High-resolution three-dimensional simulations of gas removal from ultrafaint dwarf galaxies. I. Stellar feedback

Annibale D'Ercole; C. Gareth Few; Donatella Romano; Francesco Calura

REVIEW 2 major objections 5 minor 85 references

High-resolution three-dimensional simulations of gas removal from ultrafaint dwarf galaxies. I. Stellar feedback

T0 review · 2 major / 5 minor · reviewed 2026-08-14 · deepseek-v4-flash

Challenge this review Re-run · record.json Download PDF Read on arXiv ↗ Review JSON

Pith's one-line read Stellar feedback alone cannot remove the cold gas from an ultrafaint dwarf galaxy like Boötes I, even when the total supernova energy exceeds the gas binding energy.

desk verdict Solid high-resolution simulation showing thermal feedback alone can't remove cold gas from a UFD, but the paper overreaches by concluding stellar feedback in general is insufficient when it never modeled radiation pressure. read the letter →

arxiv 1909.02329 v1 pith:DPH5MAFR submitted 2019-09-05 astro-ph.GA

Donatella Romano , Francesco Calura , Annibale D'Ercole , C. Gareth Few This is my paper · ORCID

classification astro-ph.GA

keywords ultrafaintdwarfgalaxiesstellarfeedbacksupernova-drivenoutflowsgasremovalhydrodynamicsimulationsBoötesIgalacticwindschemicalenrichment

topics Dark Matter

open problems Dark Matter

verification ladder T0 review T1 audit T2 compute T3 formal

The pith

A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.

The reading

This paper asks whether stellar feedback from a burst of star formation can, by itself, blow the cold gas out of an ultrafaint dwarf galaxy like Boötes I. Using high-resolution three-dimensional hydrodynamic simulations with an instantaneous $10^5\,M_\odot$ stellar population injecting winds and supernova energy for 30 Myr, it finds the answer is no: even though the total supernova energy exceeds the gas binding energy, no galactic-scale outflow develops and less than 20–30 percent of the cold gas is lost. The same simulations find the opposite for metal-rich ejecta: by 30 Myr, 70–80 percent of the supernova ejecta has left the simulation box. If this is right, the gas-free state of ultrafaint dwarfs cannot be explained by internal feedback alone, and environmental mechanisms such as ram-pressure stripping or tidal interactions must finish the cleaning. This matters because it challenges the common analytic assumption that a galaxy's gas is ejected once cumulative supernova energy exceeds binding energy.

What carries the argument

The load-bearing machinery is the spatially resolved, thermal injection of stellar winds and supernova energy from ten fixed OB associations inside a high-resolution, adaptive-mesh grid, with the galaxy initialized as a smooth, single-phase, non-rotating gas in hydrostatic equilibrium inside a static dark-matter potential. The comparison that carries the argument is between the integrated energy budget and the actual coupling: only a small filling factor of hot superbubbles forms, shocks from inner associations compress the central gas, and energy vents along low-density chimneys instead of accelerating the cold gas coherently. A passive metallicity tracer follows the separate fate of the enriched ejecta, showing that it leaves through the same channels while the ambient medium stays bound.

What would settle it

Run the same Boötes I-like galaxy with a clumpy, multiphase initial gas distribution and OB associations that move with the flow, keeping the total injected energy fixed; if more than half of the cold gas leaves the simulation box within 30 Myr, the claim that stellar feedback alone cannot remove the bulk of the gas would be falsified.

Watch

Extended reading notes

Core claim

On the paper's own terms, the central discovery is that stellar feedback in an ultrafaint dwarf is a selective leak, not a wind. With ten OB associations randomly scattered through a smooth, single-phase gas distribution in a dark-matter-dominated halo modeled on Boötes I, supernovae carve hot bubbles and chimneys through which metal-enriched ejecta escapes, while the cold ambient gas remains trapped. In the adiabatic run, less than 20–30 percent of the initial gas mass is lost after 30 Myr; in the run with radiative cooling, the cold gas is even less affected, and the final density profile is nearly indistinguishable from the initial one. Meanwhile, by 30 Myr the adiabatic simulation has lost 70–80 percent of the supernova ejecta, with roughly half the ejecta already gone by 20 Myr in the cooling run. The authors conclude that the simple energetic criterion—enough supernova energy to exceed the gas binding energy—does not guarantee gas removal, because the injected energy is channeled along low-density paths rather than coupled to the bulk of the gas.

Load-bearing premise

The simulations assume the gas begins as a smooth, single-phase, non-rotating cloud in hydrostatic equilibrium, with all star-forming associations fixed in place for the whole 30 Myr; if the real interstellar medium was clumpy or multiphase, or the sources moved, the retained-gas and metal-loss fractions could change.

