LMS-1/Wukong: Halo Merger Remnant

• LMS-1/Wukong is a metal-poor halo structure identified as a disrupted dwarf galaxy remnant with convergent chemodynamic evidence from Gaia and spectroscopy.
• It exhibits polar, moderately prograde orbits with tightly clustered dynamics in action–energy space, anchoring detailed studies of Milky Way merger remnants.
• Chemical analyses reveal a high-α plateau and delayed r-process enrichment, providing key constraints on nucleosynthesis and prolonged star formation in low-mass galaxies.

LMS-1/Wukong is a recently identified stellar stream and disrupted dwarf-galaxy remnant in the Galactic halo, independently discovered as “LMS-1” by Yuan et al. and “Wukong” by Naidu et al., and later recognized as the same structure because of its essentially identical dynamics (Limberg et al., 2023). Current work treats it not merely as a kinematic stream but as the debris of an ancient accretion event: a chemically coherent, highly metal-poor progenitor galaxy whose stars, associated streams, and globular clusters preserve information about early Milky Way assembly, dwarf-galaxy chemical evolution, and delayed nucleosynthetic channels (Malhan et al., 2022).

1. Identification as a Milky Way merger remnant

In the Gaia-based “global dynamical atlas” of Milky Way mergers, LMS-1/Wukong is one of six high-significance merger groups identified by clustering globular clusters, stellar streams, and satellite galaxies in action–energy space (Malhan et al., 2022). The search is carried out in the 4D space

$$\mathbf{x}_i \equiv (J_{R,i}, J_{\phi,i}, J_{z,i}, E_i),$$

using Gaia EDR3-based phase-space data and the density-based hierarchical algorithm ENLINK. The inclusion of orbital energy $E$ is emphasized because its uncertainties are smaller than those in the actions, which sharpens group separation.

Within that framework, LMS-1/Wukong is identified as Group 5. The group is described as having “a slight prograde motion” and member objects that are “very tightly clumped in $(\mathbf{J},E)$ space” (Malhan et al., 2022). For stream modeling, LMS-1 is treated with the orbit-sampling approach because it is dynamically broad or hot. This places the system among the best-defined accretion structures in current dynamical classifications of the halo.

The naming convention reflects convergent discovery rather than physical multiplicity. Later chemical work explicitly adopts “Wukong/LMS-1” as a single system and treats it as the debris of a former dwarf galaxy now stretched along a halo orbit (Limberg et al., 2023). A plausible implication is that the dual nomenclature encodes the history of discovery, whereas the present literature generally uses the combined label to indicate a single accretion structure.

2. Orbital configuration and present-day dynamical properties

LMS-1/Wukong is characterized as a polar merger group with dynamical properties in the ranges

$E\sim[-1.41,-1.19]\times10^{52}\,{\rm s^2},$

$$J_R\sim[100,605],\quad J_\phi\sim[-1560,-210],\quad J_z\sim[875,2710],$$

$L_\perp\sim[1400,3085],\quad e\sim[0.2,0.5],$

$$r_{\rm peri}\sim[5,13]\ {\rm kpc},\quad r_{\rm apo}\sim[15,25]\ {\rm kpc},\quad \phi\sim[58^\circ,85^\circ]$$

(Malhan et al., 2022). These values are interpreted in that study as indicating moderately prograde motion in its sign convention, highly inclined or polar orbits, moderate eccentricities, and a comparatively compact radial range.

For the LMS-1 stream specifically, the same work gives

$$(J_R,J_\phi,J_z)= (255^{+239}_{-149},\,-627^{+183}_{-232},\,2514^{+383}_{-263}),$$

$E=-1227^{+65}_{-39}\times10^{22}\,{\rm s^2},$

$$r_{\rm peri}=10.8^{+2.5}_{-1.8}\ {\rm kpc},\quad r_{\rm apo}=20.6^{+3.7}_{-1.9}\ {\rm kpc},$$

$E$0

with $E$1 dex (Malhan et al., 2022). In that sense, LMS-1 itself functions as an anchor object for the merger: metal-poor, halo-like, and dynamically polar.

