Galaxy evolution: reading the records that galaxies retain

A guide to different kinds of memory

In this selected coverage, a useful route through galaxy evolution runs from stellar histories inferred from images, through dust surviving a shutdown of star formation, to early disk dynamics and black hole activity. The connecting question is what an observable preserves about the past—and which assumptions enter its interpretation. Shared interest in historical reconstruction does not establish agreement on one evolutionary sequence.

Start with stars: inference versus new observations

Katachi-PRIME combines galaxy images with spectral energy distribution inference, using learned morphological priors to improve constraints over traditional five-band photometry. The authors distinguish continuous morphological memory from broader, robust memory. Main-sequence galaxies, the ordinary star-forming population, have continuous lookback windows of roughly 4–7 billion years. Starburst and green-valley populations have shorter continuous windows: in that sense, only their most recent star formation remains recoverable. Yet these populations retain extended robust memory of long-term evolutionary tracks. Disruption therefore does not mean all historical information is lost (morphology study).

Read this alongside the deep JWST continuum spectroscopy proposal. Instead of extracting more information from images, it advocates observations deep enough to constrain stellar ages, metallicities and motions, with a 1,000-hour sketch survey at redshifts above 1.5. This is a proposal, not a completed survey demonstrating those outcomes (continuum spectroscopy proposal). The approaches are not interchangeable: Katachi-PRIME's training labels inherit assumptions and uncertainties from conventional spectral energy distribution fitting, so its estimates are not independent physical ground truth (morphology study).

After quenching: dust complicates the record

The ALMA study asks what remains after quenching—the decline of star formation. Its sample comprises eight newly observed and five archival lensed quiescent galaxies at redshifts around 1–2.5. Dust-to-stellar-mass ratios span more than two orders of magnitude. Comparisons with destruction models associate some systems with rapid, harsh quenching and others with slower quenching and milder destruction. Unknown dust temperature introduces approximately ±0.3 dex systematic uncertainty in these ratios, qualifying the inferred dust masses and model comparisons (dust survival study).

The two dustiest objects exceed even the mildest destruction tracks. If truly quiescent, they point toward grain regrowth in interstellar gas and/or continued dust production by evolved stars. Separately from the temperature uncertainty, the single-band continuum observations cannot exclude obscured star formation in these objects. Thus surviving dust, replenished dust and hidden activity remain unresolved possibilities. This connects to morphological memory through reconstructing shutdown histories, but the different samples, redshifts and observables do not independently validate an image-based clock (dust survival study).

Early disks: where comparisons stop

The ALPINE-CRISTAL-JWST analysis instead measures spatially resolved dynamics at redshifts 4–6. Five of 18 star-forming galaxies are classified as disk-like, although random motions remain important. Their stellar and baryonic mass–speed relations lie below local comparisons, with offsets measured while fixing slopes to local values. The measurements are consistent with an interpretation involving evolving dark matter halo properties, not a unique causal demonstration (Tully–Fisher study).

Applying image-based historical inference to such distant systems faces a documented transfer problem: Katachi-PRIME uses a local sample below redshift 0.16, may learn dataset-specific color–mass shortcuts, and has unestablished performance at higher redshift where images are less resolved. These are limits on transferring its demonstrated results, not a requirement that every target have stable morphology (morphology study).

Black holes: shorter clocks and uncertain origins

A MaNGA analysis selects 34 fading-AGN candidates through weak nuclear activity relative to extended emission, tracing approximately 10,000 years of accretion history. Orientation affects its emission-line classifications. This short-timescale record does not by itself link black hole fading to quenching in the separate dust sample (AGN archaeology study).

The black hole seeds white paper poses a different identification problem: current observations cannot uniquely separate initial seeds from subsequent growth, and the nature of Little Red Dots remains debated. It proposes uniformly selected samples, spectroscopy, infrared imaging and repeated observations as complementary tests. Like the continuum proposal, it describes needed observations rather than reporting that those uncertainties are resolved (black hole seeds white paper).

Evidence limits

All six sources are version-1 arXiv preprints; the supplied evidence does not establish peer-review status. This guide uses abstracts and selected extracted passages, not complete papers. Extraction can corrupt equations, tables and reading order. Claims supported only by an abstract permit no additional methodological or robustness detail beyond that abstract. Missing material in excerpts does not establish absence from a paper. These connections map the selected coverage, not field-wide trends or consensus.