Editorial extensions

If this is right

The analytic rule of thumb that gas is ejected once cumulative supernova energy exceeds binding energy is not a reliable predictor for ultrafaint dwarfs; simulations must resolve where the energy actually goes.
If stellar feedback cannot remove the cold gas, explaining gas-free ultrafaint dwarfs requires external agents such as ram-pressure stripping, tidal interactions, or reionization to do the work, which the paper identifies as the next step.
Gas loss and metal loss decouple: a dwarf can shed most of its freshly produced metals while retaining most of its original cold gas, so chemical-evolution models that tie the two together need revision.
Even in the adiabatic limit, with radiative cooling switched off and an instantaneous coeval starburst, no galactic-scale outflow develops, so the failure of feedback is not a cooling artifact.

Reading between the lines

Editorial extensions of the paper, not claims the author makes directly.

Inference: a two-stage cleaning—stellar feedback leaks the metals, then an environmental process strips the remaining cold gas—would naturally produce the old, very metal-poor stellar populations of ultrafaint dwarfs, though the paper does not simulate that sequence.
Inference: because the authors use fixed, non-moving OB associations and note that this maximizes feedback effectiveness, allowing the sources to move with the flow would likely spread and dilute the energy deposition and make gas removal even harder.
Inference: a testable extension is to repeat the runs with a clumpy, multiphase initial interstellar medium; the likely outcome is that cold clumps survive while hot ejecta vents even more efficiently, preserving the qualitative conclusion while changing the quoted percentages.
Inference: the 30 Myr window covers one instantaneous burst; a longer, lower-intensity star-formation history consistent with the integrated-galactic-initial-mass-function argument would produce fewer massive stars per unit time, so the feedback failure would likely be more severe, not less.

Share X Bluesky LinkedIn Reddit HN

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, and a circularity audit.

Desk Editor's Note

read the letter

Colleague,

This one is worth reading if you work on ultrafaint dwarfs, but the headline claim overreaches a bit. What the paper does well: it runs very high-resolution (0.95 pc) AMR simulations of a Bootes I-like UFD with thermal feedback from winds and SNe, and shows—robustly, I think—that this feedback does not expel the bulk of the cold gas, even in the adiabatic limit. The separation between retained cold gas and escaping metal-rich ejecta is a genuinely useful quantitative result, and the paper is unusually honest about its limitations: the radiative run stops at 20 Myr, only one random OB association configuration is done at high resolution, and the ISM is smooth and single-phase.

The soft spot is real: the paper only includes thermal energy deposition. It explicitly says radiation pressure is not modeled. Yet the abstract and conclusions say 'stellar feedback alone' cannot remove the gas and that environmental effects are 'unavoidable.' That is a step too far. A quick momentum estimate—L t / c for a 10^5 Msun population over 30 Myr is within a factor of a few of M_gas v_esc—suggests radiation pressure could matter, and photoionization and cosmic rays are also left out. So the paper actually demonstrates that thermal feedback from winds and SNe is insufficient, not that stellar feedback in general is. That needs rewording, and probably some discussion of why radiation pressure would or wouldn't help in a low-metallicity UFD.

Other soft spots: the high-res result is a single realization, so we don't know the stochastic spread; the appendix tests different OB association counts and concentrations at lower resolution, but not the full random distribution at high res. And no code or data are public, so independent checking is limited. These are reproducibility niggles, not signs of error.

In sum: the core result is solid and the paper is a useful addition. It deserves a serious referee, and with a toned-down conclusion it would be acceptable as is. I'd send it to review if I were the editor, but I'd ask the authors to fix the overreach and add a sentence or two on the missing feedback channels.

Recommendation: engage with it, cite it if you work on dwarf feedback, and if you review it, push on the radiation-pressure caveat.

Referee Report

2 major / 5 minor

Summary. This paper presents high-resolution AMR simulations of an isolated, non-rotating Boötes I-like ultrafaint dwarf galaxy, initialized as a smooth, single-phase gas distribution in hydrostatic equilibrium with a static Burkert dark-matter halo. A coeval 10^5 M_sun stellar population, represented by ten static OB associations, injects mass and thermal energy from stellar winds and supernovae for 30 Myr. The authors run one high-resolution adiabatic run to 30 Myr and one radiative run to 20 Myr, plus lower-resolution variants with different numbers and positions of OB associations. They find that less than 20-30 per cent of the initial gas is lost, no complete blow-away occurs, and the bulk of the cold gas remains bound, whereas 70-80 per cent of the SN ejecta is lost from the box by 30 Myr in the adiabatic run. They contrast this with the analytic estimate that about 50 SNe should suffice and conclude that stellar feedback alone cannot remove the gas and that environmental mechanisms are unavoidable.

