Degeneracy and the Role of the Magnetic Quantum Number

Hold an atom in your hand — any atom, any central force — and rotate it. Nothing about its energy can change: a spherical world has no preferred direction, so the orientation of the atom's angular momentum is a fact the Hamiltonian cannot bill for. But quantum mechanics says orientations come in a discrete set: an angular momentum $\ell$ offers exactly $2\ell + 1$ projections $m$ onto whatever axis you choose. Put the two statements together and something must give — and what gives is uniqueness of the energy level. Each level of a central potential is not one state but a family of $2\ell + 1$ states, all strictly, exactly degenerate. This is the cleanest theorem in the chapter, and one of the most consequential in atomic physics. It explains why $m$ is missing from every radial equation you will ever solve; why spectral lines split into odd-numbered multiplets when a field breaks the sphere's impartiality; and why $m$ earned its name — the magnetic quantum number, invisible until a field gives it a bill to pay. The theorem also draws a boundary the next articles cross. Rotational symmetry guarantees $2\ell + 1$ and only $2\ell + 1$ : a floor set by geometry. When an atom shows more — as hydrogen spectacularly does, its $n^2$ levels collapsing across different $\ell$ — something hidden is conserving itself behind the scenes. Learning to read degeneracy as a symmetry census is the skill this article installs.

Key takeaways

For any central potential $V(r)$ , the Hamiltonian commutes with all three components of $\hat{\mathbf{L}}$ , so every bound level is at least $(2\ell+1)$ -fold degenerate in the magnetic quantum number $m$ .
The mechanism is the ladder: $[\hat{H}, \hat{L}_\pm] = 0$ means $\hat{L}_\pm$ maps an energy eigenstate to another eigenstate at the same energy, stepping $m$ across the full multiplet $-\ell, \dots, +\ell$ .
The radial equation depends on $\ell$ (through the centrifugal term) but never on $m$ — the operator reason: $\hat{L}_z$ enters $\hat{H}$ only through $\hat{\mathbf{L}}^2$ .
The $z$ -axis is your choice, not nature's: re-quantizing along a tilted axis remixes the $m$ -populations of a state but cannot move any energy.
$(2\ell+1)$ is the minimum forced by symmetry. Extra degeneracy — hydrogen's $\ell$ -collapse into $n^2$ , the oscillator's $\tfrac{1}{2}(N+1)(N+2)$ — signals a hidden conserved quantity beyond $\hat{\mathbf{L}}$ .
A uniform magnetic field breaks the sphere down to rotations about one axis and lifts the degeneracy into $2\ell+1$ equally spaced lines, $\Delta E = \mu_B B\,m$ — the splitting that named $m$ .
Key idea: degeneracy is a symmetry census — rotations guarantee exactly $2\ell+1$ , and every state beyond that count is the fingerprint of an extra conservation law.

The core insight

The radial equation knows only $\ell$ , not $m$ ; rotational symmetry guarantees that the $2\ell+1$ orientations of a given angular momentum are exactly degenerate, while any further degeneracy signals an extra conserved quantity.

Where this lives today

The $m$ -degeneracy and its controlled breaking are the working medium of the most precise devices ever built. Optical atomic clocks interrogate transitions between states chosen because of this article's census: clock transitions run between $m = 0$ sublevels, whose linear Zeeman shift $\mu_B B\,m$ vanishes identically — the degeneracy theorem selecting, out of the whole multiplet, the two states most protected from stray fields. The remaining sublevels are not wasted: their equally spaced splittings serve as built-in magnetometers, measuring the residual field so it can be servoed away. The same ladder runs quantum technology at the warm end. Optically pumped magnetometers — the field sensors in fetal magnetocardiography — track the Larmor precession of atomic multiplets, reading fields as frequencies at the $\mu_B/h = 14\ \mathrm{GHz/T}$ exchange rate. And in every atomic-physics lab, state preparation by optical pumping is nothing but herding population along the $m$ -ladder of a degenerate multiplet — motion that costs no energy, exactly as the theorem promises, until a field is switched on to hold the result in place.

