They remember how they were made

Here’s the walk-through: what a TNO is, what “cold” and “hot” mean when you’re talking about orbits rather than temperature, what the two papers measured, what the quotes say in their own words, and where the diameter numbers disagree enough that you should keep the source labels on. Metric first where NASA gives both.

What a Trans-Neptunian Object is

Trans-Neptunian Objects are typically small, faint, icy bodies orbiting the Sun beyond the orbit of Neptune. That’s the NASA Science release almost straight. Most of them are more than 100 million times dimmer than objects visible to the unaided eye. Even with Hubble and Webb, they show up only as tiny points of light. You’re not getting resolved crater maps from this survey. You’re getting brightness, color, size estimates, and orbits.

Why bother with points of light that faint? Because this class of small bodies offers the best view into an early stage of planet-building. In the early solar system, a disk of dust and pebbles in orbit around the Sun coalesced into city-sized “planetesimals” — the solid building blocks that clump together to form planets — but had not yet merged into full-sized worlds. Beyond Neptune, that second stage never happened. What got left behind is a frozen population of planetesimals. The outer disk stopped before the city-sized pieces finished becoming planets.

City-sized is NASA’s scale language for those first solid building blocks after dust and pebbles coalesce. Bigger than pebbles, smaller than planets, stuck in the stage where merging into full worlds didn’t finish beyond Neptune.

If you’ve heard “Kuiper Belt” in the same breath as Pluto, you’re in the right neighbourhood. This isn’t a Pluto biography. It’s a pencil-beam look at the faint end of that distant population — objects small enough that ground-based telescopes mostly can’t touch them, measured for the first time with Hubble and Webb working the same sky at once.

First joint power — and the deepest survey to date

NASA’s framing is blunt: for the first time, scientists used the joint power of Hubble and the James Webb Space Telescope to study some of the most far-flung bodies in the solar system. Some of these are the smallest and faintest ever directly seen. Two complementary papers published Tuesday in The Astronomical Journal analyzed the color, composition, and size distribution of 27 newly discovered tiny, dim TNOs.

The survey is the deepest TNO survey to date. Teams led by PhD candidates from the University of Victoria in Canada, under the guidance of the National Research Council of Canada, and Northern Arizona University in Flagstaff examined a patch of sky simultaneously with Hubble, observing the TNOs’ visible light, and Webb, observing their infrared light. The researchers measured the objects’ colors — which the release treats as a fingerprint of surface composition — as well as their sizes, and determined their orbits.

That simultaneous pairing matters. Hubble is strong in visible light. Webb is strong in infrared. Neither telescope alone gives you the same combination of detection depth and optical–near-infrared color on objects this faint. The NASA release says the project would not have been possible without Hubble and Webb working together to detect and characterize these TNOs. With Hubble’s sensitivity in visible light and Webb’s in infrared, the space telescopes provide more insights than either can on its own.

Hubble is a project of international cooperation between NASA and ESA. Goddard Space Flight Center in Greenbelt, Maryland, manages the telescope and mission operations. Webb is an international program led by NASA with partners ESA and CSA — the Canadian Space Agency. Those agency lines belong on the page because the instruments and the funding map are part of how a joint survey exists at all.

Cold orbits and hot orbits

In the coordinated observations, the teams studied two different types of TNOs. The labels are dynamical, not a thermometer reading of the ice.

Dynamically “cold” TNOs are on their original, relatively circular orbits around the Sun in the plane of the solar system. Think of them as the ones that mostly stayed put in the outer disk’s geometry.

Dynamically “hot” TNOs formed between the current locations of Uranus and Neptune but were pushed outward where they are today when the outer gas giants migrated early in the solar system’s history. Today they reside in highly elliptical orbits and move in and out of the plane of our solar system. Scrambled paths. Same family name. Different birth neighbourhood, if the migration story holds.

Prior to these observations, astronomers thought that small TNOs from both hot and cold populations would have undergone many collisions, changing their surfaces compared to larger TNOs. That’s the expectation the papers were built to test. Collisions chip, grind, resurface. If the tiny ones had been knocked around for billions of years, you might expect their surface colors — that composition fingerprint — to drift away from what you see on the larger siblings.