Significance. If the result holds for the modeled feedback channels, it is a useful counterpoint to simple energy-coupling arguments and supports previous hydrodynamical studies that find SN feedback in low-mass haloes to be inefficient at driving global outflows. The paper's strengths are the ~1 pc resolution in the high-resolution runs, the explicit resolution and configuration checks in Appendix A, the separate tracking of gas and metals, and the direct comparison between a binding-energy estimate and the simulated outcome. However, because the model includes only thermal energy deposition from winds and SNe, the significance for the broader question of 'stellar feedback alone' is limited; radiation pressure, photoionization and cosmic rays are not tested, and the paper's own limitation statements in Sections 4.3-4.4 and Appendix A acknowledge several of the missing ingredients.

major comments (2)

[Abstract; §2; §5] The simulations model stellar feedback only as thermal energy and mass deposition from winds and SNe: §2 states that 'no other mechanism (such as, for instance, radiation pressure) is included,' and §4.4 lists radiation, cosmic rays, and magnetic fields as missing ingredients. The abstract and §5, however, conclude that 'stellar feedback alone' cannot remove the gas and that environmental effects are 'unavoidable.' That inference is not supported by the model. A simple momentum estimate for the adopted 10^5 M_sun population, using L ~ 10^42 erg/s from Leitherer et al. (2014), gives L t / c ~ 3 x 10^46 g cm/s over 30 Myr, comparable to the binding momentum M_gas v_esc ~ 4 x 10^46 g cm/s. Radiation pressure could therefore contribute to unbinding a substantial fraction of the gas, and similar order-of-magnitude considerations apply to photoionization and cosmic rays. I request that the conclusions be restricted to thermal feedback from winds and SNe, or that the model be extended, or that a quantitative argument be given for why the omitted channels are negligible.
[§3 and Appendix A] The quantitative headline numbers (20-30 per cent gas loss and 70-80 per cent ejecta loss by 30 Myr) come from a single high-resolution realization for the fiducial configuration (A-10-04-01-P and R-10-04-01-P in Table A.1), while the lower-resolution suite explores variations in the number and placement of OB associations. The paper itself states that the exact gas-loss fraction depends more on association location than on number, and Fig. A.3 shows appreciable run-to-run spread. I would like the abstract and the §5 bullet points to present these ranges as illustrative of the sampled configurations rather than as robust predictions, or the authors to add at least one additional high-resolution realization to quantify the scatter. The qualitative conclusion that no complete blow-away occurs is supported by the existing suite; this comment concerns the precision of the quantitative claims.

minor comments (5)

[§4.1, Eqs. (3)-(4)] The definition of M0 in Eq. (4) is confusing: M0 is used in Eq. (3) as a mass in units of M_sun, but writing 'M0 = MDM/5.8 M_sun' mixes a dimensionless ratio with the M_sun unit. Please clarify the intended dimensional statement.
[§3.3] There is a typo in the sentence comparing the adiabatic and cooling metallicity profiles: 'predited' should be 'predicted'.
[§4.4 and §5] Several typos occur in these sections: 'feeback' should be 'feedback', 'simultations' should be 'simulations', and the reference 'V orobyov et al. 2015' should have the correct spacing 'Vorobyov et al. 2015'.
[Fig. A.2] The x-axis labels of Fig. A.2 appear incomplete: bins such as 10-100 pc, 200-300 pc, 500-600 pc, and 600-700 pc are not visible in the printed labels. Please ensure the full bin ranges are shown or described in the caption.
[Abstract] The abstract does not mention that the radiative high-resolution run is truncated at 20 Myr rather than 30 Myr; given that the conclusions are drawn partly from the 30 Myr adiabatic run, a brief statement of the radiative run's duration would help readers assess the comparison.

