The research is based on ten putative specimens, of which four were preserved well enough to show all the anatomical characters needed for wood taxonomy. The fossils came from a moraine on the east-northeast side of Carapace Nunatak, a site at 76°53′S, 159°25′E and 2,150 metres above sea level in the Central Transantarctic Mountains. The specimens occurred with blocks of silicified chert containing plant material from lacustrine beds of the Kirkpatrick Basalt Formation.

Radiometric dating and palynological analyses indicate an early Toarcian age, approximately 180 million years ago, for the formation. Based on the surrounding matrix, the moraine location and the associated chert, the researchers concluded that the woods were derived from the Kirkpatrick Basalt Formation and assigned them a Toarcian age. Ten possible wood specimens were recovered, but their quality varied substantially.

Four specimens—numbered 270, 271, 272 and 274—were sufficiently preserved to observe the principal characters used in the analysis. Four others, 273, 275, 276 and probably 279, were more distorted and did not permit as many reliable measurements. Specimen 277 was too poorly preserved for anatomical description, while specimen 278 was a wood impression on sedimentary rock containing small fusains rather than anatomically preserved wood.

The investigators prepared thin sections in transverse, tangential and radial planes and examined them under reflected light with stereomicroscopes, compound microscopes and a digital microscope. All the analyzable samples consisted of tracheids and parenchymatous rays. Average tracheid diameter generally ranged from 25 to 30 micrometres, depending on the specimen.

Rays were predominantly uniseriate, although some were partly biseriate, and the tallest contained no more than about 10 cells. Their walls were thin, straight and unperforated, and the cells commonly contained dark brown to black material. The end walls of the ray cells were perpendicular or subperpendicular to the rays.

Axial parenchyma was present only rarely and showed no particular organization, while resin canals were absent. Specimen 271 was identified as a branch approximately 7 by 5 centimetres in diameter and preserved several growth-ring boundaries. The other pieces lacked enough continuous wood to establish their original positions within a plant.

The most useful distinction between the samples involved pitting on the radial walls of the tracheids. Specimens 270, 271, 272 and 274 had mixed pitting, combining araucarian and abietinean forms. Where pitting was visible, araucarian pits represented about 55–60 percent of the total, and their distribution was not homogeneous throughout specimen 270.

Specimens 273, 275, 276 and probably 279 showed araucarian pitting. In this group, pits were contiguous in about 90 percent of observed cases and often flattened; when multiple rows occurred, they were arranged alternately. These differences in radial pitting were considered taxonomically significant and more reliable for classification than ray height, partly because the fragments could represent different positions within separate trees.

Ray height provided an additional statistical signal but could not reliably separate the material at the genus level. Thirty ray heights were measured in tangential or radial sections for each specimen. Individual measurements ranged from 25.5 to 310.6 micrometres, while specimen averages ranged from 61.0 to 166.6 micrometres.

A Kruskal–Wallis test found significant differences among specimens, with a chi-squared value of 71.148, seven degrees of freedom and a p-value below 0.001. Nevertheless, substantial variation among specimens prevented ray height from distinguishing the two morphogenera reliably. Using the pitting patterns and other anatomical features, the araucarian-pitting specimens were assigned to Agathoxylon , while the mixed-pitting specimens were placed in Brachyoxylon .

The araucarian group had araucarian cross-fields with numerous contiguous, unordered cupressoid to taxodioid oculipores, no spiral thickening and ray end walls oriented perpendicular or subperpendicular to the rays. The mixed group had araucarian cross-fields and earlywood oculipores that were consistently cupressoid. The authors identified only the genera, not species, because the sample was small, preservation was poor and variable, and the anatomical differences could include normal variation within a plant.

Agathoxylon has been associated with several groups of gymnosperms, including conifers, cordaites and pteridosperms, whereas Brachyoxylon has most often been linked with the extinct conifer family Hirmeriellaceae but has also been associated with Araucariaceae and Cupressaceae. The specimens also contain direct evidence of fungi. Fragmented hyphae occur in greatest concentration in the rays and adjacent cells.

The narrow forms measure 2–5 micrometres in diameter, while wider forms are at least 6 micrometres across; some of the wider structures may represent a biomimetic form. The hyphae branch perpendicularly at uneven intervals, are sparsely septate and are occasionally joined by rare clamp connections. Their presence indicates that at least one fungal organism belonged to the basidiomycetes.

Small ellipsoidal or spherical propagules occur terminally or intercalary, although similar structures could be preservation artifacts or possible tyloses rather than fungal propagules. Because the fungal remains are degraded and may include more than one species, the authors could not confidently determine their systematic affinities beyond the basidiomycete indication. Mycelia occur in several orientations and can be traced through pits from one tracheid into an adjacent tracheid.

Elsewhere, large erosional cavities create a starburst-like appearance in tracheid lumina, while troughs extend through the walls of neighboring cells. This degradation pattern resembles that produced by some living soft-rot fungi, which can make wood soft and spongy and complicate the preservation of diagnostic anatomy. The researchers propose that the trees had begun to decay before silicification, based on the fungal remains and the extensive damage.

