Cyclic Coordinates and Conserved Momenta

A conservation law is usually presented as a gift of experiment — momentum, angular momentum, energy are conserved — and left there. The calculus of variations turns each gift into a theorem with a one-line proof. The engine is the Euler–Lagrange equation $\frac{\mathrm{d}}{\mathrm{d}t}(\partial L/\partial\dot q_k)=\partial L/\partial q_k$ , a balance: the rate of change of a momentum equals a generalized force. Kill the force — arrange for a coordinate to be absent from the Lagrangian $L(q,\dot q,t)$ — and the balance collapses to $\dot p_k=0$ . The momentum conjugate to that coordinate is a constant of the motion, fixed by the initial data. Such an absent coordinate is called cyclic, or equivalently ignorable or kinosthenic: only its rate of change enters the dynamics, so its actual value can be "ignored" until a final integration recovers it. Each cyclic coordinate hands us a first integral — an equation one order lower than the equations of motion — and enough of them render a problem solvable by quadratures alone. The azimuth of a planet gives Kepler's law of equal areas; the absence of time itself gives the energy. Underneath the bookkeeping lies the deepest organizing principle in physics. A cyclic coordinate is a continuous symmetry direction of the action — shifting $q_k\to q_k+\varepsilon$ costs nothing — and by Noether's theorem every such symmetry is paired with a conserved charge. Cyclic coordinates are Noether's theorem stripped to its bare arithmetic. This article develops that arithmetic inside the calculus of variations, shows how the Routhian systematically eliminates the ignorable coordinates, and follows the same logic into field theory, where an ignorable field yields a conserved current.

Key takeaways

A coordinate $q_k$ is cyclic (ignorable) if it does not appear explicitly in the Lagrangian $L$ — only its velocity $\dot q_k$ may — so $\partial L/\partial q_k=0$ .
For a cyclic coordinate the Euler–Lagrange equation collapses to $\dot p_k=0$ : the conjugate momentum $p_k=\partial L/\partial\dot q_k$ is a constant of the motion, a first integral fixed by the initial conditions.
Each conserved momentum lowers the order of the system by one; enough of them make a problem integrable by quadratures, the Routhian procedure carrying out the reduction cleanly.
This is Noether's theorem in miniature — invariance of the action under translation of $q_k$ is the symmetry whose Noether charge is $p_k$ .
Homogeneity of space gives linear momentum, isotropy gives angular momentum ( $p_\phi=mr^2\dot\phi$ ), and time-translation invariance ( $\partial L/\partial t=0$ ) gives the energy, the Beltrami first integral.
Key idea: if $L$ does not depend on $q_k$ explicitly, the Euler–Lagrange equation for $q_k$ says exactly that its conjugate momentum $p_k=\partial L/\partial\dot q_k$ has zero time derivative — a symmetry (translation in $q_k$ ) produces a conservation law (constant $p_k$ ).

Where this lives today

Cyclic coordinates are load-bearing across modern physics. In general relativity the Killing vectors of a metric are precisely the cyclic coordinates of the geodesic Lagrangian: on the Kerr spacetime of a rotating black hole, time-translation and axial symmetries make $t$ and $\phi$ ignorable, so each orbiting star or infalling photon carries a conserved energy $E$ and angular momentum $L$ — the constants that let the Event Horizon Telescope and LIGO/Virgo teams compute photon rings and inspiral waveforms without integrating the full geodesic equations. In plasma confinement, the axisymmetry of a tokamak makes the toroidal angle cyclic, conserving the canonical toroidal angular momentum $p_\phi=mr^2\dot\phi+qA_\phi r$ , which constrains particle drifts and underpins magnetic confinement in ITER. In computational physics, symplectic and variational integrators are built to preserve exactly the momenta attached to a system's cyclic coordinates, so a solar-system simulation conserves angular momentum to machine precision over billions of timesteps where an ordinary Runge–Kutta scheme would drift. The open frontier is the same idea one level up: approximate and hidden symmetries (adiabatic invariants, Carter-like constants) that yield near-conserved momenta where no cyclic coordinate is manifest.