The physical picture

Figure

One length, many tilts — the orientations a central potential cannot bill for

Picture the angular momentum as an arrow of fixed length $\hbar\sqrt{\ell(\ell+1)}$ — the cone picture of Angular Momentum Commutation Relations. The quantum number $m$ records only the arrow's projection onto an axis you selected. A central potential is a perfectly round world: it cannot know which way the arrow tilts relative to a line you drew. So all $2\ell+1$ tilts cost the same energy. Degeneracy here is the operational signature of unused freedom: you could have drawn the axis differently, and physics had to be indifferent in advance. That indifference has a second face. If the sphere's impartiality causes the degeneracy, breaking it should lift it — and it does. Impose a uniform magnetic field: space now has one special direction, and the $2\ell+1$ states fan out into equally spaced energy rungs. Count the rungs in a spectrometer and you have measured $2\ell+1$ — an integer read off a photographic plate. This is how the magnetic quantum number earned its name decades before anyone could write $\hat{L}_z$ .

Figure

The degeneracy census — what symmetry guarantees vs what dynamics adds

The ledger picture is the article's map. In a generic central potential (left), each level carries exactly its $2\ell+1$ — no more. Hydrogen (middle) breaks the pattern: states of different $\ell$ , belonging to entirely different rotation multiplets, land on one energy, stacking $n^2$ states per level. Rotations cannot do that; the responsible party — the conserved Runge–Lenz vector, the SO(4) symmetry — is the business of Accidental Degeneracy and the SO(4) Symmetry. On the right, a field breaks the sphere and the hidden $2\ell+1$ becomes countable lines. One theorem, three exhibits.

The math: symmetry writes the degeneracy

The starting point is owned by Separation of Variables in Spherical Coordinates and imported whole: a central Hamiltonian commutes with every generator of rotations,

\hat{H} = -\frac{\hbar^2}{2m}\nabla^2 + V(r), \qquad [\hat{H}, \hat{L}_i] = 0, \quad [\hat{H}, \hat{\mathbf{L}}^2] = 0 \qquad (1)

where: $\hat{L}_i$ — the three components of orbital angular momentum; $V(r)$ — any function of the radial distance alone. This is conservation of angular momentum in operator dress, and it is the entire input: every result below follows from equation (1) plus the angular-momentum algebra, with no reference to the shape of $V(r)$ . That is why the theorem holds for hydrogen and for a neutron in a nuclear well alike. Because $\hat{H}$ , $\hat{\mathbf{L}}^2$ , $\hat{L}_z$ commute pairwise, they are labels on one simultaneous eigenbasis — the CSCO grammar of Simultaneous Diagonalization and Complete Sets of Commuting Observables — and the eigenfunctions factorize:

\psi_{n_r \ell m}(\mathbf{r}) = R_{n_r \ell}(r)\,Y_\ell^m(\theta, \phi) \qquad (2)

where: $n_r$ — the radial node count (the bookkeeping label born in Boundary Conditions and Normalization of Radial Wavefunctions); $Y_\ell^m$ — the spherical harmonics, owned by Spherical Harmonics. Now look hard at the subscript on $R$ : it carries $n_r$ and $\ell$ , but no $m$ . That omission is the theorem in embryo, and it deserves better than a shrug — Derivation B gives the operator reason. But first, the symmetry proof that never touches a differential equation.

Derivation AThe ladder proof — 2ℓ+1 partners at every level

The goal is the essential-degeneracy theorem: for any central potential, every energy eigenstate with angular momentum $\ell$ belongs to a family of $2\ell+1$ eigenstates of exactly the same energy. The strategy is pure algebra — no radial equation, no boundary conditions, nothing about $V(r)$ beyond centrality. We build the degenerate partners explicitly, by applying a symmetry operator that costs no energy, and count them with machinery already proven in Ladder Operators and the Angular Momentum Spectrum. Watch what does the work: it is only ever the commutator $[\hat{H}, \hat{L}_\pm] = 0$ .

Step 1. Assemble the ladder operators. From equation (1), $\hat{H}$ commutes with $\hat{L}_x$ and $\hat{L}_y$ separately, hence with the complex combinations

\hat{L}_\pm = \hat{L}_x \pm i\hat{L}_y, \qquad [\hat{H}, \hat{L}_\pm] = 0.

Why: commutators are linear — any polynomial in conserved generators is itself conserved. The ladder operators are symmetry operations, not just bookkeeping tricks.

