall magnetohydrodynamicsbreaks with collisions and resistivity, as in the row above; with a species pressure that is not barotropic, which supplies a baroclinic source, and with a full pressure tensor; and the object that is transported changes with the model, so “the field is frozen in” without saying into which velocity, and which field, is not a well-posed claim | exact |
| Dynamo theory | Stretch–twist–fold; flux-rope amplification; fast dynamoCowling 1933 and Zeldovich 1957 (antidynamo theorems) · Vainshtein & Zeldovich 1972 · Childress & Gilbert 1995 (monograph) | Ideal stretching ↔ amplification of B by the deformation gradient, which needs no resistivity (a uniform strain v = (γx, −γy, 0) gives B = B0eγtx̂ exactly) · stretch–twist–fold ↔ a particular construction in a bounded domain in which the fold brings the two loops into the same orientation so that their fluxes reinforce · small but finite resistivity ↔ what lets the folded flux merge rather than accumulate as ever-finer structure · fast dynamo ↔ a growth rate that stays positive as the resistivity tends to zero | The theorem is used as a mechanism: ideal transport amplifies field by stretching, exactly and without any non-ideal term; whether the growth rate stays positive as the resistivity tends to zero is the separate fast-dynamo question, in which the small resistivity can matter critically. | adds a flow with the right three-dimensional geometry; for the long-time limit, a resistivity that is small but not zero · model the kinematic induction equation with prescribed flow, then the full MHD problembreaks Cowling’s theorem excludes axisymmetric magnetic fields and Zeldovich’s planar flows, so with enough symmetry no flow sustains a field; at exactly zero resistivity a stretch–twist–fold flow folds the field ever finer rather than merging it, so the ideal case and the limit of vanishing resistivity differ, which is the fast-dynamo problem; and with back-reaction the prescribed-flow picture ends | limited |
| Turbulence (the vorticity–magnetic analogy) | The analogy between ω and B; Batchelor’s argument for equipartitionBatchelor 1950 · Moffatt 2002 (assessment) | The correspondence judged: Batchelor’s inference from a shared equation to shared dynamics · hydrodynamic vorticity in an incompressible viscous fluid ↔ magnetic field under kinematic induction in a prescribed incompressible flow: the same advection–diffusion equation, with ν ↔ η · the difference: ω = ∇×v is tied to the velocity that carries it, B is not; and in coupled MHD the vorticity equation acquires the curl of the Lorentz force, so the two equations coincide only in the kinematic setting | The restricted transport equations coincide, which by this page’s rubric is exact; the row judges the dynamical inference built on them, which does not follow. Batchelor argued from the coincidence to a turbulent dynamo with magnetic energy in equipartition with the kinetic energy of the smallest eddies; Moffatt’s assessment is that the analogy “has to be used with great care” and that the argument “is really unsound”. | adds the identification of ν with η and of ω with B; a prescribed flow, for the induction equation · model homogeneous turbulence in a conducting fluidbreaks at the constraint: ω is the curl of the field that advects it and B is free, so in the prescribed-flow problem B can be amplified where ω cannot; with back-reaction the Lorentz force enters the vorticity equation and the shared equation is gone altogether | analogy |
| Kinetic theory and stellar dynamics | Vlasov equation; collisionless Boltzmann equation; Liouville’s theoremLiouville 1838 (as a result on Jacobians, not on phase space) · Jeans 1915 · Vlasov 1938 | Phase-space density f(x, v, t) ↔ a transported scalar carried by the Hamiltonian flow on phase space · the phase-space volume element ↔ a frozen density (Liouville) · a galaxy’s stars, or a plasma’s particles ↔ the fluid whose flow does the carrying | Hamilton’s equations for the particles → material conservation of f along the flow, exactly, in the absence of collisions. | adds no collisions; a mean field (self-consistent gravity or electromagnetism) as the force · model the Vlasov–Poisson or Vlasov–Maxwell systembreaks with collisions, which move particles across the flow (the Boltzmann collision term). Phase mixing is not a break: free streaming f(x, v, t) = f0(x − vt, v) develops ever-finer velocity-space structure while satisfying the transport equation exactly, and Landau damping arises within collisionless dynamics, proved rigorously for the nonlinear Vlasov equation under stability, smallness and regularity assumptions (Mouhot & Villani 2011). Coarse-graining changes the resolved description; it does not invalidate the fine-grained equation | exact |
| Relativistic fluids and plasmas | Relativistic Helmholtz–Kelvin theorem; vorticity 2-form; relativistic flux freezing; general-relativistic MHDLichnerowicz 1967 · Bekenstein & Oron 1978 · Carter 1979 | Two results. (a) Unmagnetised isentropic perfect fluid: the enthalpy-weighted velocity 1-form w = h uμdxμ, with h = (ε + p)/n the enthalpy per particle (c = 1) ↔ the momentum per particle, and its exterior derivative dw ↔ the transported vorticity 2-form, Lie-dragged along u. (b) Ideal MHD: dF = 0 and ιuF = 0 ↔ 𝓛uF = 0, the transported magnetic flux; the fluid vorticity dw is then not transported, because the Lorentz force acts, and a circulation theorem needs a modified integrand with an electromagnetic contribution (Bekenstein & Oron 2000) | Relativistic Euler → Lie transport of dw along u, exactly, for the isentropic unmagnetised fluid. Ideal Ohm’s law and Faraday’s law → Lie transport of F along u, exactly. Bekenstein and Oron (1978) give general criteria for when a flux conservation law exists; their 2000 paper gives the generalised circulation theorem for perfect MHD. | adds for (a), an isentropic perfect fluid; for (b), ideal conductivity; a fixed or self-consistent spacetime · model relativistic hydrodynamics and ideal general-relativistic MHDbreaks as in the non-relativistic rows, with dissipation and non-ideal Ohm’s law; and the two results must not be combined into one, since the fluid vorticity of a magnetised fluid is not transported. The Newtonian circulation theorem is recovered from (a) in the limit where h reduces to the particle mass m | exact |
| Cosmology | Flux freezing in an expanding universe; comoving magnetic field; B ∝ a−2Subramanian 2016 (review, Sec. 3.1, eqs 26 and 30) | B* = a²B ↔ the transported 2-form written in comoving coordinates · the Hubble flow ↔ a deformation gradient a·I relative to a reference epoch, so that comoving areas scale as a² · a constant comoving field ↔ flux conservation through a comoving surface | Ideal MHD on a Friedmann background → constancy of a²B in the absence of resistivity and peculiar velocities, exactly; the same theorem in a different flow. | adds a homogeneous expanding background; infinite conductivity; no peculiar motion · model ideal MHD in a Friedmann–Robertson–Walker spacetimebreaks with resistivity, and at the epochs where the plasma is not a single ideal fluid; peculiar velocities do not break flux freezing but do break the expansion-only scaling, since the field is then stretched and compressed beyond the uniform expansion; the origin of the seed field is not addressed by the theorem at all. Note that B ∝ a−2 is a decay under exact flux freezing: decay of a field is not by itself evidence that transport has failed | exact |