The physical picture

Read the Euler–Lagrange equation as a force balance. The term $\partial L/\partial q_k$ is a generalized force, $p_k=\partial L/\partial\dot q_k$ is the momentum it drives, and the equation says the momentum changes at exactly the rate the force dictates. If the coordinate is cyclic, the force term is identically zero, so $p_k$ simply coasts at its initial value. The first figure sets the two situations side by side: a coordinate on a slope of $L$ feels a force and its momentum drifts; a cyclic coordinate lives on a flat direction and its momentum is frozen.

Figure

The Euler–Lagrange balance seen twice: when the coordinate appears in L there is a slope, a generalized force, and a drifting momentum; when the coordinate is cyclic the direction is flat, the force vanishes, and the conjugate momentum coasts at a fixed value

The physics behind "flatness" is symmetry. Saying $q_k$ is absent from $L$ is the same as saying the action is unchanged when you slide the whole trajectory rigidly in the $q_k$ direction: the system cannot tell where it sits along that axis, only how fast it moves, and the momentum pushing along the invariant direction never needs to change. This is Noether's theorem in its most transparent form — an ignorable coordinate is a continuous translation symmetry, and its conjugate momentum is the conserved charge. The second figure shows the archetype, the cyclic azimuth of an orbit, where the conserved quantity is angular momentum and the conservation law is something one can literally measure with a ruler — equal areas swept in equal times.

Figure

Kepler's law of equal areas as a conserved momentum: the azimuth φ is cyclic in any central potential so ℓ = m r² φ̇ is constant, forcing Earth to sweep equal areas in equal times — fast at perihelion (30.3 km/s) and slow at aphelion (29.3 km/s), with ℓ/m = r²φ̇ ≈ 4.46×10¹⁵ m²/s

The core equations

Everything starts from the Euler–Lagrange equation for a single coordinate — a second-order differential equation whose degeneration into something simpler is the entire subject of cyclic coordinates. Written with the generalized force on the right, its structure as a momentum balance is manifest.

\frac{\mathrm{d}}{\mathrm{d}t}\!\left(\frac{\partial L}{\partial \dot q_k}\right)=\frac{\partial L}{\partial q_k} \qquad (1)

where: $L$ — the Lagrangian $L(q,\dot q,t)$ ; $q_k$ — the $k$ -th generalized coordinate; $\dot q_k$ — its generalized velocity; $t$ — time, the independent variable. Read it as "rate of change of momentum = generalized force," and the role of a cyclic coordinate becomes obvious: it is the case where the right-hand side vanishes. The object on the left, before the time derivative, is the momentum canonically conjugate to $q_k$ . This is a definition, more general than "mass times velocity" — it can mix all the velocities and include field contributions such as $qA$ .

p_k\equiv \frac{\partial L}{\partial \dot q_k} \qquad (2)

where: $p_k$ — the momentum conjugate to $q_k$ ; in general a function of all the $\dot q$ and $q$ , reducing to $m\dot q_k$ only for the simplest kinetic energies. The pairing $(q_k,p_k)$ is the seed of the entire Hamiltonian formulation. Now impose the cyclic condition. If $q_k$ is absent from $L$ the force term is identically zero, and equation (1) says the conjugate momentum has zero time derivative — the central result of the article, proved in one line.

\frac{\partial L}{\partial q_k}=0\ \Longrightarrow\ \frac{\mathrm{d}p_k}{\mathrm{d}t}=0\ \Longrightarrow\ p_k=\text{const} \qquad (3)

where: $\text{const}$ — a first integral of the motion, fixed by the initial conditions. Each such constant is one equation, algebraic in the velocities, that we get for free without solving any differential equation. The archetype is the azimuthal angle in any central-force problem. Because the potential depends only on the radius, $\phi$ never appears in $L$ , and its conjugate momentum — the angular momentum — is conserved: Kepler's second law in disguise.

p_\phi=\frac{\partial L}{\partial \dot\phi}=m r^2\dot\phi=\ell=\text{const} \qquad (4)

where: $\phi$ — the azimuthal angle, cyclic in any central potential $V(r)$ ; $\ell$ — the conserved angular momentum; $m$ — the mass; $r$ — the radial distance. Note $\ell$ , not $\dot\phi$ , is the constant: the planet spins faster when nearer the Sun precisely so that $r^2\dot\phi$ stays fixed. Time itself behaves like a cyclic coordinate. When $L$ has no explicit time dependence, the energy function (Jacobi integral) is conserved — the Beltrami first integral, and the mechanism behind conservation of energy.

