Pith. Sign in

REVIEW

OASIS III: Bursty star formation drives observed galaxy sizes at $z\geq4$

PaperPDF from arXiv

data files

ReviewThe objections and verdict, as JSON Paper breakdownThe authors' claims, conditions and measurements, as JSON Signed receiptProof this review has not been altered
X Bluesky LinkedIn Reddit Hacker News
William M. Baker, Tobias J. Looser, Francesco D'Eugenio, Andrew J. Bunker, Stefano Carniani, Stéphane Charlot, Jacopo Chevallard, Mirko Curti
show 12 more authors
A. Lola Danhaive, Qiao Duan, Kevin Hainline, Jakob M. Helton, Benjamin D. Johnson, Pierluigi Rinaldi, Brant Robertson, Fengwu Sun, Yang Sun, Sandro Tacchella, James A. A. Trussler, Zihao Wu

Astrophysics of Galaxies · Submitted 30 Oct 2026 · 300 references

Pith's verdict: The anti-correlation between recent-SFR upturn and compact light is solid at z=4-9; the causal 'burstiness regulates sizes' framing is plausible but untested against young-light weighting, and the z>10 application is speculation.

T0 · AI review

How this review was made: An AI referee (deepseek/deepseek-v4-flash) read the paper on November 2, 2026 and answered a fixed list of questions about it. T0 is the starting level every paper gets. The levels describe how deeply Pith has checked the paper, never the authors.

  1. T0 AI review An AI read the paper and answered a fixed list of questions.
  2. T1 Audit Automatic checks that the references and files the paper relies on exist and match.
  3. T2 Compute The paper's code or calculations re-run with the same results.
  4. T3 Formal statement The claims written out in a formal math language a computer can read.
  5. T4 Proof Those claims proved, and the proof checked by computer.

How Pith checks papers →

· 3 major objections · 6 minor · Ask about this paper · The compiled record · Read the full review →
Challenge this review Comment (0) Read on arXiv ↗
Cite ▾
arXiv ID2610.41491 (version 1)
Pith ID6KV3DV7G, Pith's permanent record of this paper and its review
CitationA. Lola Danhaive, Andrew J. Bunker, Benjamin D. Johnson, Brant Robertson, Fengwu Sun, Francesco D'Eugenio, Jacopo Chevallard, Jakob M. Helton, James A. A. Trussler, Kevin Hainline, Mirko Curti, Pierluigi Rinaldi, Qiao Duan, Sandro Tacchella, Stefano Carniani, St\'ephane Charlot, Tobias J. Looser, William M. Baker, Yang Sun, Zihao Wu. "OASIS III: Bursty star formation drives observed galaxy sizes at $z\geq4$." arXiv:2610.41491 (2026). https://pith.science/paper/6KV3DV7G
Cite this reviewBibTeX for Pith's review of this paper
Claim paper

Why Pith says this

Highlight any text to comment on it.

Referee report

How Pith checks papers → · Full audit of this paper →

The paper

The paper claims

At z ≈ 4–9, galaxies caught in a starburst are systematically more compact than galaxies with declining star formation, and the size difference survives controlling for stellar mass and redshift.

The reading

This paper argues that at $z\approx4$–$9$, how bursty a galaxy's star formation is at the moment of observation largely sets how big the galaxy appears, and that the link is physical rather than a side effect of mass or redshift. The evidence comes from 51 galaxies in the OASIS survey, deliberately selected to include faint 'lulling' and 'mini-quenched' galaxies — systems that have recently stopped or nearly stopped forming stars — far below the star-forming main sequence, not just bright starbursts. Burstiness is the ratio $\log(\mathrm{SFR}_{\mathrm{H}\alpha}/\mathrm{SFR}_{100\,\mathrm{Myr}})$: the star-formation rate from $\mathrm{H}\alpha$ emission, tracing roughly the last 10 Myr, divided by the 100-Myr-averaged rate from full spectral fitting, so positive values mark galaxies caught in a burst. Against half-light radii (the radius enclosing half of a galaxy's light, measured from deep NIRCam imaging), this ratio shows a partial correlation of $-0.57$ ($p<0.001$) that stays significant at $p=0.001$ after controlling for stellar mass and redshift, and median stacks confirm that bursting galaxies are more centrally concentrated. The authors read this as compaction-driven central starbursts that shrink the half-light radius, with the profile relaxing as the burst fades; if the same mechanism operates at $z>10$, it would explain the size bimodality of UV-luminous galaxies without invoking a new population.