| Superconductivity | Meissner effect; London equations; “a superconductor is not merely a perfect conductor”Meissner & Ochsenfeld 1933 · London & London 1935 · Forrest 1983 (assessment) | An ideal conductor, in the comparison’s idealisation ↔ retained flux: the field present when resistance vanished stays trapped · a bulk superconductor in its equilibrium Meissner state ↔ flux expelled from the interior beyond the penetration depth, with geometry-dependent exceptions such as a hollow specimen · the London equation ∇×j = −(nse²/m)B ↔ constitutive information that the ideal-conductor comparison lacks, not a rival transport law · the fluxoid of a ring or a type-II vortex, Φ + μ0λL²∮js·dl ↔ quantised in units of h/2e; it coincides with the magnetic flux Φ only where the current contribution vanishes (Audit 13) | The counter-case. Ideal conductivity alone, in the idealised comparison, predicts retained flux; the superconductor’s equilibrium state is not reached by ideal transport from the initial field, and the distinction is the standard textbook statement of what the Meissner effect showed. | adds the ideal-conductor idealisation on one side and the equilibrium Meissner regime, bulk geometry and penetration depth on the other · model the London and Ginzburg–Landau descriptionsbreaks the premise of the comparison: the superconductor has zero resistance but its equilibrium is not the transported initial field, so the frozen-in theorem is the wrong tool; zero resistance is in any case not necessary for a particular flux to be preserved in a resistive medium; and whether the “perfect conductor” comparison is itself well posed has been questioned (see notes) | counter-case |
| Geometry | Lie transport; Lie-dragged forms; Cartan’s formula; frozen-in fields as invariantsArnold 1966 · Tur & Yanovsky 1993 · Arnold & Khesin 1998 · Frankel 2012, Ch. 4 | Frozen-in scalar, vector, covector, 2-form, density ↔ Lie-transported objects of each type, ∂tω + 𝓛vω = 0 · Helmholtz, Kelvin, Alfvén, Ertel ↔ instances · new invariants ↔ products and contractions of old ones, systematically (Tur and Yanovsky) · Euler flow ↔ a geodesic on the group of volume-preserving diffeomorphisms, with vorticity carried by the coadjoint action (Arnold) | Both directions, as definitions and theorems: the geometry says what “frozen in” means and generates every consequence that follows from transport alone. | adds smoothness; a flow map that exists for the times considered · model none; the equation, with no physical system behind itbreaks nothing, but it supplies no precondition: which objects are transported is decided by the equation of motion, outside the geometry. Its regularity assumptions (a smooth, invertible flow map) are a separate matter from whether the transport equation holds: resistive diffusion breaks flux transport while the velocity stays perfectly smooth | exact |
| Variational mechanics | Kelvin–Noether theorem; particle-relabelling symmetry; Euler–Poincaré reductionSalmon 1988 · Holm, Marsden & Ratiu 1998 | Kelvin’s circulation theorem ↔ Noether’s theorem for the symmetry of the fluid action under relabelling of particles · Ertel’s theorem ↔ the same symmetry with the stratification retained · the Kelvin–Noether theorem ↔ the abstract form covering compressible fluids and MHD | Action principle with relabelling symmetry → a circulation balance, as a theorem: the rate of change of circulation around a material loop equals a source built from the advected quantities, which vanishes in the barotropic unmagnetised case and is in general non-zero (baroclinic and magnetic forces generate circulation). Conservation is the case of vanishing source. | adds a Lagrangian formulation of the continuum; the symmetry group; the identification of the advected quantities that enter the source · model ideal continua with a variational principlebreaks nothing within the theorem; the conservative action used here does not include dissipative terms, so the theorem is silent about them. The precise correspondence between relabelling symmetry and Ertel’s theorem has been contested in the literature and is recorded here at second hand | exact |
The same transport statement appears for quantised vortex lines in superfluids (with mutual friction as the breaking term at finite temperature, and at zero temperature reconnection at the density zeros of the order parameter, where the regularity and nonvanishing-density assumptions of the classical argument fail), for the helicity and linking invariants that follow from it (Audit 12), and for the flux-conserving numerical schemes that try to preserve it on a grid. These are noted under extensions and not separately audited.
Historical relationships
The Lagrangian form of the statement, that vorticity divided by density is carried like a material line element, was written by Cauchy in his prize memoir of 1815, printed in 1827, for incompressible flow. Frisch and Villone’s 2014 study documents that Cauchy did not draw the consequences: Stokes remarked in 1883 that two of Helmholtz’s propositions “follow immediately from Cauchy’s integrals”, and Cauchy, in their account, did not perceive them. Helmholtz’s 1858 paper is the earliest source located in which the theorems are stated as theorems: vortex lines are material lines, and the strength of a vortex tube is constant along its length and in time. Kelvin’s paper on vortex motion, read in 1867 and printed in the Edinburgh Transactions for 1869, restated the result as the conservation of circulation around a material loop; Frisch and Villone note that Hankel had anticipated the circulation form in 1861. The meteorological line is separate in motivation: Rossby’s 1940 shallow-water potential vorticity and Ertel’s 1942 theorem for the stratified atmosphere, which Hoskins, McIntyre and Robertson’s 1985 review derives from the vorticity equation and the Bjerknes circulation theorem, the baroclinic extension of Kelvin’s.
The earliest located statement on the magnetic side is Alfvén’s, and I have not established whether it drew on the fluid theorems. Alfvén’s 1942 letter to Nature announced the waves that carry his name; the transport statement is in his 1943 paper in the Arkiv, where infinite conductivity is said to forbid any motion of the fluid across the lines of force, so that the matter is “fastened” to them. Blum and Choudhuri’s 2026 history, which supplies that quotation, also records that Alfvén’s argument was a heuristic later regarded as hand-waving, and locates the earliest general proof they found in a footnote to Cowling’s 1953 review “Solar electrodynamics”; that distinction between the statement and its proof is theirs and is adopted here. The word “frozen” is later usage, and I have not established where it was first used. Batchelor in 1950 drew the analogy between the induction equation and the vorticity equation explicitly and used it to argue for a turbulent dynamo; Moffatt’s 2002 assessment is that the analogy is imperfect and the argument unsound, for the reason given in the ledger. Newcomb’s 1958 paper is the earliest located source for the question of when field lines can be said to keep their identity. Cowling’s 1933 antidynamo theorem predates Alfvén’s theorem and constrains what frozen-in stretching can achieve; Zeldovich’s two-dimensional theorem of 1957 and the stretch–twist–fold picture of Vainshtein and Zeldovich in 1972 belong to the same line.
The unification is late and comes from geometry. Arnold’s 1966 paper put the Euler equations on the group of volume-preserving diffeomorphisms; Woltjer in 1958 and Moffatt in 1969 identified the helicity invariants that transport implies (Audit 12); Tur and Yanovsky in 1993 gave a general method for generating frozen-in invariants of every geometric type from one another, and Arnold and Khesin’s 1998 monograph made Lie transport the organising principle. Salmon in 1988 and Holm, Marsden and Ratiu in 1998 supplied the variational reading, in which circulation conservation is Noether’s theorem for particle relabelling. The pattern for the series, as far as the located sources show it: one structure written down in fluids by Cauchy without its consequences being drawn, stated as theorems by Helmholtz and Kelvin, stated for magnetic fields by Alfvén by a heuristic argument and proved generally by Cowling a decade later, with the vorticity–magnetic analogy drawn by Batchelor in 1950, and given a systematic common framework when the language of differential forms was applied to it a century after Helmholtz. Whether the plasma statement was independent of the fluid theorems is not established by locating early statements, and is not claimed.
