Best Assigned to Nothing Known
A 407-million-year-old column from a Scottish field has no chitin in it, which is good evidence it was not a fungus. Whether it was a lineage of its own rests on what is left over after the ruling out.
The best-preserved specimen of the most argued-about fossil on land came out of a farm field in Aberdeenshire. It is a roughly cylindrical column of organic tubes, 5.6 cm across at its widest, running obliquely through a 6.9 cm block of 407-million-year-old silica, incomplete at both ends. It was recovered from farmland at Windyfield, beside the Rhynie chert site, and passed to a core-collecting outfit called North Sea Core, which is how it ended up in front of the team that took it apart.
The column is Prototaxites, and people have been arguing about what kind of thing it is since 1859, when John William Dawson looked at the Canadian material and decided he was seeing rotted conifer trunks with fungi growing in them. Since then it has been a giant marine alga, a fungus, a rolled-up mat of liverworts, and a fungus again. The genus reaches a metre in diameter and something like eight metres tall in Devonian deposits elsewhere — which, in a landscape whose tallest vascular plants were ankle-height, makes it the first large thing that ever stood up on land. Nobody knows what kingdom it belongs to.
What the new paper does
The January paper in Science Advances — Loron, Cooper, Hetherington and colleagues at Edinburgh and elsewhere — runs two arguments at once, and they are not equally strong, which is the part I find interesting.
The chemical argument is the one that got reported. They took attenuated-total-reflectance infrared spectra from thin sections: twelve sampling spots across three Prototaxites taiti specimens, set against a reference library of 102 Rhynie chert samples (47 assembled previously, 55 new) covering fungi, arthropods, plants and bacteria from the same rock. Then they threw chemometrics at it — support vector machines, discriminant analysis, a one-class model. The classifiers separate P. taiti from fungi with 91% accuracy, and from everything chitinous (fungi plus arthropods) with 93%, with a Matthews correlation coefficient of 0.85 in both cases. The one-class model, trained only on the twelve Prototaxites spots, scores a specificity of 0.911 against every other fossil in the chert. The signal driving it is the absence of the sugar–protein fossilisation products that chitin and chitosan leave behind, and which the fungi and arthropods in the same rock do leave behind. A separate solvent extraction found no perylene — a polycyclic marker associated with ascomycetes — in the fossil, though it did find it in the peaty substrate around it.
The anatomical argument got much less attention and is, I think, the better half. Confocal imaging and a 3D reconstruction give three distinct tube types in one body: thin-walled ~10 µm tubes with septa and septal pores, making up roughly 75% of the volume; aseptate 20–40 µm tubes with double-layered walls, about 20%; and aseptate ~40 µm tubes with thick walls carrying weak annular thickenings, the remaining 5%. Scattered through the column are “medullary spots”, 200 to 600 µm across, where all three tube types connect through a dense tangle of branches down to a micron wide. The authors’ point is that fungal hyphal systems branch directionally and in order; this does not. And the annular thickenings on the big tubes look less like fungal skeletal hyphae than like the rings in plant water-conducting tissue, in an organism that is definitely not a plant.
That combination has no living analogue. It is a positive observation, it doesn’t depend on anything surviving 407 million years of chemistry, and it is the claim I’d put money on.
Where the headline gets ahead of the work
The chemical case is built on an absence, and the authors say so plainly: the simplest reading of a missing chitin signal is that there was no chitin, and therefore no fungal cell wall. “Simplest” is doing real work in that sentence. They also concede, in the discussion, that the molecular composition of P. taiti may have been uniquely more liable to alteration or replacement than that of the fungi preserved beside it — which is the whole objection, stated by the people making the argument. Silica in the chert masks parts of every spectrum. Melanin, which would be the other fungal tell, can’t be assessed at all, because its degradation products look like generic kerogen.
And then there is the sample. Twelve spots. Three specimens. One species, from one site. The extracted-molecule work — the perylene, the biomarkers — comes from a single block, NSC.36. The eight-metre trunks that make Prototaxites famous have not been through any of this; the thing that was tested is a fragment the size of a fist.