Step 2. Apply a ladder to an eigenstate. Let $\hat{H}\psi_{n_r \ell m} = E\,\psi_{n_r \ell m}$ . Then

\hat{H}\left(\hat{L}_\pm \psi_{n_r \ell m}\right) = \hat{L}_\pm \hat{H}\,\psi_{n_r \ell m} = E\left(\hat{L}_\pm \psi_{n_r \ell m}\right):

the new state is again an eigenstate of $\hat{H}$ , with the same eigenvalue $E$ . No energy was spent tilting the arrow. Why: this two-line move is the entire engine — a conserved operator maps eigenspaces of H into themselves. Everything after this step is counting.

Step 3. Identify the new state. Because $[\hat{L}_\pm, \hat{\mathbf{L}}^2] = 0$ , the ladder cannot change $\ell$ ; it steps only $m$ , with the coefficient proven in Ladder Operators and the Angular Momentum Spectrum:

\hat{L}_\pm\,\psi_{n_r \ell m} = \hbar\sqrt{\ell(\ell+1) - m(m \pm 1)}\;\psi_{n_r \ell,\, m\pm1},

nonzero except at the top and bottom rungs $m = \pm\ell$ , where the multiplet terminates by the spectrum theorem. Why: the same-radial-function claim rides inside this equation — the ladder acts only on angles, so the partner state has the identical R and the identical node count; the family shares one radial profile viewed from 2ℓ+1 tilts.

Step 4. Count. Start at $m = \ell$ and descend: $\hat{L}_-$ generates $m = \ell, \ell-1, \dots, -\ell$ — exactly $2\ell + 1$ mutually orthogonal states, all at energy $E$ :

g_\ell = 2\ell + 1.

The statement is degenerate-safe by construction: it does not assume the level was non-degenerate to begin with — within any eigenspace of $\hat{H}$ , the ladder organizes states into complete $(2\ell+1)$ -blocks (the projector phrasing of Projection Operators covers the fussy case). Why: the count is the dimension of the SO(3) irreducible representation labelled by ℓ — the theorem is representation theory wearing lab clothes, which is exactly why the shape of V(r) never entered. *Verified: on the ℓ = 1 matrix representation, the ladder coefficients match $\hbar\sqrt{\ell(\ell+1)-m(m\pm1)}$ exactly, $[\hat{H}, \hat{L}_\pm] = 0$ holds on the multiplet block, and the three-state family shares one eigenvalue to machine precision — this article's verify script; the algebra is Sakurai §3.6's argument.*

Go deeper · Derivation B · 2 minWhy m never appears in the radial equation — and why the axis is yours

Derivation BWhy m never appears in the radial equation — and why the axis is yours

The goal is the other half of the theorem, the analytic face: show that the radial problem cannot depend on $m$ even in principle, and that the choice of quantization axis is a coordinate convention with zero physical content. The strategy has two movements — first an operator observation sharper than "it separates that way," then an explicit rotation exhibit you can run numerically: tilt the axis, watch the populations remix, watch the energy refuse to move.

Step 1. Locate where angular momentum enters the Hamiltonian. The kinetic operator in spherical coordinates — the split proven in Separation of Variables in Spherical Coordinates — is

\hat{H} = -\frac{\hbar^2}{2m}\frac{1}{r^2}\frac{\partial}{\partial r}\!\left(r^2 \frac{\partial}{\partial r}\right) + \frac{\hat{\mathbf{L}}^2}{2mr^2} + V(r).

$\hat{L}_z$ appears nowhere — only $\hat{\mathbf{L}}^2$ does. Acting on the factorized state (2), $\hat{\mathbf{L}}^2$ deposits its eigenvalue $\hbar^2\ell(\ell+1)$ and the angular factor cancels, leaving the radial problem

-\frac{\hbar^2}{2m}\frac{1}{r^2}\frac{\mathrm{d}}{\mathrm{d}r}\!\left(r^2\frac{\mathrm{d}R}{\mathrm{d}r}\right) + \left[V(r) + \frac{\hbar^2\ell(\ell+1)}{2mr^2}\right]R = E\,R,

in which $m$ has no seat at the table. The eigenvalue can only be $E_{n_r \ell}$ . Why: the honest operator reason is grammatical — m is the eigenvalue of L̂_z, and L̂_z simply does not occur in Ĥ; ℓ occurs because L̂² does, through the centrifugal term. "The radial equation doesn't see m" is not an accident of separation; it is the sphere's indifference written in operator form.

