Direct answer: 江南李主帐中香 belongs to the classical 合香 (hé xiāng, layered blending) tradition and carries the name of 李煜 (Li Yu), the Southern Tang’s last ruler — “江南李主” being “the Lord of Jiangnan.” In its best-known form, agarwood powder and a trace of sandalwood are packed into hollowed goose pears, steamed three times, and left with a clean, sweet fruit character that enters the wood before burning.
Key facts: 沉香末一两，檀香末一钱，鹅梨十枚 — one liǎng of agarwood powder, one qián of sandalwood powder, ten goose pears. The 香乘 (Xiang Cheng) preserves several versions of the formula, which makes this one of the best-documented recipes in the classical canon. The attribution to Li Yu is traditional rather than documented, and the pear cultivar named in the text is not fixed today.
Key Takeaways
- Ten parts agarwood to one part sandalwood: that ratio, not the pear count, is what keeps the chief in charge.
- The manuals hold three deliveries of one idea — pear infusion, pear juice, and a resin-spice base — with agarwood central in all of them.
- The pear works as a vehicle; its juice enters the wood and lifts the top of the scent rather than adding a competing material.
- 两 and 钱 are proportions in disguise. Their gram values changed with the dynasty, so exact weights promise more than the text can deliver.
The Formula and Its Sequence
Direct answer: The flagship recipe is short enough to memorize: one liǎng agarwood powder, one qián sandalwood powder, ten goose pears, hollowed and steamed three times.
沉香末一两，檀香末一钱，鹅梨十枚 — one liǎng agarwood powder, one qián sandalwood powder, ten goose pears.
The method is compact. Hollow each pear through the top as though making a small jar, fill the cavity with the pre-mixed powders, and replace the pear’s own top as a lid. Steam the filled fruit three times — 蒸三溜, “three steamings” — then strip the peel, grind and blend the softened contents, and set the result aside to age (久窨) before burning.
Every instruction closes a failure path. Filling without sealing lets steam flood the cavity and wash the powder out. Steaming once leaves the infusion at the surface. Peeling before the third pass exposes the fruit to direct heat and drives off the volatiles the process is trying to trap. That pattern — each step preventing a specific error — is characteristic of formulas refined by practice rather than composed at a desk.
Four Roles, One Fruit
Direct answer: The blend assigns 君 (chief), 臣 (minister), 佐 (assistant), and 使 (guide) — and this recipe is unusual because the assistant is a fruit and the guide is a process rather than a material.
| Role | Material or step | What it contributes |
|---|---|---|
| 君 chief | 沉香 agarwood | The resinous body the whole formula exists to carry |
| 臣 minister | 檀香 sandalwood, one tenth of the agarwood | Warmth and roundness at the edges, without competing |
| 佐 assistant | 鹅梨 goose pear | A clean, sweet top note delivered through the juice |
| 使 guide | Steaming and aging | The process that moves the fruit into the wood and settles it |
君臣佐使 describes function, not price. A cheap material can be an assistant if its job is to shape the top note; an expensive one can be a minister if its job is to support without leading. This formula makes the logic visible because nobody can mistake a pear for a precious wood. The pear is present for one reason, does one thing, and is dosed accordingly.
That visibility is the teaching value. When every component is a rare resin, the role structure disappears into a haze of luxury and the reader learns nothing about proportion. Put a fruit in the assistant’s seat and the structure becomes legible: the chief dominates by weight, the minister supports at a fraction, the assistant shapes the surface, and the guide turns the parts into one scent.
Steam as Transfer
Direct answer: Steaming is a transfer process, not a ritual flourish — heat and water carry the pear’s juice into the agarwood’s pore structure, and three passes are how the recipe reaches depth without scorching the wood.
Follow the transfer. Steam surrounds the sealed pear and heats it through → the fruit’s juice, mostly water with dissolved sugars and faint aromatic compounds, thins and begins to move → the warm liquid wets the agarwood particles packed inside the cavity → sugars and volatiles diffuse from the wet surface into the resin-rich fibers along a concentration gradient → as the pear cools between passes, the wood draws the liquid inward by capillary action.