What carries the argument

The load-bearing quantity is the burstiness indicator $B\equiv\log(\mathrm{SFR}_{\mathrm{H}\alpha}/\mathrm{SFR}_{100\,\mathrm{Myr}})$: the logarithm of the ratio of the star-formation rate from dust-corrected $\mathrm{H}\alpha$ emission (roughly the last 10 Myr) to the rate from non-parametric spectral fitting averaged over 100 Myr, with positive values marking rising or bursting systems and negative values marking declining ones. The sizes are continuum half-light radii from single-Sérsic forward fits to dithered NIRCam exposures (the ForcePho approach), so they trace where the young stellar light sits rather than where the stellar mass sits — a distinction the paper stresses. The statistical engine is the partial correlation coefficient $\rho_{AB|C}$, which strips out stellar mass and redshift before testing the burstiness–size link, together with a linear plane fit in the three-dimensional space of size, burstiness and redshift whose 0.20 dex scatter is taken as evidence that these two variables, rather than stellar mass, set galaxy size.

Extended reading notes

Core claim

The paper's central claim, in its own words, is that there is a strong anti-correlation between $\log(\mathrm{SFR}_{\mathrm{H}\alpha}/\mathrm{SFR}_{100\,\mathrm{Myr}})$ and galaxy size that is not just caused by stellar mass or redshift. Three independent lines of evidence support it: a partial correlation of $-0.57$ ($p<0.001$) between burstiness and half-light radius; the anti-correlation surviving at $p=0.001$ when stellar mass and redshift are held fixed simultaneously; and median stacks in F200W, F277W and F356W in which the bursting subsample shows a more centrally concentrated light profile than the declining subsample. The burstiness term alone cuts the variance of the size relation by about 30%, and a two-dimensional plane in the space of size, burstiness and redshift fits with rms scatter 0.20 dex, below the 0.261 dex of the classical mass–size relation — adding stellar mass to the fit does not reduce the scatter further. The relation is driven by the bursting side, since declining galaxies show no trend with the strength of their decline, which the authors read as evidence against inside-out quenching. The authors also show that apparent size relations with ionisation-parameter tracers and with metallicity vanish once burstiness is controlled, while the burstiness–size relation survives every control including dust. The physical interpretation is compaction-driven central starbursts: a gas-rich centre shrinks the half-light radius during a burst, and the light profile relaxes as the galaxy enters its lulling or mini-quenched phase, so burstiness regulates galaxy sizes at high redshift.

Load-bearing premise

The load-bearing premise is that $\log(\mathrm{SFR}_{\mathrm{H}\alpha}/\mathrm{SFR}_{100\,\mathrm{Myr}})$ identifies a true burst episode rather than merely the normal rising shape of a young galaxy's star-formation history; the authors themselves note in Sec. 3.3 that at high redshift rising SFHs can come from continual growth, not bursts, in which case the size anti-correlation would be a light-weighting effect — young stars concentrated in the centre — rather than a burst-and-relax structural cycle.

Editorial extensions

If this is right

  • At fixed stellar mass and redshift, burstiness and redshift — not mass — set the half-light size of a $z\approx4$–$9$ galaxy; the burstiness term alone cuts the variance of the size relation by about 30%, and together with redshift by about 42%.
  • Magnitude-limited surveys over-represent bursting galaxies twice over, because they are both brighter and more compact, so completeness corrections for high-redshift population studies should include galaxy sizes.
  • If the same burstiness–size mechanism operates at $z>10$, the size bimodality of UV-luminous galaxies (compact, NIV]-strong sources versus extended ones) follows directly, with the OASIS plane predicting a ~2× size difference close to the observed 2.2×.
  • Once burstiness is included, residual size evolution with redshift becomes consistent with dark-matter halo growth ($R_e \propto H(z)^{-0.85\pm0.26}$), implying that part of the steep apparent size evolution at high redshift is an increasing fraction of bursting galaxies.
  • The size–ionisation-parameter and size–metallicity relations seen in high-redshift samples are secondary products of burstiness; each loses significance when burstiness is controlled, while the burstiness–size relation does not.