Step 2. Expose the convention. The label $m$ is defined relative to a $z$ -axis that you chose. Choosing differently is applying a rotation $\hat{R} = e^{-i\theta\,\hat{\mathbf{n}}\cdot\hat{\mathbf{L}}/\hbar}$ , and by equation (1), $\hat{R}\hat{H}\hat{R}^\dagger = \hat{H}$ : the Hamiltonian is blind to the change. A state with sharp $m$ along $z$ becomes, along the tilted axis, a superposition over the new $m'$ values — the remixing implemented by the rotation matrices of The Wigner D-Matrices — but every one of those components sits at the same $E$ . Why: degeneracy is exactly what protects the convention — if different m cost different energies, rotating your coordinate stationery would change physical predictions, which is absurd in a round world.

Step 3. Run the exhibit. On the $\ell = 1$ multiplet (basis $m = +1, 0, -1$ ), take the model Hamiltonian $\hat{H} = E_0\,\mathbb{1}$ on the block (any central potential's restriction to one level) and rotate by $\theta = \pi/3$ about $y$ using the matrix exponential of $\hat{L}_y$ . A state with populations $(1/3,\ 4/9,\ 2/9)$ across $m$ remixes to $(0.266,\ 0.524,\ 0.210)$ — visibly different occupations — while $\langle \hat{H} \rangle$ stays $E_0$ to machine precision and $\hat{R}\hat{H}\hat{R}^\dagger = \hat{H}$ exactly. Why: this is the theorem made tactile — the m-bookkeeping scrambles under your choice of axis, and nothing measurable about the energy notices. What is real is the multiplet; the m-labels are its shadow on your stationery. *Verified: rotation unitarity, population remixing at θ = π/3 (0.266/0.524/0.210 from 1/3, 4/9, 2/9), $\langle H \rangle$ invariance, and $RHR^\dagger = H$ are machine-checked in this article's verify script via expm on the ℓ = 1 representation.*

The census doctrine: minimum, not maximum

Here is the theorem's sharpest edge. Rotational symmetry guarantees $2\ell+1$ — the non-abelian-symmetry-compels-degeneracy theorem of Symmetries and Conservation Laws, consumed as proven machinery. But it guarantees only that. States of different $\ell$ live in different irreducible multiplets; no rotation connects 2s to 2p, so rotations have no opinion about their relative energies. When they nonetheless coincide — as in hydrogen, where every $\ell$ from $0$ to $n-1$ lands on $E_n$ and the level holds

g_n = \sum_{\ell=0}^{n-1}(2\ell+1) = n^2 \qquad (3)

states — the excess is called accidental degeneracy, and the name is a confession of ignorance, not an explanation. Whenever the census exceeds the symmetry's guarantee, hunt for a hidden conserved quantity: for the Coulomb potential it is the quantum Runge–Lenz vector, and the story is owned by Accidental Degeneracy and the SO(4) Symmetry, with operator credentials proven in The Laplace–Runge–Lenz Vector in Quantum Mechanics. The 3D isotropic oscillator plays the same trick with a different hidden symmetry — $g_N = \tfrac{1}{2}(N+1)(N+2)$ , owned by The Three-Dimensional Isotropic Harmonic Oscillator and its SU(3) sequel. This article's contribution is the baseline: anything beyond $2\ell+1$ is their problem to explain.

