1One metal per capsule

Part 5 looked at tablets and capsules whose active dose is a few milligrams, and found that the metal usually comes from the filler, because the filler is most of the mass. A botanical capsule is the opposite case: the mass is the plant itself — a milled root, a dried fruiting body, a leaf extract, a powdered alga. Its metals came with it, from the field, the growing substrate or the sea.

Sold in a capsule, a botanical is a food supplement under Regulation (EU) 2023/915, and Annex I gives it three numbers: lead 3.0 mg/kg (row 3.1.28), cadmium 1.0 mg/kg (row 3.2.21.1) — or 3.0 mg/kg if the product is at least 80 % dried seaweed or seaweed-derived (row 3.2.21.2) — and mercury 0.10 mg/kg (row 3.3.2). There is no arsenic row for supplements and no nickel row. Article 2(3) applies the levels to the product as placed on the market, which means to the capsule contents as sold. Part 4, section 2, explains what each limit protects against and what an exceedance costs the producer: a batch over its row cannot be sold, cannot be blended down and cannot be treated.

The difference from Part 5's tablets is the second number. Most of these plants are also sold as food, and as food they meet other rows — stricter, looser, or none at all. Dried turmeric in a spice jar has a lead row of 1.50 mg/kg; the same powder in a capsule has 3.0. Dried tea leaves have no metal row at all; a green tea extract capsule has 3.0. A fresh cultivated mushroom has a cadmium row of 0.15 mg/kg on the wet weight; the dried powder from the same harvest, in a capsule, has 1.0. The figure traces four of these routes; the family sections give the details.

2Two rulers: mg/kg in the capsule, µg per day at the label dose

The EU rows are concentrations. A capsule either contains more than 3.0 mg/kg of lead or it does not, and the label dose plays no part. The other two rulers a botanical brand meets work the other way round: they measure the daily dose. The United States Pharmacopeia sets permitted daily exposures for supplements in chapter <2232> — 5 µg of lead, 5 µg of cadmium and 15 µg of inorganic arsenic per day, as reported in the green tea study in section 4. California's Proposition 65 sets a maximum allowable dose level for lead of 0.5 µg per day, above which a warning is required, with an oral cancer level of 15 µg per day; for cadmium the oral level is 4.1 µg per day and for inorganic arsenic 10 µg per day. The Proposition 65 list carries no oral level for mercury.

Converting between the rulers is one multiplication: micrograms per day equal mg/kg times grams per day. Botanicals differ from vitamins here. A vitamin tablet delivers milligrams of active; a botanical label often says two, three, five grams a day of the plant itself. At three grams a day, a powder at the EU lead row of 3.0 mg/kg delivers 9 µg of lead — compliant in the EU, eighteen times the Proposition 65 level, and nearly twice the USP allowance. At half a gram a day the same powder delivers 1.5 µg and is well inside the USP allowance. Part 5, section 3, ran this arithmetic for a calcium tablet; here it runs through every family, and section 4 draws it for eleven real batches of one product.

3Roots and rhizomes: ashwagandha, turmeric and curcumin, ginger, maca

The metal

Which of the four metals is most likely to be over the limit in a milled root? Lead. A 2025 survey of adaptogen supplements sold in Poland — ashwagandha, rhodiola, ginseng and schisandra, as powders, tablets and dried material — found lead in every ashwagandha sample, from 0.98 to 2.35 mg/kg, in a ginseng tablet at 2.18 and in dried schisandra at 2.89 — while cadmium stayed at a few hundredths of a mg/kg and mercury near one hundredth. Not one of the eleven products exceeded the cadmium or mercury row. The same survey found nickel between 1.5 and 2.1 mg/kg in every ashwagandha sample — a metal the supplement rows do not regulate at all, and the one a pharmacopoeial buyer will still ask about.

Where it comes from

A root stands in soil, and lead arrives with it: taken up where the field is contaminated, carried as dust through washing and drying, added by equipment along the way. For turmeric there is a second, documented route. Field studies across Bangladeshi districts have shown lead chromate pigment added to polished roots to deepen the yellow — pigment that is percent-level lead, so an adulterated lot is not a marginal failure but a different substance; a follow-up study found the practice collapsed once enforcement began. For maca, surveys from the Peruvian highlands report cadmium rather than lead as the metal to watch where fields sit near mining districts, though at levels well inside the capsule rows. The Polish survey we read in full, tables included; the turmeric and maca studies we read as abstracts, so they rank the risk here but contribute no figures.

