Lin-Shu-Kalnajs Modes in Stellar Discs

• Lin–Shu–Kalnajs modes are defined as normal-mode solutions of the linearized collisionless Boltzmann–Poisson system in razor-thin stellar discs, capturing spiral density-wave behavior.
• They are derived via the WKB approximation with a reduction factor that links self-gravity, finite stellar random motions, and the local dispersion relation.
• Modern treatments extend this framework using unified linear-response theory and matrix methods to incorporate swing amplification, resonances, and transient dynamics.

Searching arXiv for recent and foundational treatments of Lin–Shu–Kalnajs modes and related stellar-disc dynamics. Relevant arXiv sources already identified in the provided corpus span foundational lecture notes, shearing-sheet analyses, unified linear-response theory, and modern matrix methods: (Binney, 2012, Binney, 2019, Meidt et al., 2023, Hamilton et al., 22 Jul 2025), and (Polyachenko et al., 30 Jun 2026). Lin–Shu–Kalnajs (LSK) modes are normal-mode solutions of the linearized collisionless Boltzmann–Poisson system for self-gravitating stellar discs, usually formulated for razor-thin, nearly axisymmetric equilibria and most commonly analyzed in the tightly wound, or WKB, limit. In that limit, spiral perturbations are represented by Fourier components of the form $e^{i(kR+m\phi-\omega t)}$, and the disc response is governed by the Lin–Shu–Kalnajs dispersion relation with its characteristic reduction factor $\mathcal{F}(s,\chi)$, which encodes the effect of finite stellar random motions. The LSK framework occupies a central place in spiral-density-wave theory because it provides both a local dispersion relation and a route to global normal-mode calculations; however, it is also now understood to be only one asymptotic sector of a broader linear-response problem that includes winding, swing amplification, corotation physics, and resonant absorption (Binney, 2012).

1. Normal-mode formulation in collisionless stellar discs

The formal starting point is the linearized collisionless Boltzmann equation coupled to Poisson’s equation for an equilibrium axisymmetric disc with distribution function $f_0(\mathbf{J})$ in action–angle variables $(\mathbf{J},\boldsymbol\theta)$: $$\frac{\partial f_1}{\partial t}+\{f_1,H_0\}=-\{f_0,\Phi_1\}, \qquad \nabla^2\Phi_1=4\pi G\int f_1\,d^3\mathbf{v}.$$ Seeking normal modes,

$$f_1(\mathbf{J},\boldsymbol\theta,t) =\sum_{\mathbf n}\hat f_{\mathbf n}(\mathbf J)\, e^{\,i(\mathbf n\cdot\boldsymbol\theta-\omega t)}, \qquad \Phi_1(\mathbf x,t)=\hat\Phi(\mathbf x)e^{-i\omega t},$$

leads to an integral eigenvalue problem for the complex frequency $\omega$ and the mode shape $\hat\Phi$ (Binney, 2012).

A closely related matrix formulation is obtained by expanding the perturbation in angle harmonics. In the four-dimensional phase space $(J_1,J_2;w_1,w_2)$, with $J_2=L$, the linearized equation for a perturbation $\mathcal{F}(s,\chi)$0 becomes

$\mathcal{F}(s,\chi)$1

and, after harmonic decomposition, yields

$\mathcal{F}(s,\chi)$2

with

$\mathcal{F}(s,\chi)$3

Together with the Poisson integral, this produces a standard eigenvalue problem

$\mathcal{F}(s,\chi)$4

which is the modern matrix form of the global normal-mode calculation (Polyachenko et al., 30 Jun 2026).

Within this formal structure, LSK modes are not defined merely by the existence of an eigenvalue problem, but by the additional asymptotic assumptions that reduce the global response to a local spiral-wave dispersion relation.

2. WKB reduction and the Lin–Shu–Kalnajs dispersion relation

The standard derivation imposes four assumptions: the spiral wave is tightly wound, so that $\mathcal{F}(s,\chi)$5; the unperturbed disc is razor-thin with surface density $\mathcal{F}(s,\chi)$6; the epicycle approximation holds; and attention is restricted to a single Fourier component proportional to $\mathcal{F}(s,\chi)$7 (Binney, 2012).

