Ancient Tooth Proteins Tell Tales

An analysis of six ancient Homo erectus tooth proteins from three locations in China, a Denisovan tooth protein, and some modern human and animal tooth proteins reveal some notable insights. Bernard's Blog has the story. Proteins in teeth are easier to recover than ancient DNA and can serve as a proxy for it in cases like these.

Interestingly, it implies that a tooth enamel protein found in 1% of humans in or near the former Denisovan range, and Denisovans have a tooth enamel protein also found in Chinese Homo erectus, probably due to introgression from Homo erectus to Denisovans and then from Denisovans to modern humans, in the view of the researchers.

John Hawk has a discussion of the paper, however, that interprets the data differently:

All six teeth share two derived amino acid changes, both in the sequence of an enamel matrix protein known as ameloblastin, or AMBN. One of these hasn’t before been seen in hominins: a change from alanine to glycine at position 253 of the sequence, or A253G. The other is 20 positions downstream, swapping in valine for the ancestral methionine, M273V. The DNA mutation encoding this change is shared by both the Denisova 3 and Denisova 25 genomes. The M273V amino acid change itself is in the Harbin and Penghu 1 dental proteomes—part of why they align with Denisovans. A number of genomes from modern people also share this change, possibly from Denisovan ancestors.

The hypothesis presented by Fu and coworkers is that Homo erectus was the source of M273V, and its presence in Denisovans is a result of introgression. In support of this idea of introgression, they note earlier research on the Denisova 3 and Denisova 25 genomes that suggests a contribution from a “superarchaic” source population. Many—including me—have speculated that this superarchaic ancestry came from H. erectus. Fu’s team may have just proved it.

But I don’t think these teeth are Homo erectus.

Their estimated ages, all around 400,000 years ago, are prime Denisovan time. Fu and coworkers find that all the teeth share a derived link with later Denisovan genomes. For me, the most likely hypothesis is that these teeth come from a population within the Denisovan branch of humanity.

This situation is basically the same as the Sima de los Huesos fossils. Those remains are around 430,000 years old. Those fossils look like Neanderthals in some subtle ways, but until DNA was recovered from them, many researchers considered them to be part of a different group, often called Homo heidelbergensis. DNA revised both the timeline and their identification.

It may seem heretical, but I think protein data may be about to do the same for fossils from East Asia.

How Common Are Hallucinated Citations In Academic Papers?

Academic papers written with AI that contain hallucinated citations are on the rise.

Large language models (LLMs) are known to generate plausible but false information across a wide range of contexts, yet the real-world magnitude and consequences of this hallucination problem remain poorly understood. 

Here we leverage a uniquely verifiable object - scientific citations - to audit 111 million references across 2.5 million papers in arXiv, bioRxiv, SSRN, and PubMed Central. We find a sharp rise in non-existent references following widespread LLM adoption, with a conservative estimate of 146,932 hallucinated citations in 2025 alone. 

These errors are diffusely embedded across many papers but especially pronounced in fields with rapid AI uptake, in manuscripts with linguistic signatures of AI-assisted writing, and among small and early-career author teams. At the same time, hallucinated references disproportionately assign credit to already prominent and male scholars, suggesting that LLM-generated errors may reinforce existing inequities in scientific recognition. Preprint moderation and journal publication processes capture only a fraction of these errors, suggesting that the spread of hallucinated content has outpaced existing safeguards. Together, these findings demonstrate that LLM hallucinations are infiltrating knowledge production at scale, threatening both the reliability and equity of future scientific discovery as human and AI systems draw on the existing literature.

Zhenyue Zhao, Yihe Wang, Toby Stuart, Mathijs De Vaan, Paul Ginsparg, Yian Yin, "LLM hallucinations in the wild: Large-scale evidence from non-existent citations" arXiv:2605.07723 (May 8, 2026).

Extended Tully-Fisher Relations

Stacy McGaugh's latest post explores in greater depth a major paper is was a co-author of in March of this year which demonstrates how tightly the baryonic Tully-Fisher relation holds from the smallest galaxies to the largest structures in the universe with very little intrinsic scatter. It also highlights that fact that while this scaling law is displaced in galaxy cluster and larger structures, that the same basic scaling law and slope, albeit displaced, holds there as well, as shown in the second chart below.

My take is that the geometry of the mass distribution or interstellar media in clusters could explain the displacement.

The Limits Of Post-Newtonian Approximations

Recognizing that non-perturbative GR effects can be important in circumstances conventionally considered to be non-relativistic, is a big step forward. Deur hit on this a long time ago, and this paper, independently, and indeed without citing to Deur, reaches the same conclusion. This could provide an answer to key unsolved problems in astrophysics, especially dark matter and dark energy phenomena.

Post-Newtonian theory is considered a reliable effective expansion of General Relativity in the weak-field and slow-motion limit. We argue that such a belief is misplaced. 

In generic many-body relativistic dynamics, the absence of globally conserved charges in the region of interest and non-integrability can drive strong sensitivity to angular-momentum exchange across inhomogeneous curvature, invalidating naive power counting in an effective theory expansion. 

Building on general lessons from effective field theory, we derive an explicit breakdown criterion that delineates when post-Newtonian truncations become unreliable despite small local potentials and velocities. This supplies a controlled systematic for weak-field mass inference, relevant to the dark matter puzzle in astrophysics and cosmology.

Marco Galoppo, Giorgio Torrieri, "When Weak Fields Arent Weak: Post-Newtonian effective theory and the Dark Matter Puzzle"  arXiv:2605.13557 (May 13, 2026) (Honorable mention, Gravity Research Foundation essay competition 2026).

Some Astrophysics Quick Hits

Many good theories start from the foundation of patterns in the empirical data.

The standard theory of galaxy formation predicts that all galaxies should contain dark matter, yet a handful of recently discovered galaxies appear to lack it, challenging our understanding of galaxy formation. We investigate whether such dark-matter deficient objects can be identified from their baryonic properties alone, analogously to the radial-acceleration relation, which tightly links baryon and dark matter distributions in spiral galaxies. 

Using a sample of ultra-diffuse and dwarf spheroidal galaxies -- systems whose baryonic properties resemble those of the confirmed dark-matter-deficient galaxies -- we systematically search for a formula to predict baryonic fractions from stellar mass, effective radius, distance to the host, and the host's baryonic mass. We find that baryonic fraction correlates most strongly with the gravitational acceleration expected from baryons alone, a(bar), or equivalently, with mean surface brightness, following an approximately a(bar)^−1 dependence. This scaling resembles the radial-acceleration relation but differs in functional form and applies to a different galaxy population. 

Strikingly, the dark-matter-deficient galaxies occupy the extreme end of the correlation. This suggests that they result from standard formation processes operating at unusual intensities rather than from exotic mechanisms. Importantly, the correlation predicts that all ultra-diffuse galaxies brighter than approximately 25 mag arcsec^−2 in the g-band should have very low dark matter content, offering a straightforward observational criterion for identifying these rare objects.

Michal Bílek, "A correlation predicting galaxies without dark matter" arXiv:2605.11070 (May 11, 2026).

Footnote: "cosmic noon" (i.e. the middle of the age of the universe measured in years) is sometimes operationally defined as "0.5 < z < 3".

Deur's approach to gravity provides a mechanism that explains the seeming observational preference for planar structures over spherical ones at galactic and larger scales.

