The finding
A new study asks a deceptively simple question in its own title — for the mosses growing on the arsenic–thallium ore at Allchar, North Macedonia, is the metal in their tissue contamination or genuine hyperaccumulation? — and mostly answers “contamination” (Jakovljević et al. 2026, Planta). The team sampled 40 bryophyte samples across 11 taxa on the most mineralised soils at the site (surface soil up to 62,800 µg g⁻¹ arsenic and 3,810 µg g⁻¹ thallium), then measured them by X‑ray fluorescence, imaged them by synchrotron micro‑XRF, and inspected surfaces by scanning electron microscopy.
The raw tissue numbers look spectacular: up to 25,000 µg g⁻¹ arsenic (~2.5% of dry weight) and 1,890 µg g⁻¹ thallium in moss tissue, with arsenic clearing the 1,000 µg g⁻¹ hyperaccumulation threshold in 34 of 40 samples and seven of eleven taxa. Taken at face value, that would be a flora of arsenic super‑accumulators. The paper’s central result is that you should not take it at face value.
The mechanism: how to tell uptake from dust
Bryophytes have no cuticle, a huge surface‑to‑volume ratio and high cation‑exchange capacity, so they trap and hold soil particles — and on ore this rich, a little trapped dust is a lot of apparent metal. The study leans on three independent discriminators rather than the concentration alone:
- Crustal tracers. Titanium and chromium barely enter plants, so they flag soil particulate. Arsenic tracked them closely (Spearman ρ(As–Ti) = 0.72, ρ(As–Cr) = 0.55), and titanium sat above 200 µg g⁻¹ in unwashed tissue — the signature of dust, not physiology.
- Electron microscopy. On Barbula convoluta, SEM–EDS found surface particles whose stoichiometry (Si, S, Ca, Fe) is plainly mineral, and a hexane wash cut the arsenic weight‑fraction of one such spot from 18% to 3.25%.
- An apolar‑solvent wash. Rinsing in hexane stripped a large share of the surface load in some species — >80% of arsenic and >90% of Ti/Cr in Hypnum cupressiforme, >50% of arsenic in Ptychostomum capillare.
The synchrotron maps agreed: enrichment concentrated in the lower parts of the plants, where soil contact and exposure are greatest.
The headline number is an artifact — but not because washing removed it
Here is the honest nuance the paper is careful about, and any peer reviewer will check. The washing did not clean the flagship sample. Hexane’s efficiency was wildly variable and species‑specific; in Barbula — the very genus carrying the ~25,000 µg g⁻¹ arsenic maximum — removal was low and in places negative (apparent concentrations rose after washing, plausibly as dislodged fines re‑lodged and increased bulk density). So the case that this flagship number is contamination rests on the mineral stoichiometry and the µXRF/Ti–Cr colocalisation, not on the wash. The corollary is sharp: several taxa still clear the arsenic threshold after washing precisely because the wash failed on them — clearing a threshold is not evidence of genuine uptake. (This “mostly artifact” reading is our editorial summary; the authors themselves phrase it as contamination being “a major challenge” and genuine accumulation “relatively low,” while still reporting threshold‑exceeding arsenic in seven taxa as unreported for the group.)
The contrast that makes the point: Pteris vittata, the founding arsenic hyperaccumulator, carries genuine frond arsenic of ~22,630 µg g⁻¹ — almost the same number as the Allchar moss — but there the arsenic is inside the plant. The concentration alone cannot tell the two apart; only localisation, washing and speciation can.
The one genuinely interesting thread is thallium
Thallium behaves in the data like the exception. It correlated negatively with arsenic, titanium and chromium, and washing produced no significant overall change — a pattern consistent with metal that is more internal than surficial. The strongest single case is the pleurocarpous moss Rhynchostegium megapolitanum: thallium above the 100 µg g⁻¹ threshold at low titanium and chromium (47 and 3 µg g⁻¹) and above the soil concentration — the paper’s best evidence for “potentially genuine” thallium hyperaccumulation. The authors also flag P. capillare as a second candidate.
Bound this hard, as the paper does. Genuine thallium accumulation is still modest: the 1,890 µg g⁻¹ maximum is an unwashed value, washed R. megapolitanum tops out near 491 µg g⁻¹, and hexane removed up to 100% of thallium from some of its samples — at least one plant’s thallium was entirely surficial. A non‑significant washing effect is a null result, not proof of internal metal, and the decisive test — X‑ray absorption spectroscopy to speciate the thallium inside the cells — has not been done. There is also an unresolved seam in the paper worth naming: the Results report thallium negatively correlated with the dust tracers, yet the Discussion asserts high thallium was “associated with high Ti and Cr, indicating contamination.” Both cannot be cleanly true; the correlation structure actually favours thallium being the less dust‑driven element.
What it means for cleanup, mining and monitoring
Agromining/phytoextraction: bryophytes are the wrong tool here. Even the credible thallium case (≤1,890 µg g⁻¹ raw, ≤491 washed) runs one to nearly two orders of magnitude below the genuine vascular thallium accumulators on the same Allchar ground — Silene latifolia near 80,000 µg g⁻¹ (the highest thallium ever recorded in a plant) and Viola arsenica, whose endogenous leaf thallium runs into the tens of thousands. With modest genuine uptake, tiny standing biomass, and assays confounded by particulates, mosses are poor phytomining or phytoextraction candidates — a conclusion we draw against the paper’s own optimistic gesture toward “recovery of contaminated sites.” (The old “no vasculature, so no harvestable metal” argument does not carry the weight: the gametophyte mat itself is the harvestable tissue, and some bryophytes do have conductive tissue. The disqualifier is the low genuine accumulation, not the anatomy.)
Biomonitoring and hyperaccumulator prospecting: report the wash. The pointed detail is that H. cupressiforme — a workhorse moss for atmospheric‑deposition surveys — shed >80% of its arsenic and >90% of its titanium/chromium on washing. The narrow, defensible lesson is not that all moss biomonitoring “over‑reports”: standard atmospheric surveys deliberately sample away from soil and analyse unwashed by design, because deposited particulate is the signal they want. It is that at metalliferous, near‑source or mining sites, unwashed terrestrial moss cannot separate soil particulate from uptake, so any hyperaccumulation screening or bioavailability claim built on unwashed near‑source tissue will be inflated — and a standardised washing step (with crustal‑element controls) must be reported alongside the numbers. This is the same discipline our own Allchar records already apply: the Viola arsenica page notes its thallium is endogenous (contamination‑controlled by synchrotron µXRF) while its leaf arsenic — barely 381 µg g⁻¹ despite the arsenic‑saturated soil — is negligible. The moss study generalises that same split to the site’s bryophyte flora.
Bounds
One site, 40 samples, 11 taxa; thresholds borrowed from vascular plants (As 1,000, Tl 100, Cu 300 µg g⁻¹; Van der Ent et al. 2013, 2021), with the authors explicit that no bryophyte‑specific criteria yet exist. Washing efficiency was erratic and sometimes negative; a couple of the published table cells are internally inconsistent; and the genuine‑vs‑mineral verdict is, by the authors’ own statement, unsettled until speciation is done. Within those limits the paper does something valuable and rare — it treats its own biggest numbers with suspicion, and gives a repeatable protocol for earning them.
Provenance: every load‑bearing figure above was read from the paper’s open‑access full text and independently cross‑checked across four vendors (three of them re‑retrieving the primary source), with zero numeric discrepancies; comparator species values are drawn from our own database entries and cited as such. See our methodology for how we source and verify.