Reading between the lines

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

  • The paper measures half-light radii, so the size–burstiness relation naturally mixes structural change with light-weighting; an editorial extension is that sizes traced by older stellar populations should show a weaker correlation, and spatially resolved maps of a few galaxies would reveal whether the young stars are genuinely central or merely bright.
  • The burstiness ratio cannot distinguish a stochastic burst from a secularly rising star-formation history, a degeneracy the authors flag; future samples that separate the two (for example via the shape of the non-parametric SFH or UV-slope variability) could show that only genuine bursts, not gradual growth, shrink half-light radii.
  • Folding galaxy size into completeness corrections should change the inferred number densities of faint lulling and mini-quenched phases relative to bursting ones, and may reduce the apparent overabundance of UV-bright galaxies at $z>10$ without invoking new physics; this is a testable prediction for future wide JWST surveys.
  • The central-starburst mechanism predicts flat or inverted metallicity gradients during bursts, consistent with the sample's metal-poor, compact bursting galaxies; resolved IFU spectroscopy could confirm whether the metal-poor gas is truly concentrated in the centre.
Original abstract

We employ JWST/NIRSpec MSA PRISM spectroscopy together with NIRCam imaging to investigate the impact of star-formation variability on galaxy sizes for a representative sample of 51 galaxies at $z\geq4$. The OASIS survey enables us to probe low-star-formation-rate (SFR) lulling and mini-quenched galaxies down to $\log_{10}(M_\star/M_\odot)=8.3$ and $\log_{10}(\mathrm{SFR}/M_\odot\,\mathrm{yr}^{-1})<-1.6$, i.e. more than one order of magnitude below the star-forming main sequence (MS). This enables us to probe objects that are often missed by shallower spectroscopic surveys, in addition to the MS and starburst systems, yielding an SFR- and mass-representative sample at high redshift. We derive galaxy half-light sizes using ForcePho, enabling us to resolve the highly compact nature of these sources, and we analyse their star-formation histories and interstellar medium (ISM) properties with the full-spectral-fitting code ppxf. We find a clear anti-correlation between burstiness, as traced by the ratio of the SFR from H$\alpha$ compared to that from non-parametric spectral fitting averaged over 100 Myr, and galaxy UV/optical half-light radii as measured with ForcePho from a combination of well-sampled NIRCam photometry. Using partial correlation coefficients we show that the anti-correlation between burstiness (as traced by the SFR ratio) and size cannot simply be explained by secondary dependencies on stellar mass or redshift. We interpret this relation as the effect of central starbursts in SFR-rising (bursting) systems, where intense central star formation reduces measured galaxy half-light sizes. This burstiness then regulates galaxy sizes at high redshift. Should the same mechanism be present at $z>10$, this could explain the size bimodality seen in the UV luminous galaxies as simply stemming from a more fundamental relationship between burstiness and size.

The compiled record

The paper's own sentences, arranged, kept beside the review and separate from it. Each label is drawn from a closed vocabulary; the number beside it is the probability the classifier gave that label. Numbers in the tables are the authors', checked by code. No sentence in this record was written by a model.

Compiled by classifier-1.13.0 via TypeSafe API, question bank abstract-p1-20260923, 2026-11-02: 1 request, 7,497 input tokens.

What the abstract itself states, as probabilities: claims first 0.04 · states limitation 0.31 · human participants 0.01 · controlled comparison 0.13 · randomized assignment 0.01 · states quantitative result 0.90 · compares with prior results 0.34 · mentions code or data availability 0.03.

The abstract, sentence by sentence

arXiv listing abstract v1. Results and the main claim are underlined; .