An update of the evidence that radio galaxies and clusters of galaxies are more common than average near the plane of the de Vaucouleurs Local Supercluster shows that in the distance range 100 to 200Mpc objects whose positions are correlated with the plane of the Local Supercluster include galaxies that are exceptionally luminous at two microns, radio galaxies, and clusters of galaxies. There can be little doubt about this property of cosmic structure. I also argue for detection of this correlation for the galaxies at 400Mpc distance that are exceptionally luminous at two microns. 

It will be interesting to learn whether these results are expected in the standard cosmology.

P. J. E. Peebles, "The Extended Plane of the Local Supercluster" arXiv:2605.11184 (May 11, 2026).

We study how constraints on the abundance of ultralight axions (ULAs) from cosmic microwave background (CMB) data depend on their nonlinear modelling. We focus on the axion mass range 10^−25 ≤ m/eV ≤ 10^−23, where the axion Jeans scale falls in the quasi-linear regime probed by CMB lensing, making constraints highly sensitive to the choice of nonlinear prescription. 

We show that the inferred constraints depend significantly on the choice of nonlinear model, which must therefore be treated carefully. Performing Markov Chain Monte Carlo (MCMC) analyses with Planck 2018, ACT DR6 and DESI DR2 BAO data, we find naive nonlinear modelling of non-cold matter can produce an artificial preference for a subdominant ULA dark matter component with mass m ≈ 10^−24 eV. This arises from a lensing-like enhancement of the CMB power spectrum.

Lauren Gaughan, Anne M. Green, Adam Moss, "Ultra-light axion constraints from Planck and ACT: the role of nonlinear modelling" arXiv:2605.12054 (May 12, 2026).

Noting this paper for future reference. Here is the chart from it that is most interesting to me:

We present results from the Big Mysteries Survey, a large-scale survey conducted through the American Physical Society's Physics Magazine on foundational and controversial topics in contemporary physics. The survey provides a snapshot of physicists' views on issues in cosmology, black-hole physics, quantum mechanics, quantum gravity, and anthropic coincidences. A central finding is that several positions often described publicly as field-wide ``consensus'' views are, in practice, supported by much narrower majorities or by pluralities rather than majorities.

Niayesh Afshordi, Phil Halper, Matteo Rini, Michael Schirber, "Big Mysteries Survey: Physicists' Views on Cosmology, Black Holes, Quantum Mechanics, and Quantum Gravity" arXiv:2605.11058 (May 11, 2026).

Physics Quick Hits

Lots of interesting papers today. Little time to write, so only minimal commentary for now.

The reactor antineutrino anomaly still isn't real.

The Reactor Antineutrino Anomaly refers to the deficit observed between the average event rate measured in reactor antineutrino experiments with respect to the theoretical prediction. This anomaly was first identified in 2011 (2.5σ) as a consequence of the Huber-Muller reactor antineutrino flux calculation. It was thought to be resolved in 2021 as a result of new reactor antineutrino flux calculations, with a reduction to about 1σ. In this work, we examine the latest reactor antineutrino flux calculation published in 2023 by a French research group. This work represents the first summation model to include a comprehensive uncertainty budget. The result indicates a revival of the Reactor Antineutrino Anomaly at the level of 2.2σ. We also consider the usual simplest explanation of the Reactor Antineutrino Anomaly by active-sterile neutrino oscillations. We present the constraints on the oscillation parameters and we derive a tension of 3.8σ with the results of gallium source experiments (Gallium Anomaly) taking into account also the solar neutrino and KATRIN bounds, that of the combined short-baseline reactor spectral ratio measurements, and that of the Daya Bay search for a sub-eV sterile neutrino. Since the tension may be due to underestimated systematic uncertainties and the main tension is between the gallium data and the other data, we finally present the results of a global analysis with enlarged gallium uncertainties, which reduce the global tension to 1.3σ.

C. Giunti, Y.F. Li, R.P. Zhang, "Revival of the Reactor Antineutrino Anomaly" arXiv:2605.10353 (May 11, 2026).

Intriguing.

The charged-lepton Koide relation remains a striking empirical regularity in Standard-Model flavor data. We prove that for any positive mass set with Koide ratio Q0, the one-particle extension Q(m1,…,mN,x) has a unique global minimum Qmin=Q0/(1+Q0) at m∗=[(∑imi)/(∑imi‾‾‾√)]2. This exact kinematic result defines a unique extension benchmark. For the measured charged leptons it gives mℓ∗=1.25534(16)GeV and Qexp4,min=0.3999978(43); in the ideal Koide limit QKℓ=2/3, the corresponding minimum is exactly 2/5. In the effective-participant language Neff≡1/Q, the optimal one-particle extension increases Neff by one, while the equal-k multiplet extension increases it by k. The one-particle Neff profile is exactly Lorentzian in a dimensionless share-mismatch coordinate u, which we interpret kinematically rather than dynamically. Using charged-lepton pole masses with the PDG~2024 own-scale MS⎯⎯⎯⎯⎯⎯⎯⎯⎯ charm mass gives Q(e,μ,τ,c)=0.4000025(64), i.e. 11.7ppm above the measured-input benchmark and 6.2ppm above 2/5. This intentionally mixed-definition comparison is treated only as a phenomenological coincidence. To calibrate it within a stated benchmark class, we perform an exhaustive common-scale scan over non-neutrino Standard Model 2-body and 3-body seeds with one added mass. The charged-lepton-plus-charm continuation ranks 33/12,720 in the raw trial set, 24/2,640 after collapsing repeated scale realizations, and 6/756 within the fermion-only collapsed subset. We present the charm case as an empirically calibrated example of the theorem, not as a dynamical flavor model.

K. Hübner, "A minimization theorem for the Koide ratio and its Standard Model calibration" arXiv:2605.09651 (May 10, 2026).

So what?

Koide's charged-lepton relation suggests that (me‾‾‾√,mμ‾‾‾√,mτ‾‾‾√) is the natural family vector. We construct an effective compact-cycle model in which this vector is sampled from one real amplitude Z(ϕ) on an internal circle, while the masses are quadratic overlaps, ma∝|Z(2πa/3)|2. The amplitude is built from the two lowest antiperiodic modes on the circle; their symmetric square is periodic and gives the minimal three-harmonic family space e^iϕ,1,e^−iϕ. A reality condition together with the requirement that the amplitude comes from the square of one two-component spinor fixes the relative weights required by Koide's 45º geometry. The remaining orientation angle is fixed by matching one C3 family shift to transport on the full circle: integrating out the higher Fourier harmonics gives the Berry dressing that enters the determinant term and selects θℓ=−2/9. Using me and mμ as inputs, the model predicts mτ=1776.97MeV.

Kirill Shulga, "Charged-Lepton Koide Geometry from a Green-Dressed Compact Family Cycle" arXiv:2605.10245 (May 11, 2026).

Similar to another recent paper.

We show how, by exploiting the process of Coherent Elastic neutrino (v) Nucleus Scattering (CEvNS), neutrinos produced by nuclear reactor experiments appear to corroborate the evidence of the so-called X17 particle, which has been invoked to explain the ATOMKI anomaly. We base our analysis primarily on CONUS+ and Dresden-II data, which, when combined with CEvNS data from COHERENT and neutrino oscillation data from IceCube, single out a unique region of couplings to neutrinos and nuclei.

