pith. sign in

arxiv: 1906.10841 · v1 · pith:B6RMWSJFnew · submitted 2019-06-26 · 🌌 astro-ph.GA · astro-ph.SR

Star formation within globular clusters:discrete multiple bursts and top-light mass functions

Kenji Bekki This is my paper

Pith reviewed 2026-05-25 16:00 UTC · model grok-4.3

classification 🌌 astro-ph.GA astro-ph.SR
keywords globular clustersstar formationmultiple stellar populationsinitial mass functionnumerical simulationsstar-gas interactionasymptotic giant branch stars
0
0 comments X

The pith

Star formation in globular clusters occurs in short bursts only above a gas mass fraction threshold, enabling discrete populations and top-light mass functions.

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

The paper uses simulations of star-gas interactions to test if bursty star formation can happen inside globular clusters. Dense clouds form, producing high-efficiency bursts only when gas fraction tops a mass- and disk-dependent threshold. Below threshold, formation stops and massive stars are suppressed, creating top-light mass functions. This internal process can account for observed discrete populations and abundance patterns. Readers care as it links cluster dynamics directly to star-formation history without external gas sources.

Core claim

Small gas clouds denser than 10^10 atoms per cubic cm form from non-turbulent gas, allowing bursty star formation with efficiency over 0.5 when gas fraction f_g exceeds threshold f_g,th. This threshold rises for lower-mass clusters and larger gas disks. Star-gas interactions plus cluster gravity suppress massive stars, yielding top-light IMFs. Massive clusters with low thresholds can form He-rich stars from AGB gas. Short bursts above threshold partly explain discrete multiple populations in GCs.

What carries the argument

The gas mass fraction threshold f_g,th within GCs, set by cluster mass and gas disk size, which determines whether star-gas interactions permit or suppress bursty high-efficiency star formation and massive star production.

If this is right

  • Discrete multiple stellar populations arise from repeated short bursts separated by intervals when the gas fraction falls below threshold.
  • Globular clusters develop top-light initial mass functions because star-gas interactions and cluster gravity suppress massive star formation.
  • He-rich stars form directly from AGB star ejecta in massive clusters that have sufficiently low f_g,th.
  • Star formation efficiency exceeds 0.5 during each burst once the gas fraction surpasses the threshold.

Where Pith is reading between the lines

These are editorial extensions of the paper, not claims the author makes directly.

  • If correct, more massive clusters should exhibit more discrete populations because their lower thresholds allow repeated bursts.
  • The model predicts that globular cluster initial mass functions differ systematically from the field, testable via luminosity functions or remnant populations.
  • External gas accretion could temporarily raise the gas fraction above threshold and restart the burst cycle.

Load-bearing premise

The numerical simulations correctly capture the development of small dense gas clouds from gas without turbulence and the resulting high star formation efficiencies when the gas fraction exceeds the threshold.

What would settle it

Detection of recent star formation in a globular cluster whose measured gas mass fraction lies well below the model's predicted threshold for its mass and disk size, or direct evidence of a standard rather than top-light initial mass function in clusters where suppression is expected.

Figures

Figures reproduced from arXiv: 1906.10841 by Kenji Bekki.