That is not what the observations showed. Instead, the small bodies look like their larger counterparts. This implies that collisions are not changing the surfaces significantly — perhaps because there are fewer collisions than expected, or because the TNOs somehow retain their primordial, pre-collision compositions. The teams are still trying to unravel that mystery. NASA says so in those words. The mystery stays open.

What Anastasia Morgan’s color paper says

Northern Arizona University PhD candidate Anastasia Morgan led the study of color and composition. Her paper in the pair is Anastasia N. Morgan et al. 2026, The Astronomical Journal 172 188 — optical–near-infrared colors for small TNOs, JWST/NIRCam detections with HST recovery.

From the abstract: the team presents optical–NIR color measurements for small (≤40 km) Trans-Neptunian Objects using coordinated and nearly simultaneous JWST and HST observations. JWST/NIRCam provided detections and NIR photometry for faint TNOs. HST/ACS and WFC3 imaging enabled recovery in the optical. Together they yield optical–NIR colors spanning 0.35–3.2 μm. Thirteen JWST-detected TNOs were recovered in the HST observations. Trailed PSF photometry was used to derive mean magnitudes, colors, and rotational lightcurves.

The color distribution of the small cold classical TNO discoveries is narrow and consistent with a single reflectance (color) sequence previously identified for larger cold classical TNOs. There is no evidence for a change in that sequence at smaller sizes. In contrast, the dynamically excited TNOs in the discovery set exhibit a broader range of colors consistent with multiple compositional classes seen at larger sizes. Lightcurve amplitudes were generally low for both dynamical groups. One previously known object in the field, 2015 GK56, displays a large-amplitude, structured lightcurve consistent with a contact binary. The results indicate that the characteristic color distribution of TNOs extends to smaller sizes than previously studied, suggesting a primordial origin rather than size-dependent collisional processing.

That last clause is the scientific spine of the hed. Primordial origin. Not a size-dependent paint job from collisions. The small ones still look like the sequence the large ones already taught us.

Morgan’s quote in the NASA release is the one that gives the hed its verb:

“You could imagine a scenario where getting knocked around and fragmented would change the surface composition, and then you would see a different surface color for tiny TNOs compared to their larger siblings. So it's really fascinating to see that the smallest objects are somehow 'remembering' and preserving the history of how they were made,”

She’s a PhD candidate at Northern Arizona University in Flagstaff. The NAU Review’s 8 September write-up puts more of her voice next to the same result. “For these objects to have a lot of the same color and shape, it’s telling us that they have probably sat there and not been very collisionally active,” she said. “The models say they should have been altered by collisions, but that’s just not what we’re seeing.”

And later, on what the mismatch means for the models: “The models tell us TNOs collided with other bodies a lot, but our observations tell us the history of the outer solar system may have been less collisional. When the data are inconsistent with the existing models, it means we need to refine and re-test the models. It also means we need to analyze more data, and we will; there’s terabytes and terabytes of it out there.”

Terabytes remaining isn’t a throwaway line. It’s a statement that Tuesday’s papers are a cut through a larger pile, not the last word from the same campaign.

What David Trilling adds

David Trilling, co-author and a professor of physics and astronomy at Northern Arizona University, puts the dynamical-hot population in one sentence in the NASA release:

“These dynamically ‘hot’ TNOs retain a signature of where they were born, even though they’ve been orbitally scrambled since then,”

Birth neighbourhood still readable in the surface fingerprint. Orbits scrambled. Color memory intact enough to say so out loud.

The NAU Review gives him a longer frame on what the overall non-variation means for solar-system history. “We thought we’d see lots of variation in the color and size of these TNOs, and we didn’t,” Trilling said. “That means there’s something not quite right about our understanding of how the solar system got to where it is now. Is it an evolutionary misunderstanding? Did we not get the upheaval right? The previous science doesn’t seem to capture how we got where we are today.”

Those are questions on the record, not answers anybody should invent. Both the “hot” and “cold” populations, NASA notes, seem to keep the same colors as when they were formed, with little change since the birth of the solar system.

