Disc of Satellite Galaxies (DoS)

• Disc of Satellite Galaxies (DoS) is an anisotropic, highly flattened satellite configuration around host galaxies, exemplified by structures in the Milky Way and Andromeda.
• Observational and kinematic diagnostics, including PCA and tensor of inertia methods, reveal pronounced flattening and aligned orbital poles in satellite systems.
• Formation scenarios range from anisotropic filamentary infall to tidal interactions, with simulations highlighting transient coherence and dynamic stability challenges.

The Disc of Satellite Galaxies (DoS) denotes a highly flattened, anisotropic arrangement of satellite galaxies around a host galaxy. In the Milky Way, the same phenomenon is often discussed as the Vast Polar Structure (VPOS), a thin plane nearly perpendicular to the Galactic disc; in Andromeda, the term usually refers to an extended, thin satellite plane with a reported vertical rms scatter of $\sim 12\,\mathrm{kpc}$ and a radial extent of $\sim 300$–$400\,\mathrm{kpc}$ (Pawlowski et al., 2012, Pawlowski et al., 2013, Bowden et al., 2013). The DoS has become a focal problem in galactic dynamics and cosmology because its interpretation depends simultaneously on geometric diagnostics, orbital information, halo shape, baryonic effects, sample selection, and the adopted formation scenario.

1. Observational definition and phenomenology

The empirical basis of the DoS problem is the repeated identification of flattened satellite configurations in the Local Group. Around the Milky Way, the classical and faint satellites were described as lying within a thin, planar structure oriented nearly perpendicular to the Milky Way disc, while the VPOS formulation extends this alignment to young halo globular clusters and stellar or gaseous streams (Pawlowski, 2011, Pawlowski et al., 2012). Around Andromeda, a thin and extended satellite plane has been reported, with 13 of 15 satellites exhibiting coherent rotation and with a radial extent of order several hundred kiloparsecs (Bowden et al., 2013).

The Milky Way VPOS is not defined purely by satellite positions. Young halo globular clusters have a fitted plane with a normal only $13^\circ$ from the DoS normal, and 7 of 14 analysed streams have normals within $32^\circ$ of the DoS normal; the probability of obtaining the observed stream clustering under isotropy was reported as only $0.3\%$ (Pawlowski et al., 2012). The structure extends from Galactocentric distances below $10\,\mathrm{kpc}$ to roughly $250\,\mathrm{kpc}$, so the DoS/VPOS terminology refers not merely to a positional flattening of satellites but, in some analyses, to a larger phase-space-correlated halo structure (Pawlowski et al., 2012).

Kinematic evidence has been central to the stronger versions of the DoS claim. Using proper motions for the 11 classical Milky Way satellites, one analysis found that 7 to 9 out of 11 satellites have aligned orbital poles, with 6 satellites within $\sim 18.5^\circ$ of the mean pole and 8 within $\sim 29^\circ$ (Pawlowski et al., 2013). That work interpreted the Milky Way VPOS as rotationally stabilized rather than merely pressure-supported. Other analyses, however, have argued that the apparent coherence depends strongly on sample definition and on how angular-momentum clustering is assessed (Maji et al., 2017, Pawlowski et al., 2017).

2. Geometrical and kinematic diagnostics

DoS studies typically quantify flattening through the minor-to-major axis ratio $\sim 300$0, the rms height above a fitted plane, or related inertia-tensor diagnostics. In the notation used in simulation studies, the axial ratios are

$\sim 300$1

with triaxiality

$\sim 300$2

Lower $\sim 300$3 indicates stronger flattening (Deason et al., 2011). Plane fitting has been implemented by Principal Component Analysis (PCA), Tensor of Inertia (TOI) methods with different radial weightings, and direct least-squares fits in Hesse normal form,

$\sim 300$4

where $\sim 300$5 is the unit normal to the plane (Maji et al., 2017, Maji et al., 2017, Zhao et al., 2023).

