3 things they don't tell you about star clusters

Florent Renaud

REVIEW 2 major objections 6 minor 37 references

3 things they don't tell you about star clusters

T0 review · 2 major / 6 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 No current simulation can predict how star clusters dissolve.

desk verdict A lucid, opinionated conference review that earns a serious referee, but its most important claim about simulation tides rests on an unquantified resolution criterion. read the letter →

arxiv 1908.02301 v1 pith:ZOAA335R submitted 2019-08-06 astro-ph.GA

Florent Renaud This is my paper · ORCID

classification astro-ph.GA

keywords starclustersglobulartidalfieldshocksclustermassfunctioncosmologicalsimulationsformationrate

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 argues that three widely used shortcuts in star cluster research are misleading. It first contends that young massive clusters are not established local analogues of the globular clusters formed in the early universe, because the present-day globular cluster mass function is universal while the young massive cluster mass function is a power law, implying the initial mass function must depend on environment. It then makes the stronger claim that no cosmological simulation currently resolves the rapidly varying tidal field that drives cluster mass-loss, so predictions of cluster survival and dissolution from galaxy-scale simulations are unreliable. Finally, it argues that the cluster formation rate does not track the star formation rate, since stars can form in pre-existing clusters and in nuclear clusters fed by gas inflows. If these points hold, interpreting star cluster populations without a resolved tidal field and an environment-dependent initial mass function is premature.

What carries the argument

The central object is the tidal field around a cluster, split into a slowly varying secular component and a rapidly varying shock component. The argument's load-bearing mechanism is a resolution criterion: physical structures that generate shocks—disk crossings at about 100 parsecs, molecular clouds at about 10 parsecs, and filaments below a parsec—must be covered by several resolution elements, requiring sub-parsec effective resolution to be captured at all. Applying this criterion separates physical tides from numerical noise in simulation derivatives.

What would settle it

Take one galaxy formation simulation and run it at both several-hundred-parsec and sub-parsec gas resolution with identical initial conditions; if the coarse run reproduces the same cluster mass-loss statistics as the fine run, the central claim fails, while a difference in tidal shocks would support it.

Watch

Extended reading notes

Core claim

The paper's most consequential assertion is that the rapidly varying component of the tidal field—the shocks a cluster experiences when crossing a galactic disk or encountering molecular clouds—cannot be resolved in any current cosmological simulation, including zoom-in runs. Since these shocks inject the energy that allows stars to escape, a simulation that cannot resolve structures down to roughly a parsec cannot predict cluster mass-loss or dissolution. The paper therefore concludes that the evolution from the unknown initial cluster mass function to the present-day one remains out of reach of current simulations, and that apparent tidal shocks measured at several-hundred-parsec resolution are numerical noise from differentiating a coarse potential. Alongside this, the paper argues that young massive clusters are not proven local analogues of young globular clusters and that the cluster formation rate is not tied to the star formation rate.

Load-bearing premise

The entire resolution critique rests on the rule that a quantity measured with less than a few resolution elements is unphysical numerical noise; if that rule is too strict, coarse simulations could still capture real tidal shocks.

Editorial extensions

If this is right

Predictions of globular cluster survival, destruction rates, and the evolution of the cluster mass function from current galaxy-scale simulations should be treated as provisional until the tidal field is resolved to sub-parsec scales.
If the initial cluster mass function is environment-dependent rather than a universal power law, then present-day globular cluster mass functions record a balance between formation conditions and evolutionary losses, not a single universal initial condition.
Using star formation rates or galaxy interaction stage to infer cluster formation rates can mislead, particularly in gas-rich, high-redshift or merging systems where the two rates decouple.
A cluster evolution model that includes a properly resolved time-varying tidal field could be used to test formation hypotheses across cosmic time, but none exists yet.