Figure 1
Figure 1. Figure 1: Illustration of the AGB scenario for the origins of discrete multiple stellar populations (e.g., 1G means the first generation of stars) and internal abundance spreads for various elements (e.g., s- and r-process). Time evolution of the total gas mass (Mg), the total mass of new stars formed from gas (Ms), the upper-mass cut-off of the IMF (mu), and the retention probability of gas that is consumed by seco… view at source ↗
Figure 2
Figure 2. Figure 2: Time evolution of the surface mass density (Σ in logarithmic scale) of gas projected onto the x-y plane for the fiducial model MN2 without star formation. The total gas mass within the GC is 3 × 103M⊙, which means that the gas mass fraction (fg) is 0.01. Time (T) that has elapsed since the start of this simulation is shown in the upper left corner of each frame. The scale bar of 0.1 pc is shown in the lowe… view at source ↗
Figure 3
Figure 3. Figure 3: The same as Fig.2 but for the x-z projection. B15). Since our main focus is not the evolution of metals and dust in galaxies and star clusters, we “switch off” the components of the code that are relevant to evolution of met￾als and dust. The code combines the method of smoothed particle hydrodynamics (SPH) with calculations of three￾dimensional self-gravitating fluids in astrophysics. Since the details of… view at source ↗
Figure 4
Figure 4. Figure 4: Distribution of gas densities (ρg) for the fiducial model MN2 without star formation at three different time steps, T = 0.81 Myr, 1.24 Myr, and 2.35 Myr. Normalized number of gas particles is given for each density bin. 1G stars and suggested that these are consistent with a for￾mation scenario in which 2G stars originate from a com￾pact configuration with strong rotation (e.g., Mastrobuono￾Battisti & Pere… view at source ↗
Figure 5
Figure 5. Figure 5: The same as [PITH_FULL_IMAGE:figures/full_fig_p009_5.png] view at source ↗
Figure 6
Figure 6. Figure 6: The same as [PITH_FULL_IMAGE:figures/full_fig_p010_6.png] view at source ↗
Figure 7
Figure 7. Figure 7: Time evolution of star formation rates for six different models with Mgc = 3.1 × 105M⊙. The initial GC size (Rgc) is the same between the models (Rgc = 10pc) except the model shown in the lower right (6.9pc). The initial gas mass is shown in the upper left corner of each panel [PITH_FULL_IMAGE:figures/full_fig_p011_7.png] view at source ↗
Figure 8
Figure 8. Figure 8: The same as [PITH_FULL_IMAGE:figures/full_fig_p011_8.png] view at source ↗
Figure 9
Figure 9. Figure 9: The same as [PITH_FULL_IMAGE:figures/full_fig_p012_9.png] view at source ↗
Figure 10
Figure 10. Figure 10: The same as [PITH_FULL_IMAGE:figures/full_fig_p012_10.png] view at source ↗
Figure 11
Figure 11. Figure 11: Number distributions (Nc) of gas cloud masses (mc) for the models with Mgc = 3.1 × 105M⊙ and fg = 0.01 (black solid line) at T = 0.47 Myr. For comparison, the same models yet without SB (dotted) and without stars and compact stellar objects (dashed) are shown. The model labeled as “W/O stars” accordingly means that there is no global gravitational field of a GC. For convenience, log(Nc + 1) is shown in th… view at source ↗
Figure 12
Figure 12. Figure 12: The same as [PITH_FULL_IMAGE:figures/full_fig_p014_12.png] view at source ↗
read the original abstract

The observed discrete multiple stellar populations and internal abundance spreads in r- and s-process elements within globular clusters (GCs) have been suggested to be explained self-consistently by discrete star formation events over a longer timescale (10^8 yr). We here investigate whether such star formation is really possible within GCs using numerical simulations that include effects of dynamical interaction between individual stars and the accumulated gas ("star-gas interaction") on star formation. The principal results are as follows. Small gas clouds with densities larger than $10^{10}$ atoms cm^{-3} corresponding to first stellar cores can be developed from gas without turbulence. Consequently, new stars can be formed from the gas with high star formation efficiencies (>0.5) in a bursty manner. However, star formation can be suppressed when the gas mass fractions within GCs (f_g) are less than a threshold value (f_g, th). This f_g, th is larger for GCs with lower masses and larger gas disks. Star-gas interaction and gravitational potentials of GCs can combine to suppress the formation of massive stars (i.e., "top-light" stellar initial mass function). Formation of He-rich stars directly from gas of massive AGB stars is possible in massive GCs due to low f_g, th (<0.01). Short bursty star formation only for f_g>f_g, th can be partly responsible for discrete multiple star formation events within GCs.

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, simulated authors' rebuttal, and a circularity audit. Tearing a paper down is the easy half of reading it; the pith above is the substance, this is the friction.

Referee Report

3 major / 2 minor

Summary. The paper reports N-body/hydrodynamical simulations of star formation inside globular clusters that include star-gas dynamical coupling. It claims that dense cores (>10^{10} atoms cm^{-3}) form even from initially non-turbulent gas, producing bursty star formation with SFE > 0.5 only when the gas mass fraction exceeds a threshold f_g,th (higher for lower-mass clusters and larger gas disks). Star-gas interactions plus the cluster potential are said to suppress massive-star formation (top-light IMF), while He-rich stars can form directly from AGB ejecta in massive clusters. These internal processes are proposed to explain observed discrete multiple populations.