In this regime, stellar orbits are decomposed into a guiding centre at $\mathcal{F}(s,\chi)$8 rotating at $\mathcal{F}(s,\chi)$9 plus small radial oscillations of frequency $f_0(\mathbf{J})$0. After Fourier–Laplace transformation in $f_0(\mathbf{J})$1 and averaging over epicyclic phases, the perturbed surface density can be written in terms of the potential amplitude through a reduction factor: $f_0(\mathbf{J})$2

$f_0(\mathbf{J})$3

or, equivalently,

$f_0(\mathbf{J})$4

For a razor-thin disc, Poisson’s equation gives

$f_0(\mathbf{J})$5

and combination of response and self-gravity yields the LSK dispersion relation

$f_0(\mathbf{J})$6

or

$f_0(\mathbf{J})$7

(Binney, 2012).

A closely related representation, derived from the Laplace-transformed Volterra kernel in a razor-thin stellar disc, is

$f_0(\mathbf{J})$8

with

$f_0(\mathbf{J})$9

This formulation makes explicit the equivalence between the local WKB approach and the more general linear-response theory (Hamilton et al., 22 Jul 2025).

The physical content of $(\mathbf{J},\boldsymbol\theta)$0 is the suppression of self-gravitating response by finite radial velocity dispersion. In the cold-disc limit, long waves approach $(\mathbf{J},\boldsymbol\theta)$1; at large $(\mathbf{J},\boldsymbol\theta)$2, short waves are stabilized as $(\mathbf{J},\boldsymbol\theta)$3 (Binney, 2012).

3. Propagation, resonances, and discrete global modes

For real $(\mathbf{J},\boldsymbol\theta)$4, the LSK dispersion relation defines neutrally stable wave propagation and permits the calculation of the radial group velocity

$(\mathbf{J},\boldsymbol\theta)$5

in the reduction-factor formulation (Binney, 2012). In the axisymmetric-sheet WKB limit, the corresponding expression is

$(\mathbf{J},\boldsymbol\theta)$6

which is Toomre’s group-velocity result (Binney, 2019).

The local dispersion relation implies a resonance structure. The Lindblad resonances occur when

$(\mathbf{J},\boldsymbol\theta)$7

and corotation occurs when

$(\mathbf{J},\boldsymbol\theta)$8

Only between the Lindblad resonances can real waves propagate; outside that interval the local solutions are evanescent (Binney, 2012). More generally, the reduction factor controls the growth or damping rate near resonances, and allowing a small imaginary part in the eigenfrequency leads to a Landau prescription for damping or growth in the lower half-$(\mathbf{J},\boldsymbol\theta)$9 plane (Hamilton et al., 22 Jul 2025).

To convert the local WKB picture into discrete global modes, one imposes boundary conditions at turning radii $$\frac{\partial f_1}{\partial t}+\{f_1,H_0\}=-\{f_0,\Phi_1\}, \qquad \nabla^2\Phi_1=4\pi G\int f_1\,d^3\mathbf{v}.$$0 and $$\frac{\partial f_1}{\partial t}+\{f_1,H_0\}=-\{f_0,\Phi_1\}, \qquad \nabla^2\Phi_1=4\pi G\int f_1\,d^3\mathbf{v}.$$1 where $$\frac{\partial f_1}{\partial t}+\{f_1,H_0\}=-\{f_0,\Phi_1\}, \qquad \nabla^2\Phi_1=4\pi G\int f_1\,d^3\mathbf{v}.$$2 changes sign. The quantization condition is

$$\frac{\partial f_1}{\partial t}+\{f_1,H_0\}=-\{f_0,\Phi_1\}, \qquad \nabla^2\Phi_1=4\pi G\int f_1\,d^3\mathbf{v}.$$3

with Maslov phase shifts of $$\frac{\partial f_1}{\partial t}+\{f_1,H_0\}=-\{f_0,\Phi_1\}, \qquad \nabla^2\Phi_1=4\pi G\int f_1\,d^3\mathbf{v}.$$4 introduced when matching across corotation and Lindblad turning points (Hamilton et al., 22 Jul 2025). The resulting radial dependence is

$$\frac{\partial f_1}{\partial t}+\{f_1,H_0\}=-\{f_0,\Phi_1\}, \qquad \nabla^2\Phi_1=4\pi G\int f_1\,d^3\mathbf{v}.$$5

This WKB quantization gives a discrete sequence of pattern speeds $$\frac{\partial f_1}{\partial t}+\{f_1,H_0\}=-\{f_0,\Phi_1\}, \qquad \nabla^2\Phi_1=4\pi G\int f_1\,d^3\mathbf{v}.$$6 and mode numbers $$\frac{\partial f_1}{\partial t}+\{f_1,H_0\}=-\{f_0,\Phi_1\}, \qquad \nabla^2\Phi_1=4\pi G\int f_1\,d^3\mathbf{v}.$$7, but its validity depends on the same tight-winding and slowly varying assumptions that underlie the local dispersion relation.