What Marielle Eduardo’s size-distribution paper says

University of Victoria PhD candidate Marielle Eduardo led the study on size distribution. Her paper is Marielle R. Eduardo et al. 2026, The Astronomical Journal 172 187 — the luminosity function of ultrafaint TNOs detected by JWST.

From the abstract: a definitive discovery of 27 trans-Neptunian objects using JWST’s Near-Infrared Camera. The team used a shift-and-stack technique and a machine-learning network geared specifically to identifying false-positive detections in JWST images produced through that shift-and-stack process. They achieved a 40% detection threshold of m_F150W2 = 28.8 mag — corresponding to about m_r ∼ 29.8 mag — across a sky area of 0.05 deg². That marks the deepest solar system survey to date, reaching magnitudes that let them explore never-before-seen regions of the TNO size distribution.

The faintest detection has m_F150W2 = 29.3 mag and a diameter of ∼10 km, assuming 15% albedo. Hold that number. You’ll want it next to NASA’s 5-kilometre claim in a minute.

Within the sample, both the cold and hot TNO subpopulations exhibit a power-law slope. The distribution of apparent magnitudes of the nominal sample is well fit by a single power law. The dynamically hot and cold subsamples in the discovery set are consistent with the same power law, suggesting that the planetesimal formation process yields similar slopes despite the differing disk conditions at the presumed formation regions of the two populations — roughly the ∼25 au and ∼45 au neighbourhoods in the abstract’s framing.

Eduardo’s quote in the NASA release is the formation-process line:

“It's very interesting that the process of planetesimal formation ends up producing the same distribution of sizes for both cold and hot populations, despite forming in different regions of the early solar system. The process seems to be insensitive to disk conditions, producing similar planetesimal sizes whether the disk is hot or cold, and dense or fluffy,”

Insensitive to disk conditions. Same size distribution from different regions. That’s a strong claim about the physics of how planetesimals assemble, and it’s sitting on a sample of 27 ultrafaint detections in 0.05 square degrees. Deep. Narrow on the sky. Powerful at the faint end.

Twenty-seven new dim worlds — and the fireflies-on-the-Moon line

Webb discovered 27 new, remarkably dim TNOs. NASA’s image for how faint one of them is: equivalent to standing on Earth and seeing a small swarm of fireflies on the Moon. The NAU Review puts a similar image in Morgan’s mouth — standing on Earth and trying to measure a swarm of fireflies on the Moon without magnification. Same metaphor family. Same point. These aren’t backyard objects.

Researchers also found fewer of these very small bodies than they expected based on some planet-formation models. That “fewer than expected” result is as load-bearing as the color memory. Models that predicted a steeper pile-up of tiny fragments don’t match the faint-end census this survey actually counted.

Now the diameter conflict, labelled, because the sources don’t say the same number and you shouldn’t mash them.

NASA Science release: the smallest one they observed has a diameter of about 3 miles (5 kilometres), which is about five times smaller than what is possible to detect with the most sensitive ground-based telescopes. Prefer that 5-kilometre figure when you’re quoting the agency release’s “smallest observed.”

Eduardo et al. abstract: the faintest detection has m_F150W2 = 29.3 and a diameter of ∼10 km assuming 15% albedo. That’s a diameter tied to an albedo assumption on the faintest magnitude, not necessarily the same object-as-labelled as NASA’s “smallest.”

NAU Review: the smallest among them about 5 to 10 kilometres, or six miles, in diameter — about five times smaller than what is possible to detect with ground-based telescopes. That range is the honest summary when you’re holding both agency and paper language at once.

Albedo assumptions move diameters. A darker surface needs a larger body to produce the same brightness; a brighter surface can be smaller. When you see 5 km next to ∼10 km in the same news week, keep the source and the albedo note on the number. Don’t average them into a fake third kilometre.

What the teams thought they would see — and didn’t

Walk the prior expectation one more time, because it’s the hinge.

Expectation: small TNOs from both populations undergo many collisions. Collisions change surfaces. Tiny TNOs should show different surface colors than larger TNOs.

Observation: small bodies look like larger counterparts. Color relationships hold. Cold classicals stay on a narrow reflectance sequence down to small sizes. Dynamically excited objects keep a broader mix that still matches the multiple compositional classes already known at larger sizes. Lightcurves mostly low-amplitude. One known contact-binary candidate in the field with a structured large-amplitude curve.