Notes by row
Fluid dynamics exact
Within the Euler equations, for barotropic flow under conservative forces, the vorticity equation ∂tω = ∇×(v×ω) is exactly the statement that the vorticity 2-form is Lie-transported, and Helmholtz’s theorems and Kelvin’s are its integral forms; Batchelor’s Section 5.3 gives the standard derivation and the conditions. Cauchy’s formula is the Lagrangian solution: ω/ρ at the present position of a parcel is the initial ω0/ρ0 at that parcel’s label, pushed forward by the deformation gradient, so it behaves as a material line element; its components are not constant along the parcel, and the quantity that is conserved is the tube strength. Kelvin’s theorem is a statement about closed loops: (∂t + 𝓛v) applied to the velocity 1-form gives the exact differential d(½|v|² − h − Φ), which integrates to zero around a closed material curve and not along an open one, so the 1-form is transported only up to an exact term while the 2-form ∇×v is transported strictly. The verdict is exact because nothing is approximated within the ideal equation. The breaks line: viscosity diffuses vorticity across material surfaces; baroclinic torque, present whenever surfaces of constant density and constant pressure do not coincide, creates circulation, which is why the theorem needs Ertel’s modification in the atmosphere; and in a turbulent flow the transport stretches vortex lines until their cross-section is small enough for viscosity to matter. That last statement is about turbulent flows, not about transport as such: a uniform translation or a rigid rotation transports vorticity exactly and creates no small scales at all.
Geophysical fluid dynamics exact
Ertel’s theorem is not the vorticity theorem contracted with a scalar, which is how the first draft of this page described it and which is wrong: in a baroclinic atmosphere the absolute-vorticity 2-form is not transported, because (∂t + 𝓛v)Ωa = ρ−2dρ∧dp, and that does not vanish. What is transported is the 3-form Ωa∧dθ, because with θ advected and an equation of state p = p(ρ, θ) the baroclinic term wedged with dθ is dρ∧dp∧dθ = 0. Dividing a transported 3-form by the transported mass form ρ d³x gives a transported scalar, q = (ωa·∇θ)/ρ, materially conserved by every parcel of air; ρq d³x is the corresponding density. McIntyre’s article makes the cancellation of the baroclinic term explicit (equations 4 to 10, p. 377), and the closure matters: a fluid with a second independent advected composition variable does not satisfy dρ∧dp∧dθ = 0 in this form and needs its own potential vorticity. Hoskins, McIntyre and Robertson’s review states the conservation principle and the separate invertibility principle in its summary, gives q (their P) as their equation (12), and derives it from the vorticity equation and the Bjerknes circulation theorem rather than from Kelvin’s, the difference being that Bjerknes’s form keeps the baroclinic term that Kelvin’s assumption removes. Rossby’s 1940 shallow-water result is the layered precursor. The verdict is exact within the inviscid, adiabatic, rotating compressible Euler equations with that closure; the hydrostatic primitive equations used in forecasting have their own form of the theorem, with their own approximations, which McIntyre states separately (p. 377, after equation 4). The breaks are diabatic heating and friction, which are the sources and sinks that forecasting has to model. Invertibility, the principle that the distribution of q determines the rest of the flow given balance and boundary conditions, is what makes the quantity useful, and it is a different theorem with different assumptions, not part of the transport statement.
Magnetohydrodynamics exact
With infinite conductivity, Ohm’s law reduces to E + v×B = 0, Faraday’s law becomes ∂tB = ∇×(v×B), and that is the same equation as the barotropic vorticity row with B for ω: magnetic flux through a surface carried by the material velocity is conserved, and B/ρ obeys Cauchy’s formula, with ρ conserved by that same velocity. Alfvén’s 1943 statement, that the matter is “fastened” to the lines of force because any motion across them would drive infinite currents, is quoted here from the secondary literature (Blum and Choudhuri 2026 and others), since the Arkiv paper was not read; Landau and Lifshitz’s Section 65 and Moffatt’s third chapter are the standard textbook treatments, cited at record level. Two qualifications belong to this row. The first is the choice of velocity. The condition for a velocity w to preserve flux is ∇·B = 0 and ∇×(E + w×B) = 0 with w sufficiently regular, and many velocities satisfy it, any two differing by a component along B and in special cases by more; Newcomb’s 1958 paper distinguishes this flux preservation from the weaker preservation of field-line connections (cited at record level, with its content taken from later papers that call it Newcomb’s criterion). What singles out the bulk velocity in ideal MHD is not convention but the model: the fluid equations define the material velocity and the ideal Ohm’s law makes it a transporting velocity. Other flux-preserving velocities exist without being material velocities, and the electron velocity of Hall MHD transports B because that model’s induction equation says so, not because it can be swapped in at will. The second is the breaks line. Finite resistivity lets flux slip through a fluid whose velocity remains perfectly smooth, and in resistive MHD it permits reconnection where the field gradients are steep; collisionless reconnection, supported by electron pressure-tensor and inertia terms in the generalised Ohm’s law, needs no resistivity at all (Kuznetsova, Hesse and Winske 2001, abstract); the Hall term by itself does not break the theorem but transfers it to the electron velocity, as the next row records. In either case the topology that ideal transport preserves is changed in a thin layer; reconnection can release stored magnetic energy, and it is central to models of solar flares. Priest and Forbes’s monograph is the standard account of the resistive case, cited at record level.
Two-fluid and Hall plasmas exact
Drop the assumption that ions and electrons move together and the theorem does not fail; it multiplies, provided each species has a barotropic scalar pressure, since a pressure that depends on more than the species density supplies a baroclinic source of exactly the kind that undoes the fluid row. Each species s, treated as an ideal fluid with such a pressure, has a canonical vorticity ∇×(msvs + qsA) that is transported by that species’ own velocity. In Hall MHD, which neglects electron inertia and keeps the Hall term, this reduces to two statements which D’Avignon, Morrison and Lingam put directly (their discussion after equation 5, and equations 26 and 31): the magnetic field’s fluxes are carried with the electron velocity, supported by a barotropic electron pressure, and the magnetic field plus the ion vorticity is carried with the ion velocity, supported by a barotropic ion pressure. Ordinary ideal MHD is not the limit in which ion inertia is dropped, as the first draft of this page said; it keeps the bulk inertia, which the ions mostly carry, and additionally neglects the Hall contribution at the scales considered. The row is exact within the ideal two-fluid equations, and its lesson for the series is the third condition above: “frozen in” is not a property of the magnetic field alone. Into which velocity, and which combination of field and vorticity, depends on the model, and a claim that omits the answer is not yet a claim.