One pass saturates the surface. A second re-wets the material and pushes the front deeper; a third extends it further. Cycling is the point — each heating opens the structure, each cooling pulls the liquid in. A single long steam is not equivalent, because continuous heat keeps the interior expanded and resistant to uptake, while cycling lets the material contract and draw.
The peel comes off afterward because it has given up its juice. Leaving it in would add spent fruit fiber to the burn and a flat, cooked note in place of the clean sweetness the fruit was there to provide.
The Ten-to-One Discipline
Direct answer: One liǎng of agarwood against one qián of sandalwood is a 10:1 ratio, and that gap is the difference between a blend with a clear chief and two woods arguing.
Run the balance as a chain. Agarwood supplies the dark, resinous spine → sandalwood at a tenth adds a soft, slightly milky warmth around the edges → the two read as one continuous scent rather than a mixture → the chief stays audible from first minute to last. Move sandalwood toward parity → its creamy tone covers the agarwood’s bitter-resin detail → the blend flattens into generic warm wood → the formula loses the characteristic it was built to display.
Ten to one is not a magic number, and the text makes no claim that it is optimal in any measurable sense. It is a discipline expressed numerically: the minister exists to support, and at a tenth of the weight it can support without speaking over the chief. The same pattern runs through the classical corpus, where the chief’s share dominates and the supporting materials divide a much smaller remainder.
For a modern blender, the practical consequence is direct. If a substitution pushes the sandalwood share upward — because the available agarwood is thin or the sandalwood unusually loud — balance is the first thing lost, and aging will not restore a chief that was dosed too low.
Three Versions, Three Deliveries
Direct answer: The 香乘 preserves the same idea in three forms — fruit infused into powder, juice steamed into wood, and a resin-and-spice base — and they are variations on a theme rather than rival recipes.
| Feature | Goose-pear | Juice-steamed | Resin-and-spice |
|---|---|---|---|
| Fruit note arrives by | Pear juice drawn into powder | Pear juice steamed into wick-sized wood | No fruit; resins and spice instead |
| Agarwood | One liǎng, powdered | One liǎng, cut to wick size | Four liǎng |
| Sandalwood | One qián | None stated | One liǎng |
| Other materials | Ten pears | One pear, minced to juice | Musk one liǎng, borneol half a liǎng, 马牙香 one fēn |
| Binder | The fruit and its juice | Reduced pear juice | Refined honey |
| Ratio reading | 10:1 agarwood to sandalwood | Single-material infusion | 4:1 agarwood to sandalwood, musk equal to sandalwood |
The second version is the sparest: agarwood cut into wick-like pieces, one pear minced and pressed to juice, and the whole steamed three times in a silver vessel until the juice is dry. No sandalwood appears in the record, which makes it a pure test of the infusion method — if the fruit note is the goal, this version isolates it.
Version three inverts the balance of support. Its charge is four liǎng of agarwood, one of sandalwood, one of musk, half a liǎng of borneol, and one fēn of 马牙香; the materials are chopped fine, left unsieved, mixed with refined honey, and burned. Musk and borneol are assertive aromatics, so the character shifts away from fruit toward something darker and warmer. Agarwood still holds four parts against one, even with a fuller supporting cast.
A further entry takes a different route: agarwood soaked in storax oil (苏合油) for a hundred days, then finished with rose water (蔷薇水). No pear and no steam cycle — just a long soak and a floral top note.