Circularity Check

0 steps flagged · score 1.0 of 10

No significant circularity: the gas-retention result is a genuine dynamical outcome that refutes the paper's own analytic estimate; the self-citations used are non-load-bearing.

full rationale

The paper's central claim — that thermal energy injected by stellar winds and supernovae fails to drive a galactic-scale outflow in a Boötes I-like ultrafaint dwarf even though the total SN energy exceeds the gas binding energy — is a genuine simulation outcome, not a restatement of its inputs. The initial conditions (Plummer gas with a ≈ 200 pc, Burkert dark-matter halo, hydrostatic equilibrium; Section 2) set the binding energy, while the energy budget is set by the adopted 650 SN progenitors and the external Leitherer et al. (2014) injection rates; nothing in these inputs forces the simulated retention of more than 70–80 per cent of the cold gas after 30 Myr in the adiabatic run. Indeed, the paper's own analytic model (Eqs. 1–7) yields E_g ≈ 5×10^52 erg and predicts that roughly 50 SNe suffice for blow-away, so the simulation's failure to produce a global outflow is a falsifiable, physically explained result (spatially inhomogeneous energy injection, venting through chimneys, partial coupling), not a definitional identity. The headline numbers (gas loss <20–30 per cent; SN-ejecta loss 70–80 per cent) are outputs compared in Fig. 6 against the external closed-box expectation from Leitherer et al. (2014). Self-citations to Romano et al. (2015) and Calura et al. (2015) are used only to motivate the Boötes I structural parameters and the OB-association grouping recipe; Appendix A explicitly varies the number, spatial concentration, and resolution of the associations and reports that the conclusion is unchanged, so these citations are not load-bearing. The one caveat is scope, not circularity: 'stellar feedback alone' in the abstract and Section 5 is broader than the modelled physics, since Section 2 states that only thermal energy deposition is included and explicitly excludes radiation pressure, and Section 4.4 acknowledges cosmic rays, magnetic fields, and stellar radiation as unmodelled ingredients; the manuscript itself cautions that 'these results have to be interpreted with caution, always bearing in mind the approximations introduced in the underlying model.' The 'unavoidable' environmental conclusion is therefore an inference conditioned on omitted mechanisms being negligible — a correctness and scope concern, not a circular reduction of the simulation result to its inputs.

Assumptions & free parameters 8 free parameters · 6 assumptions · 0 invented entities

The central claim rests on a highly idealized numerical setup, and no new physical entities are introduced. The main free parameters are initial conditions and feedback injection choices (gas mass, halo mass, Plummer radius, stellar mass, OB association size, temperature floor, OB positions), all chosen from observations or to maximize feedback efficiency rather than fitted to the target result. The analytic blow-away estimate (Eq. 1-7) is an external benchmark, not an input. The key axioms are modeling assumptions about hydrostatic equilibrium, neglect of gas self-gravity, continuous energy injection, cooling prescriptions, and free boundary conditions; the authors explicitly note several limitations in Sections 4.3-4.4 and Appendix A.

free parameters (8)

Initial gas mass Mgas = 6e6 Msun
Set from the cosmic baryon fraction times MDM; sets the gas reservoir and binding energy that feedback must overcome.
Dark matter halo mass MDM = 3.5e7 Msun
Burkert profile normalized to the Bootes I dynamical mass from Wolf et al. 2010; sets the depth of the potential well.
Gas Plummer radius a = ~200 pc
Chosen so gas follows the observed stellar light distribution of Bootes I; controls central density and cooling rates.
Burkert cut-off radius Rc = ~1.2 kpc
Halo truncation radius; affects escape velocity and the amount of mass outside the central region.
Stellar mass Mstars = 1e5 Msun
Coeval population yielding 650 SN progenitors under a Kroupa IMF; scales the total feedback energy input.
OB association radius rOB = 4 pc in high-resolution runs
Sets the initial energy injection volume; only partially satisfies the Kim and Ostriker convergence criterion.
Temperature floor Tmin = 3900 K
Mimics photoionization, photoelectric heating, and UV background heating; affects the cold gas phase and cooling.
OB association positions and N distribution = One realization listed in Table A.1
Positions drawn from a Plummer profile with a~200 pc; the appendix varies number and concentration, but the central high-resolution result is a single random realization.

assumptions (6)

domain assumption Initial gas is in hydrostatic equilibrium in the Burkert plus Plummer potential, with a smooth, single-phase, non-rotating distribution.
Invoked in Section 2 when setting pressure and temperature profiles; real UFD ISM is likely clumpy and multiphase.
domain assumption Gas self-gravity is negligible compared to the static dark matter potential.
Section 2 states that self-gravity is neglected owing to dark matter dominance; valid for MDM/Mgas ~ 6.
domain assumption Continuous injection of mass and energy from stellar winds and SNe over 30 Myr approximates discrete SN explosions in a superbubble.
Justified by the Mac Low and McCray 1988 subsonic blast wave condition, cited in Section 2.
domain assumption The Euler equations with gamma = 5/3 ideal gas and the ramses cooling functions (Sutherland and Dopita 1993; Rosen and Bregman 1995) describe the ISM thermodynamics.
Standard numerical thermodynamics assumptions; adopted in Section 2 without independent verification in this paper.
domain assumption A coeval stellar population with a canonical Kroupa IMF and Leitherer et al. 2014 yields is representative enough for the feedback calculation.
The authors explicitly note that IGIMF theory would give fewer SNe; this choice maximizes feedback (Section 2).
domain assumption Free outflow boundary conditions at the 2 kpc box do not artificially inflate gas or metal loss.
Section 2 and Figures 3-4 check escaping material against local escape velocity, but box size could still affect recycling of material on longer timescales.