A later stellar study continues to frame the system as occupying a polar, slightly prograde orbit and uses that chemodynamical locus to associate individual stars with the progenitor (Schichtel et al., 8 Sep 2025). This suggests that LMS-1/Wukong has become a stable reference structure in the orbital taxonomy of accreted halo debris.

3. Chemical abundance patterns and enrichment history

The first detailed chemical-abundance analysis of Wukong/LMS-1 covers a wide metallicity range,

$E$2

based on 14 stars observed with MIKE/Magellan spectroscopy (Limberg et al., 2023). That study concludes that Wukong/LMS-1 is “chemically indistinguishable” from the bulk of Indus and Jhelum, supporting the long-standing dynamical suggestion that these smaller streams are fragments of the same larger accretion event or parent dwarf galaxy.

A central result is that Wukong/LMS-1 exhibits a high-$E$3 plateau up to at least $E$4, with Mg and Ca remaining around $E$5–0.4 dex, and then becomes $E$6-poor toward the high-metallicity end (Limberg et al., 2023). The APOGEE star Wuk_14, at

$E$7

is lower in [Mg/Fe] and [Ca/Fe] by about 0.1–0.2 dex than the bulk of the sample. The paper interprets this as evidence for a fairly standard and simple enrichment history in which early core-collapse supernova enrichment is followed by later Type Ia Fe enrichment.

The same study confirms a carbon-enhanced metal-poor star, Wuk_4, with

$E$8

The evolutionary carbon correction is only $E$9 dex, so the CEMP classification is robust (Limberg et al., 2023). Its neutron-capture pattern is unusual: Sr, Y, and Zr are strongly enhanced, roughly

$(\mathbf{J},E)$0

whereas

$(\mathbf{J},E)$1

so it is not a classic CEMP-$(\mathbf{J},E)$2 star. The paper also notes an $(\mathbf{J},E)$3 km s$(\mathbf{J},E)$4 RV discrepancy between H3 and MIKE, suggesting binarity.

Neutron-capture evolution is especially prominent. The analysis reports that $(\mathbf{J},E)$5 rises from below $(\mathbf{J},E)$6 dex at $(\mathbf{J},E)$7 to above $(\mathbf{J},E)$8 dex at $(\mathbf{J},E)$9, while $E\sim[-1.41,-1.19]\times10^{52}\,{\rm s^2},$0 remains around $E\sim[-1.41,-1.19]\times10^{52}\,{\rm s^2},$1 dex (Limberg et al., 2023). Using $E\sim[-1.41,-1.19]\times10^{52}\,{\rm s^2},$2 versus $E\sim[-1.41,-1.19]\times10^{52}\,{\rm s^2},$3, the paper argues that the $E\sim[-1.41,-1.19]\times10^{52}\,{\rm s^2},$4-process source is delayed relative to prompt core-collapse yields and identifies Wukong/LMS-1 as the first dwarf galaxy, observed as a stream, in which the production of $E\sim[-1.41,-1.19]\times10^{52}\,{\rm s^2},$5-process elements is clearly dominated by delayed sources, presumably neutron-star mergers.

4. Star-formation history and one-zone chemical-evolution modeling

A complementary reconstruction comes from one-zone galactic chemical-evolution models with exponential infall histories applied to Wukong/LMS-1 and Gaia-Sausage Enceladus (GSE) (Johnson et al., 2022). For Wukong/LMS-1, the best-fit parameters are an exponential infall e-folding timescale

$E\sim[-1.41,-1.19]\times10^{52}\,{\rm s^2},$6

and a total star-formation duration

$E\sim[-1.41,-1.19]\times10^{52}\,{\rm s^2},$7

The paper interprets this as evidence that Wukong/LMS-1 experienced a relatively prolonged gas accretion phase and formed stars for about 3.4 Gyr in total.