Units: Ratios, Not Grams
Direct answer: The recipe’s units are 两, 钱, and 分, and their gram equivalents drifted across dynasties, so the only reliable reading is proportional.
| Unit | Relationship | Approximate Song–Ming weight |
|---|---|---|
| 两 (liǎng) | Base unit | About 37 g, with period values roughly 31–41 g |
| 钱 (qián) | One tenth of a liǎng | About 3.7 g |
| 分 (fēn) | One tenth of a qián | About 0.37 g |
Use those figures to orient yourself, not to dose. Two manuscripts separated by centuries can print identical characters for amounts a third apart, so a literal gram conversion imports an error into the blend. The proportion is what travels intact across copies: ten to one in the pear formula, four to one in the resin version, with musk level with sandalwood in the second.
So the instruction worth carrying forward is a proportion, not a weight: ten parts chief to one part minister. Anyone scaling a batch keeps the design intact that way. Following a single dynasty’s gram figure instead reproduces one manuscript’s copy of the recipe, with its own local measures baked in.
Aging, and the Hundred-Day Variant
Direct answer: 久窨 — prolonged sealing and rest — lets infused volatiles redistribute, and the storax variant shows the same principle stretched over months instead of days.
Aging works quietly. After steaming, the fruit’s sugars and aromatic compounds sit unevenly through the agarwood, concentrated where the liquid entered → sealed in a container, those compounds keep migrating along the gradient → the sharp, freshly steamed edges soften as surface and interior equilibrate → the blend reads unified rather than layered in patches. Mixing works between particles; diffusion works within them.
The hundred-day soak is the same mechanism at a different clock speed. Storax oil surrounds the wood → its aromatic compounds migrate into the resin channels over weeks → the wood’s own resin softens and blends with the incoming material → a rose-water finish adds a bright floral top note. Steam drives a water-based carrier in fast and shallow; oil soaks slowly and deep. The manual keeps both because they answer one question at different timescales.
Read 久窨 as a test as much as a step. A batch that has rested and still smells of two separate materials was probably under-infused or badly proportioned; time cannot create a fusion the process never started.
What the Attribution Can and Cannot Carry
Direct answer: The link to Li Yu is traditional and widely repeated, but no signed document ties the formula to him, and the recipe does not need one.
The formula survives in Ming-era manuals compiled centuries after the Southern Tang, which means the text passed through many hands before it reached print. “江南李主” points clearly at Li Yu — last ruler of the Southern Tang, a poet whose work defined the fall of a kingdom — and the name carries a specific atmosphere: courtly, refined, elegiac. That atmosphere is part of why the recipe is remembered.
My position, stated as editorial judgment: keep the attribution as a pointer and decline to defend it as a deed. The name tells you what world the recipe came from and what kind of scent it was meant to be. It does not tell you who wrote the proportions, and treating it as authorship sets up a test the evidence cannot pass.
This is the author’s view, not a verified fact.
The same restraint applies to the pear. 鹅梨 names an old variety, likely related to what is now called 鸭梨, but the cultivar is not fixed and cannot be reconstructed with certainty. A modern attempt uses a different fruit and produces a different top note — a limit of the source rather than a failure of the maker.
Bedchamber as a Brief
Direct answer: 帐中香 means “incense within the bed-curtains,” and the name is a functional brief: a soft, low-smoke scent for a small enclosed space rather than a smoke that fills a hall.
The setting constrains the formula. A curtained bed is small and poorly ventilated → a strong resin-heavy smoke would overwhelm it → the agarwood is softened with fruit and a trace of sandalwood → the result reads gently at close range. The same chief wood, designed for a temple hall, would be dosed and supported differently, with more smoke and less top note.
Reading the name as a specification explains choices that otherwise look arbitrary. The pear is not there to make the incense poetic; it lifts a heavy base into a range that works in a confined space. The low sandalwood dose keeps the blend from thickening the air. Even the aging step fits: a settled scent is less aggressive at close range than a freshly blended one.
Reconstructing It Today
Direct answer: The structure can be followed today, but three things cannot be recovered exactly — the pear cultivar, the original wood, and the classical weights — so a modern attempt is an interpretation rather than a reproduction.