how reviews work

0 comments

Cite this review

Pith. "Pith review of High-resolution three-dimensional simulations of gas removal from ultrafaint dwarf galaxies. I. Stellar feedback." pith.science (2026). https://pith.science/paper/DPH5MAFR

@misc{pith2026190902329,
  author       = {Pith},
  title        = {Pith review of: High-resolution three-dimensional simulations of gas removal from ultrafaint dwarf galaxies. I. Stellar feedback},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/DPH5MAFR}},
  note         = {Machine review of arXiv:1909.02329}
}

read the original abstract

The faintest Local Group galaxies found lurking in and around the Milky Way halo provide a unique test bed for theories of structure formation and evolution on small scales. Deep Subaru and Hubble Space Telescope photometry demonstrates that their stellar populations are old, and that the star formation activity did not last longer than 2 Gyr in these systems. A few mechanisms that may lead to such a rapid quenching have been investigated by means of hydrodynamic simulations, without providing any final assessment so far. This is the first in a series of papers aimed at analysing the roles of stellar feedback, ram pressure stripping, host-satellite tidal interactions and reionization in cleaning the lowest-mass Milky Way companions of their cold gas, by using high-resolution, three-dimensional hydrodynamic simulations. We simulate an isolated ultrafaint dwarf galaxy loosely modeled after Bootes I, and examine whether or not stellar feedback alone could drive a substantial fraction of the ambient gas out from the shallow potential well. In contrast to simple analytical estimates, but in agreement with previous hydrodynamical studies, we find that most of the cold gas reservoir is retained. Conversely, a significant fraction of the metal-enriched stellar ejecta crosses the boundaries of the computational box with velocities exceeding the local escape velocity and is, thus, likely lost from the system. Although the total energy output from multiple supernova explosions exceeds the binding energy of the gas, no galactic-scale outflow develops in our simulations and as such, most of the ambient medium remains trapped within the weak potential well of the model galaxy. It seems thus unavoidable that, in order to explain the dearth of gas in ultrafaint dwarf galaxies, we will have to resort to environmental effects. This will be the subject of a forthcoming paper.

Figures

Figures reproduced from arXiv: 1909.02329 by the authors.

**Figure 1.** Gas density (left), temperature (middle) and metallicity (right) maps for the Boötes I-like galaxy, in the z = 0 plane, at five representative times, t = 3, 13, 20, 25, and 30 Myr. The projected positions of the OB associations are displayed on the metallicity maps (grey circles, with sizes proportional to the number of massive stars). The snapshots refer to the galaxy simulated in the adiabatic limit at high resolu… view at source ↗

**Figure 2.** Same as [PITH_FULL_IMAGE:figures/full_fig_p005_2.png] view at source ↗

**Figure 3.** Gas radial velocity field, in the z = 0 plane, at t = 20 Myr. The maps on the left refer to the galaxy simulated in the adiabatic limit, while the maps on the right are for the run with radiative cooling. Cells coloured black in the inset maps on the bottom left of each panel highlight regions where the gas moves with velocities lower than the local escape velocity. -500 0 500 -500 0 500 y [pc] x [pc] + ADIAB + t= 3… view at source ↗

Figures from the paper (6 more)

**Figure 4.** Figure 4: Same as [PITH_FULL_IMAGE:figures/full_fig_p006_4.png]

**Figure 5.** Figure 5: Mass-weighted distributions of gas radial velocity (upper panels) and density (middle panels) at t = 20 Myr, for the runs with (right-hand panels) and without radiative losses (left-hand panels). Also shown is the distribution of metals in radial velocity bins at the s…

**Figure 7.** Figure 7: Initial (black line) and final density profiles, for the adiabatic simulation (t = 30 Myr; thin red line) and for the run with radiative cooling (t = 20 Myr; thick green line). Also shown is the density profile for the adiabatic simulation at t = 20 Myr (thick red line…

**Figure 8.** Figure 8: Same as [PITH_FULL_IMAGE:figures/full_fig_p008_8.png]

**Figure 9.** Figure 9: Metallicity profiles, for the Boötes I-like UFD simulated in the adiabatic limit (red lines) and with radiative cooling (green line). Thick lines are for t = 20 Myr, the thin one is for t = 30 Myr [PITH_FULL_IMAGE:figures/full_fig_p009_9.png]