The comparison system, GSE, is fitted with

$E\sim[-1.41,-1.19]\times10^{52}\,{\rm s^2},$8

and is inferred to have continued forming stars for roughly 1.5–2 Gyr after first infall into the Milky Way about 10 Gyr ago (Johnson et al., 2022). For Wukong/LMS-1, no similarly precise infall time is provided, but the longer $E\sim[-1.41,-1.19]\times10^{52}\,{\rm s^2},$9 is stated to be consistent with a slower buildup of the interstellar medium before quenching and disruption.

Methodologically, the fit is based on a statistically robust likelihood function using Poisson sampling from an evolutionary track and requiring no binning of the data (Johnson et al., 2022). Validation against mock data shows recovery of the input model over sample sizes $$J_R\sim[100,605],\quad J_\phi\sim[-1560,-210],\quad J_z\sim[875,2710],$$0, abundance uncertainties $$J_R\sim[100,605],\quad J_\phi\sim[-1560,-210],\quad J_z\sim[875,2710],$$1, and age uncertainties $$J_R\sim[100,605],\quad J_\phi\sim[-1560,-210],\quad J_z\sim[875,2710],$$2. A major limitation is explicit: the Wukong/LMS-1 sample used in that analysis contains no age measurements, so $$J_R\sim[100,605],\quad J_\phi\sim[-1560,-210],\quad J_z\sim[875,2710],$$3 is inferred from chemical-abundance modeling without direct age constraints.

The same work reports that the inferred outflow mass-loading factor is in reasonable agreement with the galactic-wind scaling

$$J_R\sim[100,605],\quad J_\phi\sim[-1560,-210],\quad J_z\sim[875,2710],$$4

and concludes that the differences between GSE and Wukong/LMS-1 are qualitatively consistent with trends predicted by simulations and semi-analytic models as a function of stellar mass $$J_R\sim[100,605],\quad J_\phi\sim[-1560,-210],\quad J_z\sim[875,2710],$$5 (Johnson et al., 2022). This places LMS-1/Wukong within a broader mass-dependent evolutionary sequence of accreted dwarfs.

5. Associated streams, globular clusters, and progenitor mass

The merger-group analysis associates 11 objects with LMS-1/Wukong: the streams LMS-1 itself, Phoenix, Pal 5, C-19, Indus, Sylgr, and Jhelum, plus the globular clusters NGC 5272/M 3, NGC 5053, and NGC 5024/M 53; Pal 5 also appears as both a cluster and a stream association in the summary table (Malhan et al., 2022). The study emphasizes that these are not merely kinematic coincidences: the streams and clusters are tightly grouped in $$J_R\sim[100,605],\quad J_\phi\sim[-1560,-210],\quad J_z\sim[875,2710],$$6 and are interpreted as having been accreted together in the same progenitor galaxy.

LMS-1/Wukong is singled out as the most metal-poor merger of the Milky Way because it contains the three most metal-poor streams of the Galaxy: C-19 with $$J_R\sim[100,605],\quad J_\phi\sim[-1560,-210],\quad J_z\sim[875,2710],$$7 dex, Sylgr with $$J_R\sim[100,605],\quad J_\phi\sim[-1560,-210],\quad J_z\sim[875,2710],$$8 dex, and Phoenix with $$J_R\sim[100,605],\quad J_\phi\sim[-1560,-210],\quad J_z\sim[875,2710],$$9 dex (Malhan et al., 2022). The group metallicity distribution extends from $L_\perp\sim[1400,3085],\quad e\sim[0.2,0.5],$0 to $L_\perp\sim[1400,3085],\quad e\sim[0.2,0.5],$1 dex, with median $L_\perp\sim[1400,3085],\quad e\sim[0.2,0.5],$2 dex. This is the basis for describing the progenitor as an ancient, extremely metal-poor proto-galaxy.

Detailed abundance work reinforces the connection to a larger accreted system. Wukong/LMS-1 is chemically indistinguishable from the bulk of Indus and Jhelum, and two N- and Na-rich stars in Wukong/LMS-1, Wuk_5 and Wuk_11, are interpreted as second-generation globular-cluster stars (Limberg et al., 2023). Together with an analogous Indus star, this indicates that the progenitor contained at least one globular cluster that was completely disrupted, in addition to the two intact clusters already associated with the stream, NGC 5024 and NGC 5053.