Reading between the lines

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

A direct corollary the paper leaves implicit is that comparing observed cluster populations to cosmological simulations yields conclusions only as strong as the resolution of the tidal field; until convergence is shown, apparent agreements may be coincidental.
The resolution argument suggests that improving gas resolution in simulations should change predicted cluster mass-loss rates, and likely increase them as more small-scale shocks are captured; this is testable by resolution studies.
The CFR-SFR decoupling, if real at high redshift, has implications beyond clusters: the environments that produce dense stellar systems may not trace star formation, so using star formation to predict gravitational-wave merger rates from cluster cores could be biased.
The paper's logic could be extended to observational surveys: the universality of the globular cluster mass function is consistent with an equilibrium, not a universal formation process, so survey data should be interpreted with an explicit evolution model before claiming a common origin.

Share X Bluesky LinkedIn Reddit HN

Signed reviews

No signed human review yet.

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

This is a conference proceedings piece, not a new measurement or simulation, and the reader's conditional verdict is about right. What the paper does well: it is clearly written, honest about what we don't know, and it bundles three genuine caveats that often get lost in the literature. The YMC/globular analogy point is carefully argued and the 'we still don't know' refrain is a fair brake on overreach. The CFR/SFR discussion is a useful reminder that star formation in pre-existing clusters breaks a naive proportionality. The real value, though, is the Section 2 critique of E-MOSAICS: deriving tidal fields from potential differences at scales 200 times below resolution does look dubious, and this deserves a public airing.

Now the soft spots. The section titles overstate the body: 'are not local analogues' and 'does not correlate' are stronger than the carefully hedged text underneath. That mismatch matters because the titles are what people will remember. More importantly, the central claim of Section 2 — that no cosmological simulation can predict cluster mass-loss because unresolved structures dominate the tidal field — rests on a footnote assertion that any quantity measured below a few resolution elements is 'unphysical, likely numerical noise,' supported only by a citation to the author's PhD thesis. That is a plausible rule of thumb, but it is not demonstrated here, and the stress-test note is right that the paper never quantifies how much of the mass-loss budget comes from sub-resolution shocks versus the resolved secular tide. The rebuttal that E-MOSAICS's convergence on mass-loss across smoothing scales is itself evidence that shocks are not captured assumes that resolving smaller structures would necessarily add to the tidal field; that might be true, but it is not proven. So the strongest conclusion should be 'we still don't know whether current simulations get the mass-loss right,' not 'they definitely don't.'

I would send this to a referee. It is a well-written, opinionated review that identifies a real methodological concern, and the E-MOSAICS critique, even if overstated, is sharp enough to warrant careful response from the simulators. The referee's main job would be to force the author to either soften the Section 2 claim or provide a quantitative resolution test. For a reading group, it would spark a good discussion about resolution and interpretation in galaxy-scale simulations, but it is not a must-read. I would not cite it in my own work, though I might mention it as a cautionary flag.

Recommendation: accept it as a proceedings contribution after minor revision, with the request that the author qualify the no-simulation claim and add a sentence about what a convergence test would look like. The paper is honest and well-intentioned, but the central assertion needs a better foundation.

Referee Report

2 major / 6 minor

Summary. This invited proceedings contribution (IAU Symposium 351) presents three cautionary points about the current state of star cluster research. First, present-day young massive clusters (YMCs) cannot currently be established as local analogues of young globular clusters, because the formation conditions at high redshift differ in gas fraction, turbulence, metallicity, and likely in the stellar IMF, and because the observed universality of the present-day globular cluster mass function is difficult to reconcile with a universal initial CMF under environment-dependent evolution. Second, the author argues that no current cosmological simulation resolves the small-scale, rapidly varying tidal field component (disk crossings, cloud encounters, sub-10 pc ISM structure) that drives cluster mass-loss, with E-MOSAICS singled out as an example of tidal shocks measured below the resolution limit; the paper concludes that cluster mass-loss and the evolution of the CMF remain out of reach of current simulations. Third, the cluster formation rate (CFR) need not trace the star formation rate (SFR), because star formation can occur in pre-existing clusters through the formation of multiple stellar populations and gas accretion onto nuclear clusters. The body of the paper is carefully hedged, with explicit statements such as "we still don't know", while the Section 2 conclusion is stated categorically.