The title reports an absence, the conclusion claims a lineage, and the headlines awarded a kingdom.
That escalation is visible in the publication record. The preprint, posted in March 2025, was titled “Prototaxites was an extinct lineage of multicellular terrestrial eukaryotes”. The peer-reviewed version is titled “Prototaxites fossils are structurally and chemically distinct from extinct and extant Fungi” — a careful negative claim about three fossils. The conclusion inside it still says the organism is best assigned to an extinct eukaryotic lineage. Then the coverage ran. Live Science, in a headline I had to read twice, placed it in “a completely unknown branch of animal kingdom”, which is not merely an overstatement of the paper but the opposite of it; the article body doesn’t support its own headline.
The deeper problem is that “extinct lineage” is a residual category. It is what is left over once fungi, plants, animals and the tested protists have been excluded. There is no synapomorphy on offer, no phylogenetic placement, no sister group, and there never will be a sequence — nothing survives in a 407-million-year-old chert that you can read as DNA. The paper is not claiming to have found a branch of the tree; it is reporting a shape that doesn’t fit the branches we have, and inferring a branch from the shape. That inference might well be right. It is a different kind of thing from the measurement.
Kevin Boyce at Stanford, who did more than anyone to establish the fungal reading in the first place, is unmoved on the taxonomy and sanguine about the biology: he isn’t ready to pull it out of Fungi, and says that either way it is something weird doing its own thing. Which is a fair summary of the state of play.
The thing nobody has explained
Here is what keeps nagging at me. The reason Prototaxites was reclassified as a fungus in the 2000s was not morphology — it was carbon. Boyce and colleagues found δ¹³C varying by as much as 13‰ within and between localities, far more than the vascular plants growing alongside it. An autotroph can’t do that; it shares the atmosphere with everything else. A heterotroph eating isotopically patchy substrates can, and a 2010 follow-up found a modern saprotrophic fungus in a glacial meltwater landscape showing the same kind of spread, feeding on microbial and non-vascular production.
The new paper does not touch that evidence, and it doesn’t need to — heterotrophy and fungal identity are separate claims, and plenty of non-fungi are heterotrophs. But it leaves the energetics stranded. Something eight metres tall, in a world of millimetre-to-centimetre plants, consuming organic matter it did not make, requires a substrate nobody has accounted for. Strip away the fungal label and you don’t get an answer to that; you get the same unexplained appetite attached to an organism with no relatives.
What I like about this result is that its honest end state is a named gap. Incertae sedis — placement uncertain — is a real and respectable answer in systematics, and the twelve-spot infrared dataset is not nearly enough to replace it with a kingdom. The anatomy might be. If the medullary spots and the three-tube construction hold up across other species and other sites, that is a genuinely novel body plan with no living descendants, and the absence of chitin becomes corroboration rather than the load-bearing wall. Until then the strongest sentence in the paper is its title, and the strongest thing the fossil record has to say about Prototaxites is still that we don’t know.
Sources
- Prototaxites fossils are structurally and chemically distinct from extinct and extant Fungi — Science Advances (Loron et al., 21 Jan 2026)
- Prototaxites fossils are structurally and chemically distinct from extinct and extant Fungi — PubMed Central full text
- Prototaxites was an extinct lineage of multicellular terrestrial eukaryotes — bioRxiv preprint, March 2025
- Devonian landscape heterogeneity recorded by a giant fungus — Boyce et al., Geology (2007)
- Carbon sources for the Palaeozoic giant fungus Prototaxites inferred from modern analogues — Hobbie & Boyce, Proc. R. Soc. B (2010)
- Mystery Prototaxites Tower Fossils May Represent a Newly Discovered Kind of Life — Scientific American
- Giant fungus-like organism may be a completely unknown branch of life — Live Science
- Prototaxites — Wikipedia (classification history: Dawson 1859, Carruthers 1872, Hueber 2001, Graham et al. 2010)