\frac{\mathrm{d}}{\mathrm{d}t}\!\left(\sum_k\dot q_k\frac{\partial L}{\partial\dot q_k}-L\right)=-\frac{\partial L}{\partial t},\qquad h\equiv\sum_k\dot q_k\,p_k-L \qquad (5)

where: $h$ — the energy function (Jacobi integral); for a natural Lagrangian $L=T-V$ with velocity-independent $V$ and time-independent constraints, $h=T+V$ , the total energy. If $\partial L/\partial t=0$ the right side vanishes and $h$ is conserved. The pattern generalizes verbatim to fields. If a field $\phi$ enters the Lagrangian density $\mathcal L$ only through its derivatives — the field-theoretic meaning of "cyclic" — the Euler–Lagrange equation becomes a continuity equation and a current is conserved.

\frac{\partial\mathcal L}{\partial\phi}=0\ \Longrightarrow\ \partial_\mu\!\left(\frac{\partial\mathcal L}{\partial(\partial_\mu\phi)}\right)=0\ \Longrightarrow\ \partial_\mu j^\mu=0 \qquad (6)

where: $\mathcal L$ — the Lagrangian density; $\phi$ — a field appearing only through $\partial_\mu\phi$ ; $j^\mu\equiv\partial\mathcal L/\partial(\partial_\mu\phi)$ — the conserved Noether current, whose time component integrates to a conserved charge. Global phase invariance gives electric charge; translation invariance gives the stress–energy tensor.

Derivation AFrom a cyclic coordinate to a conserved Noether charge

The goal here is to prove, from two directions, that an ignorable coordinate delivers a conserved momentum, and to make explicit that the conservation law is nothing but Noether's theorem for the simplest possible symmetry — a rigid translation of one coordinate. We first read the result off the Euler–Lagrange equation, then re-derive it from invariance of the action, so the same fact appears both as a triviality of differentiation and as a deep consequence of symmetry.

Step 1. Begin from $\frac{\mathrm{d}}{\mathrm{d}t}(\partial L/\partial\dot q_k)=\partial L/\partial q_k$ and impose $\partial L/\partial q_k=0$ . The right side vanishes, leaving $\mathrm{d}p_k/\mathrm{d}t=0$ , so $p_k=\partial L/\partial\dot q_k$ is constant along every solution. Why: this is the entire theorem in one line — the "force" that could change $p_k$ is exactly the term the cyclic condition removes.

Step 2. Now approach the same statement through the action. Consider variations $q_k\to q_k+\varepsilon$ with $\varepsilon$ constant, so $\delta q_k=\varepsilon$ , $\delta\dot q_k=0$ . The induced change is $\delta S=\varepsilon\int_{t_1}^{t_2}\frac{\partial L}{\partial q_k}\,\mathrm{d}t$ . Why: a rigid shift changes the coordinate but not its velocity, so only the $\partial L/\partial q_k$ term survives — isolating precisely the quantity the cyclic condition controls.

Step 3. Demand that this shift be a symmetry, $\delta S=0$ for arbitrary $\varepsilon$ . That requires $\partial L/\partial q_k=0$ — the cyclic condition is equivalent to translation invariance of the action in $q_k$ . Why: this identifies "coordinate absent from $L$ " with "action indifferent to sliding along $q_k$ ," the symmetry side of Noether's correspondence.

Step 4. Finally, rewrite the symmetry condition on-shell using $\partial L/\partial q_k=\mathrm{d}p_k/\mathrm{d}t$ : $0=\int_{t_1}^{t_2}\frac{\mathrm{d}}{\mathrm{d}t}\big(\frac{\partial L}{\partial\dot q_k}\big)\mathrm{d}t=p_k(t_2)-p_k(t_1)$ . Since $t_1,t_2$ are arbitrary, $p_k$ is constant. Why: Noether's theorem in miniature — the conserved charge of the translation symmetry is exactly $p_k$ , derived with nothing beyond the fundamental theorem of calculus. *Verified: consistent with Noether's theorem as stated in Stone & Goldbart §1.4 and Goldstein §13.7; both routes yield the same conserved $p_k=\partial L/\partial\dot q_k$ .*

Go deeper · Derivation B · 1 minAngular momentum and the reduction of the central-force problem

Derivation BAngular momentum and the reduction of the central-force problem

The aim of this derivation is to show how a single cyclic coordinate collapses a two-dimensional dynamical problem to an effective one-dimensional one, using the planar central-force problem — the Kepler and atomic-orbit archetype — as the vehicle. We identify the cyclic azimuth, extract its conserved momentum, use it to eliminate $\dot\phi$ , and arrive at the effective potential that governs all central motion, verifying Kepler's second law on the way.