Significance. If the numerical outcomes are robust, the work supplies a dynamical channel for multiple discrete star-formation episodes driven solely by internal gas dynamics and star-gas drag, without external accretion. The emergence of f_g,th as a simulation output rather than an imposed parameter is a methodological strength.

major comments (3)
  1. [numerical-methods section] Abstract and numerical-methods section: the central claim that small dense cores with n > 10^{10} cm^{-3} and SFE > 0.5 arise from turbulence-free gas is load-bearing for the bursty-SF and multiple-population conclusions, yet the manuscript supplies no description of the hydro solver, artificial viscosity, cooling function, softening length, or initial density profile.
  2. [results section] Results on f_g,th: the threshold is presented as an emergent output, but no resolution study, variation of feedback implementation, or convergence test with respect to particle number or time-stepping is reported; without these the dependence of f_g,th on cluster mass and disk size cannot be assessed as physical rather than numerical.
  3. [results section] IMF-suppression claim: the assertion that star-gas interaction plus the cluster potential produces a top-light IMF inherits the same numerical uncertainties; the manuscript does not specify how stellar masses are assigned or sampled in the runs, nor does it show quantitative IMF slopes or comparisons to observed GC IMFs.
minor comments (2)
  1. [abstract] Notation: f_g,th is used in the abstract before being defined; a brief parenthetical definition on first appearance would improve readability.
  2. The manuscript would benefit from a short table summarizing the suite of runs (cluster mass, f_g, disk size, outcome) to make the threshold dependence explicit.

Simulated Author's Rebuttal

3 responses · 0 unresolved

We thank the referee for the thorough and constructive report. The comments correctly identify several areas where the manuscript lacks necessary detail on numerical methods and supporting tests. We will revise the paper to address these points by expanding the methods description, adding convergence studies, and providing quantitative IMF analysis. Our point-by-point responses follow.

read point-by-point responses

  1. Referee: [numerical-methods section] Abstract and numerical-methods section: the central claim that small dense cores with n > 10^{10} cm^{-3} and SFE > 0.5 arise from turbulence-free gas is load-bearing for the bursty-SF and multiple-population conclusions, yet the manuscript supplies no description of the hydro solver, artificial viscosity, cooling function, softening length, or initial density profile.

    Authors: We agree that the numerical-methods section is incomplete and that these details are essential for assessing the formation of dense cores. In the revised manuscript we will add a dedicated subsection describing the hydrodynamical solver, artificial viscosity prescription, cooling function, gravitational softening lengths, and the functional form of the initial gas density profiles. These parameters were part of the simulation setup but were inadvertently omitted from the submitted text. revision: yes

  2. Referee: [results section] Results on f_g,th: the threshold is presented as an emergent output, but no resolution study, variation of feedback implementation, or convergence test with respect to particle number or time-stepping is reported; without these the dependence of f_g,th on cluster mass and disk size cannot be assessed as physical rather than numerical.

    Authors: The referee is correct that no resolution or convergence tests are shown. We will perform and report a resolution study (varying particle number by factors of 4 and 8) together with tests of time-stepping criteria and, where relevant, feedback implementation. A new subsection will demonstrate that the reported dependence of f_g,th on cluster mass and disk size remains stable under these variations, thereby supporting its physical origin. revision: yes

  3. Referee: [results section] IMF-suppression claim: the assertion that star-gas interaction plus the cluster potential produces a top-light IMF inherits the same numerical uncertainties; the manuscript does not specify how stellar masses are assigned or sampled in the runs, nor does it show quantitative IMF slopes or comparisons to observed GC IMFs.

    Authors: We acknowledge that the IMF analysis is presented only qualitatively. In the revision we will (i) state explicitly how stellar masses are assigned or sampled during the runs, (ii) include binned IMF histograms with fitted power-law slopes for the different cluster models, and (iii) overlay these slopes against literature values for observed globular-cluster IMFs. These additions will quantify the claimed top-light character and allow direct comparison with observations. revision: yes

Circularity Check

0 steps flagged

No circularity: results from forward numerical integration of hydrodynamics and gravity

full rationale

The paper's claims (bursty SF above f_g,th, dense cores >10^10 cm^{-3}, top-light IMF, He-rich stars) are outputs of N-body/hydro simulations that include star-gas drag and potentials. No load-bearing step reduces by construction to a fitted parameter, self-definition, or self-citation chain; thresholds and efficiencies are simulation results, not inputs renamed as predictions. The derivation is self-contained against external benchmarks and does not invoke uniqueness theorems or ansatzes from prior author work.