Johan Rathsman, Joakim Cederkäll, Yasar Hicyilmaz, Else Lytken, Stefano Moretti, "The X17 Existence Hinted at by Nuclear Reactor Neutrinos" arXiv:2605.10689 (May 11, 2026) (Short version of 2603.15246 using a different model for the X17).

Neutrinos do not have negative mass, so something isn't quite right in the model to estimate its masses from cosmology.

Recent baryon acoustic oscillation (BAO) distance measurements, when combined with Cosmic Microwave Background (CMB) observations in the ΛCDM framework, lead to a preference for negative neutrino masses. We investigate whether this neutrino mass anomaly can be alleviated by a class of astrophysically motivated reionization histories. Using a frequentist analysis, we find that some reionization histories can move the best-fit value of ∑mν to a positive value and bring ∑mν ≃ 0.06 eV into the 95% confidence interval. To separate the effect of the total optical depth from that of the details of the reionization history, we compare a high-τ history with a two-step tanh-like reionization history of the same τ. The resulting Δχ2(∑mν) profiles are nearly identical. This indicates that the effect is mainly driven by the total optical depth, while the details of the reionization history play only a minor role.

Yi Cheng Dai, Wei Liao, "Reionization History and Neutrino Mass" arXiv:2605.10116 (May 11, 2026).

As expected.

Modified Newtonian Dynamics (MOND) is a paradigm that can do away with dark matter at galaxy scales, but displays a residual missing mass discrepancy in galaxy clusters. Prompted by the updated JWST-based gravitational lens model of the Bullet Cluster, I confirm here that this cluster exhibits the same residual missing mass discrepancy as other clusters of similar mass in the MOND context. Moreover, this missing mass should be mostly collisionless, since it is centred on the galaxies of the Bullet Cluster.

Benoit Famaey, "On the residual missing mass of the Bullet Cluster" arXiv:2605.10022 (May 11, 2026).

Color me skeptical. It will take a closer look to poke holes in it, however. I suspect that while it may point out problems in toy-model MOND and some other models, this data could actually point the way towards a better modified gravity theory rather than towards dark matter particles which have myriad problems of their own that are ignored in this study.

Modified gravity theories such as Modified Newtonian Dynamics (MOND) and Scalar-Tensor-Vector Gravity (STVG) have been proposed as alternatives to dark matter, but decisive tests have been hindered by degeneracies between baryonic structure and gravitational laws. Here we break this degeneracy using independent, high-precision constraints: the Milky Way radial rotation curve, vertical phase-space spirals from Gaia, and a broken-exponential stellar disk. A joint reconstruction of the radial and vertical gravitational fields reveals a structural inconsistency in modified gravity -- no model can simultaneously reproduce both observations. Our results strongly disfavor MOND at >13σ and STVG at >4σ. In contrast, dark matter halo models naturally explain the observations, providing a self-consistent test of gravity on galactic scales.

Zheng-long Wang, Yue-Lin Sming Tsai, Lan Zhang, Yin Wu, Haining Li, Xiang-Xiang Xue, Hongsheng Zhao, Yi-Zhong Fan, "Milky Way Dynamics Favor Dark Matter over Modified Gravity Models" arXiv:2605.10857 (May 11, 2026).

A Notable Coincidence Related To The Proton Mass And Charge Radius

There is a functional relationship between the mass of the proton and the charge radius of the proton that is consistent with experimental measurements of those quantities, that doesn't have an obvious cause.

The simple proton mass and charge radius relationship

From @dandb at Physics Stack Exchange on May 5, 2016 (ten years ago). This can also be stated another way:

The charge radius of the proton is almost exactly four times the reduced Compton wavelength of the proton.

The reduced Compton wavelength is a natural representation of mass on the quantum scale and is used in equations that pertain to inertial mass, such as the Klein–Gordon and Schrödinger equations.

Equations that pertain to the wavelengths of photons interacting with mass use the non-reduced Compton wavelength. A particle of mass m has a rest energy of E = mc^2. The Compton wavelength for this particle is the wavelength of a photon of the same energy.

The reduced Planck's constant, h-bar, is Planck's constant divided by 2π. So, this relationship could also be stated as r = 2h/πmc, for Planck's constant h, the proton charge radius r, and the proton mass m.

This relationship is consistent with experimental measurements made to 0.05% precision

The uncertainty in the "predicted" value of the charge radius of the proton from this relationship, which is 0.84124 to five significant digits, is negligible, because the speed of light (c) and the reduced Planck's constant (h-bar) are quantities used to define SI units of measurement which are thus known "exactly" in terms of SI units of measurement, and the mass of the proton is known to the exquisite precision of about one part per hundred billion. See the Particle Data Group table of physical constants.

The Particle Data Group world average value is currently 0.8409(4) fm, i.e. a one sigma range of 0.8405 to 0.8413 This is a relative uncertainty of 0.048% (i.e. about one part per two thousand).

The PDG value is also consistent with a February 11, 2026 measurement of the charge radius of the proton with a relative uncertainty of 0.18% published in the prestigious peer reviewed journal Nature found it to be rp = 0.8406(15) fm, i.e. a one sigma range of 0.8392 to 0.8421 fm. 

So, the conjectured relationship is consistent with the experimentally measured value of the charge radius of the proton. 

At the time that this Physics Stack Exchange post was written, there was a discrepancy between the electron measurement of the proton charge radius and the muon measurement of the proton charge radius, but that has since been resolved. The muon measurement was found to be correct, and the electron measurement was found to have been incorrect due to experimental measurement errors not fully reflected in the stated uncertainty of the measurement.

This "prediction" is also notable because it is a testable hypothesis. As measurements of the proton charge radius grow more precise, we can find out if the experimentally measured value continues to be consistent with this prediction.

For example, if this hypothesis is merely numerology with no deeper meaning, it would be highly likely that it would grow less consistent with the experimental measurement if the experimental measurement's precision were increased by a factor of ten. And, in fact, experiments to do that are on the agenda of the physics community.

Analysis

What makes this relationship surprising?

Since the charge radius of the proton and the mass of the proton are both, in principle, derived quantities in the Standard Model, that this isn't actually a "coincidence" so much as it is a simple relationship arising from Standard Model physics whose source isn't trivially obvious.

The reason that it isn't trivially obvious is that the calculation of the mass and charge radius of the proton in the Standard Model are primarily functions at leading order of (1) the QCD coupling constant (which describes the strength of the "strong force") evaluated with non-perturbative QCD, (2) the mass of the up quark, (3) the mass of the down quark, and (4) the electromagnetic coupling constant. Yet, none of these experimentally measured physical constants have a functional relationship to Planck's constant or the speed of light.

There are comparatively minor contributions to these quantities that tweak their value beyond the leading order values from the masses of the other quarks (especially the strange quark), the weak force coupling constant, the W boson mass, and the CKM matrix elements (especially the  two elements of the nine elements in the matrix involving up-down quark transitions and up-strange quark transitions).

The reason that this relationship is surprising is that there is no known functional relationship between the reduced Planck's constant or the speed of light, and the other experimentally measured determinants of the proton mass and the proton charge radius (such as the Standard Model coupling constants, the quark masses, and the CKM matrix elements).