Step 1. Write the Lagrangian for mass $m$ in a central potential $V(r)$ : $L=\tfrac12 m(\dot r^2+r^2\dot\phi^2)-V(r)$ . The angle $\phi$ appears only through $\dot\phi$ ; since $\partial L/\partial\phi=0$ , it is cyclic. Why: isotropy of the potential — depending on $r$ but not direction — is exactly what makes $\phi$ ignorable, so rotational symmetry is built into the coordinates.

Step 2. Form the conjugate momentum and invoke the cyclic rule: $p_\phi=\partial L/\partial\dot\phi=mr^2\dot\phi\equiv\ell=\text{const}$ , so $\dot\phi=\ell/(mr^2)$ . Why: this conserved angular momentum expresses $\dot\phi$ in terms of $r$ and the constant $\ell$ , removing $\phi$ entirely. The statement $\tfrac12 r^2\dot\phi=\ell/2m=\text{const}$ is Kepler's law of equal areas.

Step 3. Write the radial equation $m\ddot r-mr\dot\phi^2=-\,\mathrm{d}V/\mathrm{d}r$ and substitute $\dot\phi=\ell/(mr^2)$ : $m\ddot r=-\frac{\mathrm{d}V}{\mathrm{d}r}+\frac{\ell^2}{mr^3}$ . Why: the cyclic momentum reappears as a repulsive centrifugal term $\ell^2/mr^3$ , converting the coupled $(r,\phi)$ system into a single equation for $r(t)$ — the reduction the conserved momentum buys us.

Step 4. Recognize the right side as $-\mathrm{d}V_{\rm eff}/\mathrm{d}r$ with the effective potential $V_{\rm eff}(r)=V(r)+\ell^2/2mr^2$ , and, since $\partial L/\partial t=0$ , add the energy first integral $E=\tfrac12 m\dot r^2+\ell^2/2mr^2+V(r)$ . Why: two first integrals ( $\ell$ and $E$ ) make the two-degree-of-freedom problem completely integrable — $r(t)$ follows from a single quadrature and $\phi$ from a second, with no differential equation left to solve. *Verified: reproduces Goldstein §3.1–3.2 and Landau & Lifshitz §14; the areal-velocity law $\tfrac12 r^2\dot\phi=\ell/2m$ matches Kepler's second law, and the effective potential is the standard central-force result.*

Go deeper · Derivation C · 1 minRouthian reduction — systematically ignoring the ignorable

Derivation CRouthian reduction — systematically ignoring the ignorable

This derivation records the mechanical procedure for eliminating cyclic coordinates: Routh's partial Legendre transform, which trades each ignorable velocity for its conserved momentum while leaving the remaining coordinates in Lagrangian form. The payoff is a reduced variational problem in the non-cyclic coordinates alone, with the conserved momenta as constant parameters, plus a final quadrature that recovers the ignored coordinates.

Step 1. Split the coordinates into non-cyclic $q$ and cyclic $\phi$ (with $\partial L/\partial\phi=0$ ), and identify $p_\phi=\partial L/\partial\dot\phi=\text{const}$ . Solve this algebraic relation for $\dot\phi$ as a function of $p_\phi$ and the non-cyclic variables. Why: because $p_\phi$ is constant, this is not a differential equation — it expresses the ignorable velocity in terms of a number and the surviving variables.

Step 2. Define the Routhian by a Legendre transform in the cyclic velocities only, $R(q,\dot q,p_\phi)=p_\phi\dot\phi-L$ , and substitute $\dot\phi$ from Step 1 so $R$ contains neither $\phi$ nor $\dot\phi$ . Why: this is a hybrid — Hamiltonian in the cyclic pair $(\phi,p_\phi)$ but still Lagrangian in $(q,\dot q)$ — and transforming only the ignorable velocities makes the reduction exact.