The number

In a capsule, the rows are 3.0 for lead, 1.0 for cadmium, 0.10 for mercury. None of the eleven surveyed products was over the lead row — but the worst four sat at 72 to 96 % of it, the band where the verdict turns on the measurement uncertainty; Part 3 is about exactly that band. Sold as food instead, dried turmeric and ginger are root and rhizome spices with a lead row of 1.50 mg/kg — half the capsule allowance — and the fresh roots carry 0.80. Ashwagandha and maca sold as plain food powders have no row that names them. A curcumin extract has no row of its own either: Article 3 applies the turmeric numbers through the concentration factor of the extraction, which the producer must justify — section 4 shows the one such factor the regulation sets out itself.

The dose arithmetic: the surveyed ashwagandha tablet at 2.35 mg/kg, taken at one gram a day, delivers 2.4 µg of lead — under the EU row, under the USP allowance, and nearly five times the Proposition 65 level. A traditional five-gram daily dose of a root powder at 0.98 mg/kg delivers 4.9 µg — still EU-compliant, and at the edge of the USP allowance.

When to test

Each raw-material lot, per origin, before encapsulation. For turmeric, test each lot rather than each supplier: adulteration is a lot-level event, and a lead result far above the spice row is the signature. Ask for chromium on the same run; lead chromate carries both.

4Leaf and extract: green tea extract, moringa

The metal

Lead again. The one study of encapsulated green tea supplements read in full for this article — eleven batches from six brands, measured by graphite-furnace AAS and published in 2026 — found lead from 0.22 to 4.54 mg/kg, cadmium never above 0.28 and arsenic never above 0.49. Lead is the only element in that table that approaches any ruler, and two of the eleven batches are over every ruler at once: 4.50 and 4.54 mg/kg is over the EU capsule row of 3.0, and at the labelled maximum of 1.4 grams a day those batches deliver 6.3–6.4 µg, over the USP allowance of 5.

Where it comes from

The tea plant takes lead into the leaf from acid soils and from deposition, and an extract is the leaf, concentrated. The same study makes the batch-to-batch point plainly: four of the six brands never measured above 0.7 mg/kg of lead, and the two failing batches came from a single brand whose third batch measured 2.0 — the difference is the raw material, not the category. For moringa we state a gap rather than fill it: we found no market survey of moringa leaf powders that we could read in full; the one candidate study, from Benin, we could not obtain. Moringa is therefore ranked with green tea by analogy — a dried leaf, sold in grams per day — and not by measurement.

The number

The capsule rows are unchanged. As food, dried tea leaves appear nowhere in the metals section — the regulation names Camellia sinensis only under pyrrolizidine alkaloids and perchlorate — so a tin of loose-leaf tea has no lead row while the capsule from the same harvest has 3.0. For the extract, Article 3 applies the concentration factor, and here the regulation sets the factor out itself: 4, for powdered tea extracts, in the pyrrolizidine-alkaloid row — the model for what a producer's own Article 3 file should look like.

The figure below draws the study's own table as daily doses. Seven of the eleven batches sit above California's 0.5 µg warning level at the label dose; two are over the USP line and the EU row as well; four are under all three. A brand can read its own certificate against the same three lines with one multiplication.

When to test

Per bulk lot of extract, before encapsulation, and per finished batch when the material changes hands in between. Ask the laboratory to report the result twice: as mg/kg against the EU row, and as µg per day at your label dose against the USP and Proposition 65 lines. It is the same measurement.

5Functional mushrooms: reishi, lion's mane, cordyceps, chaga

The metal

For mushrooms the answer splits by how the material was grown. For wild-collected cordyceps the metal is arsenic, and the numbers are large: across three full studies of wild Ophiocordyceps sinensis from Tibetan-plateau markets, total arsenic ran from 2 to 15.5 mg/kg, with means around 8 to 10. For the cultivated species European brands actually sell — Cordyceps militaris, reishi, lion's mane — the same laboratory series measured C. militaris at 0.036–0.047 mg/kg, two hundred times below the wild material. In cultivated fruiting bodies the metal to watch is cadmium, because fungi concentrate cadmium from whatever they grow on; but we write "watch", not "over the limit", because we found no market survey of reishi, lion's mane or chaga supplements that we could read in full.