Three possible explanations

The stack exchange thread linked above contains some speculations as to why this is true, some more credible than others, but they are only speculations. For example, Michell Porter notes that:

Via P.R. Silva (eqn 6), I have run across a heuristic model of the nucleon in which M = 4/R (in natural units). Here R is the radius of the bag in the "bag model". See Xiangdong Ji, "Mass of the hadron", slide 20. I have not found where this argument originates, but a remark in a 1994 paper by Ji (see paragraph beginning "In the chiral limit...", on the final page) hints at it.

One possibility, which is to some extent the default one, is that this numerical coincidence of these two values has no deep meaning or connection and doesn't point to anything. In other words, this relationship just happens to hold for one hadron out of hundreds, for one of a large set of possible combinations of other physical constants that have no actually physical relationship to each other.

Another reason that this could be true is that the contributions of the experimentally measured constants cancel out in the combination of the proton mass and the proton charge radius, since the same experimentally measured constants enter into both calculations.

If true, this would suggest that should be a way of calculating the proton charge radius from first principles that more transparently and obviously reveals this cancelation.

This would be very interesting, would provide us to a deeper understand of the Standard Model and hadron physics. 

It would also suggest that this relationship ought be to generalizable in some way to the relationship between hadron mass and hadron charge radius for many hadrons (hadrons are composite particles made up of quark and/or gluons bound by the strong force of the Standard Model).

A calculation in this form would also have practical use, because the first principles Standard Model calculation of the proton mass has less than one part per thousand precision (vastly less than the precision of the experimentally measured value). And, in general, this would provide a quick and easy way to calculate hadron charge radii (which are no more precise than first principles calculations of hadron masses using current methods, see also here) which could then be compared to experimental measurements of hadron charge radii.

A third possibility, which would be even more grand, is that the values of the physical constants of the Standard Model that go into calculating the mass and charge radius of the proton actually have some deep functional connection to Planck's constant and the speed of light that has not previously been recognized or hypothesized.

The (Weak) Evidence For Extra Higgs Bosons

The 95 GeV bump is suspiciously close to the mass of the Z boson plus the mass of the b-quark. The 152 GeV bump is close to the mass of two W bosons reduced by the mass of two b-quarks. 

The implication of these coincidences is that there could be an explanation along the lines of diphoton signals that are missing decay products that prevent them from accurately reflecting the true source of the diphoton signals. 

Also, given the modest statistical significant of these alleged resonances, it could be that these are simply the product of statistical flukes in the background estimations or some other sort of measurement errors.

After the Higgs discovery, the question of whether particles beyond those of the Standard Model exist is more pressing than ever. In this context, the scalar sector is particularly promising, since it lies at the core of the internal problems of the Standard Model, while extensions of it allow us to resolve them and can provide explanations for Dark matter, non-zero neutrino masses, inflation etc. 

In these proceedings, we review the indications for new Higgs bosons at the electroweak scale with masses of ≈95 GeV and ≈152 GeV. These excesses are most significant in the di-photon channel but are supported by weaker-than-expected limits in other decay modes. 

While for the 95 GeV candidate the production mechanism is mostly unknown, the (hypothetical) 152 GeV Higgs is dominantly produced in association with leptons, (b) jets and missing energy, pointing towards the Drell-Yan production of an SU(2)L triplet with Y=0. Interestingly, this model predicts t→H±b with H±→WZ, which resembles the signature of tt¯Z production in the Standard Model and is in fact preferred by current data. 

Finally, we investigate the possibility that the significant tensions between the Standard Model predictions and the measurements in differential top-quark distributions are due to contamination from new physics involving both the 152 GeV and the 95 GeV scalar.

Andreas Crivellin, et al., "Indications for New Higgs Bosons" arXiv:2605.04233 (May 5, 2026) (Proceedings of the Corfu Summer Institute 2025 "School and Workshops on Elementary Particle Physics and Gravity" (CORFU2025)).

Cracking Linear Elamite

Four years ago, French archaeologist François Desset reportedly cracked the 4,000 year old Linear Elamite script, "[m]ade up of 77 signs – diamonds, curves, and other geometric patterns – the writing system comes from the Bronze Age civilisation of Elam" in Southwest Iran that collapsed long ago. The script was rediscovered in 1903, but has eluded decipherment until now. 

Elam is shown in red.

The Elamite language has partially been known long before this breakthrough from inscriptions of the language made in an Elamite cuneiform script, which was adapted from Akkadian cuneiform. This language was still spoken in the first century CE, and probably went extinct around the eleventh century CE. Most linguists consider it to be a language isolate (and in my review of this literature, I have not found that Elamo-Dravidian hypothesis to be a credible one).

Elamite is regarded by the vast majority of linguists as a language isolate,[31][32] as it has no demonstrable relationship to the neighbouring Semitic languages, Indo-European languages, or to Sumerian, despite having adopted the Sumerian-Akkadian cuneiform script.

An Elamo-Dravidian family connecting Elamite with the Brahui language of Pakistan and Dravidian languages of India was suggested in 1967 by Igor M. Diakonoff[33] and later, in 1974, defended by David McAlpin and others.[34][35] In 2012, Southworth proposed that Elamite forms the "Zagrosian family" along with Brahui and, further down the cladogram, the remaining Dravidian languages; this family would have originated in Southwest Asia (southern Iran) and was widely distributed in South Asia and parts of eastern West Asia before the Indo-Aryan migration.[36] Recent discoveries regarding early population migration based on ancient DNA analysis have revived interest in the possible connection between proto-Elamite and proto-Dravidian.[37][38][39][40] A critical reassessment of the Elamo-Dravidian hypothesis has been published by Filippo Pedron in 2023.[41]

Václav Blažek proposed a relation with the Semitic languages.[42]

In 2002 George Starostin published a lexicostatistic analysis finding Elamite to be approximately equidistant from Nostratic and Semitic.[43]

None of these ideas have been accepted by mainstream historical linguists.[31]

Desset accomplished this primarily by using proper names to decode the meaning of those signs and applying that method to ten new Linear Elamite texts inscribed on vases, according an April 28, 2026 story from France 24. This story of breakthrough is also reported in National Geographic, January 2026, pp. 110-131, entitled, "Decoding the Lost Scripts of the Ancient World", by Joshua Hammer. Hat top to Language Log. The Smithsonian magazine also has a recent article on the topic. It isn't entirely clear to me why this development is making headlines now, four years after the leading article on the topic was published (which I blogged at the time, see also an earlier post on a related topic).

Wikipedia (at the link above) explains that:

In 2022, Desset et al. (2022) argued that Linear Elamite is an alpha-syllabary, which would make it the oldest known purely phonographic writing system.[5] However, they admit that some logograms may have been used, although only rarely and not systematically, arguing that Elamite scribes rejected logographic writing in the 3rd millennium BCE.[30] Other researchers, such as the linguist Michael Mäder, dispute this, arguing that only around 70 percent of Linear Elamite characters are likely to be purely phonographic and that the remainder are logograms, as evidenced by mathematical analyses of Linear Elamite inscriptions.[3][31]

His 2022 article is the capstone of the project. Its abstract states:

Linear Elamite writing was used in southern Iran in the late 3rd/early 2nd millennium BCE (ca. 2300–1880 BCE). First discovered during the French excavations at Susa from 1903 onwards, it has so far resisted decipherment. The publication of eight inscribed silver beakers in 2018 provided the materials and the starting point for a new attempt; its results are presented in this paper. A full description and analysis of Linear Elamite of writing, employed for recording the Elamite language, is given here for the first time, together with a discussion of Elamite phonology and the biscriptualism that characterizes this language in its earliest documented phase.