Step 3. The surviving coordinates obey ordinary Euler–Lagrange equations with $-R$ as the Lagrangian, $\frac{\mathrm{d}}{\mathrm{d}t}\frac{\partial(-R)}{\partial\dot q}-\frac{\partial(-R)}{\partial q}=0$ , in which $p_\phi$ enters only as a fixed constant. Why: the reduced problem has one fewer degree of freedom, and the conserved momentum shows up as a centrifugal-type term — exactly the $\ell^2/2mr^2$ of the central-force derivation.

Step 4. Once $q(t)$ is known, recover the ignored coordinate by the single quadrature $\phi(t)=\phi_0+\int(\partial R/\partial p_\phi)\,\mathrm{d}t$ , using $\dot\phi=\partial R/\partial p_\phi$ . Why: the cyclic coordinate never coupled back, so it is reconstructed by one integration after the hard work is done — the literal sense of "ignorable." *Verified: matches the Routhian procedure in Goldstein §8.3 and Whittaker, Analytical Dynamics §38; the reduced equations reproduce the central-force effective potential when applied to $L=\tfrac12 m(\dot r^2+r^2\dot\phi^2)-V(r)$ .*

Limiting cases

The single statement "cyclic coordinate ⇒ conserved momentum" specializes to every elementary conservation law of mechanics. Each row below is the same theorem applied to a different symmetry.

Case	Result	Meaning
Free particle, cartesian	$L=\tfrac12 m\dot x^2$ , so $p_x=m\dot x=\text{const}$	$x$ is cyclic; homogeneity of space gives conserved linear momentum
Central force, planar polar	$L=\tfrac12 m(\dot r^2+r^2\dot\phi^2)-V(r)$ , so $p_\phi=mr^2\dot\phi=\text{const}$	$\phi$ is cyclic; isotropy gives angular momentum and Kepler's equal-area law
Time-independent $L$	$\partial L/\partial t=0\Rightarrow h=\sum_k\dot q_k p_k-L=\text{const}$	time acts as a cyclic variable; the Beltrami energy first integral is conserved
Ignorable field (field theory)	$\partial\mathcal L/\partial\phi=0\Rightarrow\partial_\mu j^\mu=0$	a field entering only through derivatives yields a conserved Noether current
Cyclic coordinate, general	$\partial L/\partial q_k=0\Rightarrow p_k=\partial L/\partial\dot q_k=\text{const}$	the universal statement of which all the above are special cases

Common trapCyclic means the coordinate is missing, not that its velocity is zero

The commonest error is to read "cyclic $q_k$ " as " $\dot q_k=0$ ." It means the opposite: $q_k$ is free to run, and usually does (a planet's azimuth increases forever). What is fixed is the conjugate momentum $p_k=\partial L/\partial\dot q_k$ , generally a nontrivial function of the velocities — for the azimuth it is $p_\phi=mr^2\dot\phi$ , not $\dot\phi$ itself, so $\dot\phi$ visibly changes around an eccentric orbit while $p_\phi$ does not. Two further cautions. First, $p_k$ is conserved only on-shell, along solutions, not as an identity. Second, whether a coordinate is cyclic depends on the coordinate choice: the same free particle has $p_x$ conserved in cartesian coordinates but $p_r=m\dot r$ not conserved in polar coordinates, because $r$ appears in the kinetic term. Symmetry is coordinate-independent; its manifestation as a cyclic coordinate is not.

Worked examples

1 · Kepler's equal-area law for Earth, checked with real numbers (fully worked)

Earth's orbit has perihelion $r_p=1.471\times10^{11}$ m, $v_p=30.29$ km/s, and aphelion $r_a=1.521\times10^{11}$ m, $v_a=29.29$ km/s. Verify that $p_\phi/m=r^2\dot\phi$ is the same at both ends, and reconstruct the orbit's area from it.