Implication, in NASA’s own fork: collisions are not changing surfaces significantly — fewer collisions than expected, or primordial compositions retained somehow. Mystery open.

Morgan’s NAU line about models saying collisionally altered surfaces, and data saying otherwise, is the same hinge in campus English. Trilling’s “something not quite right about our understanding of how the solar system got to where it is now” is the same hinge in faculty English. Eduardo’s similar size distributions across hot and cold, despite different formation regions, is the size-axis version of the surprise.

You could ask whether “remembering” is a metaphor too soft for a science page. Morgan used it in quotes. The NASA hed used it. It’s here because the sources used it — not because photometry needed poetry.

Why beyond Neptune still matters for planets like Earth

The planetesimal story isn’t only an outer-solar-system curiosity. Dust and pebbles become city-sized building blocks. Building blocks merge into planets — or, beyond Neptune, they don’t. The frozen population is a leftover stage that the inner system finished and the outer system kept.

Morgan, in the NAU Review: “TNOs hold some of the history of the formation and evolution of the outer solar system. They tell us what was here at the beginning, and more broadly, they give us clues into how other planetary systems are formed.”

That’s the comparative-planetology stake without inventing an exoplanet detection in this survey. If planetesimal sizes come out similar whether the disk is hot or cold, dense or fluffy — Eduardo’s framing — then the assembly physics may be more universal than a single disk’s local weather. If colors at ≤40 km still track the sequences of larger bodies, then collisional grinding may not be the dominant paintbrush at these sizes in this region. Both results constrain models people will run for other disks.

The migration backdrop stays what the release already says: hot TNOs formed between current Uranus and Neptune locations and were pushed outward when the outer gas giants migrated. The migration is the backdrop for the dynamical classes. The new data are the colors and the faint-end sizes.

How faint is faint — and how small is the patch

0.05 square degrees is a tiny postage stamp on the celestial sphere. The Eduardo abstract is explicit about that area and about the 40% detection threshold at m_F150W2 = 28.8. Deep surveys often trade width for depth. This one went deep.

Shift-and-stack is the practical reason faint movers become findable. You take many frames, shift them at the rates distant solar-system objects would move, stack so a real TNO builds up as a point while stars and galaxies smear or subtract away, then filter false positives — here with a machine-learning network tuned to the junk that shift-and-stack itself produces. The abstract credits that pipeline for the definitive set of 27.

Faint solar-system objects move. Stars and galaxies, for practical purposes in a short visit, don’t move the same way. Stacking at many rates manufactures artifacts that look like detections if you’re not ruthless about vetoes. Eduardo et al. put a machine-learning network on that junk specifically — false positives born from the shift-and-stack process itself. The 40% detection threshold at m_F150W2 = 28.8 over 0.05 deg² is the completeness hinge. It means the survey is quoting a magnitude where detection probability hits 40 percent, not a claim that every object brighter than 28.8 was caught. Corresponding roughly to m_r ∼ 29.8 puts the depth in a more familiar optical magnitude language for people who live in ground-based catalogs.

The faintest detection at m_F150W2 = 29.3 is past that threshold, as faintest detections often are — a real object that cleared the filters even where average completeness is falling. Diameter ∼10 km at 15% albedo is the abstract’s conversion for that faintest magnitude. Change the albedo, change the kilometre. NASA’s separate “smallest observed about 5 kilometres” line remains the agency release’s claim for the smallest in the set they highlight. Both can sit on the same page if you refuse to average them into a fake third number.

Hubble’s role on the color paper is recovery and optical leverage: thirteen of the JWST detections recovered in HST imaging, enough to build optical–NIR colors across 0.35–3.2 μm. Simultaneous, coordinated, nearly simultaneous — the papers and the NASA release all stress the pairing. Visible plus infrared. Size distribution plus composition fingerprint.

Ground-based telescopes, even the most sensitive ones, lose at roughly five times larger than the smallest NASA cites. That’s why the fireflies-on-the-Moon line isn’t colour writing. It’s an attempt to make magnitude 29-class objects feel as absurd as they are.