Axiom & Free-Parameter Ledger

2 free parameters · 2 axioms · 0 invented entities

The central claim rests on outcomes of hydrodynamical simulations incorporating star-gas interactions. Free parameters are simulation-derived thresholds rather than ad-hoc fits. No new physical entities postulated; relies on standard gravitational dynamics and gas cooling assumptions.

free parameters (2)
  • f_g,th
    Gas mass fraction threshold below which star formation is suppressed; emerges from simulations and varies with GC mass and gas disk size.
  • density threshold
    10^10 atoms cm^{-3} for identifying first stellar cores; used to trigger star formation in the model.
axioms (2)
  • domain assumption Small gas clouds with densities >10^10 atoms cm^{-3} can develop from gas without turbulence.
    Invoked to enable bursty star formation in the simulations.
  • domain assumption Star formation efficiencies exceed 0.5 in a bursty manner when conditions are met.
    Stated as a principal result of the gas dynamics modeled.

pith-pipeline@v0.9.0 · 5789 in / 1506 out tokens · 48548 ms · 2026-05-25T16:00:42.478956+00:00 · methodology

discussion (0)

Sign in with ORCID, Apple, or X to comment. Anyone can read and Pith papers without signing in.

Reference graph

Works this paper leans on

77 extracted references · 77 canonical work pages · 3 internal anchors

  1. [1]

    Adamo, A., et al., 2012, MNRAS, 426, 1185

    work page 2012

  2. [2]

    Armstrong, B., For, B.-Q., Bekki, K., 2018, MNRAS, 481, 3651

    work page 2018

  3. [3]

    M., Zepf, S

    Ashman, K. M., Zepf, S. E., 2001, AJ, 122, 1888

    work page 2001

  4. [4]

    Balin, J., 2018, ApJ, 863, 99 (B18)

    work page 2018

  5. [5]

    Bastian, N., Lardo, C., 2018, ARA&A, 56, 83

    work page 2018

  6. [6]

    R., Burkert, A., 1998, MNRAS, 288, 1060

    Bate, M. R., Burkert, A., 1998, MNRAS, 288, 1060

    work page 1998

  7. [7]

    Bekki, K., 2006, MNRAS, 367L, 24

    work page 2006

  8. [8]

    Bekki, K., 2010, ApJ, 724, L99 (B10)

    work page 2010

  9. [9]

    Bekki, K., 2011, MNRAS, MNRAS, 412, 2241 (B11)

    work page 2011

  10. [10]

    Bekki, K., 2013, 432, 2298 (B13)

    work page 2013

  11. [11]

    Bekki, K., 2015, MNRAS, 449, 1625 (B15)

    work page 2015

  12. [12]

    Bekki, K., 2018 A&A in press (arXiv:1807.02309)

    work page internal anchor Pith review Pith/arXiv arXiv 2018

  13. [13]

    E., 2006, ApJ, 637, L109

    Bekki, K., Norris, J. E., 2006, ApJ, 637, L109

    work page 2006

  14. [14]

    W.; Lattanzio, J

    Bekki, K.; Campbell, S. W.; Lattanzio, J. C.; Norris, J. E., 2007, MNRAS, 377, 335

    work page 2007

  15. [15]

    Bekki, K., Chiba, M., 2007, ApJ, 665, 1164

    work page 2007

  16. [16]

    D., 2009, MNRAS, 394, 124

    Bekki, K., Mackey, A. D., 2009, MNRAS, 394, 124

    work page 2009

  17. [17]

    Bekki, K., Yong, D., 2012, MNRAS, 419, 2063

    work page 2012

  18. [18]

    Bekki, K., Tsujimoto, T., 2016, ApJ, 831, 70

    work page 2016

  19. [19]

    Bekki, K., Tsujimoto, T., 2017, ApJ, 844, 34 (BT17)

    work page 2017

  20. [20]

    Bekki, K., Jerabkova, T., Kroupa, P., 2017, MNRAS, 471, 2242 (BJK17)

    work page 2017

  21. [21]

    Bianchini, P., et al., 2018, MNRAS, 481, 2125

    work page 2018

  22. [22]

    Binney, J., Tremaine, S., 1987 in Galactic Dynamics

    work page 1987

  23. [23]

    Bruzual, G., Schweizer, F., 2014, MNRAS, 441, 2754

    work page 2014

  24. [24]

    G., Lucatello, S., 2009, A&A, 505, 117

    Carretta, E., Bragaglia, A., Gratton, R. G., Lucatello, S., 2009, A&A, 505, 117

    work page 2009

  25. [25]

    Carretta, E., et al., 2014, ApJL, 795, 28

    work page 2014

  26. [26]

    Carretta, E., et al., 2015, ApJ, 810, 148

    work page 2015

  27. [27]