Dynamo theory limited
The dynamo problem uses the theorem as a mechanism rather than as a conservation law, and three things have to be kept apart, which the first draft of this page ran together. First, ideal amplification: if flux is transported, stretching a flux tube amplifies the field inside it, and this needs no resistivity whatever. The uniform strain v = (γx, −γy, 0) with B = B0eγtx̂ solves the ideal induction equation exactly and grows exponentially; it is a stretching flow in an unbounded domain, not a dynamo, but it disposes of any claim that amplification requires the ideal condition to be broken. Second, the stretch–twist–fold construction of Vainshtein and Zeldovich, in which a bounded flow stretches, twists and folds a flux tube so that the folded loops lie in the same orientation and their fluxes reinforce, doubling the flux through a cross-section at each cycle; at exactly zero resistivity the folded loops are never smoothed into a single tube and the structure becomes ever finer, so what the repeated construction sustains is a separate question from whether stretching amplifies. Third, the fast-dynamo problem, whether the growth rate stays positive as the resistivity tends to zero, in which the small resistivity matters critically and the limit and the ideal case differ; Childress and Gilbert’s monograph is the standard treatment, and the statement that the limit needs care is recorded without a pinpoint. The antidynamo theorems bound the second and third: Cowling’s (1933) excludes axisymmetric magnetic fields, which is a constraint on the field and not on every axisymmetric flow, and Zeldovich’s (1957) excludes planar flows. The verdict is limited because the row’s correspondence is between the theorem and a long-time, bounded-domain growth problem that the ideal theorem alone does not settle.
Turbulence, the vorticity–magnetic analogy analogy
This is the row where “the same equation” is exactly true and “the same physics” is exactly false, and it is the reason the series insists on the difference. The setting has to be stated carefully, because the first draft of this page did not. The vorticity of an incompressible viscous fluid and the magnetic field under kinematic induction in a prescribed incompressible flow obey the same advection–diffusion equation, with ν in the place of η. That is the comparison Batchelor drew; it is not a statement about coupled magnetohydrodynamics, where the vorticity equation acquires the curl of the Lorentz force and the two equations no longer coincide. Batchelor’s 1950 paper argued from the coincidence to a turbulent dynamo in which the magnetic energy comes into equipartition with the kinetic energy of the smallest eddies. Moffatt’s 2002 assessment, from which the quotations in the ledger are taken, is that the analogy is imperfect because vorticity is constrained by ω = ∇×u to the very velocity that convects it, whereas B is free of any such constraint, and that Batchelor’s argument is unsound as a result: in the prescribed-flow problem a given flow can amplify B where it cannot amplify its own vorticity, and with back-reaction the comparison ends. The verdict is analogy because the correspondence judged is Batchelor’s inference from the shared equation to shared dynamics; the restricted equations themselves coincide, which by this page’s rubric is exact, and the dictionary column says which is being judged. The row is the cleanest example in the series of a shared equation that is not a shared problem.
Kinetic theory and stellar dynamics exact
On the six-dimensional phase space of a single particle, Hamilton’s equations define a flow, and that flow preserves the phase-space volume; the density f of particles in phase space is therefore carried as a transported scalar, which is the collisionless Boltzmann equation of stellar dynamics (the term “Jeans equations” is best kept for the velocity-moment equations derived from it) and Vlasov’s equation for a plasma. The names attach to the same statement in different fields, and the row is exact. Two cautions on the history. The 1838 note of Liouville that is universally cited is a result about the Jacobian of a system of ordinary differential equations; it does not mention phase space, and the theorem that carries his name was given its mechanical reading later, a point made in the secondary literature and recorded here at that level. And the Vlasov and Jeans papers are cited at record level only. The breaks line is collisions, which move particles across the flow and are the Boltzmann collision term. Phase mixing is not a break, and the first draft of this page was wrong to list it with the failures of the fluid rows. Free streaming, f(x, v, t) = f0(x − vt, v), develops ever-finer structure in velocity space while satisfying the collisionless equation exactly at every time, and Landau damping, the decay of an electric field in a collisionless plasma, arises within that collisionless dynamics, self-consistent field included, and not from transport alone; Mouhot and Villani proved it for the nonlinear Vlasov equation under their stability, smallness and regularity assumptions (2011, abstract). Coarse-graining changes the resolved description and produces an apparent entropy; it does not retrospectively invalidate the fine-grained equation, and any collision rate however small is a separate physical term with its own scale.
Relativistic fluids and plasmas exact
Two results share the row and the first draft of this page combined them wrongly. For an unmagnetised isentropic perfect fluid, the object that is transported is the exterior derivative of the enthalpy-weighted velocity 1-form, w = h uμdxμ; the velocity 1-form is already covariant, and with h = (ε + p)/n the enthalpy per particle in units with c = 1, huμ is the momentum per particle, which is what the circulation theorem is about (Bekenstein and Oron 2000, equation 3.19, as cited in review). Lichnerowicz’s 1967 lectures are the earliest located source for the theorem in that form (record level), and the Newtonian theorem is recovered where h reduces to the particle mass m. For magnetic flux the argument is separate and short: dF = 0 always, and the ideal Ohm’s law ιuF = 0 then gives 𝓛uF = 0 by Cartan’s formula, so the electromagnetic 2-form is Lie-dragged along the fluid worldlines. What does not follow is that dw is transported in a magnetised fluid: the Lorentz force acts on the fluid, and a circulation theorem for perfect MHD needs a modified integrand with an electromagnetic contribution, which Bekenstein and Oron give in their 2000 paper (abstract read; equation numbers as cited in review). Their 1978 paper, whose abstract states that it gives “the general criteria for a flux conservation law to exist”, is the source for the general criteria; Carter’s 1979 treatment is cited at record level. Each result is exact within its own ideal equations, and each breaks as the non-relativistic rows do.
Cosmology exact
In an expanding universe the flow is the Hubble expansion, and the theorem says that the flux through a comoving surface is conserved, which for a surface whose area grows as a² means B ∝ a−2. Subramanian’s review defines the comoving field B* = a²B as its equation (26) and states after equation (30) that in the absence of resistivity and peculiar velocities B* is constant; that passage was read. The row is exact because it is the same theorem in a different flow, and it is included because the cosmological literature tends to present the scaling as its own result. The deformation gradient of the Hubble flow is a·I relative to a reference epoch, so areas scale as a² and the flux through a comoving surface is constant. The breaks are the usual ones plus the ones the background adds; peculiar velocities do not break flux freezing but do break the expansion-only scaling, since they stretch and compress the field beyond the uniform expansion, and at early epochs the plasma is not a single ideal fluid. Two things the row is careful about: B ∝ a−2 is a decay that happens under exact transport, so the decay of a field is not by itself evidence that transport has failed; and where the seed field came from is a separate question that the theorem does not touch.