Desset's main publications on the subject are as follows: 

Desset, François (2018a). "Linear Elamite Writing". In Álvarez-Mon, Javier; Basello, Gian Pietro; Wicks, Yasmina (eds.). The Elamite World. Abingdon, Oxon: Routledge. pp. 397–415. ISBN 978-1-315-65803-2.

Desset, François (2018b). "Nine Linear Elamite Texts Inscribed on Silver "Gunagi" Vessels (X, Y, Z, F', H', I', J', K' and L'): New Data on Linear Elamite Writing and the History of the Sukkalmaḫ Dynasty". Iran. 56 (2): 105–143. doi:10.1080/05786967.2018.1471861. S2CID 193057655.

Desset, François (2020a). "Breaking the Code: The Decipherment of Linear Elamite, a Forgotten Writing System of Ancient Iran (3rd Millenium BC)". Canal-U.

Desset, François (2020b). A New History of Writing on The Iranian Plateau – via YouTube.

Desset, François (1 September 2021). "On The Decipherment of Linear Elamite Writing". The Postil (Interview). Interviewed by Robert M. Kerr.

Desset, François; Tabibzadeh, Kambiz; Kervran, Matthieu; Basello, Gian Pietro; Marchesi, Gianni (2022). "The Decipherment of Linear Elamite Writing". Zeitschrift für Assyriologie und vorderasiatische Archäologie. 112 (1): 11–60. doi:10.1515/za-2022-0003. ISSN 0084-5299. S2CID 250118314.

Surfaceology

A new technique called "surfaceology" (described in the linked Quanta magazine article) provides a profoundly more efficient method than the path integrals implied by Feynman diagrams to calculate the probability of Standard Model interactions. 

It is also useful in doing calculations in "double copy" approaches to quantum gravity, in which on does a calculation in QCD and "squares" it, to get an answer for a parallel problem in quantum gravity. 

Surfaceology flows from the same line of reasoning as the amplituhedron of theoretical physics superstar Nima Arkani-Hamed (which only works for supersymmetry theories) and was devised by a junior member of his research group, Carolina Figueiredo, in 2022, with a pair of preprints (here and here) first published in September of 2023. But, it works for real Standard Model particles and not just for simplified theoretical physics models.

Further developments in the winter of 2023-2024 described outcomes that were considered with many calculations in Feynman diagram calculations that eventually revealed that these outcomes were effectively impossible called "hidden zeros." Figueiredo and Arkani-Hamed, along with Qu Cao, Jin Dong, and Song He, posted theses findings in a series of preprints.

More efficient calculations that this method facilitates could turn many particle physics and quantum gravity problems that were theoretically possible to calculate, but as a practical matter, impossible to numerically work out, into practically solvable problems, and can very difficult calculations vastly easier to solve.

Hat tip to 4Gravitons.

(opens a new tab

The Standard Model Still Works (Again)

The LHCb experiment at the Large Hadron Collider (LHC) has made a statistically significant observation (although not an absolutely certain discovery) a rare decay of a particular kind of positively charged bottom quark meson (to a positively charged pion and an electron-positron pair, which is an example of what is called a semi-leptonic decay because it is a mix of a hadron, the pion, and leptons like electrons and positrons) with a frequency of one decay per 40 million decays of this kind of meson (a kind of meson which, itself, doesn't make up a large share of mesons produced at LHCb). 

This just happens to be statistically consistent with the frequency of this kind of decay of this kind of meson that the Standard Model predicts of B(B+→ π+ℓ+ℓ−) = (2.04 ± 0.21) × 10^−8, which is about one per 50 million decays. The same decay, but with muons, was first seen in 2012 at a branching fraction of one per 55 million decays that was also statistically consistent with the Standard Model expectation (which is the same for electrons and for muons due to lepton universality).

The first evidence for the decay B+→π+e+e− is reported using proton-proton collision data recorded by the LHCb experiment at centre-of-mass energies of 7, 8 and 13 TeV, corresponding to an integrated luminosity of 9 fb^−1. 

A signal excess with a significance of 3.2σ is observed and the branching fraction is measured to be B(B+→ π+e+e−) (2.4+0.9−0.8+0.4−0.2) × 10^−8, where the first set of uncertainties is statistical and the second is systematic. The result is consistent with the Standard Model expectation.

LHCb collaboration, "First evidence of the decay B+→π+e+e−" arXiv:2604.26784 (April 29, 2026).

Combining the statistical and systemic uncertainties, the total uncertainty is about 2.4 ± 0.9 x 10^-8, which a larger branching fraction (i.e. more events) actually slightly favored over a smaller one (i.e. fewer events), relative to the best fit value.

The deviation from the Standard Model expectation in the muon measurement was about 0.7 sigma (in the opposite direction of the deviation in the electron experiment, from the best fit value), while the deviation from the Standard Model expectation in the electron measurement was about 0.4 sigma. This suggests that the systemic uncertainty estimate in the Standard Model prediction and in the experiments was probably conservatively somewhat high.

This particular hadron decay isn't extremely significant (hadrons are either mesons like the B+ or baryons like the proton). But comparing the decay rate of a positively charged pion with a muon-antimuon pair to the decay rate of a positively charge pion with an electron-positron pair is a good test of "lepton universality" (i.e. the Standard Model rule that electrons, muons, and tau leptons have properties that are identical except for their masses). For several years there were experimental anomalies that made it appear that lepton universality was violated, but those anomalies were recently resolved in favor of the Standard Model prediction that lepton universality is not violated.

There are about a hundred plain vanilla mesons and baryons in the Standard Model like the B+ meson studied here, and some of the heavier ones have perhaps hundreds of decay modes with a predicted branching fraction of less than one decay per billion decays. So, the universe of Standard Model predicted meson decays to look for is somewhere on the order of 10,000.

The B+ meson has two "valance quarks" an up quark and an anti-b quark. It has a rest mass of 5279.26 ± 0.17 MeV/c^2 (about 5.6 times the mass of a proton and a little less massive than a Lithium-6 atom). It has total angular moment (a.k.a. "spin") of 0 and odd (i.e. negative) parity, which means that it is a "pseudo-scalar" meson. It is ephemeral, it has a mean lifetime of (1.638 ± 0.004) × 10^−12 seconds (i.e. a little more than a trillionth of a second). It has more than two dozen measured decay modes that happen in more than one in a million decays, and the vast majority of the time B+ mesons decay to particles that include some kind of charm quark hadron. It has hundreds of decay modes more probable than this one.

The Standard Model was devised in the early 1970s, the b quark was discovered at Fermilab in 1977. The full set of fundamental particles (except the Higgs boson, which was discovered at Fermilab in 2012 and the discovery that the neutrinos were massive), was in place in 1995, more than three decades ago. 

The Standard Model prediction for the frequency of this particular B+ meson decay was cited in connection with the first observation of the parallel muon decay in 2012 and in 2015, and derive from a 2008 paper (i.e. it was made more than 14 years before this decay was observed just as predicted).

'The Higgs boson and the neutrino masses don't (meaningfully) enter into the calculation of the branching fractions of the B+ meson, so the only thing that has changed in the Standard Model since 1995 that is relevant to this calculation is that the measurements of some of the fundamental physical constants involved in the calculation, especially the relevant CKM matrix elements (as noted at page 13 of the 2008 paper) have gotten more precise over that time. (The accuracy with which we know another non-fundamental physical constant, called the "form factor" of the B+ meson, which is too hard to calculate from first principles at this point, has also improved and is material to this calculation.)