Significance. If the Section 2 claim is correct, it has broad consequences: inferences about cluster destruction, survival timescales, and the evolution of the cluster mass function that are drawn from galaxy-scale and cosmological simulations (including E-MOSAICS) would rest on an unverified basis, and the paper sets a concrete, falsifiable resolution target (effective sub-parsec resolution with several resolution elements covering cold, <10 pc ISM features). The paper is a perspective rather than a new calculation, but it makes its claims testable: the Section 2 resolution requirement can be checked by controlled convergence studies, and the Section 3 prediction of a CFR-SFR mismatch in mergers can be confronted with observations. I credit the author for engaging directly with the E-MOSAICS convergence argument in a footnote rather than ignoring counter-evidence, and for repeatedly flagging the limits of the community's knowledge. The main weakness is that the paper's most consequential claim rests on an unquantified assertion about numerical noise below the resolution limit, which makes the strength of the conclusion disproportionate to the evidence presented.

major comments (2)

[§2, footnote and concluding paragraph (p. 4-5)] The section's conclusion — "to date no cosmological simulation captures the tidal field with sufficient precision to predict the mass-loss of clusters" — rests on the footnote assertion that "any quantity measured with less than a few resolution elements (and all the more below resolution limit, like in E-MOSAICS) is unphysical, and likely relates to numerical noise induced by the derivation", which is supported by a citation to the author's PhD thesis (Renaud 2010) rather than demonstrated here. This premise is load-bearing: if it is false, the critique of E-MOSAICS and the global negative conclusion lose their force. The author should either provide a quantitative justification (e.g., a convergence test of the tidal field measurement itself, or an estimate of the numerical noise floor relative to the physical amplitudes of disk-crossing and cloud-encounter shocks) or soften the claim to state that the required resolution has not yet been demonstrated, rather than that existing measurements are unphysical. The rest of the paper is carefully hedged, and this categorical step stands out.
[§2, E-MOSAICS convergence footnote (p. 4)] The rebuttal to the E-MOSAICS convergence test states that resolving smaller structures "would add their contribution to the net tidal field". This requires two quantitative inputs that the paper does not supply: that structures at the relevant scales (<10 pc clouds and finer) exist with the assumed properties in the simulated galaxies, and that their contribution dominates the cluster mass-loss budget relative to the resolved secular tide. Without such an estimate (e.g., from analytic shock-driven mass-loss prescriptions or from the higher-resolution Li et al. 2017 simulations), the interpretation of mass-loss convergence across smoothing scales as evidence of uncaptured physics, rather than of a physically converged treatment, cannot be evaluated. I recommend adding a short quantitative discussion of the relative mass-loss contributions from the resolved versus unresolved tidal components, or explicitly framing the claim as a resolution requirement for future work.

minor comments (6)

[Section 1 and Section 3 titles] The section titles "Young massive clusters are not local analogues of young globular clusters" and "The cluster formation rate does not correlate with the star formation rate" are stronger than the corresponding body text, which explicitly states "we still don't know" and "quantifying the importance of these mechanisms is necessary before reaching definite conclusions". I recommend rewording the titles to match the hedged content (e.g., "... are not established as local analogues ..." and "... need not correlate ...").
[§2] There are several typos in Section 2: "E-MOASAIC" should be "E-MOSAICS" (in both the text and the footnote), "a effectiv e sub-parsec resolution" should be "an effective sub-parsec resolution", and the phrase "It is needless to say" is awkward and could simply be removed.
[Title and Abstract] The proceedings title contains "W ay" and the abstract contains "V era Rubin Survey Telescope", which appear to be compilation or OCR artifacts and should be corrected during editing. In addition, the acronym VRST is non-standard; the facility is normally referred to as LSST or the Vera C. Rubin Observatory.
[§1, bullet list] In the bullet list, "An holistic understanding of the IMF is still lacking" should read "A holistic understanding", and the phrase "the stellar winds (of which energy depends on the metallicity)" is awkward and should be rephrased.
[§3] The claim that "simulations suggest that the CFR significantly deviates from the SFR at different degrees at different stages of a galactic interaction (see e.g. Fig. 11 of Renaud et al. 2015)" is not supported by a reproduced figure; since this is a proceedings paper, citing an internal figure is acceptable, but including the relevant panel or an independent confirmation would help the reader assess the claim.
[References] The paper relies heavily on the author's own prior work (Renaud 2010, 2014, 2015, 2017; Gieles & Renaud 2016) for several of the load-bearing physical and numerical points; this is not improper, but independent references would strengthen the arguments for readers outside the author's immediate collaborations, notably for the resolution/noise criterion and the CFR-SFR mismatch.

