What the regulation requires

The rules are 21 CFR 101.4(h) and, for source ingredients shown inside the Supplement Facts panel, 21 CFR 101.36(d)(1). They are more specific than most people expect.

Botanical common names on a supplement label must be "consistent with the names standardized in Herbs of Commerce, 1992 edition." That is a reference book published by a trade association, incorporated into federal regulation by name and edition.

The common name must then be followed by two things:

The part of the plant. The regulation gives its own example: "'Garlic bulb' or 'Garlic (bulb)'." The part must be named in English: "flower," not "flos." There is one exception: the designation is not required for algae.

The Latin binomial, in parentheses: "except that this name is not required when it is available in the reference entitled: Herbs of Commerce" for the common name on the label.

Read that exception carefully, because it inverts the usual assumption. The Latin name is required precisely when the common name is not in that book. A label with no Latin name is not necessarily hiding anything; it may simply be using a standardized common name that the regulation accepts on its own.

There is one more wrinkle: the regulation still points at the 1992 first edition. The publisher has since issued a second edition in 2000 and a third in 2023, with thousands of species and hundreds of Ayurvedic and pinyin names added. US labels remain legally keyed to a book from 1992.

Three cases where the species matters

Cinnamon

The clearest example, because the difference is not academic.

Ceylon cinnamon is Cinnamomum verum, described by the National Center for Complementary and Integrative Health as "'true' cinnamon," grown mainly in Sri Lanka. Cassia cinnamon, which NCCIH identifies as Cinnamomum aromaticum, is "the most common type sold in North America."

The difference that matters is coumarin. NCCIH states that "some cassia cinnamon products contain high levels of coumarin," that "interactions between coumarin, a chemical found in cassia cinnamon, and the liver have been reported," and that "prolonged use of cassia cinnamon could be an issue for sensitive people, such as those with liver disease." Ceylon cinnamon, by contrast, "may contain a trace amount of coumarin."

Two products, both labeled cinnamon, with a meaningful difference in a compound that matters for some people. The Latin name is how you tell them apart.

(A small note for accuracy: cassia cinnamon is called both Cinnamomum aromaticum and Cinnamomum cassia in different sources. These are two names for the same plant, not two plants.)

Ginseng

Asian ginseng is Panax ginseng. American ginseng is Panax quinquefolius. And "Siberian ginseng" is Eleutherococcus senticosus, a different genus entirely, in a different part of the plant family, sharing the marketing word and not much else.

A product labeled simply "ginseng" has not told you which. And the research literature on each is separate, so a claim built on studies of one species sitting next to a bottle of another is a category error, however sincerely made.

Elderberry, where the plant part is the safety question

Elderberry is Sambucus nigra, and the part used is the dark purple berry.

NCCIH is direct about the rest of the plant: "Raw or unripe elderberries and other parts of the elder tree, such as the leaves and stem, contain poisonous cyanide-producing substances that can cause nausea, vomiting, and severe diarrhea; cooking eliminates this toxin."

That is a case where "which part" is not a quality question but a safety one, and it is exactly the question the plant-part requirement exists to force onto the label.

Why the plant part matters more generally

Different parts of the same plant are different material, and the published work on botanical adulteration documents this repeatedly.

Asian ginseng: root extracts are sometimes diluted with leaf extracts. Both contain ginsenosides, but in different ratios: the relative concentrations of ginsenosides Rb1 and Rc are higher in the roots, while leaf extracts carry more Rd and Re. A test looking only at total ginsenosides can be satisfied by the wrong plant part.

Ashwagandha: root extracts are sometimes diluted or substituted with leaf extracts, since many of the same withanolides occur in both. Leaves and stems carry more withaferin A, and the aerial parts contain flavonol glycosides absent from the roots.

Eleuthero: root extracts may be adulterated with aerial parts, and the researchers are candid that this one is hard to catch: no clear distinction criteria have been established, because the chromatographic profiles of roots, stems and leaves look quite similar.

Milk thistle: the part used is the fruit, and one documented adulterant is "exhausted" seed: material that has already been extracted once and is being sold again.

So "ashwagandha 500 mg" and "ashwagandha root extract 500 mg" are not the same claim. The first is vaguer than the regulation contemplates; the second tells you what you are buying.

How common is adulteration?

Here we have to be careful, because the most widely repeated number in this field came from a paper that no longer stands.

A 2013 study using DNA barcoding on North American herbal products reported that most products tested contained species not listed on the label. That paper was retracted in July 2024. The retraction notice states that an investigation by the authors' university "found evidence of data fabrication in relation to this article," and that the editor "no longer has confidence in the presented data."

That paper is where the widely circulated claim about herbal mislabeling originated. Whatever figure you have seen quoted from it should not be repeated.

The honest number is more modest and better founded. A 2023 review by researchers involved in the main US botanical adulteration monitoring program reports that two separate reviews of the published literature "found the same percentage of adulterated materials (ca. 27%) independent of researchers using genetic or chromatographic and spectroscopic assays of authentication."