The physical constant whose improved precision matters most in this context are the CKM matrix elements for the b quark to up quark transition probability in W boson interactions and the top quark to down quark transition probability in W boson interactions, which are low: about 0.14% and 0.007% respectively. 

The respective 3% and 2% uncertainties in the world average measurement of these physical constants are probably some of the leading sources of the roughly 10% uncertainty in the Standard Model prediction of the frequency of this B+ meson decay branching fraction. It is hard to say exactly how much of a share of the uncertainty in the predicted value is from this source, however, because while the respective papers linked above provide an error budget chart for the uncertainties in their experimental measurements, none of the papers provide an exact error budget chart for their Standard Model predictions for this decay frequency, probably because this was considered too elementary to publish. 

Computer processing capacity has also improved greatly since then, which makes these calculations much less cumbersome to actually make.

In isolation, this experimental confirmation of the Standard Model prediction could be just a lucky fluke, although a quite remarkable one, even on its own. But together with thousands of other measured hadron branching decay fractions, the Standard Model is really unstoppable. 

Experimental result anomalies where there are deviations from the Standard Model prediction are few, far between, modest in statistical significance, and usually go away quickly for closer inspections and more experiments and analysis. Experiments testing the Standard Model in contexts other than hadron decay branching fractions that involve completely different kinds of calculations are just as consistently correct. It is an extremely robustly tested theory.

Even if there are beyond the Standard Model physics gaps that are missing from the Standard Model, it is very close to the truth. The open parameter space for deviations from it are very small.

Theoretical X17 Considerations And Related Conjectures

Could the X17 resonance, if it is even real, be an electromagnetically bound light quark-light antiquark meson?

This explanation is much more attractive than a new fundamental particle, as it wouldn't involve beyond the Standard Model physics, and would instead involve a low energy electromagnetically bound up-antiup or down-antidown pair of quarks.

It has to be electromagnetically bound, rather than strong force bound, because a neutral light quark-antiquark pair bound by the strong force, i.e. a neutral pion, has a mass of about 135 MeV, mostly due to the binding energy of the gluons confining them in a hadron. 

This said, this theory has a big problem. 

Why aren't the light quarks confined in a QCD bound hadronic state? 

The only times quarks are not in QCD bound hadronic states that have so far been observed are shortly after top quarks form (because they almost always decay before they can hadronize, although we just learned in 2025 that in rare cases a top anti-top quark pair can form toponium in a QCD bound state the persists very briefly) and in quark-gluon plasma at temperatures corresponding to about 1-2 GeV (i.e. 11-23 trillion Kelvin).

The invariant mass spectrum of e+e− pairs produced in high-energy Pb-emulsion collisions at 160 A GeV at CERN SPS exhibits a complex structure of many resonances resting on top of a broad enhancement at invariant masses below 50 MeV, with the prominent resonance at 19 ±1 MeV providing independent support for the hypothetical X17 particle. 

We show that this complex structure may be coherently described as signatures for the neutral color-singlet qq¯ quark matter in both its deconfined and confined phases. That is, the broad enhancement may arise from thermal annihilation of QED(U(1))-deconfined quarks and antiquarks into e+e− pairs at the phase transition temperature Tc(QED), theoretically estimated to be 4.75 ± 1.2 MeV from the transitional equilibrium condition. The observed 3±1 and 7±1 MeV resonances may correspond to the QED(U(1))-deconfined dd¯ and uu¯ Coulomb bound states near their quark rest masses, respectively, whereas the observed 19 ± 1 MeV resonance may correspond to the QED(U(1))-confined isoscalar QED meson. 

The approximate agreement between the theoretical and the experimental spectrum suggests that both QED(U(1))-confined and QED(U(1))-deconfined neutral color-singlet qq¯ quark matter may have been produced in these high-energy Pb-emulsion collisions. We propose future experiments to confirm or refute these findings.

Cheuk-Yin Wong, "Possible Evidence for Neutral Color-Singlet qq¯ Quark Matter from High-Energy Pb-Emulsion Collisions" arXiv:2604.23473 (April 25, 2026) (21 pages).

Some conjectures

What would work without breaking the rules of the Standard Model, however, is if the 3 and 7 MeVs were light quark-antiquark pairs that were produced and immediately annihilated before  they could hadronize, and if the 19 MeV resonance was an electromagnetically bound positron-electron state (i.e. positronium). Positronium has a ground state mass of 1.022 MeV  (twice the 0.511 MeV mass of an electron or positron), however, with excited states varying in mass by single digit eV amounts per state, which wouldn't generate a single resonance at 17-19 MeV. 

Another possibility is that the observed 3 ± 1 MeV resonances may correspond to the QED(U(1))-deconfined uu¯ Coulomb bound state near its quark rest masses, that the 7 ± 1 MeV resonances correspond to the QED(U(1))-deconfined dd¯ Coulomb bound state and also to uu¯uu¯ Coulomb bound state near their respective quark rest masses, and that the observed 19 ± 1 MeV resonance may correspond to the QED(U(1))-deconfined dd¯dd¯ Coulomb bound state.

The light quark masses, according to the Particle Data Group (admittedly at the 1-2 GeV energy scale and not the low single digit to tens of MeVs energy scale) is as follows:

The rest mass of four d-quarks is 18.8 MeV, which is right where the resonance is observed.

In this hypothesis, these resonances fail to hadronize because the e+e− pairs that produced one or two light quark-antiquark pairs didn't have enough mass-energy to form a 135 MeV neutral pion, so they instead formed one or two deconfined quark-antiquark pairs that quickly annihilate again because the system had enough energy to create the quarks, but not enough energy to create the bound system of quarks and gluons necessary to form a pion. This has the virtue, again, of not requiring any BSM fundamental particles or new forces.

A four quark solution requires angular momentum that wouldn't normally be present in a simple e+e− pair, but if there were two e+e− pairs in close proximity, both with only modest kinetic energy, which is plausible in the context of the complex overall environment of the high-energy Pb-emulsion collisions generating the data here, or the interactions of the full fledged multi-nucleon atoms present in other contexts where there are claimed sightings of the X17 resonance, a coincidence of two low energy e+e− pairs would be expected with some calculable frequency.

This explanation would still be ground breaking, as it would represent a third circumstance, previously unknown and not predicted, where quarks are (briefly) deconfined. But it would be far less radical than most of the alternative explanations.

Does a(0) Evolve Over Time?

The radical acceleration relation (RAR) which is implied by MOND but isn't necessary caused by MOND, holds true for all low-z observations (i.e. nearby galaxies). But this study concludes that while the RAR still holds in intermediate age galaxies (i.e. those that are farther away), that Milgrom's constant a(0) for these galaxies has a numerical value that is a factor of two greater than what it is for low-z galaxies.

The radial acceleration relation (RAR) is a tight empirical correlation between the observed radial acceleration (a_tot) and the baryonic radial acceleration (a_bar) measured across galaxy radii: these two accelerations start to deviate significantly from each other below a characteristic acceleration scale, a0. So far, observational studies of the RAR have predominantly focused on galaxies in the local Universe, leaving its evolution with cosmic time largely unexplored. 

