Wild and Cultivated Dang Gui: What the Classics Say, and Why Source Matters

Wild and Cultivated Dang Gui: What the Classics Say, and Why Source Matters

One name, several roots

Dang gui (Chinese angelica) is among the most valued roots in the tradition — and precisely because it is valued, "dang gui" was never one single thing. The classical materia medica knew this. It does not simply name the herb and stop; it grades the root — by where it grew, by the shape of its head and tail, by its colour and its smell — and it ranks the sources against each other without the authorities ever quite agreeing. It even records, in its own words, the shift from a root gathered in the mountains to one grown in tended fields.

This article transcribes those identifying lines, each traced to its source. It is a record of how the classics described telling one dang gui from another — a demonstration of source-traceability pointed at the problem of identity and provenance. It is not a buying guide, it carries no dosage, and it makes no claim about what dang gui does. The question is narrower: what did the texts say to look at, and can you check the quote? (For the classic case where the texts name each substitute one by one, see how the classics verify real ginseng; dang gui is the other kind of problem — not a different plant sold under the name, but the same name spread across roots of very different grade.)

Wild in the mountains, or grown in the plain — a distinction the classics already drew

Start with the name itself. In the 【释名】 section, Su Song works out the old synonyms through the Erya and lands on a distinction of habitat:

然則當歸，芹類也。在平地者名芹，生山中粗大者，名當歸也。

Dang gui, he reasons, is "of the celery kind" — the plant "growing in the plain is called qin (celery); the coarse, large one growing in the mountains is called dang gui." Wild-in-the-mountains versus grown-in-the-plain is drawn here as a difference of identity, not a modern afterthought.

And then, in the same breath, that distinction is contested — by Kou Zongshi, who says cultivation has overtaken it:

今川蜀皆以畦種，尤肥好多脂，不以平地、山中為等差也。

"In Sichuan today they all grow it in beds (畦種), and it is especially plump and rich — no longer graded by 'plain' versus 'mountain.'" So the wild-versus-cultivated question is not something a modern reader brings to the text. The text is already having the argument: one voice grades by wild habitat, another says field-cultivation has made that grading moot. Both are kept, each attributed.

The genuine root, read off its head and tail

What did the classics say genuine, top-grade dang gui looked like? Tao Hongjing (陶弘景) fixes the prized form and gives it its name:

今隴西汧陽黑水當歸，多肉少枝氣香，名馬尾當歸……歷陽所出者，色白而氣味薄，不相似，呼為草當歸。

The best, from Longxi, is "fleshy, with few side-branches, and fragrant" — this is the horsetail dang gui (馬尾當歸). What comes from Liyang is "white in colour, thin in smell and taste, not comparable" — dismissed as mere grass dang gui (草當歸). Su Gong (苏恭) sharpens it into a two-form key, using another herb — chuanxiong — as the visual reference:

有二種：一種似大葉芎藭者，名馬尾當歸，今人多用；一種似細葉芎藭者，名蠶頭當歸，即陶稱歷陽者，不堪用。

Two kinds: one "resembling broad-leaved chuanxiong, called horsetail dang gui, which people mostly use," and one "resembling fine-leaved chuanxiong, called silkworm-head dang gui (蠶頭當歸)" — the very thing Tao Hongjing had called Liyang's, and "unfit for use." Li Shizhen (李时珍) then draws the whole grade into a single line you can hold a root against:

以秦歸頭圓尾多色紫氣香肥潤者，名馬尾歸，最勝他處；頭大尾粗色白堅枯者，為鑱頭歸。

The prized horsetail dang gui has "a round head and many tails, purple in colour, fragrant, plump and moist"; the lesser spade-head dang gui (鑱頭歸) has "a big head, coarse tail, white colour, hard and dry." Round-plump-purple-moist against big-coarse-white-dry: the whole tell is in the head, the tail, the colour, and the smell of the root in your hand — not in the word "dang gui" stamped on it.

Whose origin is best? The authorities never settle it

If the grade is read partly off origin, you would want the texts to agree on which origin is best. They do not — and that disagreement is itself the record. Line them up:

• The oldest layer, the Mingyi Bielu, simply locates it: 「當歸生隴西川谷。」 — dang gui "grows in the river-valleys of Longxi."
• Tao Hongjing prizes the Longxi horsetail form (above).
• Su Gong ranks a different origin first: of the sources he lists, "Dangzhou is best" (以宕州者最勝).
• Su Song ranks yet another: 「以蜀中者為勝。」 — "Sichuan is best."
• Li Shizhen, centuries later, prizes the Qin (Shaanxi/Gansu) horsetail root — his 秦歸馬尾歸, "surpassing all other places."