Step 1. At both apses the velocity is purely tangential, so the specific angular momentum is $\ell/m=r\,v$ . Evaluate: $(\ell/m)_p=r_p v_p=(1.471\times10^{11})(3.029\times10^4)=\mathbf{4.456\times10^{15}}$ m²/s and $(\ell/m)_a=r_a v_a=(1.521\times10^{11})(2.929\times10^4)=\mathbf{4.455\times10^{15}}$ m²/s. The two agree to better than 0.03%, confirming that the cyclic azimuth conserves $p_\phi=mr^2\dot\phi$ even though $r$ and $v$ each change by ~3% around the orbit.

Step 2. The areal velocity is half the specific angular momentum: $\mathrm{d}A/\mathrm{d}t=\tfrac12 r^2\dot\phi=\tfrac12(\ell/m)=\mathbf{2.23\times10^{15}}$ m²/s. This is the quantitative form of Kepler's second law — the radius vector sweeps area at a constant rate, faster-but-shorter at perihelion, slower-but-longer at aphelion, so the shaded sectors in the second figure are equal.

Step 3. Integrate over one orbital period $T=3.156\times10^{7}$ s (one year): the total swept area is $A=(\mathrm{d}A/\mathrm{d}t)\,T=\mathbf{7.04\times10^{22}}$ m². As an independent check, the geometric area of the ellipse is $\pi a b$ with $a=1.496\times10^{11}$ m and $b=a\sqrt{1-e^2}=1.4958\times10^{11}$ m ( $e=0.0167$ ), giving $\pi a b=\mathbf{7.03\times10^{22}}$ m². The two agree to 0.1%, closing the loop: a single conserved momentum, born of one cyclic coordinate, reproduces both Kepler's second law and the full orbital area.

2 · The conical pendulum from a cyclic azimuth (fully worked)

A bob on a massless rod of length $L=1.00$ m swings as a spherical pendulum, with polar angle $\theta$ from the downward vertical and azimuth $\phi$ . Show $\phi$ is cyclic, find the effective potential for $\theta$ , and compute the steady precession rate for conical motion at $\theta_0=30^\circ$ (take $g=9.81$ m/s²).

Step 1. The Lagrangian is $L=\tfrac12 mL^2(\dot\theta^2+\sin^2\theta\,\dot\phi^2)+mgL\cos\theta$ . The azimuth $\phi$ appears only through $\dot\phi$ , so it is cyclic: $p_\phi=mL^2\sin^2\theta\,\dot\phi=\text{const}$ . The vertical component of angular momentum is conserved — the meaning of the ignorable azimuth.

Step 2. Use the conserved $p_\phi$ to eliminate $\dot\phi$ . The $\theta$ motion is governed by the effective potential $V_{\rm eff}(\theta)=\dfrac{p_\phi^2}{2mL^2\sin^2\theta}-mgL\cos\theta$ — a centrifugal barrier supplied by the conserved momentum plus gravity. This is the Routhian reduction of the third derivation, applied to the pendulum.

Step 3. Steady conical motion sits at the minimum of $V_{\rm eff}$ , where $\mathrm{d}V_{\rm eff}/\mathrm{d}\theta=0$ . Differentiating gives $mgL\sin\theta_0=\dfrac{p_\phi^2\cos\theta_0}{mL^2\sin^3\theta_0}$ ; substituting $p_\phi=mL^2\sin^2\theta_0\,\dot\phi$ yields the clean result $\dot\phi^2=\dfrac{g}{L\cos\theta_0}$ , which diverges as $\theta_0\to90^\circ$ .

Step 4. With $\theta_0=30^\circ$ , $\cos30^\circ=0.8660$ : $\dot\phi=\sqrt{9.81/(1.00\times0.8660)}=\sqrt{11.33}=\mathbf{3.37}$ rad/s, a precession period $T_\phi=2\pi/\dot\phi=\mathbf{1.87}$ s. The conserved specific momentum is $p_\phi/m=L^2\sin^2\theta_0\,\dot\phi=(1.00)(0.25)(3.37)=\mathbf{0.842}$ m²/s. The cyclic azimuth turns a two-angle problem into a one-line condition for steady precession.

Physics of the portrait

Cyclic coordinates were a working tool of analytical mechanics for a century — Lagrange, Hamilton, Jacobi and Routh all exploited "ignorable" coordinates — before anyone saw the general principle beneath them. That principle is Emmy Noether's, and it is why every conserved momentum in this article is really a statement about symmetry.