Breaking the sphere: the field takes the census

If degeneracy comes from indifference, splitting comes from bias. Add a uniform field $\mathbf{B} = B\hat{z}$ : the perturbation couples to the orbital magnetic moment, $\hat{H}' = (\mu_B B/\hbar)\,\hat{L}_z$ , which commutes with $\hat{L}_z$ but not with $\hat{L}_x, \hat{L}_y$ — the symmetry collapses from all rotations to rotations about $\mathbf{B}$ . Each $m$ -state, previously anonymous, now pays its own bill:

\Delta E = \mu_B B\,m, \qquad \mu_B = \frac{e\hbar}{2m_e} = 57.88\ \mu\mathrm{eV/T} \qquad (4)

where: $\mu_B$ — the Bohr magneton, the natural unit of orbital magnetism (the anchor of Magnetic Moment and the Gyromagnetic Ratio, cited). The splitting is linear in $m$ : the $2\ell+1$ sublevels fan out in exactly equal steps $\mu_B B$ — a ladder whose rung count is the degeneracy and whose spacing measures the field. Equal spacing is the theorem's fingerprint: the perturbation is proportional to the label the symmetry had been hiding. The preview stops here deliberately — full Zeeman machinery (selection rules, line patterns, the spin interplay that makes real spectra "anomalous") belongs to the atomic-physics lanes, and the deeper $m$ -machinery of matrix elements to The Wigner-Eckart Theorem. What this article owns is the census logic: an odd count of equally spaced lines, $2\ell+1$ of them, is rotational symmetry announcing its multiplet size.

Limiting cases

Case	Result	Meaning
Generic central potential	$g_\ell = 2\ell+1$ per level	the symmetry floor: exactly what rotations guarantee, nothing more
$\ell = 0$ levels	$g = 1$ , no degeneracy	a scalar multiplet — nothing to tilt, nothing to be indifferent about
Hydrogen $1/r$	$g_n = n^2$	different $\ell$ collapse: hidden SO(4) — owned by Accidental Degeneracy and the SO(4) Symmetry
3D isotropic oscillator	$g_N = \tfrac{1}{2}(N+1)(N+2)$	the other famous accident — owned by the oscillator articles
Uniform field $B\hat{z}$	$\Delta E = \mu_B B\,m$ : $2\ell+1$ equal steps	sphere broken to a cylinder; the census becomes countable lines
Slightly non-central $V$	$m$ -degeneracy survives if axial symmetry survives; otherwise splits	each surviving symmetry protects its own slice of the multiplet

Worked examples

1 · A p-level census (confidence builder)

An electron in a central potential occupies the $\ell = 1$ level with state $\psi = R(r)\left(\sqrt{3}\,Y_1^1 + 2i\,Y_1^0 - \sqrt{2}\,Y_1^{-1}\right)/3$ . What are the probabilities of each $m$ outcome, what is $\langle \hat{L}_z \rangle$ , and what does an energy measurement give?

Step 1. Check normalization. The amplitudes are $c_{+1} = \sqrt{3}/3$ , $c_0 = 2i/3$ , $c_{-1} = -\sqrt{2}/3$ ; the moduli squared sum to $(3 + 4 + 2)/9 = 1$ . The state is properly normalized (the spherical harmonics are orthonormal, so the radial factor carries its own $\int|R|^2 r^2\,\mathrm{d}r = 1$ ).

Step 2. Read off the census: $P(m{=}{+}1) = \mathbf{1/3}$ , $P(m{=}0) = \mathbf{4/9}$ , $P(m{=}{-}1) = \mathbf{2/9}$ . The complex phase on $c_0$ and the sign on $c_{-1}$ are invisible to the probabilities.

Step 3. Average the projection: $\langle \hat{L}_z \rangle = \hbar\left[(+1)\tfrac{1}{3} + 0\cdot\tfrac{4}{9} + (-1)\tfrac{2}{9}\right] = \mathbf{\hbar/9}$ — a slight net tilt along $+z$ .

Step 4. Measure the energy: the state is a superposition across $m$ only — all three components share the radial factor $R$ and hence the eigenvalue $E_{n_r,1}$ . The energy measurement returns $E_{n_r,1}$ with certainty, variance zero. That certainty is the degeneracy theorem: superposing within a multiplet costs nothing in energy sharpness. (Verified: census, $\langle \hat{L}_z\rangle = \hbar/9$ , and $\mathrm{var}(E) = 0$ machine-checked.)

2 · Taking the census with a magnet (exam level)

A d-level ( $\ell = 2$ ) of an atom (orbital moment only) is placed in $B = 0.5\ \mathrm{T}$ . How many sublevels appear, with what spacing, and what frequency resolution does a spectrometer need to resolve adjacent components?