**Figure 10.** Figure 10: Same as [PITH_FULL_IMAGE:figures/full_fig_p009_10.png]

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

85 extracted references · 57 canonical work pages

[1]

, " * write output.state after.block = add.period write newline

ENTRY address archiveprefix author booktitle chapter edition editor howpublished institution eprint journal key month note number organization pages publisher school series title type volume year label extra.label sort.label short.list INTEGERS output.state before.all mid.sentence after.sentence after.block FUNCTION init.state.consts #0 'before.all := #1 ...
[2]

write newline

" write newline "" before.all 'output.state := FUNCTION n.dashify 't := "" t empty not t #1 #1 substring "-" = t #1 #2 substring "--" = not "--" * t #2 global.max substring 't := t #1 #1 substring "-" = "-" * t #2 global.max substring 't := while if t #1 #1 substring * t #2 global.max substring 't := if while FUNCTION word.in bbl.in " " * FUNCTION format....
[3]

L., & Norman , M

Abel , T., Bryan , G. L., & Norman , M. L. 2000, , 540, 39

2000
[4]

V., Leitner , S

Agertz , O., Kravtsov , A. V., Leitner , S. N., & Gnedin , N. Y. 2013, , 770, 25

2013
[5]

2018, , 70, S4

Aihara , H., Arimoto , N., Armstrong , R., et al. 2018, , 70, S4

2018
[6]

2008, , 384, 1231

Baumgardt , H., Kroupa , P., & Parmentier , G. 2008, , 384, 1231

work page 2008
[7]

2015, , 807, 50

Bechtol , K., Drlica-Wagner , A., Balbinot , E., et al. 2015, , 807, 50

work page 2015
[8]

2018, , 476, 4565

Bellazzini , M., Armillotta , L., Perina , S., et al. 2018, , 476, 4565

work page 2018

Show all 85 references

[9]

J., Koposov , S

Belokurov , V., Irwin , M. J., Koposov , S. E., et al. 2014, , 441, 2124

2014
[10]

B., Evans , N

Belokurov , V., Zucker , D. B., Evans , N. W., et al. 2007, , 654, 897

2007
[11]

B., Evans , N

Belokurov , V., Zucker , D. B., Evans , N. W., et al. 2006, , 647, L111

2006
[12]

2015, , 807, 154

Bland-Hawthorn , J., Sutherland , R., & Webster , D. 2015, , 807, 154

2015
[13]

M., Tumlinson , J., & Geha , M

Brown , T. M., Tumlinson , J., & Geha , M. e. a. 2014, , 796, 91

2014
[14]

1995, , 447, L25

Burkert , A. 1995, , 447, L25

1995
[15]

G., Romano , D., & D'Ercole , A

Calura , F., Few , C. G., Romano , D., & D'Ercole , A. 2015, , 814, L14

2015
[16]

A., da Silva , A

Caproni , A., Lanfranchi , G. A., da Silva , A. L., & Falceta-Gon c alves , D. 2015, , 805, 109

2015
[17]

C., Magnier , E

Chambers , K. C., Magnier , E. A., Metcalfe , N., et al. 2016, ArXiv e-prints [ [arXiv] 1612.05560 ]

2016 arXiv
[18]

V., & Wise , J

Corlies , L., Johnston , K. V., & Wise , J. H. 2018, , 475, 4868

2018
[19]

de los Reyes , M. A. C. & Kennicutt , Robert C., J. 2019, , 872, 16

2019
[20]

2016, , 460, 1270

DES Collaboration . 2016, , 460, 1270

2016
[21]

Dopita , M. A. & Smith , G. H. 1986, , 304, 283

1986
[22]

L., & Mac Low , M.-M

Emerick , A., Bryan , G. L., & Mac Low , M.-M. 2019, , 482, 1304

2019
[23]

2016, , 826, 148

Emerick , A., Mac Low , M.-M., Grcevich , J., & Gatto , A. 2016, , 826, 148

2016
[24]

2018, Nature Astronomy

Faucher-Gigu \`e re , C.-A. 2018, Nature Astronomy

2018
[25]

Ferguson , H. C. & Binggeli , B. 1994, , 6, 67

1994
[26]

2016, , 832, 21

Homma , D., Chiba , M., Okamoto , S., et al. 2016, , 832, 21

2016
[27]

2018, , 70, S18

Homma , D., Chiba , M., Okamoto , S., et al. 2018, , 70, S18

2018
[28]