They also acknowledge that some deterioration could have occurred before burial or during the taphonomic process, so the relative timing of decay and mineral replacement is not established with certainty. Comparisons with other Jurassic woods show that the Carapace Nunatak material is anatomically distinct from some previously described specimens but similar to others. Early Jurassic trunks from the Mesa Range in North Victoria Land had been assigned to Protocupressinoxylon , a name treated as a nomenclatural synonym of Protobrachyoxylon .

Those trunks had rays reaching 40 cells, no axial parenchyma and cross-field pitting with three to six non-contiguous cupressoid pits, unlike the crowded pits in the new specimens. The conifer stump from nearby Coombs Hills shared uniseriate rays, thin unpitted ray-cell walls and araucarioid cross-fields with the Carapace Nunatak woods. However, the stump had smaller tracheids, rays reaching 15 cells and a mean ray height of 66 micrometres, and it lacked data on the frequency of the two radial pitting types.

The researchers therefore said that similarity with some of the new woods could not be ruled out rather than treating it as a definitive match. Taken together, the new material indicates that at least three named wood morphotypes occurred in Victoria Land during the Early Jurassic: Protobrachyoxylon , Agathoxylon and Brachyoxylon , with the poorly compared Coombs Hills stump potentially representing additional diversity. The identifications also fit other fossils from Carapace Nunatak.

The site contains abundant Classopollis pollen, often associated with Hirmeriellaceae, and the permineralized pollen cone Classostrobus elliotii , assigned to that family. The Brachyoxylon wood provides additional evidence for Hirmeriellaceae at the locality. Another pollen cone, Chimaerostrobus minutus , combines characters reminiscent of Voltziales and Araucariaceae, and the authors note that the Agathoxylon wood may be associated with one of those groups, although that relationship is not demonstrated.

The apparent scarcity of anatomically studied Jurassic wood in Antarctica may reflect environmental constraints, volcanic activity and limited sampling; it may also be influenced by the greater ease of collecting recognizable leaves, cones and pollen than wood requiring sectioning and microscopy. The findings provide anatomical evidence that multiple conifer groups were present in Antarctica during an interval in which several modern conifer lineages were diverging and diversifying.

Subject of Research: Anatomical diversity, conifer affinities and fungal decay in Early Jurassic silicified wood from Carapace Nunatak, Antarctica

Article Title: Early Jurassic silicified woods from Carapace Nunatak, South Victoria Land, Antarctica

Article References: Toumoulin, A., Decombeix, A.-L., Harper, C. J., & Serbet, R. (2023). Early Jurassic silicified woods from Carapace Nunatak, South Victoria Land, Antarctica. Fossil Record, 26(1), 103-115. https://doi.org/10.3897/fr.26.102570

Image Credits: AI Generated

DOI: 10.3897/fr.26.102570

Keywords: Antarctica, Early Jurassic, fossil wood, conifers, silicification, fungi, Toarcian, Victoria Land, paleobotany, Gondwana, Early, Jurassic

Silicon is not classed as an essential element for most plants, yet for many crops it behaves like one. Plants take it up from the soil solution as monosilicic acid and deposit it in their tissues as amorphous silica, a process called silicification. That mineral armor strengthens stems, helps plants resist fungal pathogens and chewing insects, and buffers against drought, salinity and metal toxicity. The world’s most widely grown cereals, including rice, wheat, barley, maize and sugarcane, typically store more than 10 grams of silicon per kilogram of dry tissue, which is why silicon fertilizers are now applied in agroecosystems worldwide. But those fertilizers depend on finite mineral reserves, and they are often scarce or inaccessible in tropical regions, precisely where soils tend to be poorest in plant-available silicon.

The problem, the researchers note, is that most soil silicon is locked away. It sits inside primary and secondary silicate minerals, sorbed onto iron oxides, or embedded in weathering-resistant phases such as quartz, rather than floating free in the soil solution as monosilicic acid. Whether a crop can actually access silicon therefore depends on the rate at which these poorly available pools are dissolved. Laboratory work has long shown that low-molecular-weight organic anions, especially carboxylates such as citrate, oxalate and malate, accelerate the dissolution of silicate minerals. In one classic experiment, quartz dissolved roughly ten times faster in a citrate solution than in pure water. Recent theoretical work has proposed that root-released carboxylates might similarly unlock soil silicon in the field, raising an intriguing agricultural possibility: could a carboxylate-pumping legume feed silicon to a high-silicon cereal growing next to it?

Chickpea was an obvious candidate. Previous research on 20 chickpea accessions had documented large variation in how much carboxylate these plants release into their rhizosheath, the tight sleeve of soil that clings to living roots. The new study exploited that natural diversity. In a controlled glasshouse experiment at the University of Western Australia, wheat was grown either as two individuals per pot or paired with one of the 20 chickpea accessions, using a silicon-poor, iron-oxide-rich quartz sand representative of more than two-thirds of Western Australia’s cropping soils. After two months, the team measured wheat biomass, leaf and stem silicon concentrations using portable X-ray fluorescence, rhizosheath carboxylate amounts by high-performance liquid chromatography, and foliar manganese, which serves as an established proxy for carboxylate release, since manganese uptake in plants is poorly regulated and strongly tracks rhizosphere availability.