Superconductivity counter-case
The row is a counter-case, included because it is the standard demonstration that infinite conductivity and the frozen-in theorem are not the same thing as superconductivity. An ideal conductor, in the idealisation the comparison uses, cooled in a magnetic field would keep the flux it had when its resistance vanished. Meissner and Ochsenfeld found in 1933 that a bulk superconductor in its equilibrium state expels the field from its interior instead, beyond a penetration depth at the surface and with geometry-dependent exceptions; Forrest’s 1983 reassessment, which was read, states the distinction between a superconductor and a perfect conductor in those terms and discusses both the penetration and the flux retained through a hollow specimen. The London equations of 1935 supply constitutive information about the superconducting state that the ideal-conductor comparison lacks, and have expulsion as a consequence; they are not a rival transport law. Tinkham’s textbook, where the same distinction opens the subject, is cited at record level. The verdict is “counter-case” because the transport equation is not the governing equation of the superconductor: its equilibrium is not the transported initial field. Three cautions. Zero resistance is not a universally necessary condition for flux to be preserved, since particular fields survive in resistive media, so the comparison rests on its stated idealisation. The “perfect conductor” of the comparison has itself been argued over (Fiolhais, Essén and co-authors dispute the textbook account), and the row records the standard distinction without adjudicating that argument. And what is quantised for a ring or a type-II vortex is the fluxoid, Φf = Φ + μ0λL²∮js·dl = n h/2e, which includes a current contribution and coincides with the magnetic flux Φ only where that contribution vanishes (Brandt and Clem 2004, abstract); that is Audit 13.
Geometry exact
The geometric statement is the one in the equation box: an object of any tensor type is frozen in when its Lie derivative along the flow, plus its partial time derivative, vanishes, and Cartan’s formula makes the Lie derivative of a form computable from d and the contraction with v. Frankel’s fourth chapter gives the formula and its consequences for differentiating integrals over moving regions; the application to vorticity and flux is in his examples rather than in a titled section, and the chapter is cited at record level. Tur and Yanovsky’s 1993 paper, whose abstract was read, gives the general method for deriving invariants of dissipationless media and the relations between frozen-in scalars, vectors and covectors, so that new invariants are generated from old ones by products and contractions. Ertel’s theorem is the example on this page: the absolute-vorticity 2-form is not transported in baroclinic flow, but its wedge product with the transported 1-form dθ is, because the baroclinic source dρ∧dp vanishes once wedged with dθ under the closure p = p(ρ, θ). Arnold’s 1966 paper, whose abstract was read, put the Euler equations on the group of volume-preserving diffeomorphisms as geodesic flow, and Arnold and Khesin’s monograph, cited at record level, made transport and its topological consequences the organising principle of the subject. The row is exact because these are definitions and theorems. What the geometry does not supply is the precondition: it says what transport implies, not which objects are transported, and that decision belongs to the equation of motion. Its regularity assumption, a smooth invertible flow map, is a separate condition from the transport equation itself: resistive diffusion breaks the transport of magnetic flux while the transporting velocity remains perfectly smooth, so a failure of transport is not the same as a failure of regularity, and the first draft of this page conflated them.
Variational mechanics exact
Kelvin’s theorem is not an accident of the Euler equations; it is what a symmetry gives. The fluid action is unchanged if the particles are relabelled, and Noether’s theorem turns that symmetry into a statement about circulation. The precise statement is a balance, not a conservation law: the general Kelvin–Noether theorem of Holm, Marsden and Ratiu (their Theorem 4.1, with the fluid and MHD cases in Corollary 6.3 and equation 7.18, as cited in review; the abstract was read) gives the rate of change of circulation around a material loop as a source built from the advected quantities, and that source is non-zero in general, since baroclinic and magnetic forces generate circulation. Conservation is the case in which the source vanishes, which for the barotropic unmagnetised fluid it does. Salmon’s 1988 review is the standard source for the relabelling argument and is cited at record level. The verdict is exact because it is a theorem. Two qualifications. The conservative action used here does not include dissipative terms, so the variational reading is silent about the breaks line rather than contradicted by it. And the precise correspondence between relabelling symmetry and Ertel’s theorem, as opposed to Kelvin’s, has been contested in the literature (Charron and Zadra, among others, dispute the standard association), which is recorded here at second hand and not adjudicated.
The dictionary
Terms that name the same object across rows, within each row’s ideal equations, except where an entry says otherwise.
- The transported object
- ω, a Lie-transported form · barotropic vorticity 2-form (fluids); the velocity 1-form only up to an exact differential · potential temperature scalar and the 3-form Ωa∧dθ, hence the scalar q (geophysics); the absolute-vorticity 2-form is not transported in baroclinic flow · magnetic 2-form B (MHD, cosmology) · canonical vorticity per species, given barotropic species pressures (two-fluid) · phase-space density scalar and volume form (kinetics) · dw for the unmagnetised isentropic fluid, and F for the ideal conductor, separately (relativity)
- The conserved integral
- ∫C(t)ω over a material chain · circulation around a material loop (Kelvin) · flux through a material surface (Helmholtz’s tube strength; Alfvén) · potential vorticity per parcel (Ertel) · number of particles in a comoving phase-space volume (Liouville, Vlasov) · flux through a comoving surface (cosmology)
- The Lagrangian solution
- (ω/ρ)(X(a,t), t) = ∇aX·(ω0/ρ0)(a), the initial field, expressed in material coordinates, pushed forward by the deformation gradient ∇aX · Cauchy’s vorticity formula (1815, incompressible) · the same for B/ρ, with ρ conserved by the transporting velocity (every MHD text) · a·I for the Hubble flow, so that areas scale as a²
- The ideal condition
- what buys the transport statement from the equation of motion · inviscid, barotropic, conservative forces (fluids) · adiabatic and frictionless with p = p(ρ, θ) (geophysics) · infinite conductivity (MHD) · ideal species equations with barotropic pressures (two-fluid) · no collisions (kinetics) · isentropic perfect fluid, or ideal Ohm’s law, for the two separate results (relativity)
- The transporting velocity
- mathematically non-unique: v + λω transports ω as well as v does, and any regular w with ∇×(E + w×B) = 0 preserves the flux of B · physically selected by the model: the material velocity of ideal MHD, through its Ohm’s law · the electron velocity, for B in Hall MHD · the ion velocity, for B plus ion vorticity · the Hamiltonian flow, for f · the fluid worldlines, in relativity · flux preservation ↔ a stronger condition than field-line-connection preservation (Newcomb)
- The breaking term
- viscosity ν∇²ω · baroclinic torque ∇ρ×∇p/ρ² · diabatic heating and friction (sources of q) · resistivity η∇²B, and the reconnection it permits in resistive MHD; pressure-tensor and inertia terms, which permit reconnection without it · the collision term (not phase mixing, which is exact transport) · mutual friction (superfluids at finite temperature); at zero temperature no term is added, and reconnection at the density zeros of the order parameter is a failure of the regularity and nonvanishing-density assumptions instead · terms that act most strongly at small scales, which may be created by the flow, present in the initial field or imposed from outside
- Topological consequence
- for a transported 2-form under a regular invertible flow: linkage and knottedness of its lines are preserved · helicity ∫v·ω or ∫A·B conserved, under the appropriate domain, boundary and gauge conditions (Woltjer, Moffatt; Audit 12) · a transported scalar carries no such invariant · reconnection ↔ a change of topology, which requires either a term added to the transport equation or a failure of the hypotheses (a regular invertible flow; a nonvanishing density where the argument needs one) under which preservation was established
- Counter-case
- the superconductor: zero resistance in its equilibrium Meissner state without retained flux, because that equilibrium is not reached by transport
Verdict
Three claims are usually run together. The ledger supports the first once the transported object is identified correctly, supports the second with a qualification that is easy to overlook, and supports the third only as a statement about particular solutions.
