The cheapest way to find a hyperaccumulator is not to walk a mountain — it is to walk into a herbarium and analyse plants someone else already collected and identified. Morgenstern et al. (2026) in Environmental Monitoring and Assessment did exactly that for Angola, whose flora “remains unexplored … with respect to metal accumulation traits.” They pulled 234 samples from 132 species (59 genera, 31 families) off sheets in the Dresden and Hamburg herbaria — specimens collected since 2011 across six provinces — digested each in nitric acid and quantified ten elements by ICP-OES. Note the method: this is acid digestion and optical-emission spectrometry, not the surface X-ray fluorescence often used to screen herbarium sheets, so the numbers are bulk-tissue concentrations rather than a surface read. The screen expands the African metallophyte map — 28 species newly listed as aluminium hyperaccumulators and one new manganese hyperaccumulator — and quietly shows why that map is harder to monetise than it looks.
What crossed the line
Aluminium dominated. Against the 1,000 µg g⁻¹ (0.1%) aluminium-hyperaccumulation mark, 199 of 234 samples sat below 1,000, and only seven cleared 10,000. The genus Landolphia (Apocynaceae) is the standout: accumulators in five of the eight Landolphia species examined, with L. camptoloba reaching 45,066 µg g⁻¹ — 4.5% of leaf dry mass — and both it and Anisophyllea quangensis exceeding 2%. All four Spermacoce species screened accumulated at low-to-medium levels. Manganese was rarer: only Landolphia congolensis crossed the 10,000 µg g⁻¹ manganese line, at up to 20,896 µg g⁻¹, roughly twice the threshold and in the range of confirmed Mn hyperaccumulators in our database such as Phytolacca acinosa and Celosia argentea. Every aluminium or manganese hyperaccumulator named here was previously undescribed for that element — L. congolensis had earlier been recorded as an aluminium hyperaccumulator, but at a far lower < 2,000 µg g⁻¹ (Baumgärtel et al.).
The mechanism reads off the plant’s plumbing
What the numbers describe is internal tolerance, not exclusion. In aluminium accumulators the paper’s mechanism is the textbook one: organic-acid ligands (oxalate, malate, citrate) bind Al(III), which is then sequestered into vacuoles by membrane transporters, translocated in the xylem, and parked in leaf epidermis, cuticle or chloroplasts; manganese follows a parallel citrate/oxalate-plus-vacuolar-transporter route. The organ data are a suggestive tell: fruits and inflorescences carried on average only ~21% (aluminium) and ~34% (manganese) of the matched leaf concentration, whereas copper — a phloem-mobile micronutrient — stayed roughly level (~111% of leaf). That contrast is consistent with aluminium and manganese being delivered to leaves in the transpiration stream and not readily remobilised in the phloem — hyperaccumulation as a leaf-directed storage strategy — though the study measures the ratios, not the transport, so read it as pattern, not proof. Either way, it is why leaves, not fruits, are both the tissue to sample and, in principle, the tissue to harvest. The authors also map their hits onto Angola’s soil types: most aluminium accumulators sit on highly weathered, acidic Arenosols and Ferralsols in which aluminium and iron are the abundant metals — an association they note fits the aluminium-dominated result, though (see below) they had soil type from maps, not measured soil metal.
The catch: it found the wrong metals
Here is the non-obvious result. The screen was stacked toward finding accumulators — species were chosen from genera already in the Global Hyperaccumulator Database — yet cadmium, cobalt, chromium, lead and zinc stayed mostly below detection, and several database-listed cobalt/copper hyperaccumulators (Acalypha cupricola, Anisopappus chinensis, Buchnera henriquesii) showed no elevated metal in these sheets at all. The Katanga Copperbelt just across the border in the DRC is the world’s richest copper/cobalt metallophyte flora — home to Haumaniastrum katangense and Aeolanthus biformifolius — but on Angola’s weathered substrates the very same genera read as aluminium accumulators, not copper ones (H. katangense here carried >7,000 µg g⁻¹ Al in its flowers). What a plant actually accumulates depends on the metal its soil offers, not on taxonomy alone: a genus’s place in the hyperaccumulator database does not guarantee the metal where the substrate lacks it, and Angola’s soils offer aluminium and manganese.