Circularity Check

0 steps flagged · score 2.0 of 10

No circular derivation: the paper is a critical review whose claims rest on resolution arguments and external observations, not on an input-output loop.

full rationale

This is a review/position paper, not a derivation, so there is no fitted parameter or equation whose output equals an input. The strongest claim (Section 2: no cosmological simulation captures the tidal field precisely enough to predict cluster mass-loss) is supported by a resolution argument (structures of 0.1-100 pc are unresolved) and by an external benchmark (E-MOSAICS resolves at several 100 pc and derives tides at 1/200 of resolution). The footnote invoking Renaud 2010 for the statement that sub-resolution tidal quantities are unphysical is a self-citation, but it is a supporting methodological reference rather than a premise that already contains the conclusion. The paper also anchors Section 1 to the observed universality of the globular cluster mass function. The E-MOSAICS convergence argument in the footnote is a logical criticism (resolving smaller structures would add contributions), not a tautology. Concerns about whether the resolution criterion is adequately quantified belong to correctness risk, not circularity. Therefore no circular step is identified; the minor self-citation warrants at most a low score.

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

This is a review/opinion paper, so the ledger contains no free parameters and no invented entities. The claims rest on a set of background assumptions imported from the cited literature, mostly domain assumptions about cluster observables, cluster evolution, and numerical resolution. The most fragile entry is the resolution-scale assertion used against E-MOSAICS, which is not justified in this paper.

assumptions (6)

domain assumption The present-day globular cluster mass function is universal.
Invoked in Section 1 to argue that if YMCs were analogues then the initial CMF must be non-universal, because evolution is environment-dependent. Cited to Vesperini 2001 and treated as an established observation.
domain assumption Cluster mass-loss and evolution depend on the galactic environment through the tidal field.
Used in Section 1 to infer that an environment-dependent evolution would imprint on the CMF, and in Section 2 to motivate the importance of tidal shocks. Supported by cited work such as Gnedin & Ostriker 1999 but not derived here.
domain assumption Tidal shocks from small-scale structures below 10 pc are a significant driver of cluster mass loss.
Central to Section 2's claim that unresolved shocks make cosmological simulations unreliable. Based on Gnedin & Ostriker 1999 and Gieles & Renaud 2016.
domain assumption Numerical quantities derived from a potential at scales much smaller than the resolution are unphysical noise.
The key premise of the E-MOSAICS critique in Section 2. Cited to Renaud 2010 without a proof or demonstration in this paper.
domain assumption All stars form in clusters.
Opening premise of Section 3, cited to Lada & Lada 2003, used to frame the relationship between cluster formation and star formation.
domain assumption Multiple stellar populations in globular clusters imply multiple generations of star formation within the same cluster.
Used in Section 3 to argue the SFR and CFR can decouple. The paper hedges with 'if the differences between the populations translate into differences in age' and cites Bekki et al. 2017.

how reviews work

0 comments

Cite this review

Pith. "Pith review of 3 things they don't tell you about star clusters." pith.science (2026). https://pith.science/paper/ZOAA335R