Origins · 1 minEmmy Noether and how this was discovered

Emmy Noether

1882–1935 · symmetry and conservation laws

Amalie Emmy Noether was the greatest algebraist of her generation and, almost incidentally, the author of the theorem that unifies this entire subject. Invited to Göttingen in 1915 by Hilbert and Klein, she was set on a puzzle in Einstein's new general relativity: why energy conservation behaved so strangely there. Her answer, published in 1918 as Invariante Variationsprobleme, was far more than a fix. She proved that every continuous symmetry of an action functional corresponds to a conserved quantity, and conversely — translation gives momentum, rotation gives angular momentum, time-translation gives energy, phase symmetry gives charge. The cyclic coordinates that Routh and Jacobi had manipulated as a computational convenience were suddenly revealed as the visible tip of this correspondence: an ignorable coordinate is simply a translation symmetry written in coordinates, and its conjugate momentum is the Noether charge. Faculty resistance on the grounds of her sex kept her from lecturing under her own name until 1919; driven from Göttingen in 1933 for being Jewish, she taught at Bryn Mawr until her sudden death in 1935. Einstein called her the most significant creative mathematical genius produced since the higher education of women began, and her theorem is now the organizing principle of all theoretical physics, from classical mechanics to the Standard Model.

History

1788

— Lagrange's Mécanique analytique casts dynamics in generalized coordinates, exposing coordinates that never appear in the equations

1834

— Hamilton reformulates mechanics from a single variational principle, sharpening the role of conjugate momenta

1836

— Jacobi develops the energy integral and the method of ignoring coordinates in the reduction of dynamical problems

1877

— Routh introduces the Routhian to systematically eliminate ignorable coordinates in his treatise on the stability of motion

1918

— Noether proves that every continuous symmetry of the action yields a conserved quantity, unifying all cyclic-coordinate conservation laws

Going further

Texts. Goldstein, Poole & Safko, Classical Mechanics, 3rd ed., §2.6 (cyclic coordinates), §3.1–3.2 (central force), §8.3 (Routhian) · Landau & Lifshitz, Mechanics, §§7–15 · Stone & Goldbart, Mathematics for Physics, Ch. 1 (calculus of variations and Noether) · Arfken & Weber, Mathematical Methods for Physicists, calculus-of-variations chapter · Whittaker, A Treatise on the Analytical Dynamics of Particles and Rigid Bodies, §38 (ignoration of coordinates).

Provenance. Equations and derivations cross-checked against Stone & Goldbart (Ch. 1), Goldstein/Poole/Safko 3rd ed. (§2.6, §3.1–3.2, §8.3, §13.7), and Landau & Lifshitz Mechanics (§§7–15). A verifier should confirm: (i) the cyclic law $p_\phi=mr^2\dot\phi$ and its equivalence to Kepler's areal law; (ii) the Earth-orbit numbers $r_p v_p=4.456\times10^{15}$ and $r_a v_a=4.455\times10^{15}$ m²/s and area $7.0\times10^{22}$ m² $\approx\pi a b$ ; (iii) the conical-pendulum result $\dot\phi^2=g/(L\cos\theta_0)$ giving $\dot\phi=3.37$ rad/s at $\theta_0=30^\circ$ ; (iv) the Beltrami relation $\mathrm{d}h/\mathrm{d}t=-\partial L/\partial t$ and the Noether identification of $p_k$ . Orbital data are standard tabulated values for Earth.

Prerequisites. Euler–Lagrange Equation · First Variation and Stationarity · Functionals of Several Functions and Higher Derivatives. Leads to. Beltrami Identity and First Integrals · Hamilton's Principle of Stationary Action · Constrained Variation and Lagrange Multipliers · Geodesics as Variational Problems.

Ask the tutor: Ask me to take any Lagrangian you like, spot its cyclic coordinates, and reduce it to quadratures — or to show how the same move gives energy, momentum, and the Noether current from one principle.

What is the difference between a cyclic coordinate and a conserved momentum?