We employ JWST/NIRSpec MSA PRISM spectroscopy together with NIRCam imaging to investigate the impact of star-formation variability on galaxy sizes for a representative sample of 51 galaxies at $z\geq4$. The OASIS survey enables us to probe low-star-formation-rate (SFR) lulling and mini-quenched galaxies down to $\log_{10}(M_\star/M_\odot)=8.3$ and $\log_{10}(\mathrm{SFR}/M_\odot\,\mathrm{yr}^{-1})<-1.6$, i.e. more than one order of magnitude below the star-forming main sequence (MS). This enables us to probe objects that are often missed by shallower spectroscopic surveys, in addition to the MS and starburst systems, yielding an SFR- and mass-representative sample at high redshift. We derive galaxy half-light sizes using ForcePho, enabling us to resolve the highly compact nature of these sources, and we analyse their star-formation histories and interstellar medium (ISM) properties with the full-spectral-fitting code ppxf. We find a clear anti-correlation between burstiness, as traced by the ratio of the SFR from H$\alpha$ compared to that from non-parametric spectral fitting averaged over 100 Myr, and galaxy UV/optical half-light radii as measured with ForcePho from a combination of well-sampled NIRCam photometry. Using partial correlation coefficients we show that the anti-correlation between burstiness (as traced by the SFR ratio) and size cannot simply be explained by secondary dependencies on stellar mass or redshift. We interpret this relation as the effect of central starbursts in SFR-rising (bursting) systems, where intense central star formation reduces measured galaxy half-light sizes. This burstiness then regulates galaxy sizes at high redshift. Should the same mechanism be present at $z>10$, this could explain the size bimodality seen in the UV luminous galaxies as simply stemming from a more fundamental relationship between burstiness and size.

background contribution method result interpretation main claim

What the classifier reports for each sentence. The number is the probability it gave the label.
#sentencerolethe authors' own workstates the main conclusionstrength
s0 We employ JWST/NIRSpec MSA PRISM spectroscopy together with NIRCam imaging to investigate the impact of star-formation variability on galaxy sizes for a representative sample of 51 galaxies at $z\geq4$. method 0.65 0.96 0.03 asserted 0.77
s1 The OASIS survey enables us to probe low-star-formation-rate (SFR) lulling and mini-quenched galaxies down to $\log_{10}(M_\star/M_\odot)=8.3$ and $\log_{10}(\mathrm{SFR}/M_\odot\,\mathrm{yr}^{-1})<-1.6$, i.e. more than one order of magnitude below the star-forming main sequence (MS). method 0.53 0.89 0.03 asserted 0.99
s2 This enables us to probe objects that are often missed by shallower spectroscopic surveys, in addition to the MS and starburst systems, yielding an SFR- and mass-representative sample at high redshift. contribution 0.79 0.71 0.02 asserted 1.00
s3 We derive galaxy half-light sizes using ForcePho, enabling us to resolve the highly compact nature of these sources, and we analyse their star-formation histories and interstellar medium (ISM) properties with the full-spectral-fitting code ppxf. method 1.00 0.97 0.02 asserted 0.98
s4 We find a clear anti-correlation between burstiness, as traced by the ratio of the SFR from H$\alpha$ compared to that from non-parametric spectral fitting averaged over 100 Myr, and galaxy UV/optical half-light radii as measured with ForcePho from a combination of well-sampled NIRCam photometry. result 1.00 0.97 0.59 asserted 1.00
s5 Using partial correlation coefficients we show that the anti-correlation between burstiness (as traced by the SFR ratio) and size cannot simply be explained by secondary dependencies on stellar mass or redshift. result 0.96 0.96 0.10 asserted 1.00
s6 We interpret this relation as the effect of central starbursts in SFR-rising (bursting) systems, where intense central star formation reduces measured galaxy half-light sizes. interpretation 1.00 0.60 0.49 asserted 0.76
s7 This burstiness then regulates galaxy sizes at high redshift. interpretation 0.93 0.54 0.79 asserted 1.00
s8 Should the same mechanism be present at $z>10$, this could explain the size bimodality seen in the UV luminous galaxies as simply stemming from a more fundamental relationship between burstiness and size. interpretation 0.90 0.10 0.06 hedged possibility 0.96

What kind of paper the classifier reports this is: empirical study 1.00 · evidence from observation or experiment 0.91.