That matters commercially, and the authors do the sum you should. Aluminium and manganese are cheap — about US$2,706 and US$3,168 a tonne in 2022. Recovering $1,000 of aluminium from L. camptoloba would mean harvesting and processing roughly 8,200 kg of dry leaf; $1,000 of manganese from L. congolensis, about 15,800 kg. Set that against nickel agromining, where a ~1–2% Ni crop like Odontarrhena chalcidica targets a metal worth several times more per tonne — nickel traded around US$20,000–25,000 a tonne in 2022, roughly six-to-nine times the aluminium and manganese prices above. So these records enlarge the botanical inventory but not the agromining opportunity: their value is scientific — model plants for Al/Mn tolerance — or at best a co-product to the latex and edible fruit these Landolphia already yield, not a metal crop in their own right.
What a dried sheet cannot tell you
Three cautions bound every record here, and the authors are candid about them. First, no soil: they could not obtain soil from the collection sites, so there is no bioconcentration factor. A herbarium “hyperaccumulator” is a shoot concentration over a threshold — not proof the plant concentrates metal against a normal substrate rather than merely growing on metal-rich ground, a real ambiguity for aluminium on Al-dominated Ferralsols. Second, underestimation, not just contamination: 59 of 234 digests left an undissolved silicate precipitate, so some reported values are likely too low; conversely, iron above 1,000 µg g⁻¹ flagged possible adhering dust, and several aluminium “hyperaccumulators” are demoted to merely potential on that basis. Third, one sheet, enormous variance: accumulation is strongly site-dependent — among species with multiple specimens, 17 of 31 (54.8%) crossed the threshold, and L. camptoloba ranged from below quantification to 45,066 µg g⁻¹ across sites, a several-hundred-fold spread. A single herbarium sheet is therefore a weak basis to christen a species a hyperaccumulator. These are candidates awaiting field validation, and the paper says so.
Where it connects, and the quieter risk
The cleanest addition to our database is the manganese story: L. congolensis at ~20,900 µg g⁻¹ belongs in the conversation with confirmed manganese hyperaccumulators — logged, on the evidence above, as a candidate record until soil-normalised. The cobalt and copper blanks are a useful negative for the cobalt and copper hubs: the Copperbelt signal does not automatically cross a national border. And there is a quieter, non-phytomining risk the paper flags but cannot resolve. These are wild and traditionally-used plants in a food-insecure region, and the paper notes that Landolphia and Anisophyllea fruits are eaten and sold. The organ data are reassuring for fruit — aluminium there runs ~21% of the leaf value, and the authors report that although elevated aluminium has been found in L. congolensis fruit, no health risk has been reported (citing Baumgärtel et al. 2023) — but leaves used as vegetables or medicine could deliver a genuine aluminium or manganese dose. That is a monitoring question, not a remediation one, and it is the kind of finding a metallophyte survey turns up almost by accident.
Bottom line. One archive, one snapshot, no soil. Bound the claims: this is a targeted herbarium screen, not a flora-wide survey, and its “hyperaccumulators” are shoot-threshold candidates — several possibly underestimated, several demoted for contamination, none soil-normalised. What it establishes cleanly is twofold: acid-digest screening of existing collections is a fast, cheap way to enlarge the metallophyte map of an under-studied flora; and on Angola’s weathered acid soils that map fills with aluminium and manganese — scientifically interesting, poor phytomining feedstock, and a live dietary-exposure question — while the nickel, cobalt and copper that pay for agromining stay stubbornly below the line.
Provenance: every figure here was read from the open-access full text (CC BY) and independently re-verified against the primary source — a fresh re-retrieval of the full article plus separate numeric audits on other systems, with no numeric disagreement. The phytomining comparison, the database connections and the flagged risks are our framing. See our methodology for how we source and check.