    G., Lucatello, S., D’Orazi, V., 2012, ApJ, 750, L14

    Carretta, E., Bragaglia, A., Gratton, R. G., Lucatello, S., D’Orazi, V., 2012, ApJ, 750, L14

    work page 2012

  28. [28]

    Carretta, E., et al., 2018, A&A, 615, 17

    work page 2018

  29. [29]

    N., 2011, ApJ, 726, 36 (CS11) Da Costa, G

    Conroy, C., Spergel, D. N., 2011, ApJ, 726, 36 (CS11) Da Costa, G. S., Held, E. V., Saviane, I., Gullieuszik, M., 2009, ApJ, 705, 1481

    work page 2011

  30. [30]

    E., Ngoumou, J., Ercolano, B., Bonnell, I

    Dale, J. E., Ngoumou, J., Ercolano, B., Bonnell, I. A, 2014, MNRAS, 442, 694 D’Antona, F., Caloi, V., Montalban, J., Ventura, P., Grat- ton, R., 2002, A&A, 395, 69 D’Antona, F., Vesperini, E., D’Ercole, A., Ventura, P.,

    work page 2014

  31. [31]

    P., Marino, A

    Milone, A. P., Marino, A. F., Tailo, M., 2016, MNRAS, 458, 2122 (D16) D’Ercole, A., Vesperini, E., D’Antona, F., McMillan, S. L. W., & Recchi, S. 2008, MNRAS, 391, 825 (D08) D’Ercole, A., D’Antona, F., Vesperini, E., 2016, MNRAS, 461, 4088 (DDV16)

    work page 2016

  32. [32]

    1997, ApJL, 474, 19 Dominik et al

    Parmeggiani, G. 1997, ApJL, 474, 19 Dominik et al. 2012, ApJ, 759, 52

    work page 1997

  33. [33]

    G., 2017, ApJ, 836, 80

    Elmegreen, B. G., 2017, ApJ, 836, 80

    work page 2017

  34. [34]

    For, B.Q., Bekki, K., 2017, MNRAS, 468, L11

    work page 2017

  35. [35]

    2015, ApJL, 807, 4

    Fukui, Y., et al. 2015, ApJL, 807, 4

    work page 2015

  36. [36]

    2017, PASJ, L69, 5

    Fukui, Y., et al. 2017, PASJ, L69, 5

    work page 2017

  37. [37]

    2018, PASJ, 70, 60

    Fukui, Y., et al. 2018, PASJ, 70, 60

    work page 2018

  38. [38]

    Goldreich, P., Lynden-Bell, D., 1965, MNRAS, 130, 97

    work page 1965

  39. [39]

    Goudfrooij, P., et al., 2014, ApJ, 797, 35

    work page 2014

  40. [40]

    G., Carretta, E., Bragaglia, A., 2012, A&ARv, 20, 50

    Gratton, R. G., Carretta, E., Bragaglia, A., 2012, A&ARv, 20, 50

    work page 2012

  41. [41]

    A User Guide from Low to High Redshift

    Greggio, L., Rensini, A., 2011, Stellar populations. A User Guide from Low to High Redshift

    work page 2011

  42. [42]

    Inoue, T., Fukui, Y., 2013, ApJL, 774, 31

    work page 2013

  43. [43]

    I., et al

    Johnson, C. I., et al. 2015, AJ, 150, 63

    work page 2015

  44. [44]

    I., 2010, MNRAS, 403, 1413 (K10)

    Karakas, A. I., 2010, MNRAS, 403, 1413 (K10)

    work page 2010

  45. [45]

    Kim, J., Lee, Y-W., 2018, ApJ in press (arXiv:1807.01317)

    work page internal anchor Pith review Pith/arXiv arXiv 2018

  46. [46]

    R., 1998, ApJL, 501, 205

    Klessen, Ralf S.; Burkert, Andreas; Bate, M. R., 1998, ApJL, 501, 205

    work page 1998

  47. [47]

    Kroupa, P., et al., 2013, Springer Science+Business Media Dordrecht, 115

    work page 2013

  48. [48]

    2013, MNRAS, 433, 1941

    Lardo, C., et al. 2013, MNRAS, 433, 1941

    work page 2013

  49. [49]

    Leigh, N. W. C., Boker, T., Maccarone, T. J., Perets, H. B., 2013, 429, 2997 (L13)

    work page 2013

  50. [50]

    F., Milone, A

    Marino, A. F., Milone, A. P., Piotto, G., Villanova, S., Be- din, L. R., Bellini, A., Renzini, A., 2009, A&A, 505, 1099

    work page 2009

  51. [51]

