An amendment that is supposed to make metals unavailable
Liming is phytoremediation’s default opposite move. Raise soil pH with CaCO₃ and cationic metals — Cd, Zn, Ni, Pb — sorb or precipitate, plant uptake falls, and regulators approve. Arsenic, however, is an anionic metalloid, and Pteris vittata — the Chinese brake fern, the first plant ever shown to hyperaccumulate arsenic and still the workhorse of the arsenic remediation literature — is a calcicolous lithophyte that naturally thrives on lime-rich ground. Guan, Deng, Xiao, Liu and Ma (2026) asked what liming actually does in that pairing, and the answer runs against the entire amendment playbook: in a pot trial, 1% CaCO₃ roughly doubled the arsenic that ended up in fern fronds.
The experiment and the numbers
The team grew P. vittata for two months in a contaminated neutral soil containing 100 mg kg⁻¹ As, comparing an As-only treatment against As + 1% CaCO₃ (doi:10.1016/j.seh.2026.100238). On top of the growth promotion that As itself induced, CaCO₃ raised biomass by 48%, to 11.5 g plant⁻¹. Nutrient acquisition improved broadly (43–125%, depending on the element), with frond Ca, Fe and Mn reaching 9210, 4009 and 68.2 mg kg⁻¹ respectively. The headline: frond phosphorus and arsenic both rose by 81–104%, to 3729 mg P kg⁻¹ and 3942 mg As kg⁻¹ — frond arsenic at ~0.39% of dry mass, roughly four times the conventional 1000 mg kg⁻¹ hyperaccumulation threshold (Ma et al. 2001), though well short of the species’ documented maxima (~2.3% under favourable conditions).
The mechanism: carbon, pH, and the phosphate door
The proposed chain is specific and testable. CaCO₃ kept the rhizosphere at neutral pH while dissolved organic carbon there rose 123% above bulk soil — attributed by the authors, with appropriate caution, to the larger root system (“probably due to the increased root biomass”). That carbon pulse coincides with Ca/Fe/Mn solubilization in the rhizosphere, and downstream the fern mounted a 2.3–3.3-fold upregulation of the phosphate transporters PvPht1;3 and PvPht1;4 in roots. This is the mechanistic heart of the result: arsenate (As(V)) is a structural phosphate analogue, and the Pht1 family is precisely the entry door it hijacks. Liming did not liberate As by brute chemistry; it fed the fern’s phosphorus-acquisition machinery, and the arsenic rode the phosphate co-transport system in. The authors’ own phrasing is “supported”, not “caused” — the transporter data are correlative expression results, not uptake assays in heterologous systems.
Why this is non-obvious, and what it costs
The inversion matters because amendment advice is usually written for cationic metals and then applied to arsenic sites by default. For P. vittata systems, a cheap, already-approved agricultural input apparently amplifies removal rather than suppressing it — more biomass per plant and a higher frond concentration, the two multiplicative terms of any phytoextraction budget. There is also a direct governance hook: phosphate fertilisation is known to compete with arsenate for Pht1 transporters, so this result sharpens the field-deployed tension that P-starved ferns extract more As but starve, while P-fed ferns grow but may take up less As. Carbonate appears to thread that needle by boosting P nutrition and Pht1 expression together, but the paper does not test that interaction — it is the obvious next experiment.
The genuine risk sits in the same numbers. A 123% rhizosphere dissolved-organic-carbon pulse in a soil holding 100 mg kg⁻¹ As is a mobilisation event; it pays off only where root density is high enough to capture the liberated pool. On a sparsely rooted field margin or during establishment, the same amendment could transiently raise porewater arsenic and export it sideways. And this is one species, one soil, one pot trial: the result is calcicolous-specific by construction, the “and/or” treatment structure in the abstract leaves the unamended/zero-As controls unconfirmed, and per-pot or per-hectare removal totals are not reported in the abstract — no absolute extraction rate should be quoted until the full text is checked.
Bottom line
For arsenic — the metalloid that behaves like a phosphate, not like a metal — the liming reflex is wrong in both directions: calcium carbonate does not immobilise it, and in the roots of P. vittata it appears to hand the contaminant to the very transporters that pump it into the frond. If the Pht1 link holds in field soils, “lime first, then plant” becomes a candidate protocol for brake-fern remediation; if the DOC mobilisation outruns the roots anywhere, the same protocol becomes a leaching hazard. Which of those is true is a field-trial question this pot study cannot answer.
Provenance: based on the publisher’s abstract and metadata (gold OA, CC BY-NC-ND 4.0, Soil & Environmental Health 100238, published 1 October 2026); the ScienceDirect full text was captcha-blocked from our network, so all claims are bounded to the abstract and per-claim hedges (“probably”, “supported”, “and/or”) are preserved as printed. Every load-bearing number was verified verbatim against the primary source by three independent reviewers on separate vendors (Claude / GPT / Kimi, each self-retrieving the source via OpenAlex/Crossref/Unpaywall); all 8 checked claims were confirmed by all three with zero numeric discrepancies. See /methodology/.