What is stable is the ratio, the sequence, and the role assignments. Agarwood powder with a tenth of its weight in sandalwood is workable at any scale, and hollowing, sealing, and steaming three times can be repeated with ordinary equipment — the text’s silver vessel is a heat-conduction choice, not a magic ingredient.
What cannot be matched is the material. Old-growth agarwood with dense resin is not plantation wood, and the 鹅梨 is gone as a named cultivar. Judge the result against the formula’s logic — does the chief dominate, does the fruit sit above rather than beside, does the blend read as one thing — rather than against a memory of how the original might have smelled. [need-source] on exact modern substitutions for 鹅梨.
FAQ
Q: What kind of incense is the Jiangnan Lord’s bedchamber blend?
Tradition hands the recipe down under 李煜 (Li Yu), the last Southern Tang ruler, and the intended setting is a curtained bedchamber. The best-known version steams agarwood powder with a small measure of sandalwood inside hollowed goose pears, giving the wood a clean fruit sweetness before it burns. Several variants appear in the 香乘, which is why the recipe is so well documented. The Li Yu link is a name inherited, not an authorship established by evidence.
Q: Why is the assistant a pear rather than another wood?
Because the job is a top note, and fruit delivers that more directly than resin would. A second precious wood would compete with the agarwood chief in the same register; a pear sits above it with a clean, slightly cool sweetness that never challenges the base. The choice also makes the role system visible, since no reader can mistake a pear for a luxury material. Function, not rarity, decides a role in 合香.
Q: Why steam the pears three times instead of once?
Each cycle drives the juice a little deeper. The first pass wets the surface of the agarwood particles, the cooling that follows draws liquid inward, and the next pass re-wets and pushes the front further. One long steam is not equivalent, because continuous heat keeps the material expanded and resistant to uptake. Three passes are the text’s way of reaching depth without scorching the wood.
Q: Are the gram conversions reliable?
Only as a rough guide. One liǎng was around 37 g in the Song–Ming period, with period values spanning roughly 31–41 g, and 钱 and 分 scale down from there. Because the unit drifted, a literal conversion reproduces one copy of the text rather than the recipe. Work from the proportion — ten parts agarwood to one of sandalwood in the pear version — and the design survives the change in measures.
Q: Can this be made at home, and will it smell the same?
The procedure can be repeated with ordinary equipment, and the ratio can be scaled. The scent will not match, because the named pear cultivar is no longer identifiable and today’s agarwood differs from the historical material. Treat a modern attempt as an interpretation: check whether the chief dominates, whether the fruit sits above the base, and whether the blend reads as one scent after resting. Those tests come from the formula and do not depend on reproducing a lost ingredient.
Sources
- 香乘 (Xiang Cheng), a Ming compilation by 周嘉胄 — source for the 帐中香 entries: pear-steamed, juice-steamed, and resin-and-spice. Full text on Wikisource.
- 陈氏香谱 (Chen’s Incense Manual) — an earlier collection that keeps a second formula under the same title.
- 本草綱目 (Bencao Gangmu) — the materia medica listing agarwood, sandalwood, and the other materials used in the variants. Full text on ctext.org.
Related Resources
- Chinese incense recipes (香方) — the translated library this page belongs to, with the unit conventions used across the texts.
- Agarwood incense guide — the chief material, its grades, and what separates resinous wood from filler.
- The 合香 blending system — how 君臣佐使 works as a design method rather than a price ladder.
- Incense authentication database — how agarwood’s value invites substitution, and what to check.
Summary
- 江南李主帐中香 sets ten parts agarwood against one part sandalwood, with ten goose pears delivering the fruit note.
- Its assistant is a fruit and its guide is a procedure, which exposes the 君臣佐使 structure more plainly than a formula built from rare woods.
- The manuals hold three deliveries — pear-infused, juice-steamed, and resin-and-spice — and one long storax soak, all organised around agarwood.
- Dynastic weights drifted, so the ratio is the durable instruction and grams are only an approximation.
- Heat moves juice into the wood and rest settles it; neither can compensate for a chief dosed too low.