Step 1. Count: $2\ell + 1 = \mathbf{5}$ sublevels, $m = -2, -1, 0, +1, +2$ — an odd count, the multiplet's signature.

Step 2. Space them: $\Delta E = \mu_B B = 57.88\ \mu\mathrm{eV/T} \times 0.5\ \mathrm{T} = \mathbf{28.94\ \mu\mathrm{eV}}$ between adjacent rungs — all four gaps identical, because the shift is linear in $m$ .

Step 3. Convert to frequency: $\Delta E/h = 28.94\ \mu\mathrm{eV} / (4.1357 \times 10^{-9}\ \mathrm{eV\,ns}) = \mathbf{7.00\ \mathrm{GHz}}$ . A spectrometer resolving better than $7\ \mathrm{GHz}$ near the line (about $0.01\ \mathrm{nm}$ at optical wavelengths) sees the quintet.

Step 4. Read the ledger backwards: five equally spaced lines announce $\ell = 2$ without any model of the atom's interior — the degeneracy census performed with a magnet and a grating. (The observed optical pattern also involves selection rules and spin — the atomic-lane story; the level splitting itself is exactly this arithmetic.)

Origins · 1 minJoseph John Larmor and how this was discovered

Joseph John Larmor

1857–1942 · Cambridge

A quarter-century before quantum mechanics could define $m$ , Larmor proved the theorem that would make it visible. His 1897 precession theorem: place any classical system of orbiting charges in a uniform magnetic field, and — to first order — the motion is the same motion viewed from a frame rotating at $\omega_L = eB/2m_e$ . A field does not distort orbits; it rotates them, all at one universal rate. When Zeeman found lines splitting in 1896, and Lorentz computed the classical triplet, Larmor's theorem explained the regularity: the field couples only to the rotation it superposes, which is why the splitting comes in equal steps — one $\omega_L$ for every component. In quantum language, "the field superposes a uniform rotation" becomes "the perturbation is proportional to $\hat{L}_z$ ," and equal precession becomes the equally spaced ladder $\Delta E = \mu_B B\,m = \hbar\omega_L m$ of this article. The $2\ell+1$ lines a spectroscopist counts are Larmor's one rotation, quantized into its allowed projections. His radiation-power formula — the other 1897 result — is a different story, told in the electromagnetism track; here he is the man who knew, before quantum mechanics existed, that a magnetic field's first move is always a rotation.

History

1896

— Zeeman observes spectral lines broadening and splitting in a magnetic field; Lorentz's classical theory predicts a triplet.

1897

— Larmor proves the precession theorem: a uniform field superposes a single universal rotation ω_L = eB/2mₑ on any orbiting-charge system.

1916

— Sommerfeld introduces the magnetic quantum number m and space quantization: the projection of angular momentum is discrete.

1922

— The Stern-Gerlach experiment photographs space quantization directly — discrete beam components where classical physics predicted a smear.

1926

— Wave mechanics derives the (2ℓ+1)-fold degeneracy from rotational symmetry; m becomes the label of the spherical harmonics.

Going further

Texts. Griffiths & Schroeter, Introduction to Quantum Mechanics, 3rd ed., Ch. 4 · Sakurai & Napolitano, Modern Quantum Mechanics, 2nd ed., §3.6 · Shankar, Principles of Quantum Mechanics, 2nd ed., §12.5–13 · Cohen-Tannoudji, Diu & Laloë, Quantum Mechanics, Vol. I, Ch. VII · Townsend, A Modern Approach to Quantum Mechanics, 2nd ed., Ch. 10.