Mass estimates depend on method and associated cluster census. Using the globular-cluster members, the dynamical-atlas study estimates approximately

$L_\perp\sim[1400,3085],\quad e\sim[0.2,0.5],$3

(Malhan et al., 2022). A later chemical study, using the empirical relation between globular-cluster number and halo virial mass and the presence of at least three globular clusters, infers

$L_\perp\sim[1400,3085],\quad e\sim[0.2,0.5],$4

with a stellar mass of order $L_\perp\sim[1400,3085],\quad e\sim[0.2,0.5],$5 and a total mass corresponding to $L_\perp\sim[1400,3085],\quad e\sim[0.2,0.5],$6 of the present-day Milky Way (Limberg et al., 2023). One paper explicitly notes that its higher value relative to earlier LMS-1 work results from identifying more globular-cluster associations (Malhan et al., 2022). This suggests that progenitor-mass inference is sensitive to the adopted membership inventory.

6. Individual-member studies, nucleosynthetic implications, and disambiguation

The star HE2159-0551 has been proposed as a likely LMS-1/Wukong member on both chemical and kinematic grounds (Schichtel et al., 8 Sep 2025). It is a very metal-poor giant with

$L_\perp\sim[1400,3085],\quad e\sim[0.2,0.5],$7

and a distinctive heavy-element pattern: $L_\perp\sim[1400,3085],\quad e\sim[0.2,0.5],$8

$L_\perp\sim[1400,3085],\quad e\sim[0.2,0.5],$9

The more than 1 dex spread among these neutron-capture species, especially the contrast between low Ba and comparatively elevated Zr, is identified as the most striking chemical feature.

That study argues that HE2159-0551 satisfies the LMS-1/Wukong chemodynamical selection criteria, falls within or adjacent to the Wukong/LMS-1 region in the $$r_{\rm peri}\sim[5,13]\ {\rm kpc},\quad r_{\rm apo}\sim[15,25]\ {\rm kpc},\quad \phi\sim[58^\circ,85^\circ]$$0 plane and action diamond, and does not satisfy Helmi-stream criteria (Schichtel et al., 8 Sep 2025). Although it is thick-disc-like in a Toomre-diagram sense, the favored interpretation is an accretion origin tied to LMS-1/Wukong rather than an in-situ thick-disk origin. On nucleosynthesis, the authors reject both a classical AGB-driven main $$r_{\rm peri}\sim[5,13]\ {\rm kpc},\quad r_{\rm apo}\sim[15,25]\ {\rm kpc},\quad \phi\sim[58^\circ,85^\circ]$$1-process origin and a strong main $$r_{\rm peri}\sim[5,13]\ {\rm kpc},\quad r_{\rm apo}\sim[15,25]\ {\rm kpc},\quad \phi\sim[58^\circ,85^\circ]$$2-process explanation, and instead favor a weak-$$r_{\rm peri}\sim[5,13]\ {\rm kpc},\quad r_{\rm apo}\sim[15,25]\ {\rm kpc},\quad \phi\sim[58^\circ,85^\circ]$$3 or $$r_{\rm peri}\sim[5,13]\ {\rm kpc},\quad r_{\rm apo}\sim[15,25]\ {\rm kpc},\quad \phi\sim[58^\circ,85^\circ]$$4-process producing Sr–Zr but not Ba, possibly mixed with a low level of underlying $$r_{\rm peri}\sim[5,13]\ {\rm kpc},\quad r_{\rm apo}\sim[15,25]\ {\rm kpc},\quad \phi\sim[58^\circ,85^\circ]$$5-process material.

A common source of confusion is the reuse of the name “Wukong” in unrelated arXiv literature. For example, “PDF-WuKong” is a multimodal LLM for long PDF question answering, and its paper does not establish that it is the same as LMS-1/Wukong; the safest reading given there is that it is a separate model name sharing the “WuKong” branding (Xie et al., 2024). In astrophysical usage, by contrast, LMS-1/Wukong denotes the halo stream and disrupted dwarf-galaxy remnant described above.