Cold classicals, excited populations, and a contact binary in the field

Keep the dynamical dictionary straight when you read Morgan’s color results next to Eduardo’s luminosity function. Cold classicals are the dynamically cold set — original, relatively circular orbits in the plane. Dynamically excited is the broader label that covers the hot population and related scrambled orbits. Morgan’s abstract is careful: small cold classical discoveries sit on a narrow color sequence that matches larger cold classicals, with no evidence of a break at smaller sizes. Dynamically excited discoveries spread across a broader color range that still matches the multiple compositional classes already known among larger excited TNOs.

That split isn’t a contradiction with Eduardo’s similar size distributions. Color can remember birth chemistry while size slopes look alike across regions. Different axes. Same survey. The cold sequence staying narrow down to ≤40 km is the strongest single argument in the color paper against size-dependent collisional repainting for that population. The excited population’s broader mix is still “like the large ones,” not “a new tiny-only paint job.”

2015 GK56 is the exception that proves the lightcurve rule was written carefully. Most amplitudes in the sample were generally low. GK56, already known and sitting in the field, shows a large-amplitude, structured lightcurve consistent with a contact binary — two lobes stuck together, rotating as a pair. The abstract doesn’t promote every small TNO into a binary. It flags one structured curve.

Optical–NIR color spanning 0.35–3.2 μm is the wavelength lever. Hubble’s optical recovery plus Webb’s near-infrared photometry is how you turn two faint points of light into a composition fingerprint instead of a single-band brightness. Thirteen recoveries out of the JWST-detected set are enough to populate that comparison for the small-size bin. The other discoveries still feed Eduardo’s luminosity function even when optical recovery isn’t complete for every object.

Fewer small bodies than some models wanted

Planet-formation models aren’t a single spreadsheet. Some of them predicted more tiny debris at the faint end once collisions had ground larger planetesimals down. The survey found fewer of those very small bodies than expected based on some of those models. NASA states that without naming a specific competing paper in the release text. The result is directional: the faint end isn’t as crowded as those particular expectations.

Pair that shortfall with the color memory and you get a double constraint. Not only do the small survivors look compositionally like their larger siblings; there also aren’t as many of the smallest survivors as grinding-heavy stories wanted. Trilling’s NAU questions — evolutionary misunderstanding? upheaval not got right? — sit next to that double constraint without anyone picking a favourite theory.

Morgan’s “less collisional” outer-solar-system history is one reading the data allow. Retained primordial compositions despite collisions is the other fork NASA leaves open. Refining and re-testing models is the operational next step she names, along with analysing more of the terabytes already in hand. Tuesday is a publication date, not a campaign end.

Power-law slopes for cold and hot subsamples looking consistent is the size-distribution surprise in one clause. A single power law fitting the nominal sample’s apparent-magnitude distribution is the other. Eduardo’s quote about hot or cold, dense or fluffy disks producing similar planetesimal sizes is the plain-language version of that statistical result. Formation physics that doesn’t care as much as theorists hoped about which annulus of the early disk you were born in.

Answer from this week’s papers, compressed: yes on the color family resemblance; yes on similar size-distribution behaviour across cold and hot; no on the crowded faint-end debris pile some models wanted; open on the exact collision physics.

The agency and campus map

Media contacts on the NASA Science page: Claire Andreoli at Goddard; Ann Jenkins and Christine Pulliam at the Space Telescope Science Institute in Baltimore. Artwork credit on the release: NASA, ESA, Leah Hustak (STScI). Editor listed on the page: Andrea Gianopoulos. Location: NASA Goddard Space Flight Center. Last updated 8 September 2026.

Campus side: Northern Arizona University in Flagstaff for Morgan and Trilling; University of Victoria PhD candidates under National Research Council of Canada guidance for the Eduardo-led size work. Jill Kimball is the NAU Communications contact on the Review piece. Complementary studies, same Tuesday in The Astronomical Journal — volume 172, papers 187 and 188.

Hubble has been operating for over three decades and continues to make discoveries that shape fundamental understanding of the universe — NASA’s own institutional sentence. Webb is framed as the world’s premier space science observatory, solving mysteries in the solar system and beyond. For this story, the relevant fact is narrower: both telescopes on one patch of sky, visible and infrared, catching planetesimals that ground glass can’t.