Using high signal-to-noise data from the MUSE Hubble Ultra Deep Field survey, we investigate the RAR with a sample of 79 star-forming galaxies (complete above M* >10^8.8 Msun) at intermediate redshifts (0.33 < z <1.44). We estimate the observed intrinsic acceleration and the baryonic acceleration from a disk-halo decomposition that incorporates stellar, gas, and dark matter components, with corrections for pressure support, using 3D forward modelling. 

We find a RAR in our intermediate-z sample offset from the local relation, with a higher characteristic acceleration scale, a0(z~1) = 2.38+/-0.1* 10^-10 m/s^2, and a larger intrinsic scatter (~0.17 dex). Dividing the sample into redshift bins and refitting the RAR in each bin, we find a characteristic acceleration scale that systematically increases with z. Parametrizing the z-dependence as a0(z)= a0(0) + a1 * z, we obtain a1 = 1.59 +/- 0.1 * 10^-10 m/s^2, providing evidence for a z-evolution. 

We find similar results using various dark matter halo profiles as well as the Modified Newtonian Dynamics framework in our 3D forward modelling. Our results show that the RAR persists at intermediate redshift, with statistically significant redshift evolution of the characteristic acceleration, pointing to a possible evolution of the baryon-missing mass connection over cosmic time.

B. I. Ciocan, N. F. Bouché, J. Fensch, D. Krajnović, J. Freundlich, H. Desmond, B. Famaey, R. Techi, "MUSE-DARK III: The evolution of the radial acceleration relation at intermediate redshifts" arXiv:2604.22613 (April 24, 2026) (Accepted in A&A).

For reference z=0.33 is about 3.7 to 3.8 billion years ago, z=1 is about 7.7 to 8 billion years ago, and z=1.44 is about 9 to 10 billion years ago. The universe is about 13.8 billion years old. A variation of 0.17 dex is about ± 48%. The intrinsic scatter in the recent time SPARC galaxy sample is about ± 8% (0.034 dex), which is about is small as possible given the precision of the astronomy instrumentation involved. Milgrom's constant is about a(0) ≈ 1.2 × 10^−10 m/s^2.

Ciocan (2026), above, and the cluster data, both point to something very like MOND, except that a(0) evolves under certain circumstances to higher values. 

Missing baryonic matter (i.e. matter made up of ordinary atoms) is, at least, a partial explanation and one that could evolve other time. Indeed, it should evolve over time, because over time more baryonic matter ends up in stars, which are easy for astronomers to see, rather than interstellar gas and dust, which are hard for astronomers to see (and hence often called "missing" when it isn't seen and couldn't be seen even if it was there with current instrumentation). Still, missing baryonic matter may not be the entire explanation, because the magnitude of the change in a(0) may not be big enough, and changes in the naively measured value of Milgrom's constant shouldn't be very uniform since some galaxies are forming starts more actively than others (although this may be reflected in the greater dispersion of Milgrom's constant measurements in older samples).

Deur (who bibliography is linked in the sidebar) argues that the missing piece for cluster scale phenomena is the geometry of the mass distributions, by an appealing analogy to similar phenomena in QCD (which is attractive theoretically because in many respects gravity behaves like QCD squared). (QCD stands for quantum chromodynamics which is the Standard Model theory of the strong force that holds hadrons together and indirectly through hadron mediated forces accounts for the nuclear binding force that binds atomic nuclei together.)

Stacy McGaugh at Triton Station has another post about MOND v. dark matter particles (DM) and why the evidence favors something like MOND but the sociology of astrophysics favors dark matter particles.

The search for a final explanation of dark matter phenomena continues, and while toy-model MOND isn't the final solution, it does a remarkably good job over a very wide range of masses. McGaugh is surely right that the final solution looks a lot more like MOND than it does like most DM models, because for DM to describe the universe we see, we need a theory that explains how DM particles consistently form in a way entirely predicted by baryonic mass distribution, which contrary to protests that it has, it hasn't.

Even if a(0) changes over time, it provides a vastly smaller degree of freedom in how galaxy dynamics can vary than DM, especially if the variation is systemic between galaxies and galaxy clusters, or between galaxies over billions of years of time, and not just random.

South American Genetic History

I remain skeptical that the Australasian ancestry is as ancient as claimed. It is much less than 2%, maybe a hundred times less, and the regional variation in its frequency is far too great for it to be ancient. I suspect an origin in Polynesian sea farers that may be obscured by natural selection against some signature Polynesian genes. 

I have not yet seen any really solid evidence that it has been present for 10,000+ years, or any explanation for the extremely varied frequency of these genes in the populations where they are found, indicating a very recent dispersal to these populations that hasn't had time for these frequencies to harmonize. 

South American ancient DNA samples are few and far between at that time depth, and this paper has some of them, but a lot of the references supporting this analysis are in supplementary materials and extended data, or references to other papers, and the nature of the ancient DNA sample is one I haven't been able to look at closely yet. As I've only had time to cut, paste, and highlight, rather than enough to do a proper critical analysis of this claim. 

I'll add more analysis in an update in this post, if time permits, which may or may not happen (I'm busy preparing for an upcoming jury trial).

[A] study published today in Nature reveals these migrations were anything but simple. Examining ancient and modern genomes collected from across South America and beyond, the team found that genetically diverse groups populated the continent in at least three separate pulses. And some people or communities carried with them possibly advantageous genes acquired from long-ago Australasian ancestors. . . .

His team published a complementary study today in Current Biology, finding evidence of unexpected genetic diversity and otherwise invisible migrations in 52 ancient genomes from Argentina and Uruguay.

In the Nature study, Tábita Hünemeier, a geneticist at the University of São Paulo and the Institute of Evolutionary Biology, collaborated with researchers and Indigenous communities across Latin America to sequence 128 whole genomes from living people from north Mexico to southern Argentina. The team then analyzed them alongside existing databases and previously published ancient genomes.

Previous work had identified the first two waves of settlement in South America, the earliest of which included people related to the Anzik child, who was buried in Montana 12,700 years ago. A second dispersal followed about 9000 years ago and ultimately contributed more to the genomes of most ancient and modern South Americas, including those Posth studied.

Hünemeier and her team found evidence of a third dispersal, whose genetic signature first appears in their data about 1300 years ago and then spreads widely across the continent and even into the Caribbean. The newcomers show hints of being related to Mesoamericans from Mexico and Central America, but so far, researchers don’t know exactly where they came from or who were their closest relatives. “Without the source population and more direct evidence [of a third pulse] from ancient DNA, it’s hard to really wrap our heads around” how and when a third migration might have happened, Posth says.

The study also digs deeper into a mystery that has bedeviled the genetic history of the Americas for over a decade: How did traces of Australasian ancestry end up in some ancient and modern South American genomes? Genetic variants from this lineage make up only about 2% of ancestry in the people who carry it, but that proportion has stayed remarkably consistent over the past 10,000 years. “This signal is found again and again and again,” Posth says. “It must mean something.”

Hünemeier suspects people carrying this ancestry were among several distinct populations that lived for thousands of years in Beringia, the now-drowned landmass that connected eastern Siberia to Alaska, and that it eventually spread southward into the Americas from there. (This Australasian ancestry, sometimes known as Population Y or the Ypykuéra signal after the Tupi word for “ancestor,” is different from the genetic sequences some Polynesian populations share with South American ones. Scientists continue to debate how that more recent gene flow happened—for example, whether Polynesian voyagers may have reached western South America about 800 years ago—but the findings from the Nature paper have no bearing on that mystery.)