2017, , 848, 85

Jeon , M., Besla , G., & Bromm , V. 2017, , 848, 85

2017
[29]

2018, , 620, A39

Je r \'a bkov \'a , T., Hasani Zonoozi , A., Kroupa , P., et al. 2018, , 620, A39

2018
[30]

1992, , 391, 502

Katz , N. 1992, , 391, 502

1992
[31]

& Ostriker , E

Kim , C.-G. & Ostriker , E. C. 2015, , 802, 99

2015
[32]

V., & Khokhlov , A

Klypin , A., Gottl \"o ber , S., Kravtsov , A. V., & Khokhlov , A. M. 1999, , 516, 530

1999
[33]

R., et al

Komatsu , E., Dunkley , J., Nolta , M. R., et al. 2009, , 180, 330

2009
[34]

Krause , M. G. H., Charbonnel , C., Bastian , N., & Diehl , R. 2016, , 587, A53

2016
[35]

2001, , 322, 231

Kroupa , P. 2001, , 322, 231

2001
[36]

Lada , C. J. & Lada , E. A. 2003, , 41, 57

2003
[37]

Laevens , B. P. M., Martin , N. F., Bernard , E. J., et al. 2015, , 813, 44

2015
[38]

Larson , R. B. 1974, , 169, 229

1974
[39]

2014, , 212, 14

Leitherer , C., Ekstr \"o m , S., Meynet , G., et al. 2014, , 212, 14

2014
[40]

& Ferrara , A

Mac Low , M.-M. & Ferrara , A. 1999, , 513, 142

1999
[41]

& McCray , R

Mac Low , M.-M. & McCray , R. 1988, , 324, 776

1988
[42]

2006, , 371, 643

Marcolini , A., D'Ercole , A., Brighenti , F., & Recchi , S. 2006, , 371, 643

2006
[43]

2015, , 450, 504

Martizzi , D., Faucher-Gigu \`e re , C.-A., & Quataert , E. 2015, , 450, 504

2015
[44]

McKee , C. F. & Williams , J. P. 1997, , 476, 144

1997
[45]

1999, , 524, L19

Moore , B., Ghigna , S., Governato , F., et al. 1999, , 524, L19

1999
[46]

2002, , 571, 40

Mori , M., Ferrara , A., & Madau , P. 2002, , 571, 40

2002
[47]

Ni , Y., Di Matteo , T., Feng , Y., Croft , R. A. C., & Tenneti , A. 2018, , 481, 4877

2018
[48]

2012, , 744, 96

Okamoto , S., Arimoto , N., Yamada , Y., & Onodera , M. 2012, , 744, 96

2012
[49]

Plummer , H. C. 1911, , 71, 460

1911
[50]

I., Agertz , O., & Collins , M

Read , J. I., Agertz , O., & Collins , M. L. M. 2016, , 459, 2573

2016
[51]

& Hensler , G

Recchi , S. & Hensler , G. 2013, , 551, A41

2013
[52]

Roberts , M. S. 1957, , 69, 59

1957
[53]

2017, , 835, 136

Robles-Valdez , F., Rodr \' guez-Gonz \'a lez , A., Hern \'a ndez-Mart \' nez , L., & Esquivel , A. 2017, , 835, 136

2017
[54]

2015, , 446, 4220

Romano , D., Bellazzini , M., Starkenburg , E., & Leaman , R. 2015, , 446, 4220

2015
[55]

& Bregman , J

Rosen , A. & Bregman , J. N. 1995, , 440, 634

1995
[56]

& Ferrara , A

Salvadori , S. & Ferrara , A. 2009, , 395, L6

2009
[57]

S., Fattahi , A., et al

Sawala , T., Frenk , C. S., Fattahi , A., et al. 2016, , 456, 85

2016
[58]

& Br \"u ggen , M

Scannapieco , E. & Br \"u ggen , M. 2010, , 405, 1634

2010
[59]

1959, , 129, 243

Schmidt , M. 1959, , 129, 243

1959
[60]

2015, , 451, 4238

Shanks , T., Metcalfe , N., Chehade , B., et al. 2015, , 451, 4238

2015
[61]

& Tenorio-Tagle , G

Silich , S. & Tenorio-Tagle , G. 2001, , 552, 91

2001
[62]

Simon , J. D. 2019, arXiv e-prints [ [arXiv] 1901.05465 ]

2019 arXiv
[63]

M., Bryan , G

Simpson , C. M., Bryan , G. L., Hummels , C., & Ostriker , J. P. 2015, , 809, 69

2015
[64]