The results were striking. Wheat grown with chickpea produced more leaf, stem, root and total aboveground biomass than wheat grown with another wheat plant, and it also carried higher silicon concentrations in its leaves and a far greater total silicon content aboveground. Crucially, a follow-up experiment ruled out the simplest alternative explanation, competition. Wheat grown entirely alone in a pot still had lower leaf and stem silicon concentrations, and lower total silicon content, than wheat paired with chickpea, showing that the cereal genuinely benefited from its neighbor rather than merely suffering less competition for a limited resource. Tellingly, when pots were fertilized with abundant silicon, the chickpea advantage disappeared, suggesting that facilitation matters most exactly where silicon availability is limiting.

The rhizosphere data lent cautious support to the carboxylate hypothesis. Pots containing wheat and chickpea held significantly more rhizosheath carboxylates per unit of root dry weight than pots with two wheat plants, and accessions previously classified as high carboxylate releasers produced more carboxylates than low releasers. Wheat leaf silicon concentration was marginally correlated with rhizosheath carboxylate amounts, and wheat aboveground silicon content was positively, if weakly, associated with chickpea foliar manganese. Yet the mechanistic picture remained incomplete. Wheat leaf silicon showed no clear relationship with chickpea foliar manganese, a stronger proxy for carboxylate exudation, and within wheat itself, leaf silicon tracked leaf manganese, hinting that silicon and manganese may be mobilized by overlapping rhizosphere processes rather than by chickpea exudates alone.

The authors are candid about these limitations. Extracting and quantifying root exudates from an intertwined root system shared by two species is notoriously difficult; carboxylates are rapidly degraded by microbes or sorbed onto mineral surfaces between release and measurement, so the harvested pool may capture only a snapshot of a dynamic process. Other unmeasured exudates, notably the phytosiderophores that grasses themselves release, could also mobilize silicon, as has been shown for iron-deficient barley drawing silicon and iron from clay minerals. Shifts in rhizosphere pH, altered competition for manganese, or even physiological and gene-expression responses triggered simply by having a legume neighbor could all contribute. The correlation evidence, in other words, suggests carboxylates play a part, but the researchers emphasize that further studies are needed to confirm the mechanism and weigh it against alternatives.

The study delivered a second, unexpected discovery: chickpea itself varies enormously in silicon. Foliar silicon concentrations differed significantly among the 20 accessions, and, counterintuitively, the low-carboxylate group accumulated more silicon in their leaves than the high-carboxylate releasers. A principal component analysis revealed that high-silicon chickpeas formed a distinct syndrome: they were smaller, more branched, invested proportionally more biomass in roots, and exuded fewer carboxylates. The authors suggest a dilution effect may partly explain the pattern, since fast-growing accessions produce biomass faster than they take up silicon, effectively lowering concentration. The finding confirms that silicification varies not only within cereals like rice and wheat but also within a major legume crop, an area largely ignored because silicon research has traditionally centered on grasses, even though mounting evidence shows silicon also benefits legumes, potentially by alleviating metal toxicity and salinity and by promoting root nodulation in the rhizobia symbiosis.

For agriculture, the implications are tantalizing. If the facilitation observed in pots holds under field conditions, wheat-chickpea intercropping, a practice already common in many dryland farming systems, could enhance crop silicification without any added fertilizer, improving resistance to pests, diseases, drought and salinity while potentially lifting yields. The variation among chickpea accessions in their capacity to boost wheat silicon uptake suggests a breeding angle as well: lines that are especially good silicon facilitators could become targets for selection in silicon-limited regions. The authors also flag follow-up questions, including whether silicon supply alters wheat root architecture and water relations under drought, whether other carboxylate-releasing legumes such as soybean show similar facilitative effects, and how the broader community of rhizosphere organisms shapes silicon mobility. Field trials will be the decisive test, since soil chemistry and biology in the open ground differ substantially from controlled pot conditions. For now, the study adds a compelling chapter to a growing recognition that the microscopic traffic of organic molecules around roots, and the quiet cooperation between neighboring crops, can move elements once thought immovable.

Subject of Research: Legume-cereal intercropping and rhizosphere-mediated silicon mobilization in crops

Article Title: Chickpea enhances silicon accumulation in neighboring wheat: a role for rhizosheath carboxylates?

Article References: de Tombeur, F., Yan, L., Tang, D., Plouzeau, L., Gille, C. E., Thorne, S., Lambers, H., Violle, C., Hartley, S., & Pang, J. (2026). Chickpea enhances silicon accumulation in neighboring wheat: a role for rhizosheath carboxylates?. Plant and Soil. https://doi.org/10.1007/s11104-026-09070-w

Image Credits: AI Generated

DOI: 10.1007/s11104-026-09070-w

Keywords: chickpea, wheat, silicon, silicification, rhizosheath carboxylates, intercropping, rhizosphere, root exudates, plant facilitation, soil chemistry, crop stress resilience, Plant and Soil