What was examined, and what was not

sources read
  • arXiv listing abstract v1
not examined
  • full text
  • figures and tables
  • references and citing sentences
  • relations to other papers
levels run
  • P abstract
not run
  • S sections
  • T body sentences
  • C citations
  • R cross-paper relations

Every quoted sentence on this record is the authors'. Every label is from Pith's closed vocabulary, with the probability shown beside it. Every number in the measurements table is the authors', checked by code. Record as JSON

Figures
Figure 1. Specific star-formation rate vs redshift for our sample
Figure 2. Effective half-light radius from ForcePho multi-band fit￾ting vs burstiness (log(SFRHα/SFR100Myr)) colour-coded by stel￾lar mass. The red points correspond to the bin medians, with the errors corresponding to the standard error on the mean. It shows a clear relation between burstiness and the size as measured by the half-light radius. The partial correlation coefficients between size and log(SFRHα/SFR100Myr) show a statistically significant anti-correlation of -0.57. Galaxies with greater burstiness (higher log(SFRHα/SFR100Myr)) are systematically smaller. photometry, 7 were found…
Figure 3. log(SFRHα/SFR100Myr) vs stellar mass colour coded by half-light radius (left), and burstiness vs redshift colour-coded by half-light radius (right). The arrow points in the direction of the greatest increase in the colour coded quantity, in this case size (for more details see Baker et al. 2022). The left hand figure shows that size depends primarily on log(SFRHα/SFR100Myr) (inversely) and is almost independent of mass, whilst the right hand figure shows that size depends on both burstiness and redshift. the half-light effective radius of the galaxies against their…
Figure 4. Partial correlation coefficients between all quantities and the size whilst controlling for all other quantities in the fig￾ure. The black error bars correspond to bootstrap random sam￾pling 1000 times and taking the 16th and 84th percentiles of the resulting distribution, whilst the red error bars are taking the standard error on the mean. The strongest (anti-)correlation is with the log(SFRHα/SFR100Myr) ratio, followed by the redshift, which are the only ones significant enough to reject the null hypothesis. This highlights the strong anti-correlation between log(SFRHα/SFR100Myr)…
Figure 5. Left: stacks of the OASIS sample galaxies divided into those with rising or falling SFHs in F200W, F277W and F356W.
Figure 6. Effective half-light radius vs stellar mass, colour coded by burstiness (i.e. the size-mass relation). The black line gives our best-fit size-mass-redshift relation for the whole sample scaled to z=5.5, whilst blue and red dashed lines show this best fit model evaluated at the median burstiness for the rising and falling sam￾ples respectively. The size mass relations from McClymont et al. (2025a) and Miller et al. (2025) are overplotted. Kawinwanichakij et al. 2025; Miller et al. 2026). Simulations vary as to whether they can produce the relation at high-z with some finding clear…
Figure 8. Effective half-light radius vs log([Ne iii] λ3869/[O ii] λλ3727,29) (left) and log([O iii] λ5007/[O ii] λλ3727,29) (right), colour￾coded by stellar mass. Both line ratios are strong tracers of the ionisation parameter, and both show a strong anti-correlation with galaxy size. This indicates that galaxies with higher ionisation parameter (consistent with those undergoing a burst of star-formation) are systematically more compact. outskirts this can still have the effect of increasing the half-light radius due to a relaxation of the light profile, i.e. it will no longer remain…
Figure 9. Schematic diagram showing our interpretation of the observed anti-correlation between burstiness and half-light radius
Figure 10. Effective half-light radius vs redshift. Inspired by the figure of Donnan et al. (2026), we separate out z > 10 galaxies into NIV] detected or undetected. We show that extrapolating the behaviour of the OASIS sample to high redshift would predict a size difference between the galaxies in a burst phase, as traced by NIV] (amongst other properties), and those undergoing a de￾cline, whereby those undergoing a burst are more compact. 4.4. Implications for galaxy population studies The uncovered relation between galaxy size and burstiness has important implications for our studies of…

Showing 9 of the paper's 10 figures. See all figures in the PDF ↗ Figures reproduced from arXiv:2610.41491 by the authors.

References

Discussion

Disagree with the review, or know something it missed? Leave a comment, highlight any passage above to comment on it, or challenge this review formally.