A new open-access study of thallium phytomining reports shoot concentrations high enough to reclassify the plant as ore — and, buried in its soil chemistry, a result that should make anyone who screens contaminated land by “available” metal uneasy. Regini and colleagues (Plant and Soil, 2026, CC BY 4.0) grew Silene latifolia and the classic thallium hyperaccumulator Biscutella laevigata on soil from the Allchar (Crven Dol) As–Tl ore body in North Macedonia and on artificially thallium-spiked soil, then harvested after 30 days. This is a controlled greenhouse pot experiment, one harvest, not a field trial — bound every number below accordingly.
The result
Grown on the natural Allchar substrate, the S. latifolia ecotype from Allchar reached a mean shoot concentration of 12,600 µg Tl g⁻¹ dry weight, with an individual maximum of 18,500 µg g⁻¹ — 126 to 185 times the 100 µg g⁻¹ thallium hyperaccumulation threshold, or roughly 1.3–1.9% of the dry shoot. The abstract’s “up to 12,600” is actually the mean; 18,500 is the recorded peak. On the spiked substrate the same plants reached 3920 µg g⁻¹. Per plant, that is 6.08 mg Tl on Allchar soil versus 1.48 mg on the spiked mix — the biomass edge on the richer soil widens the gap beyond the concentration ratio alone.
The extractable fraction pointed the wrong way
The two soils were not comparable, and the mismatch is the interesting part. The natural Allchar substrate held 4021 µg Tl g⁻¹ total but only 2.02 µg g⁻¹ DTPA-extractable — about 0.05% of the total was in the “available” fraction. The spiked substrate held 11.3 µg g⁻¹ total, essentially all of it DTPA-extractable (freshly dosed as TlNO₃). So the soil that gave three times more shoot thallium had roughly 356× more total metal but about 5.6× less extractable metal.
A shoot at 12,600 µg g⁻¹ cannot be built from a 2.02 µg g⁻¹ labile pool; over 30 days the plant and its rhizosphere drew on the large non-labile reservoir the standard extractant does not see. The practical warning is blunt: DTPA-extractable metal, the usual bioavailability proxy for ranking and screening contaminated or mineralised land, under-predicted phytoextraction here, and a lab trial on freshly spiked soil under-estimated real-ore-body performance about three-fold. For phytomining prospecting, the accessible total pool and the plant’s own uptake capacity did the work — not the extractable fraction. (The paper reports no bioconcentration or translocation factor, so those metrics cannot be sourced to it.)
Provenance did more work than species choice
The sharpest contrast is within one species: on the same Allchar soil, the Allchar ecotype of S. latifolia vastly out-accumulated the Ganges (France) ecotype, whose shoot thallium was “notably lower.” The authors’ own conclusion is that the variation “between S. latifolia accessions … can be exploited for plant selection.” Our database already flags Silene latifolia as a facultative thallium hyperaccumulator; this study says provenance, not just the choice of species, sets the ceiling.
Two honest caveats temper the headline. First, S. latifolia out-accumulated B. laevigata in this experiment only — but that B. laevigata was a Ganges accession grown on foreign Allchar soil for 30 days, not on home ground. Both plants are best described as facultative thallium hyperaccumulators (the paper treats them as genuine hyperaccumulators, not one obligate and one facultative), and B. laevigata’s own known maxima are higher: our record lists 15,200 µg g⁻¹ for it, and the thallium hub carries field values to ~79,200 µg g⁻¹. This is an ecotype-and-short-pot result, not a species ranking. Second, this pot study (Regini et al. 2026) is from the same group as the field survey behind our existing Silene record; the short-pot maximum of 18,500 µg g⁻¹ should not be read against field-collected maxima as if the two were measured the same way.
The biostimulant trap
The team also tested Spirulina as a biostimulant. It raised biomass across accessions — but it cut shoot thallium concentration in both S. latifolia ecotypes (by up to ~50% on Allchar soil, ~70% on Ganges), so per-plant yield in the champion fell from 6.08 to 3.84 mg. Thallium yield rose only in B. laevigata, where concentration went up rather than down. The agromining lesson is the recurring one on this site: biomass gain is not metal-yield gain. An amendment that grows a bigger plant while diluting the metal can lower the harvestable yield — so the number that matters is concentration × biomass, not biomass.
The toxicity ledger
Thallium is among the most acutely mammalian-toxic heavy metals: Tl⁺ mimics K⁺, is absorbed readily, and accumulates. A crop carrying ~1.3–1.9% thallium in its shoots is ore-grade material and an inherent biohazard of thallium agromining — not an accident of the experiment but the intended product, which means the handling, ash management, and containment burden must be engineered like ore processing, and food-chain transfer (not measured in this contained pot study) is a real downstream concern wherever such plants might be grown. The economics are honest in the paper: it projects 2–40 kg Tl per hectare per harvest and states plainly that thallium phytomining “has yet to be tested at field scale.” The opportunity — remediating a genuinely dangerous contaminant while recovering a saleable, if thinly-traded, metal — is real. It is also, on this evidence, a 30-day pot result awaiting a field.
Provenance note: every load-bearing figure above was verified verbatim against the CC BY 4.0 version of record by three independent model vendors and an independent web re-retrieval; the full text was accessed directly. See methodology.