4. Winding, swing amplification, and the limits of the LSK picture

The main limitation of the LSK framework is that it assumes a single fixed $$\frac{\partial f_1}{\partial t}+\{f_1,H_0\}=-\{f_0,\Phi_1\}, \qquad \nabla^2\Phi_1=4\pi G\int f_1\,d^3\mathbf{v}.$$8 for a normal mode, whereas differential rotation shears non-axisymmetric disturbances. In a differentially rotating disc,

$$\frac{\partial f_1}{\partial t}+\{f_1,H_0\}=-\{f_0,\Phi_1\}, \qquad \nabla^2\Phi_1=4\pi G\int f_1\,d^3\mathbf{v}.$$9

so an initially open spiral tightens with time and eventually leaves the WKB regime or is damped near resonances (Binney, 2012). In the shearing sheet, the local radial wavenumber evolves as

$$f_1(\mathbf{J},\boldsymbol\theta,t) =\sum_{\mathbf n}\hat f_{\mathbf n}(\mathbf J)\, e^{\,i(\mathbf n\cdot\boldsymbol\theta-\omega t)}, \qquad \Phi_1(\mathbf x,t)=\hat\Phi(\mathbf x)e^{-i\omega t},$$0

with $$f_1(\mathbf{J},\boldsymbol\theta,t) =\sum_{\mathbf n}\hat f_{\mathbf n}(\mathbf J)\, e^{\,i(\mathbf n\cdot\boldsymbol\theta-\omega t)}, \qquad \Phi_1(\mathbf x,t)=\hat\Phi(\mathbf x)e^{-i\omega t},$$1 Oort’s first constant (Binney, 2019).

This winding dynamics is the basis of swing amplification. In the shearing sheet, a leading disturbance can be transiently amplified by factors of ten or more before it winds up, and this explains why cool stellar discs are responsive systems that amplify ambient noise (Binney, 2012). The shearing-wave framework extends this by treating tight-winding Lin–Shu density waves and swing-amplified material patterns as limits of a common non-steady problem, with growth driven by the term $$f_1(\mathbf{J},\boldsymbol\theta,t) =\sum_{\mathbf n}\hat f_{\mathbf n}(\mathbf J)\, e^{\,i(\mathbf n\cdot\boldsymbol\theta-\omega t)}, \qquad \Phi_1(\mathbf x,t)=\hat\Phi(\mathbf x)e^{-i\omega t},$$2 and interpreted either as swing amplification or as the Lynden-Bell–Kalnajs donkey effect at corotation (Meidt et al., 2023).

The shearing-sheet analysis also alters the interpretation of wave propagation near corotation. Rather than a single packet bounded by a forbidden zone, two wavepackets emerge inside corotation: one each side of the inner Lindblad resonance (Binney, 2019). Disturbances do not avoid an annulus around corotation, and as wavepackets approach Lindblad resonances they are subject to Landau absorption, resonant trapping, or partial reflection from the velocity-space distortion left by previous waves (Binney, 2019).

For this reason, the LSK dispersion relation and even global normal-mode calculations provide a very incomplete understanding of the dynamics of stellar discs (Binney, 2012). A plausible implication is that LSK modes are best regarded as asymptotic organizing structures rather than exhaustive descriptions of long-lived spiral morphology.

5. Relation to unified linear-response theory

Modern work places LSK modes inside a broader response framework for razor-thin stellar discs. In this formulation, the perturbed potential and surface density are expanded in a biorthogonal basis, and the modal amplitudes satisfy a Volterra integral equation,

$$f_1(\mathbf{J},\boldsymbol\theta,t) =\sum_{\mathbf n}\hat f_{\mathbf n}(\mathbf J)\, e^{\,i(\mathbf n\cdot\boldsymbol\theta-\omega t)}, \qquad \Phi_1(\mathbf x,t)=\hat\Phi(\mathbf x)e^{-i\omega t},$$3

with kernel

$$f_1(\mathbf{J},\boldsymbol\theta,t) =\sum_{\mathbf n}\hat f_{\mathbf n}(\mathbf J)\, e^{\,i(\mathbf n\cdot\boldsymbol\theta-\omega t)}, \qquad \Phi_1(\mathbf x,t)=\hat\Phi(\mathbf x)e^{-i\omega t},$$4

(Hamilton et al., 22 Jul 2025).