@misc{pith2026190802301,
  author       = {Pith},
  title        = {Pith review of: 3 things they don't tell you about star clusters},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/ZOAA335R}},
  note         = {Machine review of arXiv:1908.02301}
}

read the original abstract

Dense stellar systems in general and star clusters in particular have recently regained the interest of the extragalactic and even cosmology communities, due to the role they could play as actors and probes of re-ionization, galactic archeology and the dark matter content of galaxies, among many others. In the era of the exploitation and the preparation of large stellar surveys (Gaia, APOGEE, 4MOST, WEAVE), of the detection of gravitational waves mostly originating from dense regions like the cores of clusters (Ligo, LISA), and in an always more holistic view of galaxy formation (HARMONI, Euclid, LSST, soon to be known as the Vera Rubin Survey Telescope, VRST), a complete theory on the formation and evolution of clusters is needed to interpret the on-going and forthcoming data avalanche. In this context, the community carries an effort to model the aspects of star cluster formation and evolution in galactic and even cosmological context. However, it is not always easy to understand the caveats and the shortcuts taken in theories and simulations, and their implications on the conclusions drawn. I take the opportunity of this document to highlight three of these topics and discuss why some shortcuts taken by the community are or could be misleading.

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

37 extracted references · 30 canonical work pages

[1]

E., Messa, M., et al

Adamo, A., Ryon, J. E., Messa, M., et al. 2017, ApJ, 841, 131

2017
[2]

2009, MNRAS, 392, 2 94

Agertz, O., Lake, G., Teyssier, R., et al. 2009, MNRAS, 392, 2 94

work page 2009
[3]

2017, MNRAS, 471, 2242

Bekki, K., Jeˇ r´ abkov´ a, T., & Kroupa, P . 2017, MNRAS, 471, 2242

work page 2017
[4]

M., Alatalo, K., et al

Cappellari, M., McDermid, R. M., Alatalo, K., et al. 2012, Na ture, 484, 485 den Brok, M., Peletier, R. F., Seth, A., et al. 2014, MNRAS, 44 5, 2385

work page 2012
[5]

& Adamo, A

Dessauges-Zavadsky, M. & Adamo, A. 2018, MNRAS, 479, L118

work page 2018
[6]

& Basu, S

Dib, S. & Basu, S. 2018, A&A, 614, A43

work page 2018
[7]

Elmegreen, B. G. 2010, ApJ, 712, L184

work page 2010
[8]

M., Elmegreen, B

Elmegreen, D. M., Elmegreen, B. G., Kaufman, M., et al. 2006, ApJ, 642, 158

work page 2006

Show all 37 references

[9]

2014, A&A, 570, A2

Feldmeier, A., Neumayer, N., Seth, A., et al. 2014, A&A, 570, A2

2014
[10]

& Kroupa, P

Fellhauer, M. & Kroupa, P . 2003, Ap&SS, 284, 643

2003
[11]

& Kroupa, P

Fellhauer, M. & Kroupa, P . 2005, MNRAS, 359, 223

2005
[12]

M., Tumlinson, J., et al

Geha, M., Brown, T. M., Tumlinson, J., et al. 2013, ApJ, 771, 2 9

2013
[13]

& Renaud, F

Gieles, M. & Renaud, F. 2016, MNRAS, 463, L103

2016
[14]

Gnedin, O. Y . & Ostriker, J. P . 1999, ApJ, 513, 626

1999
[15]

2016, MNRAS, 461, 36 20

Guillard, N., Emsellem, E., & Renaud, F. 2016, MNRAS, 461, 36 20

2016
[16]

Hayward, C. C. & Hopkins, P . F. 2017, MNRAS, 465, 1682

2017
[17]

C., Kennicutt, Jr., R

Keel, W. C., Kennicutt, Jr., R. C., Hummel, E., & van der Hulst , J. M. 1985, AJ, 90, 708

1985
[18]

Kruijssen, J. M. D., Pfeffer, J. L., Crain, R. A., & Bastian, N . 2019, MNRAS, 486, 3134

2019
[19]

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

2003
[20]

& Gnedin, O

Li, H. & Gnedin, O. Y . 2018, arXiv e-prints, arXiv:1810.11036

2018 arXiv
[21]

Y ., Gnedin, N

Li, H., Gnedin, O. Y ., Gnedin, N. Y ., et al. 2017, ApJ, 834, 69 Milosavljevi´ c, M. 2004, ApJ, 605, L13

2017
[22]