    F., et al

    Marino, A. F., et al. 2011, A&A, 532, 8

    work page 2011

  52. [52]

    Marino, A. F. et al. 2015, MNRAS, 450, 815

    work page 2015

  53. [53]

    Marino, A. F. et al. 2018, ApJ, 859, 81

    work page 2018

  54. [54]

    B., 2013, ApJ, 779, 85

    Mastrobuono-Battisti, A., Perets, H. B., 2013, ApJ, 779, 85

    work page 2013

  55. [55]

    J., Wadsley, J

    Maxwell, A. J., Wadsley, J. Couchman, H. M. P.; Sills, A., 2014, MNRAS, 439, 2043

    work page 2014

  56. [56]

    Mckenzie, M., Bekki, K., 2018, MNRAS, 479, 3126

    work page 2018

  57. [57]

    C., 1978, ApJ, 225, 380

    Meyers, P. C., 1978, ApJ, 225, 380

    work page 1978

  58. [58]

    P., et al

    Milone, A. P., et al. 2013, A&A, 555, 143

    work page 2013

  59. [59]

    P., et al

    Milone, A. P., et al. 2017, MNRAS, 464, 3636

    work page 2017

  60. [60]

    F., Frenk, C

    Navarro, J. F., Frenk, C. S., White, S. D. M., 1996, ApJ, 462, 563 (NFW)

    work page 1996

  61. [61]

    F., 2006, MNRAS, 373, 1039

    Nelson, A. F., 2006, MNRAS, 373, 1039

    work page 2006

  62. [62]

    F., et al., 2007, MNRAS, 381, 1450

    Neto, A. F., et al., 2007, MNRAS, 381, 1450

    work page 2007

  63. [63]

    Niederhofer, F., et al., 2016, MNARS in press (arXiv:1612.00400) Pflamm-Altenburg, J., Kroupa, P., 2009, MNRAS, 397, 488

    work page internal anchor Pith review Pith/arXiv arXiv 2016

  64. [64]

    P., 1999, A&AS, 134, 241

    Phillips, J. P., 1999, A&AS, 134, 241

    work page 1999

  65. [65]

    Piotto, G., et al., 2005, 621, 777

    work page 2005

  66. [66]

    Prantzos, N., Charbonnel, C., 2006, A&A, 458, 135

    work page 2006

  67. [67]

    Renzini, A., et al., 2015, MNRAS, 454, 4197

    work page 2015

  68. [68]

    U., 2011, ApJ, 732, L17

    Roederer, I. U., 2011, ApJ, 732, L17

    work page 2011

  69. [69]

    Saigo, K., Matsumoto, T., Hanawa, T., 2000, ApJ, 531, 971

    work page 2000

  70. [70]

    E., & Prosser, C

    Langer, G. E., & Prosser, C. F. 1997, AJ, 114, 1964 Sobeck et al. 2011, AJ, 141, 175

    work page 1997

  71. [71]

    2015, Nature, 523, 318 (T15)

    Tailo, M., et al. 2015, Nature, 523, 318 (T15)

    work page 2015

  72. [72]

    K., Klein, R

    Truelove, J. K., Klein, R. I., McKee, C. F., Holliman, J. H. II., Howell, L. H., Greenough, J. A., 1997, ApJL, 489, 179

    work page 1997

  73. [73]

    Ventura, P.; D’Antona, F., 2009, A&A, 499, 835

    work page 2009

  74. [74]

    J., D’Antona, F., D’Ercole, A., 2018, MNRAS, 476, 2731

    Vesperini, E., Hong, J., Webb, J. J., D’Antona, F., D’Ercole, A., 2018, MNRAS, 476, 2731

    work page 2018

  75. [75]

    C., Hill, V., Sobeck, J., & Carretta, E

    Worley, C. C., Hill, V., Sobeck, J., & Carretta, E. 2013, A&A, 553, 47 c⃝ 2005 RAS, MNRAS 000, 1– ?? Star formation in dense stellar systems 19

    work page 2013

  76. [76]

    F., Milone, A

    Marino, A. F., Milone, A. P., & Shingles, L. J. 2014, MN- RAS, 441, 3396

    work page 2014

  77. [77]

    2017, A&A, 601, 29 c⃝ 2005 RAS, MNRAS 000, 1– ??

    Zaprtas, E., et al. 2017, A&A, 601, 29 c⃝ 2005 RAS, MNRAS 000, 1– ??

    work page 2017