First: Helmholtz’s, Kelvin’s, Alfvén’s and Ertel’s theorems, the collisionless Boltzmann equation and their relativistic and cosmological forms share one geometric language, Lie transport by a flow, but the transported object must be identified correctly, and it differs from row to row in a way the names hide. Barotropic vorticity and the ideal-MHD magnetic field are transported 2-forms. Kelvin’s circulation is conserved around closed material loops because the velocity 1-form is transported only up to an exact differential. Ertel’s potential vorticity is a transported scalar, obtained from a transported 3-form and the mass form, and it survives in baroclinic flow precisely where the vorticity 2-form does not. The collisionless distribution function is a transported scalar on phase space. Within each row’s ideal equations, and for the object so identified, the correspondence is exact, and the geometric rows say why the transport-only consequences then follow without further physics.
Second: the geometric type, the transporting flow and the governing equations jointly determine what is conserved, and none of the three can be left implicit. The Lagrangian invariant of vorticity is ω/ρ, carried as a material line element, not ω; the transporting velocity is not mathematically unique for any of these objects and is selected by the physical model, the material velocity of ideal MHD by its Ohm’s law and the electron velocity of Hall MHD by its induction equation; in a two-fluid plasma the transported object changes with the model and needs barotropic species pressures. Shared transport equations do not establish shared dynamics or a common mechanism: the vorticity–magnetic analogy has the same restricted equation on both sides and different physics, because one object is tied to the flow that carries it and the other is not.
Third: that the theorem, being exact, describes what these systems do. It describes what they do while the ideal condition holds. Transport can stretch and fold structure into scales where weak non-ideal effects become important, and in turbulent flows, in reconnecting current sheets and in the dynamo it does; whether a given flow does so, and whether the result is dissipation, reconnection or something else, depends on the solution, its scales and times, and the model, and is not a further consequence of Lie transport. A uniform translation creates no small scales; phase mixing and magnetic amplification proceed while ideal transport remains exact; a field can decay, as the cosmological row shows, without transport failing at all. The first draft of this page claimed that transport ends its own validity in every row, and that claim is withdrawn. What remains is narrower and still worth saying: where non-ideal effects matter, the scales at which they act may be created by the flow, inherited from the initial conditions or imposed from outside (a resistive fluid at rest with B = B0 sin(kx) e−ηk²t ŷ decays at a scale no flow created), and a page that reports the theorem without the breaks line has reported the less interesting half.
What this means for a hydrodynamic theory of quantum mechanics or gravity. Any theory that models a field as a structure carried by a flow inherits this page, but not as a bundle. Transport preserves the geometric structure of the object under a sufficiently regular invertible flow, and which invariants follow depends on what the object is: a transported 2-form has conserved fluxes and, under the appropriate domain, boundary and gauge conditions, conserved linkage and helicity; a transported scalar has neither; quantised circulation needs a phase or order-parameter structure that transport alone does not supply. Nothing about a singular limit comes for free either. So the obligation on such a theory is specific: to say which geometric object its equations transport, by which velocity, under which ideal condition, and to identify which invariant that establishes; and then, for any process in which that invariant is claimed to change, reconnection or a change of linking, to say what is modified: either a term added to the transport equation, with the scale on which it acts, or a named failure of the hypotheses under which topology preservation was established. Decay is not such a process by itself, since a field can decay under exact transport. The Meissner row is the caution on the other side: zero resistance is not the same as retained flux, and a medium that is ideal in one sense may reach its equilibrium by a route that ideal transport does not describe.
What this page does not claim
It does not claim that any of the physical rows is exact as a description of its system; each is exact within an ideal equation. It does not claim that the absolute-vorticity 2-form is transported in a baroclinic fluid, or that Ertel’s potential vorticity is a density; it is a scalar, and it exists because a 3-form is transported where the 2-form is not. It does not claim that the velocity 1-form is transported; only its closed-loop integrals are conserved. It does not claim that transport creates small scales in every flow, or that phase mixing, magnetic amplification or the decay of a field are evidence that transport has failed; they are not. It does not claim that reconnection or amplification requires resistivity. It does not claim that Alfvén derived his theorem from Helmholtz’s, used the word “frozen”, or gave a general proof; the 1943 wording is “fastened”, the argument was heuristic, and the earliest general proof located by Blum and Choudhuri is Cowling’s of 1953. It does not claim that Cauchy stated Helmholtz’s theorems; he wrote the formula from which they follow and, on Frisch and Villone’s account, did not draw them. It does not claim that Liouville’s 1838 note is about phase space. It does not claim that the vorticity–magnetic analogy is more than kinematic, or that a superconductor is a perfect conductor, or that magnetic flux and fluxoid are the same quantity. It does not claim that the invertibility of potential vorticity is part of Ertel’s theorem, or that the relabelling-symmetry derivation of Ertel’s theorem is settled, or that the Kelvin–Noether theorem is a conservation law rather than a balance. It does not claim that every transported object inherits flux, linkage, helicity and quantisation together. It does not audit helicity, quantised circulation or the numerical preservation of flux, which are noted below. And it does not claim that the transporting velocity is mathematically unique for any of these objects; it is selected by the model.
Extensions of the structure, and what is known about each
Every entry below was searched before being written.
Next in the series
Audit 11 is the Bianchi identities: the statement that the boundary of a boundary is zero, d² = 0, which appears as the identity dF = 0 in electromagnetism (charge conservation follows separately, by applying d to the sourced equation d⋆F = J), as the contracted Bianchi identity that makes the Einstein equations consistent, and as the integrability conditions of any gauge theory. It is the second conservation-law structure in the series, and unlike this one it is an identity rather than a theorem, which is why the distinction between the identity and the field equations it is applied to will be the page’s main business.
References
Bibliographic details for every entry (authors, title, year, journal, volume, page) were verified against the publisher’s record or an equivalent index, with two exceptions marked at the entry: the page range of Carter 1979 and the editor attribution of Cowling 1953. The flag beside each says how far the content itself was consulted. Where a claim in the text rests on a specific passage, the section, equation or page is given in the text.