The inferred strength of a DoS depends materially on the adopted statistic. For the Milky Way, analyses of 11 classical satellites yielded thinner planes than analyses of larger samples: one study reported $\sim 300$6 for 11 satellites and $\sim 300$7 for 39 satellites, with rms thickness increasing from $\sim 300$8 to $\sim 300$9 (Maji et al., 2017). A related simulation-based study reported that using only the 11 most massive satellites can yield $400\,\mathrm{kpc}$0–$400\,\mathrm{kpc}$1 and rms height $400\,\mathrm{kpc}$2–$400\,\mathrm{kpc}$3, whereas including all 106 satellites within $400\,\mathrm{kpc}$4 raises $400\,\mathrm{kpc}$5 to $400\,\mathrm{kpc}$6 under PCA and to $400\,\mathrm{kpc}$7 with TOI weighted by $400\,\mathrm{kpc}$8, with rms height increasing to $400\,\mathrm{kpc}$9 (Maji et al., 2017).

Orbital poles and angular-momentum vectors are the second major diagnostic. For individual satellites or debris particles,

$13^\circ$0

and the direction of $13^\circ$1 defines the orbital pole (Pawlowski, 2011). Because small samples can exaggerate pole clustering, Monte Carlo baselines are essential. One study found that for 11 randomly oriented vectors, obtaining 6 objects within a $13^\circ$2 cone occurs by chance with probability $13^\circ$3, and 5 objects with probability $13^\circ$4 (Maji et al., 2017). The same paper emphasized that even a perfectly isotropic distribution returns a median $13^\circ$5 for $13^\circ$6, so low-$13^\circ$7 samples systematically exaggerate planarity.

3. Halo geometry, orbital families, and dynamical persistence

The long-term survival of a thin DoS depends on the host potential. A constructive framework was provided by analytic triaxial generalisations of Navarro-Frenk-White haloes obtained by adding spherical harmonics to the NFW potential: $13^\circ$8

$13^\circ$9

and, for triaxial haloes,

$32^\circ$0

$32^\circ$1

These models retain the NFW-like density behaviour $32^\circ$2 at small radii and $32^\circ$3 at large radii, while allowing axis ratios that vary with radius (Bowden et al., 2013).

In these triaxial potentials, closed periodic orbits that can support long-lived discs exist only in planes perpendicular to the long and short axes of the halo. Epicyclic theory gives the orbital ellipticity as

$32^\circ$4

with $32^\circ$5 and $32^\circ$6 the angular and epicyclic frequencies and $32^\circ$7 the quadrupole amplitude (Bowden et al., 2013). This formalism underlies a sharp stability result: a thin satellite disc can persist over cosmological times if and only if it lies in the planes perpendicular to the long or short axis of a triaxial halo, or in the equatorial or polar planes of a spheroidal halo (Bowden et al., 2013).

Outside those special orientations, thickening is rapid on galactic timescales. The vertical scaleheight was described as growing exponentially,

$32^\circ$8

and the disc thickness doubles on $32^\circ$9 timescales; even a misalignment of $0.3\%$0 can produce noticeable thickening (Bowden et al., 2013). A separate hydrodynamical zoom-in study reached a different but related conclusion from the kinematic side: in a Milky-Way-like galaxy containing a DoS, the angular momentum vector of the whole satellite system differed from the fitted DoS normal and from the normal direction of the velocity dispersion, the bulk velocity flow was largely perpendicular to the DoS, and the structure was therefore interpreted as infalling sub-halos rather than a rotationally supported system (Zhao et al., 2023). Taken together, these results imply that a geometrically thin DoS need not be dynamically disk-like.

4. DoS in cosmological and hydrodynamical simulations

Within $0.3\%$1CDM-based simulations, satellite anisotropy is common, but strong rotational support is not. In the GIMIC simulations, the disc of the host galaxy is well aligned with the inner halo at $0.3\%$2, whereas substantial disc–halo misalignments occur at $0.3\%$3, with misalignments greater than $0.3\%$4 in approximately $0.3\%$5 of systems (Deason et al., 2011). The satellite population aligns preferentially with the shape and angular momentum of the outer halo, not with the central galaxy, and roughly $0.3\%$6 of systems with ten bright satellites within $0.3\%$7 show a polar spatial alignment relative to the disc (Deason et al., 2011). Only a small fraction, $0.3\%$8, show significant rotational support with $0.3\%$9, which that study attributed to group infall (Deason et al., 2011).

