Nobody buys a plant. A buyer of botanical extracts buys a number, and the number is a convention agreed by the trade rather than a property of the plant. Ginkgo leaf contains over 140 isolated compounds; the material on the shelf is described by two of them. Grape seed sold as "95% OPC" by one route of analysis can be defended at a materially lower figure by another. The specification sheet, not the headline percentage, is what protects a formulation, and six extracts between them illustrate five recurring traps.

The marker is a convention, not a molecule

Where the ingredient is a single substance the question is easy: chlorogenic acid is chlorogenic acid. Where it is a plant, someone had to decide which compound represents the whole, and that decision became the specification. Two quotations for the same headline percentage can therefore describe different materials, and the only way to compare them is to compare the marker, the plant part and the analytical method behind the number.

UV and HPLC figures are not interchangeable

The distinction bites hardest on the polyphenol extracts. A UV reading counts everything absorbing at the chosen wavelength, which for a proanthocyanidin extract includes gallic acid and the polymeric tannins; an HPLC reading counts individual monomers. One lot can be simultaneously "95%" and considerably less, with nobody having mis-stated anything. Quote both, and ask for the chromatogram: a supplier who will not send one has told you something useful.

Milk thistle: what the silymarin number leaves out

Milk thistle extract, CAS 84604-20-6, is taken from the seed of Silybum marianum and standardised on silymarin, a group of flavonolignans of which silybin is the best known. Standard grades run 80% to 98% silymarin by HPLC — a light yellow to brown powder from ethanol or supercritical CO₂ extraction, with residual ethanol below 50 ppm against ICH Q3C and heavy metals held to lead ≤ 1 ppm and arsenic ≤ 0.5 ppm. Two grades fifteen points apart are not the same product commercially: the 80% powder moves in drums as a commodity, while the 98% material carries a minimum order near 10 kg and a two to three week lead time. Decide which the formulation needs before negotiating — the answer changes who can supply you.

Ginkgo: two markers, and an acid nobody puts in the headline

Ginkgo biloba extract, CAS 90045-36-6, is the one entry the pharmacopoeia describes with two unrelated markers at once — flavonol glycosides at 24% or more and terpene lactones at 6% or more, the latter being bilobalide and the ginkgolides. The two families are chemically unrelated, so one tells you nothing about the other: a lot can reach its flavonoid figure comfortably while running short on lactones. The frame is EP 10.0 and USP 43, stricter than a house standard. The parameter that quietly decides acceptance in regulated markets is ginkgolic acid — derivatives in the leaf that carry sensitising potential and are capped at a few parts per million. An excellent flavonoid figure with no ginkgolic acid statement is an unfinished quotation. The powder is light brown and water-soluble.

Rhodiola: the ratio matters as much as the total

Rhodiola rosea extract, CAS 97404-52-9, is standardised on salidroside at 3% or more together with rosavins at 3% or more, again by HPLC, as a brown water-soluble powder. The plant grows wild between roughly 1 800 and 2 500 metres, which makes the crop a harvest rather than a plantation. A lot can present a healthy total while leaning heavily on salidroside and running light on rosavins, so ask for both figures separately rather than a combined claim.

Ashwagandha: root, leaf, and the carrier you did not order

Ashwagandha extract, CAS 90147-43-6, comes from the root of Withania somnifera and is standardised on withanolides between 2.5% and 10% by HPLC, as a brown to tan powder, with residual ethanol below 100 ppm under FDA 21 CFR 172.510. Two questions separate quotations. The plant part: leaf-derived material carries withanolides too and costs a fraction of root material, while the pharmacopoeial article is the root — put it in writing. Then the carrier — a powder formulated to 5% is usually the extract triturated with maltodextrin, which is part of what you are buying, so its percentage belongs on the specification.

Grape seed: buying the pair of numbers, not the larger one

Grape seed extract, CAS 84929-27-1, from Vitis vinifera seed, is the clearest case of the analytical gap above: oligomeric proanthocyanidins at 95% or more by UV, with catechin and epicatechin held to 5% or less by HPLC. Both numbers on one sheet marks a supplier who understands the material. Production is water extraction followed by column chromatography, giving a reddish brown powder soluble in water and ethanol, with moisture below 5% by Karl Fischer. That figure deserves attention: the powder is hygroscopic, and a colour shift towards brown in storage is visible evidence that oxidative degradation has begun.

Elderberry: the least stable family here

Elderberry extract, CAS 84603-58-7, is standardised on anthocyanins at 25% or more by HPLC with flavonoids at 5% or more by UV — a dark purple 80 mesh powder, water-soluble, commonly produced by supercritical CO₂ extraction to keep solvent residue below 50 ppm. Anthocyanins are the most fragile family in this group: pH, heat and light all degrade them, which makes colour strength a working proxy for how much has already been lost. Ask what stability data accompanies the lot and keep a retention sample; a dispute about an unstable active is unwinnable without one.

What to put on the enquiry

Six lines make botanical quotations comparable: the marker and the analytical method, with a chromatogram; the plant part; the extraction solvent and its residue limit; the carrier used for standardisation and its percentage; the limits for heavy metals, pesticide residues and microbiology; and a retention sample held against the lot. Packaging is consistent across the range — 1 kg nitrogen-flushed foil bags inside 25 kg PE-lined drums, with a deoxidiser for the polyphenol-rich powders — and shelf life is 24 months stored cool, dry and dark. None of this makes botanical extracts hard to buy; it makes them hard to buy on a percentage alone.

The rest of the shelf

The same discipline travels across the catalogue. Chlorogenic acid from green coffee, CAS 327-97-9, is a genuine single molecule and therefore the easy end of the business, while moringa leaf extract, CAS 93165-54-9, standardised on polyphenols at 20% or more, needs the carrier question asked. Tannic acid, CAS 1401-55-4, and galla chinensis extract sit in the astringent and tanning tradition, where the assay convention is long settled. Aloe vera extract, CAS 85507-69-3, is bought on polysaccharide content rather than a single marker. Guarana, CAS 84929-28-2, and yerba mate, CAS 68916-96-1, are natural caffeine sources where the caffeine figure is the specification. Acai, CAS 879496-95-4, camu camu, CAS 50-81-7 — largely a natural vitamin C article — and pomegranate extract complete the antioxidant shelf. Different plants, one first question: which marker, by which method.