Putting Tuesday’s papers on one table

Object class: Trans-Neptunian Objects — small, faint, icy, beyond Neptune; most more than 100 million times dimmer than unaided-eye objects.

Method: deepest TNO survey to date; simultaneous Hubble visible and Webb infrared on a patch of sky; colors as surface-composition fingerprints; sizes; orbits.

Sample: 27 newly discovered tiny dim TNOs from Webb, characterized jointly; Morgan paper recovers 13 JWST detections in HST for optical–NIR colors on small (≤40 km) bodies.

Dynamical classes: cold — original relatively circular orbits in the plane; hot — formed between current Uranus and Neptune locations, pushed out during giant-planet migration, highly elliptical, moving in and out of the plane.

Color result: small bodies follow the same relationships as larger family members; cold classicals narrow and consistent with a single reflectance sequence; no evidence of change at smaller sizes; dynamically excited objects broader, matching multiple compositional classes at larger sizes; lightcurve amplitudes generally low; 2015 GK56 large-amplitude structured lightcurve consistent with contact binary; color distribution to smaller sizes suggests primordial origin rather than size-dependent collisional processing.

Size result: overall size distributions for both populations surprisingly similar; planetesimal formation process appears insensitive to disk conditions; fewer very small bodies than some planet-formation models expected; Eduardo survey depth m_F150W2 = 28.8 at 40% detection threshold over 0.05 deg²; faintest detection m_F150W2 = 29.3, ∼10 km at 15% albedo per abstract; NASA smallest observed about 5 kilometres.

Quotes to keep: Morgan on remembering how they were made; Trilling on hot TNOs retaining a birth signature after orbital scrambling; Eduardo on similar size distributions from different early-disk regions.

Open mystery: why surfaces aren’t collisionally reset at small sizes — fewer collisions than expected, or primordial compositions retained.

How to read the papers

If you only skimmed a hed about fireflies on the Moon, you might’ve thought the news was merely that Webb can see very faint rocks. The primary text is sharper. The news is that the faint rocks still look like their larger siblings, that hot and cold populations share size-distribution behaviour they weren’t guaranteed to share, and that the faint end is less crowded than some models wanted.

If you only skimmed “they remember,” you might’ve thought the team claimed literal memory. They didn’t. They claimed surface colors that preserve formation history against the expectation of collisional repainting. Metaphor in the quote. Photometry in the papers.

If you only skimmed campus excitement about refining models, you might’ve missed the quantitative spine: 27 objects, 0.05 deg², magnitude 28.8 threshold, thirteen optical recoveries, ≤40 km color sample, power-law consistency across dynamical classes. Excitement without the ledger isn’t enough.

Where this stands

As of NASA Science’s 8 September 2026 release and the same day’s pair of Astronomical Journal papers, Hubble and Webb have been used together for the first time on Trans-Neptunian Objects, including some of the smallest and faintest ever directly seen. Twenty-seven new dim TNOs sit in the discovery set. Small-body colors track larger-family relationships. Cold and hot size distributions look alike enough for Eduardo to call the formation process insensitive to disk conditions. Fewer tiny bodies showed up than some planet-formation models expected. The collision-versus-primordial mystery remains open. Morgan says the smallest objects are somehow remembering how they were made. Trilling says the hot ones still carry a birth signature after being orbitally scrambled. The data pile still has terabytes to go.

Primary sources: NASA Science, “NASA’s Hubble, Webb Find Far-out Solar System Objects ‘Remember’ Past,” 8 September 2026 — https://science.nasa.gov/missions/hubble/nasas-hubble-webb-find-far-out-solar-system-objects-remember-past/. Marielle R. Eduardo et al. 2026, AJ 172 187 — https://iopscience.iop.org/article/10.3847/1538-3881/ae907f. Anastasia N. Morgan et al. 2026, AJ 172 188 — https://iopscience.iop.org/article/10.3847/1538-3881/ae9084. Corroboration: NAU Review, “Far-out space objects ‘remember’ their past,” 8 September 2026 — https://in.nau.edu/news/trans-neptunian-objects/.