Hydrodynamic calculations alter both abundance and anisotropy relative to dark-matter-only runs. In one matched comparison, the hydrodynamic simulation contained 106 luminous subhalos within $10\,\mathrm{kpc}$0 at $10\,\mathrm{kpc}$1, whereas the dark-matter-only counterpart contained 21,220 subhalos (Maji et al., 2017, Maji et al., 2017). The hydrodynamic satellite distribution was more anisotropic, with $10\,\mathrm{kpc}$2 at $10\,\mathrm{kpc}$3 compared to $10\,\mathrm{kpc}$4 in the dark-matter-only case, but the kinematic fractions did not indicate coherent rotation: among 77 satellites moving within the fitted DoS plane, 18 were corotating and 19 counter-rotating, and across cosmic time the fractions of corotating and counter-corotating satellites remained comparable (Maji et al., 2017).

Time evolution in such simulations generally proceeds from near-isotropy at high redshift to stronger anisotropy at low redshift. One study reported $10\,\mathrm{kpc}$5 at $10\,\mathrm{kpc}$6–$10\,\mathrm{kpc}$7, decreasing to $10\,\mathrm{kpc}$8 by $10\,\mathrm{kpc}$9 (Maji et al., 2017). Another Milky-Way-like hydrodynamic zoom-in found a DoS aspect ratio of $250\,\mathrm{kpc}$0–$250\,\mathrm{kpc}$1, thickness $250\,\mathrm{kpc}$2–$250\,\mathrm{kpc}$3 within the virial radius, and inclination $250\,\mathrm{kpc}$4–$250\,\mathrm{kpc}$5 relative to the stellar disc, often close to perpendicular (Zhao et al., 2023). These properties were interpreted as consequences of anisotropic infall and the triaxial nature of the dark matter halo rather than evidence for a dynamically cold, long-lived rotating disk (Deason et al., 2011, Zhao et al., 2023).

5. Formation scenarios

One class of explanations places the DoS within ordinary hierarchical assembly. In this view, anisotropic accretion along filaments and along the major axes of dark matter haloes imprints flattened satellite distributions and a modest prograde bias. In GIMIC, satellites were reported to be more flattened than the dark matter itself, with median sphericity $250\,\mathrm{kpc}$6 compared to $250\,\mathrm{kpc}$7, and to show a co-rotating fraction of about $250\,\mathrm{kpc}$8 relative to the disc and $250\,\mathrm{kpc}$9 relative to the halo (Deason et al., 2011). This interpretation treats the DoS as a fossil of directional accretion and incomplete phase mixing.

A more specific standard-gravity merger model proposes that a secondary galaxy brings its own satellites into a merger, after which tidal forces and dynamical friction spread them into an extended, flattened, commonly rotating structure around the remnant (Smith et al., 2015). In that framework, near-circular mergers with $\sim 18.5^\circ$0 and a small initial $\sim 18.5^\circ$1-range of the secondary’s satellites are most conducive to forming a DoS; the simulated outcomes were diameters of $\sim 18.5^\circ$2–$\sim 18.5^\circ$3, thicknesses of $\sim 18.5^\circ$4–$\sim 18.5^\circ$5, flattening $\sim 18.5^\circ$6–$\sim 18.5^\circ$7, and survival for more than $\sim 18.5^\circ$8–$\sim 18.5^\circ$9 (Smith et al., 2015). Prograde satellites survive preferentially, whereas retrograde satellites tend to become radial and be destroyed, providing a dynamical filter for rotation coherence (Smith et al., 2015).

A different family of models attributes DoS members to tidal debris or tidal dwarf galaxies. Stellar-dynamical calculations with disc-galaxy encounters showed that both co- and counter-orbiting tidal debris arise naturally in mergers and fly-bys, with retrograde fractions ranging from a few percent up to $\sim 29^\circ$0 in some cases and, in fly-bys, even nearly $\sim 29^\circ$1 depending on parameters (Pawlowski, 2011, Pawlowski et al., 2011). The fly-by case displays a two-phase behaviour: retrograde material forms first as debris streams back along the tidal tail, then later material is captured on prograde orbits (Pawlowski, 2011). This framework has been used to explain why the Milky Way DoS contains a mostly co-orbiting population together with a counter-orbiting member such as Sculptor (Pawlowski, 2011, Pawlowski et al., 2011).