Alfvén, H. (1942). Existence of electromagnetic-hydrodynamic waves. Nature 150, 405–406. abstract
Alfvén, H. (1943). On the existence of electromagnetic-hydrodynamic waves. Ark. Mat. Astron. Fys. 29B(2), 1–7. secondary (the “fastened” passage, p. 4, quoted by Blum & Choudhuri 2026 and others)
Alfvén, H. (1976). On frozen-in field lines and field-line reconnection. J. Geophys. Res. 81, 4019–4021. record
Arnold, V. I. (1966). Sur la géométrie différentielle des groupes de Lie de dimension infinie et ses applications à l’hydrodynamique des fluides parfaits. Ann. Inst. Fourier 16(1), 319–361. abstract
Arnold, V. I. & Khesin, B. A. (1998). Topological Methods in Hydrodynamics. Springer (Applied Mathematical Sciences 125). record (chapter titles)
Barenghi, C. F., Donnelly, R. J. & Vinen, W. F. (1983). Friction on quantized vortices in helium II. A review. J. Low Temp. Phys. 52, 189–247. abstract
Batchelor, G. K. (1950). On the spontaneous magnetic field in a conducting liquid in turbulent motion. Proc. R. Soc. Lond. A 201, 405–416. record
Batchelor, G. K. (1967). An Introduction to Fluid Dynamics. Cambridge University Press. Sec. 5.3, Kelvin’s circulation theorem and vorticity laws for an inviscid fluid. record (table of contents)
Bekenstein, J. D. & Oron, E. (1978). New conservation laws in general-relativistic magnetohydrodynamics. Phys. Rev. D 18, 1809–1819. abstract
Bekenstein, J. D. & Oron, A. (2000). Conservation of circulation in magnetohydrodynamics. Phys. Rev. E 62, 5594–5602. abstract (eqs 3.32–3.34 as cited in review)
Blum, A. S. & Choudhuri, A. R. (2026). Combining electrodynamics with hydrodynamics: the origins and development of magnetohydrodynamics. Arch. Hist. Exact Sci. 80, art. 16. full text (the passages on Alfvén’s 1943 argument and on Cowling’s 1953 proof)
Brandt, E. H. & Clem, J. R. (2004). Superconducting thin rings with finite penetration depth. Phys. Rev. B 69, 184509. abstract
Carter, B. (1979). Perfect fluid and magnetic field conservation laws in the theory of black hole accretion rings. In C. Hazard & S. Mitton (eds), Active Galactic Nuclei, Cambridge University Press. record (pages not verified)
Cauchy, A.-L. (1827). Théorie de la propagation des ondes à la surface d’un fluide pesant d’une profondeur indéfinie. Mém. présentés par divers savans à l’Acad. Roy. Sci. Inst. France 1, 5–318 (prize memoir of 1815). secondary (via Frisch & Villone 2014)
Charron, M. & Zadra, A. (2017). On the dynamical consequences of the relabeling symmetry. arXiv:1704.04214. abstract
Childress, S. & Gilbert, A. D. (1995). Stretch, Twist, Fold: The Fast Dynamo. Springer (Lecture Notes in Physics Monographs 37). abstract
Cowling, T. G. (1933). The magnetic field of sunspots. Mon. Not. R. Astron. Soc. 94, 39–48. record
Cowling, T. G. (1953). Solar electrodynamics. In G. P. Kuiper (ed.), The Sun, University of Chicago Press, 532–591. secondary (via Blum & Choudhuri 2026, who cite the footnote on pp. 542–543; the editor’s name as given there is not confirmed)
D’Avignon, E. C., Morrison, P. J. & Lingam, M. (2016). Derivation of the Hall and extended magnetohydrodynamics brackets. Phys. Plasmas 23, 062101. full text (arXiv version; the statement of which field is carried by which velocity, and the barotropic-pressure conditions after eq. 5 and at eqs 26 and 31)
Donnelly, R. J. (1991). Quantized Vortices in Helium II. Cambridge University Press. record
Ertel, H. (1942). Ein neuer hydrodynamischer Wirbelsatz. Meteorol. Z. 59, 277–281. secondary
Evans, C. R. & Hawley, J. F. (1988). Simulation of magnetohydrodynamic flows: a constrained transport method. Astrophys. J. 332, 659–677. record
Forrest, A. M. (1983). Meissner and Ochsenfeld revisited. Eur. J. Phys. 4, 117–120. full text
Frankel, T. (2012). The Geometry of Physics, 3rd ed. Cambridge University Press. Ch. 4, The Lie derivative (Secs 4.2–4.3). record (table of contents)
Frisch, U. & Villone, B. (2014). Cauchy’s almost forgotten Lagrangian formulation of the Euler equation for 3D incompressible flow. Eur. Phys. J. H 39, 325–351. full text (Secs II.C and IV)
Helmholtz, H. (1858). Über Integrale der hydrodynamischen Gleichungen, welche den Wirbelbewegungen entsprechen. J. reine angew. Math. 55, 25–55. English translation by P. G. Tait, Phil. Mag. (4) 33, 485–512 (1867). record
Holm, D. D., Marsden, J. E. & Ratiu, T. S. (1998). The Euler–Poincaré equations and semidirect products with applications to continuum theories. Adv. Math. 137, 1–81. abstract
Hoskins, B. J., McIntyre, M. E. & Robertson, A. W. (1985). On the use and significance of isentropic potential vorticity maps. Q. J. R. Meteorol. Soc. 111, 877–946. full text (summary; Sec. 1(b), eqs 1, 6 and 12)
Jeans, J. H. (1915). On the theory of star-streaming and the structure of the universe. Mon. Not. R. Astron. Soc. 76, 70–84. record
Koplik, J. & Levine, H. (1993). Vortex reconnection in superfluid helium. Phys. Rev. Lett. 71, 1375–1378. record
Kuznetsova, M. M., Hesse, M. & Winske, D. (2001). Collisionless reconnection supported by nongyrotropic pressure effects in hybrid and particle simulations. J. Geophys. Res. 106(A3), 3799–3810. abstract
Landau, L. D. & Lifshitz, E. M. (1984). Electrodynamics of Continuous Media, 2nd ed. Pergamon. Sec. 65, The equations of motion for a fluid in a magnetic field. record (table of contents)
Landau, L. D. & Lifshitz, E. M. (1987). Fluid Mechanics, 2nd ed. Pergamon. Sec. 8, The conservation of circulation. record (table of contents)
Lichnerowicz, A. (1967). Relativistic Hydrodynamics and Magnetohydrodynamics: Lectures on the Existence of Solutions. W. A. Benjamin. record
Liouville, J. (1838). Note sur la théorie de la variation des constantes arbitraires. J. Math. Pures Appl. (1) 3, 342–349. record (its relation to phase space from Nolte 2010, secondary)
London, F. & London, H. (1935). The electromagnetic equations of the supraconductor. Proc. R. Soc. Lond. A 149, 71–88. record
McIntyre, M. E. (2015). Potential vorticity. In G. R. North, J. Pyle & F. Zhang (eds), Encyclopedia of Atmospheric Sciences, 2nd ed., vol. 2, 375–383. Academic Press. secondary (eqs 4–10, p. 377, as cited in review; the author’s own copy is online)
Mahajan, S. M. & Yoshida, Z. (1998). Double curl Beltrami flow: diamagnetic structures. Phys. Rev. Lett. 81, 4863–4866. abstract
Meissner, W. & Ochsenfeld, R. (1933). Ein neuer Effekt bei Eintritt der Supraleitfähigkeit. Naturwissenschaften 21, 787–788. record
Moffatt, H. K. (1969). The degree of knottedness of tangled vortex lines. J. Fluid Mech. 35, 117–129. abstract
Moffatt, H. K. (1978). Magnetic Field Generation in Electrically Conducting Fluids. Cambridge University Press. Ch. 3. secondary
Moffatt, H. K. (2002). G. K. Batchelor and the homogenization of turbulence. Annu. Rev. Fluid Mech. 34, 19–35. full text (pp. 27–28)
Mouhot, C. & Villani, C. (2011). On Landau damping. Acta Math. 207, 29–201. abstract
Newcomb, W. A. (1958). Motion of magnetic lines of force. Ann. Phys. (N.Y.) 3, 347–385. record (the criterion as cited by later papers)
Nolte, D. D. (2010). The tangled tale of phase space. Physics Today 63(4), 33–38. full text
Priest, E. & Forbes, T. (2000). Magnetic Reconnection: MHD Theory and Applications. Cambridge University Press. record