Where it comes from

A mushroom is a filter for its substrate. The mechanism is well described: the fruiting body's metal content tracks the growing substrate — sawdust, grain, straw, compost — and species differ in how strongly they carry each element across; cadmium, mercury, lead and arsenic are the hazards the reviews name. The same species can therefore be clean from one grower and not from another, and a species name on a certificate says nothing about the substrate behind it. Chaga is the odd one out in the family: it is wild-harvested from birch trunks, not cultivated, so it inherits the wild-collection risk profile rather than the substrate one.

For wild cordyceps the arsenic story carries a documented disagreement, and we state it as one. One speciation study found about 9 % of the total arsenic inorganic; a second, on 34 market samples, reported nearly all of it as inorganic; a third found 90–100 % of it as unidentified organic arsenic species and inorganic arsenic dominant only in cultivated material. The three used different extraction and column methods, and the field has not settled. Anyone selling wild-collected cordyceps is selling that open question.

The number

The capsule rows say nothing about any of this: there is no arsenic row, so a wild cordyceps powder at 8 mg/kg of total arsenic breaks no EU supplement limit at all. The daily-dose arithmetic is different: at a three-gram daily dose it delivers about 25 µg of total arsenic, and whether that clears the USP allowance of 15 µg of inorganic arsenic, or California's 10, depends entirely on which speciation result you believe. Sold fresh as food, the rows are narrow: the lead rows name exactly three cultivated species — button mushroom, oyster, shiitake, at 0.30 — plus wild fungi at 0.80, so a fresh cultivated reishi or lion's mane has no lead row as food; cadmium has a catch-all at 0.15 for cultivated species other than Agaricus and 0.50 for wild, on the wet weight. A dried powder is judged through Article 3's drying factor, which the producer must justify. Fresh fruiting bodies are roughly nine-tenths water — general composition, not a measurement — so the factor is large and the fresh-weight rows are stricter than they look.

When to test

Cultivated material: each substrate change and each new grower, for the panel, with cadmium read first. Wild-collected material — chaga and any true wild cordyceps: each lot, and for cordyceps ask for arsenic speciation, not the total alone. The screening logic of Part 1, section 6, applies: a total below the inorganic threshold settles the question without a speciation run; a total of several mg/kg does not.

6Kelp and seaweed

The metal

Arsenic has the numbers; cadmium has the row. In EFSA's 2023 occurrence dataset — around two thousand analytical results per metal — dried kombu averaged about 55 mg/kg of total arsenic and dried wakame about 31, an order of magnitude above anything earlier in this series. Cadmium in dried seaweed averaged 1.1 to 1.7 mg/kg across the reported categories — above the general supplement row of 1.0 — while lead stayed under 0.9 and mercury was low throughout.

Where it comes from

Brown algae concentrate the elements of the water they grow in, arsenic and cadmium included; there is no field and no substrate to change — only the species and the harvest ground. The arsenic is mostly not the regulated kind: in the same dataset the inorganic fraction in kombu averages about 2.7 mg/kg against the 55 total, a twentieth. The exception the report singles out is hijiki, whose arsenic runs high specifically in the inorganic form — the reason several food authorities have long advised against eating it at all. For Laminaria and Saccharina the report treats iodine alongside the metals — a subject for another article.

The number

Cadmium is where the row itself is unusual. A capsule that is at least 80 % dried seaweed or seaweed-derived is judged at 3.0 mg/kg of cadmium (row 3.2.21.2) instead of 1.0 — a row that exists because typical dried seaweed would fail the general one. The threshold cuts both ways, as Part 1 noted: a blend that is 70 % kelp is judged at 1.0, and at the EFSA averages that is not a comfortable margin. Sold as food, dried seaweed has no cadmium, lead or arsenic row at all; the one metal row it does have is nickel — 30 mg/kg on the dry matter, 40 for wakame — a metal with no supplement row at all. As everywhere in this series, there is no arsenic row for the capsule. At the kombu average a kelp capsule carries about 55 µg of total arsenic per daily gram — invisible to the EU rows, and modest on the USP and California rulers once speciated: about 1.4 µg of inorganic arsenic from a 500 mg dose. Hijiki is a different matter entirely.

When to test

Per harvest lot: cadmium against whichever row the recipe actually triggers — the 80 % question is answered by the formulation, not the laboratory — and arsenic with speciation, because for seaweed the total alone overstates the regulated risk twentyfold. Name the species on the order; kombu, wakame and hijiki are three different answers.

7One table to keep

The four families reduce to one table: the metal, its origin, the rows on both sides of the food–supplement line, and the lot to test.