Provenance. Equation forms and the $[\hat H,\hat L_i]=0$ argument follow Sakurai §3.6, Shankar §12.5–13, and Cohen-Tannoudji Ch. VII. A verifier should recompute: (1) $\sum_{\ell=0}^{n-1}(2\ell+1)=n^2$ ; (2) the Rydberg energy $13.6\,\mathrm{eV}$ and $\mu_B\approx9.274\times10^{-24}\,\mathrm{J/T}$ (CODATA); (3) the ladder-operator coefficient $\hbar\sqrt{\ell(\ell+1)-m(m\pm1)}$ ; (4) the oscillator degeneracy $\tfrac12(N+1)(N+2)$ . The $SO(4)$ assignment of the $n^2$ multiplet is standard (Cohen-Tannoudji Complement on the hydrogen atom; Shankar §13). Edition note (2026-07-08). Per the chapter's ownership map, the base article's second derivation — hydrogen's $n^2$ count and its identification as accidental — is split between its owners: the counting arithmetic $\sum(2\ell+1) = n^2$ is performed in The Hydrogen Atom: Coulomb Radial Equation (whose solve produces the $\ell$ -collapse it counts), and the explanation via the conserved Runge–Lenz vector and SO(4) is owned by Accidental Degeneracy and the SO(4) Symmetry; this article states both with owners named, keeping the symmetry census (minimum $2\ell+1$ ) as its own jurisdiction. The symmetry-implies-degeneracy generality is consumed from Symmetries and Conservation Laws; the $\mu_B$ anchor from Magnetic Moment and the Gyromagnetic Ratio; Zeeman pattern theory is deferred to the atomic-physics serials.

Prerequisites. Separation of variables in spherical coordinates · the radial equation · effective potential and the centrifugal barrier. Leads to. Accidental degeneracy and the SO(4) symmetry · the Laplace–Runge–Lenz vector in quantum mechanics · hydrogenic wavefunctions and quantum numbers.

Is the (2ℓ+1) degeneracy exact, or approximate like most degeneracies in real atoms?

Exact, so long as the potential is exactly central and no external field is present. It is protected by symmetry, not fine-tuning: any spherically symmetric perturbation shifts whole multiplets rigidly without splitting them. What splits the multiplet is only anisotropy — external fields, crystal environments, or internal couplings (spin-orbit) that single out directions. That robustness is why observed splittings are such clean probes of the symmetry-breaking agent.

If m is just a choice of axis, is it even physical?

The label is conventional; the structure is physical. What is real: the multiplet size 2ℓ+1, the equal energies, and — once an axis is singled out by an actual field or measurement — the discrete set of projections along it. Choosing a different axis relabels states by remixing amplitudes (Derivation B's exhibit) but predicts identical physics. The situation is exactly like vector components: "the x-component of momentum" depends on your axes, yet momentum is physical.

Does the theorem require bound states, or does it hold for scattering states too?

The ladder argument never used normalizability — only [Ĥ, L̂±] = 0 — so it applies to continuum solutions as well: scattering states of the same energy and ℓ come in the same (2ℓ+1) multiplets, which is why partial-wave analysis can treat each ℓ-channel with a single radial function. The bound-state case is where the census is most visible, because there the levels are sharp lines you can split and count.

Why does degeneracy attach to ℓ but the SPLITTING attach to m?

Because the two quantum numbers answer different questions. The energy of a central problem is set by the radial physics, which feels ℓ through the centrifugal barrier — different ℓ means a genuinely different one-dimensional problem. The label m records only orientation, which costs nothing while the world is round. A field makes the world unround along one axis, and the cheapest anisotropic coupling available is proportional to L̂z — so the previously free label m becomes the billing unit. In short: ℓ prices the shape, m prices the tilt, and tilt is free until something breaks the sphere.

Hydrogen's n² degeneracy — is it really "accidental" or is that name outdated?

The name is traditional and, taken literally, wrong: nothing in physics is an accident. It survives as jargon for "degeneracy exceeding what the obvious geometric symmetry guarantees." For the pure Coulomb potential the excess is fully explained by a hidden conserved vector (the Runge–Lenz vector) enlarging the symmetry from SO(3) to SO(4) — the subject of Accidental Degeneracy and the SO(4) Symmetry. The practical usage: "essential" degeneracy is protected by rotations and robust; "accidental" degeneracy is protected by a more delicate symmetry and shatters the moment the potential deviates from its special form.

Do the 2ℓ+1 degenerate states have the same wavefunction shape?

They share the identical radial profile R(r) — the ladder operators act only on angles, so node count and radial extent are common to the whole multiplet. Their angular shapes differ: Y_ℓ^ℓ hugs the equator while Y_ℓ^0 concentrates toward the poles. Summed over the full multiplet, the angular shapes conspire to a spherically symmetric total density (Unsöld's closure, told with Spherical Harmonics) — a filled multiplet has no direction, as the symmetry demands.