Within this theory, the Volterra, Landau, and van Kampen representations are equivalent, and several classic results emerge as limiting cases: Lindblad–Kalnajs density waves, swing amplification, Lin–Shu–Kalnajs modes, and groove instabilities (Hamilton et al., 22 Jul 2025). The short-wavelength asymptotic expansion recovers the WKB kernel and hence the LSK dispersion relation precisely in the limit $$f_1(\mathbf{J},\boldsymbol\theta,t) =\sum_{\mathbf n}\hat f_{\mathbf n}(\mathbf J)\, e^{\,i(\mathbf n\cdot\boldsymbol\theta-\omega t)}, \qquad \Phi_1(\mathbf x,t)=\hat\Phi(\mathbf x)e^{-i\omega t},$$5 (Hamilton et al., 22 Jul 2025).

A complementary unification is provided by the shearing-wave cubic dispersion relation

$$f_1(\mathbf{J},\boldsymbol\theta,t) =\sum_{\mathbf n}\hat f_{\mathbf n}(\mathbf J)\, e^{\,i(\mathbf n\cdot\boldsymbol\theta-\omega t)}, \qquad \Phi_1(\mathbf x,t)=\hat\Phi(\mathbf x)e^{-i\omega t},$$6

which, in the steady-mode limit with $$f_1(\mathbf{J},\boldsymbol\theta,t) =\sum_{\mathbf n}\hat f_{\mathbf n}(\mathbf J)\, e^{\,i(\mathbf n\cdot\boldsymbol\theta-\omega t)}, \qquad \Phi_1(\mathbf x,t)=\hat\Phi(\mathbf x)e^{-i\omega t},$$7 and the imaginary term neglected, reduces to

$$f_1(\mathbf{J},\boldsymbol\theta,t) =\sum_{\mathbf n}\hat f_{\mathbf n}(\mathbf J)\, e^{\,i(\mathbf n\cdot\boldsymbol\theta-\omega t)}, \qquad \Phi_1(\mathbf x,t)=\hat\Phi(\mathbf x)e^{-i\omega t},$$8

and thereby recovers the Lin–Shu–Kalnajs dispersion relation in 3D (Meidt et al., 2023). The significance of these unified treatments is not that they replace LSK modes, but that they delimit the regime in which the classical normal-mode language is asymptotically valid.

6. Modern computation of global LSK-type modes

The contemporary numerical treatment of collisionless-disc modes generalizes the classical action–angle matrix method. For distribution functions with a sharp edge at $$f_1(\mathbf{J},\boldsymbol\theta,t) =\sum_{\mathbf n}\hat f_{\mathbf n}(\mathbf J)\, e^{\,i(\mathbf n\cdot\boldsymbol\theta-\omega t)}, \qquad \Phi_1(\mathbf x,t)=\hat\Phi(\mathbf x)e^{-i\omega t},$$9,

$\omega$0

the generalized linear-matrix equation acquires a boundary-integral term but no increase in matrix size (Polyachenko et al., 30 Jun 2026). This is important for unidirectional discs with sharp angular-momentum boundaries.

The same formulation naturally incorporates gravitational softening by replacing the Poisson kernel denominator $\omega$1 with

$\omega$2

so the softening length $\omega$3 enters directly through the matrix elements (Polyachenko et al., 30 Jun 2026). This enables direct comparison with softened $\omega$4-body calculations.

For the Kuzmin–Toomre disk with Kalnajs DF index $\omega$5, the five most unstable $\omega$6 modes were computed with a sharp-edge calculation using grid $\omega$7, $\omega$8, and $\omega$9, giving

$\hat\Phi$0

A convergence study in $\hat\Phi$1 yielded

$\hat\Phi$2

while comparison to an independent log-spiral method gave the “exact” values

$\hat\Phi$3

so the extrapolated matrix method agreed to $\hat\Phi$4 in each component (Polyachenko et al., 30 Jun 2026).

These calculations illustrate that the classical LSK formalism remains operational as a computational framework, even though its physical interpretation must now be embedded in a more general theory of transient, resonant, and shearing spiral structure.