In modified-gravity work, DoS structures have been modeled as the aftermath of a Milky Way–Andromeda flyby. In a QUMOND simulation, the galaxies were tuned to reproduce observed positions, velocities, and disc orientations, producing a first pericenter passage $\sim 29^\circ$2 ago at a minimum separation of $\sim 29^\circ$3 and relative velocity of $\sim 29^\circ$4 (Bílek et al., 2021). A tidal tail from the Milky Way was captured by Andromeda and formed a cloud of particles resembling the Andromeda DoS in size, orientation, rotation, and mass, while also inducing a Milky Way warp of the observed magnitude and orientation (Bílek et al., 2021). That study also stressed a caveat: only stellar particles were included, and gas physics was not modeled, so bound tidal dwarf formation was not directly demonstrated (Bílek et al., 2021).

6. Statistical significance, external-galaxy evidence, and rarity

The DoS literature has been shaped as much by methodology as by dynamics. Several studies emphasize that conclusions require an explicit null hypothesis, typically isotropically distributed satellite positions and velocities, together with realistic treatment of survey footprints, obscuration by the Galactic disc, and proper-motion uncertainties (Pawlowski et al., 2017). One methodological warning concerns weighted plane fits: weighting by radius, such as $\sim 29^\circ$5 or $\sim 29^\circ$6, can reduce the inferred planar signal, while comparing observed samples limited to a few hundred kiloparsecs with simulated samples extending to $\sim 29^\circ$7 can generate meaningless contrasts (Pawlowski et al., 2017, Maji et al., 2017).

Claims for coherent rotation in external-galaxy satellite systems have likewise proven sensitive to sample definition. In Sloan Digital Sky Survey analyses, significant spatial anisotropy was detected, with about $\sim 29^\circ$8 more satellites at $\sim 29^\circ$9 than at $\sim 300$00 around the brightest primaries, in good agreement with Millennium simulations populated by semi-analytic galaxies (Cautun et al., 2014). However, the excess of diametrically opposed satellite pairs with anticorrelated radial velocities was highly sensitive to small changes in the selection criteria, and no corresponding excess of same-side pairs with correlated velocities was found, contrary to the expectation for rotating disks of satellites (Cautun et al., 2014). The conclusion of that study was that the detection of coherent rotation in current observational samples is not robust.

Rarity assessments have also diversified. The conventional measure treats the Milky Way as DoS-like if its present-day minor-to-major axis ratio satisfies $\sim 300$01, but a recent analysis argued that a single present-day $\sim 300$02 is inadequate because the intrinsic $\sim 300$03 distribution is broad and time-variable (Seo et al., 2024). Using a “satellite distribution generator” and $\sim 300$04 spatially and kinematically analogous systems (SKASs), that work found the Milky Way’s $\sim 300$05 probability distribution to have width $\sim 300$06 and re-estimated the rarity of the Milky Way DoS in IllustrisTNG50-1 as $\sim 300$07–$\sim 300$08 (Seo et al., 2024). It further argued that the Milky Way is exceptional because both the orbital poles and the radial distances of the 11 classical satellites are more plane-friendly than in simulated host–satellite systems (Seo et al., 2024).

The resulting picture is not a single consensus model but a structured controversy. One branch of the literature interprets the DoS as transient anisotropic infall, often reproduced in hydrodynamic $\sim 300$09CDM simulations without rotational support (Maji et al., 2017, Maji et al., 2017, Zhao et al., 2023). Another emphasizes the Milky Way’s orbital-pole coherence, the inclusion of streams and globular clusters in the VPOS, and the apparent rarity of similarly plane-friendly systems, arguing that standard subhalo accretion is insufficient and that tidal or interaction-based origins are better matched to the data (Pawlowski et al., 2012, Pawlowski et al., 2013, Seo et al., 2024). This suggests that the DoS is best understood not as a single observable with a single interpretation, but as the intersection of geometry, phase-space structure, host-halo shape, and assembly history.