From Science.

Both of these papers are open access.

The Nature article and its abstract:

Indigenous peoples of America represent the last principal expansion of humans across the globe, yet their genetic history remains one of the least explored. Although these populations have inhabited the continent for thousands of years, their evolutionary history remains largely unresolved, owing to the limited availability of genomic data. 

Here we present data on 128 high-coverage Indigenous American genomes and show they harbour extensive and previously uncharacterized genetic diversity, reflecting at least three dispersals into South America, followed by regional differentiation and long-term continuity. 

We identified widespread natural selection signals in genes associated with immunity, metabolism, reproduction and development, which were shaped by adaptation to diverse environmental conditions. 

Notably, several genomic regions exhibit a remarkable allele sharing with Australasian populations, probably originating from an ancient admixture event and partly maintained by selection for more than 10,000 years. 

We also detected distinct contributions from archaic humans with adaptive introgression affecting key biological functions. The limited overlap between the regions of Australasian affinity and archaic ancestry indicates independent evolutionary origins of these signals. These findings challenge simplified models of continental settlements and show a more dynamic and complex evolutionary history for the Indigenous peoples in America.

Castro e Silva, M.A., Nunes, K., Ribeiro, M.R. et al. :The evolutionary history and unique genetic diversity of Indigenous Americans." Nature (April 22, 2026). https://doi.org/10.1038/s41586-026-10406-w

The section on archaic and Australasian ancestry in the body text of the Nature article states (with citations omitted):

Affinity with Australasians and archaics

Some Indigenous American populations show elevated genetic affinity to present-day Australasians relative to other groups contradicting a single non-Arctic Indigenous American clade. This affinity is best explained by admixture between the ancestors of Indigenous Americans and an unsampled ancient Asian population, termed Ypykuéra (here referred to as Ypykuéra ancestry), partially related to a sister clade of present-day Australasians.

We assessed genetic affinity between ancient and modern Indigenous Americans and present-day Australasians, the closest living proxies for Ypykuéra ancestry. We applied F-statistics to modern Indigenous American pairwise comparisons and to comparisons including ancient individuals.

Several Indigenous groups, including the Awajún, Ayoreo, Guarani, Karitiana, Sirionó, Suruí and Tsimané, show significant excess genetic affinity to Australasians relative to other present-day populations (Z > 3). These groups span eastern and western South America and the Chaco, with the strongest enrichment in the southwestern Amazon, where five of these seven populations are located.

A second analysis detected at least one individual with significant affinity in all examined clusters, except Arctic and northern North American groups, which were excluded from this analysis because of partial or complete ancestry from independent Siberian dispersals. The earliest signal occurs in the 10,400-year-old Sumidouro individual. Signals persist from the Early Holocene to the present, increasing in frequency during the Late Holocene, especially in the Andes, Pacific Coast and western South America. The partially discontinuous spatiotemporal pattern probably reflects variation in prevalence within and among populations. Taken together, these findings indicate that this ancestry was present during the initial peopling of America and that it may have contributed more strongly to Late Holocene and present-day genetic diversity.

We tested whether Ypykuéra-related ancestry in Indigenous Americans reflects shared ancestry with Australasians by means of archaic hominins (Neanderthals and/or Denisovans). We compared D(Mbuti, Onge; Mixe, X) with D(Mbuti, Neanderthal or Denisova; Mixe, X), where X denotes Indigenous American groups. Mixe served as a Mesoamerican reference to match earlier studies reporting Australasian affinity. No correlation was detected between Australasian and Neanderthal (Spearman’s r = −0.006, P = 0.971) or Denisovan affinity (r = −0.1002, P = 0.5372). By contrast, Neanderthal and Denisovan affinities were strongly correlated (r = 0.6572, P = 7.2 × 10^−6), consistent with homogeneous archaic ancestry in the founding populations.

An alternative hypothesis proposes that Australasian affinity reflects retention of the Ypykuéra component in isolated groups with high internal genetic similarity. Such populations and genomic regions, characterized by elevated ROH, would be less affected by admixture that could dilute signals of ancient Population Y ancestry. This hypothesis is not supported by our data, which show no correlation between Australasian affinity and inbreeding (FROH) (Spearman’s r = 0.2503; P = 0.1192). Moreover, ROH hotspots, defined as regions with ROH density greater than three standard deviations above the mean, show little overlap with loci of Australasian affinity, with only about 6% of such positions coinciding.

We tested whether Indigenous American affinity to present-day Australasians also includes ancient Hòabìnhian individuals, proposed ancestors of mainland Southeast Asian hunter-gatherers, including the Onge Using D(Mbuti, Y; Mixe, X), with Y as Onge or Hòabìnhian individuals (La368, La364) and X as Indigenous American populations, we evaluated correlations in affinity to Onge and Hòabìnhian individuals. La368 forms a sister branch to Onge, whereas La364 is modelled as Australasian-related plus Austronesian ancestry, sister to Ami. We observed significant correlations for La368 (Spearman’s r = 0.6444; P = 1.1856 × 10^−5) and La364 (Spearman’s r = 0.6208; P = 2.8848 × 10^−5). These results support a shared ancestry component between Indigenous Americans and Australasians that extends deep into the past.

The Current Biology article and its abstract:

• Expansion of ancestry into the Pampas, Uruguay, and Patagonia from the Middle Holocene

• Repeated mobility from southern Andean and southern Patagonian-related populations

• Genetic differentiation between the Upper and Lower Paraná River Delta ∼600 years ago

• Coastal dispersal from southern Brazil to eastern Uruguay via mound-builder societies

The Southern Cone represents the southernmost region of South America settled by humans. Although ancient genomes from southern Patagonia have been sequenced, genomes from the central Southern Cone (CSC) remain temporally and spatially sparse. Archaeology documents major cultural transformations during the Middle and Late Holocene, yet their relationship with demographic processes has been debated. 

We present genome-wide data from 52 individuals spanning 6,000 years, originating from four regions of the CSC in present-day Argentina and Uruguay: the central and southern Pampas, Northwest Patagonia, the Paraná River Delta and Lower Uruguay River, and the eastern lowlands of Uruguay. 

Genomic evidence from the Pampas reveals the presence of at least three distinct ancestries during the Middle Holocene. Although genetic contacts with southern Patagonian groups were sporadic, we identified the expansion of an ancestry of unknown geographic origin by 5,500 years ago (ya), which increased during the Late Holocene. This ancestry arrived in Northwest Patagonia by at least 600 ya and co-existed locally with a southern Andean genetic profile until colonial times. 

Genetic structure differentiates populations along the Paraná River Delta and Lower Uruguay River by 1,500 ya. 

Individuals from the eastern lowlands of Uruguay show genetic links with Sambaqui-associated populations from the southern coast of Brazil, suggesting the role of human dispersals in connecting tropical lowland cultural traditions. 

Our work documents the diffusion of genetically distinct groups across all studied regions and provides compelling evidence that large-scale human movements contributed to the remarkable cultural diversity of CSC populations during the Middle and Late Holocene.

Kim-Louise Krettek, et al., "The shared genomic history of Middle- to Late-Holocene populations from the Southern Cone of South America" Current Biology (April 22, 2026).