They are two sides of one fact, related by the Euler–Lagrange equation. A cyclic coordinate is a property of the Lagrangian: the coordinate $q_k$ does not appear in $L$ , so $\partial L/\partial q_k=0$ . A conserved momentum is a property of the motion: $p_k=\partial L/\partial\dot q_k$ does not change in time. The bridge is equation (1), $\dot p_k=\partial L/\partial q_k$ ; when the coordinate is cyclic the right side is zero, so the momentum is conserved. In short, "cyclic" describes the input (the Lagrangian) and "conserved" the output (the trajectory), and the equation of motion turns one into the other.

Why are the terms cyclic, ignorable, and kinosthenic all used?

They emphasize different aspects of the same idea. "Cyclic" comes from the archetypal angle: the azimuth $\phi$ is periodic, and rotating through a full cycle returns the system to an equivalent state, which is why the coordinate drops out of $L$ . "Ignorable," used heavily by Routh, stresses the practical payoff: once you have its conserved momentum you can ignore the coordinate, reduce the problem, and recover it by a final quadrature. "Kinosthenic," the rarest, is from Greek roots for motion and strength, marking that only the coordinate's motion — its velocity — carries dynamical content, never its position. All three name the identical condition $\partial L/\partial q_k=0$ .

If a coordinate is cyclic, does that mean nothing depends on it at all?

No — only the Lagrangian is indifferent to it; the actual motion is not. A planet's azimuth $\phi$ is cyclic, yet its position genuinely depends on $\phi(t)$ , which increases steadily around the orbit. What "cyclic" guarantees is that the dynamics cannot tell where you are along that coordinate, only how fast you move through it, so $\phi$ is fixed by initial conditions and a final integration rather than by a force law. This is why the Routhian procedure works: the ignorable coordinate is inert during the reduction and is reconstructed afterward by one quadrature, $\phi(t)=\phi_0+\int(\partial R/\partial p_\phi)\,\mathrm{d}t$ .

How is energy conservation a special case of this, when time is not a coordinate?

Time plays the role of a "cyclic parameter" rather than a cyclic coordinate, and the conserved quantity is the energy function $h=\sum_k\dot q_k p_k-L$ instead of a momentum. The logic is parallel: if $L$ has no explicit $t$ -dependence, translating the origin of time leaves the action unchanged, and Noether's theorem returns $h$ as the conserved charge. Equation (5) shows this directly: $\mathrm{d}h/\mathrm{d}t=-\partial L/\partial t$ , which vanishes precisely when $t$ is absent from $L$ . For a natural Lagrangian $L=T-V$ with time-independent constraints, $h=T+V$ , so time-translation symmetry is literally the origin of energy conservation — the content of the companion Beltrami identity.

Why is the conserved momentum for a charged particle the canonical, not the mechanical, one?

Because the Lagrangian for a particle in an electromagnetic field, $L=\tfrac12 m\dot{\mathbf r}^2+q\dot{\mathbf r}\cdot\mathbf A-q\varphi$ , contains the vector potential $\mathbf A$ , the conjugate momentum picks up a field term: $\mathbf p=\partial L/\partial\dot{\mathbf r}=m\dot{\mathbf r}+q\mathbf A$ . When a coordinate is cyclic — say the azimuth in an axisymmetric magnetic field — it is this canonical momentum $p_\phi=mr^2\dot\phi+qA_\phi r$ that is conserved, not the mechanical $mr^2\dot\phi$ alone. The mechanical part is not conserved because the field continuously exchanges momentum with the particle; only the combination including the field's contribution stays fixed. The same canonical momentum is what quantum mechanics promotes in $\mathbf p\to-i\hbar\nabla$ .

Does the field-theory version really work the same way?

Yes, essentially verbatim, with the conserved number replaced by a conserved current. The analog of "coordinate absent from $L$ " is a field $\phi$ entering the Lagrangian density only through its derivatives, so $\partial\mathcal L/\partial\phi=0$ . The field Euler–Lagrange equation then reads $\partial_\mu(\partial\mathcal L/\partial(\partial_\mu\phi))=0$ , a continuity equation $\partial_\mu j^\mu=0$ for the current $j^\mu=\partial\mathcal L/\partial(\partial_\mu\phi)$ , whose time component integrates to a conserved charge — the direct descendant of $p_k$ . Global phase invariance yields electric charge; spacetime-translation invariance yields the stress–energy tensor. The particle result of this article is simply the zero-dimensional case.