Springel , V., Yoshida , N., & White , S. D. M. 2001, , 6, 79

2001
[65]

Sutherland , R. S. 2010, , 327, 173

2010
[66]

Sutherland , R. S. & Dopita , M. A. 1993, , 88, 253

1993
[67]

2002, , 385, 337

Teyssier , R. 2002, , 385, 337

2002
[68]

Teyssier , R., Pontzen , A., Dubois , Y., & Read , J. I. 2013, , 429, 3068

2013
[69]

Thacker , R. J. & Couchman , H. M. P. 2000, , 545, 728

2000
[70]

J., Bullock , J

Tollerud , E. J., Bullock , J. S., Strigari , L. E., & Willman , B. 2008, , 688, 277

2008
[71]

E., Belokurov , V., et al

Torrealba , G., Koposov , S. E., Belokurov , V., et al. 2016, , 463, 712

2016
[72]

K., Broos , P

Townsley , L. K., Broos , P. S., Feigelson , E. D., et al. 2006, , 131, 2140

2006
[73]

Vader , J. P. 1986, , 305, 669

1986
[74]

Vincenzo , F., Matteucci , F., Vattakunnel , S., & Lanfranchi , G. A. 2014, , 441, 2815

2014
[75]

I., Recchi , S., & Hensler , G

Vorobyov , E. I., Recchi , S., & Hensler , G. 2015, , 579, A9

2015
[76]

2015, , 799, L21

Webster , D., Bland-Hawthorn , J., & Sutherland , R. 2015, , 799, L21

2015
[77]

F., Pace , A

Wheeler , C., Hopkins , P. F., Pace , A. B., et al. 2018, arXiv e-prints [ [arXiv] 1812.02749 ]

2018 arXiv
[78]

S., et al

Wheeler , C., O \ n orbe , J., Bullock , J. S., et al. 2015, , 453, 1305

2015
[79]

White , S. D. M. & Rees , M. J. 1978, , 183, 341

1978
[80]

R., West , A

Willman , B., Blanton , M. R., West , A. A., et al. 2005, , 129, 2692

2005
[81]

H., Abel , T., Turk , M

Wise , J. H., Abel , T., Turk , M. J., Norman , M. L., & Smith , B. D. 2012, , 427, 311

2012
[82]

D., Bullock , J

Wolf , J., Martinez , G. D., Bullock , J. S., et al. 2010, , 406, 1220

2010
[83]

Yadav , N., Mukherjee , D., Sharma , P., & Nath , B. B. 2017, , 465, 1720

2017
[84]

2017, , 607, A126

Yan , Z., Jerabkova , T., & Kroupa , P. 2017, , 607, A126

2017
[85]

G., Adelman , J., Anderson , Jr., J

York , D. G., Adelman , J., Anderson , Jr., J. E., et al. 2000, , 120, 1579

2000

Pith tools

Reviewed August 14, 2026 · model on record in the stance chip above.

[1] [1]

, " * write output.state after.block = add.period write newline

ENTRY address archiveprefix author booktitle chapter edition editor howpublished institution eprint journal key month note number organization pages publisher school series title type volume year label extra.label sort.label short.list INTEGERS output.state before.all mid.sentence after.sentence after.block FUNCTION init.state.consts #0 'before.all := #1 ...

[2] [2]

write newline

" write newline "" before.all 'output.state := FUNCTION n.dashify 't := "" t empty not t #1 #1 substring "-" = t #1 #2 substring "--" = not "--" * t #2 global.max substring 't := t #1 #1 substring "-" = "-" * t #2 global.max substring 't := while if t #1 #1 substring * t #2 global.max substring 't := if while FUNCTION word.in bbl.in " " * FUNCTION format....

[3] [3]

L., & Norman , M

Abel , T., Bryan , G. L., & Norman , M. L. 2000, , 540, 39

2000

[4] [4]

V., Leitner , S

Agertz , O., Kravtsov , A. V., Leitner , S. N., & Gnedin , N. Y. 2013, , 770, 25

2013

[5] [5]

2018, , 70, S4

Aihara , H., Arimoto , N., Armstrong , R., et al. 2018, , 70, S4

2018

[6] [6]

2008, , 384, 1231

Baumgardt , H., Kroupa , P., & Parmentier , G. 2008, , 384, 1231

work page 2008

[7] [7]

2015, , 807, 50

Bechtol , K., Drlica-Wagner , A., Balbinot , E., et al. 2015, , 807, 50

work page 2015

[8] [8]

2018, , 476, 4565

Bellazzini , M., Armillotta , L., Perina , S., et al. 2018, , 476, 4565

work page 2018