Rossby, C.-G. (1940). Planetary flow patterns in the atmosphere. Q. J. R. Meteorol. Soc. 66 (Suppl.), 68–87. record
Salmon, R. (1988). Hamiltonian fluid mechanics. Annu. Rev. Fluid Mech. 20, 225–256. record
Subramanian, K. (2016). The origin, evolution and signatures of primordial magnetic fields. Rep. Prog. Phys. 79, 076901. full text (Sec. 3.1, eqs 26 and 30)
Thomson, W. (Lord Kelvin) (1869). On vortex motion. Trans. Roy. Soc. Edinburgh 25, 217–260 (read 1867; the part is dated 1868 by the publisher). abstract
Tinkham, M. (1996). Introduction to Superconductivity, 2nd ed. McGraw-Hill. record
Tur, A. V. & Yanovsky, V. V. (1993). Invariants in dissipationless hydrodynamic media. J. Fluid Mech. 248, 67–106. abstract
Vainshtein, S. I. & Zeldovich, Ya. B. (1972). Origin of magnetic fields in astrophysics (turbulent “dynamo” mechanisms). Sov. Phys. Usp. 15, 159–172. abstract
Vlasov, A. A. (1938). On vibration properties of electron gas. Zh. Eksp. Teor. Fiz. 8, 291 (in Russian); English translation, The vibrational properties of an electron gas, Sov. Phys. Usp. 10, 721–733 (1968). secondary
Woltjer, L. (1958). A theorem on force-free magnetic fields. Proc. Natl. Acad. Sci. USA 44, 489–491. record
Zeldovich, Ya. B. (1957). The magnetic field in the two-dimensional motion of a conducting turbulent liquid. Sov. Phys. JETP 4, 460–462. record
Change log
0.3 · 11 Sep 2026 Focused correction pass after a third external review. Surviving universal claims removed: the opening no longer says the ideal condition fails on small scales in every case or that a whole family of consequences follows for free, and the verdict now says that the scales at which non-ideal effects act may be created by the flow, inherited from the initial conditions or imposed from outside. Geometry note: Ertel’s theorem described as the vanishing of the baroclinic source once wedged with dθ, not as arising from Helmholtz’s. Geophysical model line names inviscid adiabatic rotating compressible Euler flow with the closure, with the hydrostatic primitive-equation versions stated separately. Reconnection described as requiring either an added term or a failure of the hypotheses under which preservation was established; the density zeros of a superfluid order parameter are the latter, not a breaking term. Stretch–twist–fold: folding brings the loops into the same orientation so that fluxes reinforce; the fast-dynamo question stated by its definition, without a large-scale requirement. MHD breaks line split: the Hall term transfers freezing to the electron velocity, while electron inertia and pressure-tensor terms break that too and support collisionless reconnection. Fluxoid quantisation stated correctly, in units of h/2e, coinciding with magnetic flux where the current term vanishes. Relativistic row: h defined as enthalpy per particle, huμ as momentum per particle, reducing to the particle mass. Smaller: (∂t + 𝓛v) written out where a time-dependent form is meant; Cauchy’s formula described as a push-forward of the initial field; Mouhot and Villani qualified by their assumptions and Landau damping attributed to collisionless dynamics rather than transport alone; solar-flare sentence corrected; history sentence reconciled with Batchelor’s 1950 analogy; legend and references note tidied; the method note now carries the same exception as the references note. Published.
0.2 · 11 Sep 2026 Revised after external review; the geometric dictionary and the main conclusion were wrong in places and are corrected. Ertel: the absolute-vorticity 2-form is not transported in baroclinic flow; the transported object is Ωa∧dθ, and potential vorticity is a transported scalar, not a density, requiring the closure p = p(ρ, θ). Kelvin: the velocity 1-form is transported only up to an exact differential, so circulation is conserved on closed loops only; the Kelvin–Noether theorem stated as a balance with a source. The universal claim that transport ends its own validity withdrawn: uniform flows create no small scales, phase mixing and Landau damping are exact collisionless transport, dissipation does not require loss of regularity, and decay is not evidence of failed transport. Dynamo: ideal stretching amplifies without resistivity (uniform-strain example); ideal amplification, stretch–twist–fold and the fast-dynamo limit separated; Cowling’s theorem stated as a constraint on axisymmetric fields; collisionless reconnection added. Two-fluid: barotropic species pressures required; the relation between Hall MHD and ideal MHD corrected. Relativistic row split into the unmagnetised vorticity result and the electromagnetic flux result, with Bekenstein and Oron 2000 for the MHD circulation theorem. Transporting velocity: non-unique for vorticity too; the MHD bulk velocity selected by the model, not by convention; flux preservation distinguished from line-connection preservation. Turbulence row restated as hydrodynamic vorticity against kinematic induction, with the verdict attached to Batchelor’s inference. Superconductivity relabelled a counter-case, scoped to the equilibrium Meissner regime, with flux and fluxoid distinguished and the London equation described as constitutive. Closing note and verdict rewritten to say which invariants follow from which transported object. Smaller corrections: Cauchy’s formula as a pull-back; the Hubble deformation gradient a·I; “collisionless Boltzmann” for “Jeans’s equation”; zero-temperature superfluid reconnection; the Audit 11 teaser; “one origin” and “found separately” withdrawn; Cowling’s 1953 proof added from Blum and Choudhuri. Still draft.
0.1 · 11 Sep 2026 First draft on the series template. Sources verified before drafting; six points corrected during verification and carried into the text: Alfvén’s 1943 wording is “fastened”, not “frozen”; Cauchy’s formula is for incompressible flow and he did not draw Helmholtz’s theorems from it; Hoskins, McIntyre and Robertson derive Ertel’s theorem from the Bjerknes circulation theorem, not Kelvin’s; Liouville’s 1838 note does not concern phase space; Batchelor’s material-line result is in Section 5.3, not 5.2; Landau and Lifshitz’s Section 8 is titled “The conservation of circulation”. A planned citation of Anile’s monograph for flux freezing was dropped because its contents do not support it.
Erratum · 12 Sep 2026 The dates on the entries above were first printed as 13–14 September 2026 in error; they now give the day on which each revision was made and the page published, Sydney time.
Method note. This page was drafted with AI assistance (Anthropic’s Claude) working from my brief and revised with me. Bibliographic details were verified against publisher records for every reference before the text was written, except for the gaps explicitly identified in the references; the flag beside each says how far the content itself was consulted, and where a claim rests on a particular passage the section, equation or page is given in the text. Claims that a topic was unexplored were searched before being kept. The verdicts are mine and I expect some of them to be argued with; corrections are welcome and will be logged. Content on this site is licensed CC BY-SA 4.0.