Around a quarter of materials examined, by two independent analytical approaches. That is a serious number, and it is a good deal smaller than the withdrawn one. The difference between an honest problem and an exaggerated one matters for whether anyone believes the rest of what you are told.

The same review adds an important qualifier: adulteration "usually represent[s] a form of economic fraud," and "most types of adulteration do not constitute a safety risk, although there are some notable exceptions."

What adulteration looks like

The documented examples are more inventive than "they substituted a cheaper plant."

• Turmeric extracts spiked with synthetic curcumin, so the marker compound reads correctly
• Ginkgo leaf spiked with rutin to hit a flavonoid specification
• St. John's wort spiked with a cocktail of four food dyes, including brilliant blue, which absorbs light at 590 nanometers, the same wavelength at which hypericin is measured
• Black cohosh substituted with related Asian species
• Bilberry and elder berry extended with black rice extract, which supplies the anthocyanins a test looks for
• Cranberry extended with black rice or hibiscus for anthocyanins, and grape seed, peanut skin or pine bark for proanthocyanidins
• Saw palmetto cut with vegetable oils

The reviewers state the pattern outright: "In most cases, adulteration is done by providing botanical ingredients that appear to comply with specifications and standardization requirements for specific marker/active compounds but, in fact, simply exploit the lack of specificity of the test method used to measure them."

The adulterant is chosen to pass the test, which means a certificate of analysis showing the marker compound is present, on its own, is not proof of identity.

What about DNA testing?

DNA barcoding sounds like it should settle this. It is useful, and it is also more limited than its reputation.

The core problem is that extraction destroys DNA. A team at FDA's own Center for Food Safety and Applied Nutrition compared chemical and DNA barcoding methods across 112 market samples of ginkgo, soy, valerian, yohimbe and St. John's wort, and noted that concerns have been raised about "the appropriateness of using DNA barcoding techniques with finished botanical products." Separate research recommends against using universal primers on processed plant material "as a sole means of species identification," favoring a combination of targeted DNA methods, chemical analysis and microscopy.

Real-world amplification failure illustrates the issue: in one study, DNA suitable for analysis could not be extracted from three of forty supplements tested; in another, from three of thirty-seven.

And DNA tests can be defeated deliberately. The adulteration reviewers note that genetic methods are "relatively easy to fool, i.e., by using a different part of the labeled plant (e.g., Asian ginseng leaf rather than root) or by adding inert materials (lactose, maltodextrin) that do not contain any DNA." A DNA test confirming Panax ginseng does not confirm that the material is root.

One appealing detail from the same review: there are no known examples of adulterants selected specifically to deceive botanical microscopy. The old-fashioned method is, in that narrow sense, the hardest one to game, though it only works on material that still has recognizable structure.

What good identity testing looks like

Federal manufacturing rules require a manufacturer to "conduct at least one appropriate test or examination to verify the identity of any component that is a dietary ingredient." What counts as appropriate depends on the material.

• Whole or cut botanical: macroscopic examination plus microscopy
• Powders: microscopy plus chromatographic fingerprinting
• Extracts: chromatographic and spectroscopic fingerprints, since morphology is gone
• Essential oils: specialized techniques including stable-isotope and enantiomer ratio analysis, which detect synthetic or diluted oils
• DNA methods: valuable for raw and whole material, and for targeted species questions, but not as a sole method on processed extracts

The more processed the material, the less morphology and DNA can tell you, and the more the chemical fingerprint has to carry.

A word on "standardized"

Standardization is often presented as the premium option. It is useful, and it is narrower than it sounds.

The NIH Office of Dietary Supplements is unusually blunt about it: "U.S. law does not require dietary supplements to be standardized. In fact, there is no legal or regulatory definition of the term in the United States." And on the markers themselves: "Ideally, the chemical markers chosen for standardization would also be the constituents that are responsible for a botanical's effect in the body… However, the constituents responsible for the effects of most botanicals are not known."

So standardization does help with batch-to-batch consistency, which is worth something real. What it does not do is establish species identity, since markers "are quite often molecules that are found in many plants," and, as the adulteration examples show, a standardization specification is frequently the exact target the fraud is aimed at.

What to look for on a label

1. A Latin binomial, where the common name is not one the regulation accepts on its own
2. The plant part, in English: root, leaf, fruit, bulb, rhizome, aerial parts
3. The preparation (whole powdered herb, dried extract, liquid extract) and the extract ratio or standardization if there is one
4. Consistency between the label and the research you have seen cited. If a claim rests on studies of a root and the bottle contains aerial parts, or on one species and the bottle contains another, the citation and the product do not match.
5. Testing you can verify. Ask which identity method was used on the botanical: macroscopy, microscopy, chromatography, DNA, or some combination. A company that authenticates its botanicals properly knows exactly how to answer that, and the specificity of the answer is the signal.