Five authorities, four different "best" origins, spread across a thousand years. A database that quietly picked one and called it the provenance of dang gui would be erasing exactly this — the fact that "authentic origin" (dao di) was a moving, argued thing, and that the record preserves the argument.

Cultivation, recorded as it happened

The most striking line for the wild-versus-cultivated question is Li Shizhen's own, plain and commercial:

今陝、蜀、秦州、汶州諸處人多栽蒔為貨。

"Today in Shaanxi, Sichuan, Qinzhou, Wenzhou and such places, people mostly plant and grow it as merchandise (栽蒔為貨)." Set beside Kou Zongshi's 畦種 ("bed-cultivation") from centuries earlier, you can read the shift straight off the page: dang gui was moving from a root gathered wild in the mountains to a crop tended in fields and sold — and the texts logged it as it went, without ever declaring the cultivated root the equal, or the lesser, of the wild one. The tension is left open.

There is even a note that origin was held to change the root's very character. Li Shizhen relays Han Mao (韓懋), a Ming physician, who held the Sichuan-grown root "firm" in strength (力剛) and the Qin-grown one "gentle" (力柔) — a claim, recorded as a claim, that the same named herb was not the same material depending on where it came from. We keep it as exactly that: a historical statement about provenance and identity, tagged to who said it, not advice and not a modern finding.

The flavour line disagrees with itself, too

For completeness, the same "contested from the start" pattern that runs through ginseng and licorice runs through dang gui's nature-and-flavour line as well:

甘，溫，無毒。（《本經》）／ 辛，大溫。（《別錄》）

The Ben Jing calls it "sweet, warm"; the Bielu calls it "acrid, greatly warm." Wu Pu's note then piles on five more attributions (Shennong and Huangdi read it "sweet," Qibo and Leigong "acrid," Li Dangzhi "slightly warm"). Even the base descriptors were never single-voiced — and, as with ginseng, we keep all the voices rather than flatten them into one tidy word.

How MateriaTrace handles it: every grade traced

So "wild versus cultivated dang gui" turns, in the classics, into a stack of specific, checkable records:

1. habitat — mountain-wild versus plain-grown, a distinction Su Song draws and Kou Zongshi says cultivation has overtaken;
2. form — horsetail (round head, many tails, purple, plump, fragrant) versus silkworm-head / spade-head / grass (big, coarse, white, hard, dry);
3. origin — Longxi, Dangzhou, Sichuan, Qin, each ranked "best" by a different authority, never reconciled;
4. cultivation — recorded in the texts' own words (畦種, 栽蒔為貨) as it displaced the wild root;
5. character by provenance — the claim, tagged to Han Mao, that where it grew changed the material.

Here is how we hold that: each line is kept verbatim, attributed to the authority who said it (弘景, 蘇恭, 蘇頌, 時珍, 韓懋), placed in its section of the Bencao Gangmu, and traced to the source. Where the authorities disagree — which origin is best, whether cultivation matters — we keep the disagreement visible rather than choosing a winner. We add no modern verdict on what dang gui is or does; we record how the classics described recognising it.

Worth stating plainly: this is a record of how classical texts described distinguishing grades and origins of dang gui, kept for study and traceability. It is not advice on buying, identifying for purchase, using, or dosing any herb, and it makes no claim about dang gui's effects. MateriaTrace records what sources say and makes no treatment recommendation; any question about a herb belongs with a qualified professional.

Why source matters — and why the label was never enough

It would be easy to write one line — "dang gui: root of Angelica sinensis, from Gansu" — and be done. But that line throws away what the classics actually built: one name spread across roots of sharply different grade, a form-key you can read off the head and tail, four "best" origins that never reconciled, and a cultivation shift the texts recorded as it happened. Those survive only if the exact lines survive, each attributed, each traceable — so that a later reader can see, as Su Gong and Li Shizhen did, that "dang gui" on a label tells you the name and not yet the root.

See the roots themselves: dang gui (Chinese angelica); the chuanxiong the texts used as the visual reference for its two forms; and tu dang gui, a different plant that carries the "dang gui" name and has to be kept apart from it. For how the classics name each substitute of a prized herb one by one, read about verifying real ginseng and the herbs most often faked.

If this "every line clicks back to its own source" approach is useful to you, see what else is in MateriaTrace on the homepage.

MateriaTrace is a source-traceable, bilingual knowledge graph of Chinese medicine. It records how classical and modern sources describe each entry — each field traceable to its primary source, contradictions kept side by side — and makes no treatment or usage recommendations. Every line quoted in this article can be traced to its classical source.