2018, Nature Astronom y, 2, 478

Motte, F., Nony, T., Louvet, F., et al. 2018, Nature Astronom y, 2, 478

2018
[23]

2019, MNRAS, 485, 3887

Ohlin, L., Renaud, F., & Agertz, O. 2019, MNRAS, 485, 3887

2019
[24]

& Gilmore, G

Parmentier, G. & Gilmore, G. 2007, MNRAS, 377, 352

2007
[25]

W., Struck, C., Smith, B

Peterson, B. W., Struck, C., Smith, B. J., & Hancock, M. 2009, MNRAS, 400, 1208

2009
[26]

& Baumgardt, H

Pfeffer, J. & Baumgardt, H. 2013, MNRAS, 433, 1997

2013
[27]

Pfeffer, J., Kruijssen, J. M. D., Crain, R. A., & Bastian, N. 2 018, MNRAS, 475, 4309 Portegies Zwart, S. F., McMillan, S. L. W., & Gieles, M. 2010, ARA&A, 48, 431

2010
[28]

2010, PhD thesis, arXiv:1008.0331

Renaud, F. 2010, PhD thesis, arXiv:1008.0331

2010 arXiv
[29]

2018, New A Rev., 81, 1

Renaud, F. 2018, New A Rev., 81, 1

2018
[30]

2017, MNRAS, 465, 3622

Renaud, F., Agertz, O., & Gieles, M. 2017, MNRAS, 465, 3622

2017
[31]

2019, A&A, 625, A 65

Renaud, F., Bournaud, F., Agertz, O., et al. 2019, A&A, 625, A 65

2019
[32]

2015, MNRAS, 446, 2038

Renaud, F., Bournaud, F., & Duc, P .-A. 2015, MNRAS, 446, 2038

2015
[33]

2014, MNRAS, 442, L33

Renaud, F., Bournaud, F., Kraljic, K., & Duc, P .-A. 2014, MNRAS, 442, L33

2014
[34]

Romeo, A. B. 2019, arXiv e-prints, arXiv:1905.05752

2019 arXiv
[35]

Romeo, A. B. & Mogotsi, K. M. 2017, MNRAS, 469, 286

2017
[36]

Romeo, A. B. & Mogotsi, K. M. 2018, MNRAS, 480, L23

2018
[37]

D., Ostriker, J

Tremaine, S. D., Ostriker, J. P ., & Spitzer, Jr., L. 1975, ApJ , 196, 407 V esperini, E. 1998, MNRAS, 299, 1019 V esperini, E. 2001, MNRAS, 322, 247 V esperini, E. & Heggie, D. C. 1997, MNRAS, 289, 898

1975

Pith tools

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

[1] [1]

E., Messa, M., et al

Adamo, A., Ryon, J. E., Messa, M., et al. 2017, ApJ, 841, 131

2017

[2] [2]

2009, MNRAS, 392, 2 94

Agertz, O., Lake, G., Teyssier, R., et al. 2009, MNRAS, 392, 2 94

work page 2009

[3] [3]

2017, MNRAS, 471, 2242

Bekki, K., Jeˇ r´ abkov´ a, T., & Kroupa, P . 2017, MNRAS, 471, 2242

work page 2017

[4] [4]

M., Alatalo, K., et al

Cappellari, M., McDermid, R. M., Alatalo, K., et al. 2012, Na ture, 484, 485 den Brok, M., Peletier, R. F., Seth, A., et al. 2014, MNRAS, 44 5, 2385

work page 2012

[5] [5]

& Adamo, A

Dessauges-Zavadsky, M. & Adamo, A. 2018, MNRAS, 479, L118

work page 2018

[6] [6]

& Basu, S

Dib, S. & Basu, S. 2018, A&A, 614, A43

work page 2018

[7] [7]

Elmegreen, B. G. 2010, ApJ, 712, L184

work page 2010

[8] [8]

M., Elmegreen, B

Elmegreen, D. M., Elmegreen, B. G., Kaufman, M., et al. 2006, ApJ, 642, 158

work page 2006