Skip to main content
Support

Browse by category

All categories
Home

Fermented milk from China

Lao, su and tihu (酪, 酥, 醍醐), historical Chinese milk products酪

Early Chinese texts rank milk products in a chain, milk to lao (soured milk), to raw su, to cooked su, to tihu, the finest. The Qimin Yaoshu gives working recipes for lao and su; tihu is known only from short notes quoted in the Bencao Gangmu, so its step here is a reconstruction from those notes, not a tested recipe.

Fully documentedFermented milk11 days

Recipe

Amounts from a sourceAmounts are the texts' own, in old Chinese measures. The Qimin Yaoshu gives half a spoon of old lao to 1 sheng of boiled milk, hot water equal to half the lao and then cold water equal to the hot water; the Bencao Gangmu gives 3–4 sheng of tihu from 1 dan of good su.
Cow's or sheep's milk[qmys-ws]
Old lao as the starterHalf a spoon per sheng of boiled milk[qmys-ws]
Good su (rendered butter), for tihu1 dan of good su gives 3–4 sheng of tihu[bcgm-ws]
Starterback-slopped with old lao for the soured milk

Total time across the recorded steps: 11 days

Start
Drag the bar or press Play
shapes, not to countLactic acid bacteria38 °CpH 6.6 → 3.8Milk sugarLactic acid

What changes: Boiled milk set with old lao sours overnight into lao

Step 1 of 4
Sour boiled milk into lao
Time
19 h
Reported
Temperature
38 °C
Reported
Salt
None given
Acidity
pH 6.6 → 3.8
Measured
Which microbes the steps name, step by step
Lactic acid bacteria

Full recipe under qmys-lao-soured-milk. The Qimin Yaoshu does not name a starter beyond old lao, so the reproduction used a commercial sour milk. It found the lao to be 86.8 % moisture, 3.3 % protein, 3.8 % fat and 5.0 % carbohydrate, and judged it a sour milk. Output: lao (soured milk).

Flavour
Not documented
Texture
Not documented
Colour
Not documented

Microbes at work in this step

lactic bacteria from the old lao sour the milk

Method

Measured measured in a studySet fixed by a specification or tested recipeReported one source reports itTypical a general value, not sourced for this fermentUnconfirmed not yet checked against a source

At a glance

Early Chinese texts rank milk products in a chain, milk to lao (soured milk), to raw su, to cooked su, to tihu, the finest. The Qimin Yaoshu gives working recipes for lao and su; tihu is known only from short notes quoted in the Bencao Gangmu, so its step here is a reconstruction from those notes, not a tested recipe. Made in 4 steps over 11 days. The longest step, render raw su into cooked su, takes 10 days (Jia Sixie 2026). In tarag, a Mongolian soured milk related to lao, Lactobacillus delbrueckii was the predominant species (74.6 % average abundance across 26 samples), with L. helveticus and Streptococcus thermophilus also found does most of the fermenting.

Key numbers
Main step10 daysReported[qmys-ws]Render raw su into cooked su
Temperature38 °CReported[hirata2010]Sour boiled milk into lao
AciditypH 6.6 → 3.8Measured[hirata2010]Sour boiled milk into lao

What makes this one different

Only the first step is a fermentation; su and tihu are fat separated from it. Tihu is known from short notes, not a full recipe.

Related ferments

Safety

The Qimin Yaoshu says a damp jar makes the lao fail to set, so a jar that has held lao before is fired over embers until dry and cooled before use. Setting the lao too warm makes it sour and too cold keeps it from setting; a starter that is too sour or too much sweet lao also sours it. The same chapter adds that dried lao and strained lao last for years but taste stale, and that strained lao not fried breeds worms and cannot be kept past summer.

History and sources

Every number is labelled with how it is known. How this atlas is made

In 2010 Hirata and colleagues reproduced the Qimin Yaoshu recipes and identified lao as a sour milk and su as butter and butter oil; they could not identify the dried lao. The Nirvana Sutra, in its southern 36-juan Chinese text, likens the stages of teaching to milk becoming lao, raw su, cooked su and tihu, and calls tihu the best, curing every illness. Of the writers quoted in the Bencao Gangmu, Tao Hongjing cites this Buddhist chain, and Su Gong calls tihu the essence of su.

  1. [qmys-ws] Jia Sixie, Qimin Yaoshu (6th century), vol. 6, ch. 57 養羊, 作酪法 and 抨酥法; Chinese text on Wikisource (齊民要術/卷第六), read 2026-10archive
  2. [bcgm-ws] Li Shizhen, Bencao Gangmu (1578), 獸之一, entry 醍醐 (quoting Tao Hongjing, the Tang Bencao of Su Gong, Han Baosheng and Kou Zongshi); Chinese text on Wikisource, read 2026-10archive
  3. [hirata2010] Hirata M., Yoneda Y., Ariga H., Uchida K., Motoshima H., Hanada M., Kawai M. 2010, 『斉民要術』に基づいた東アジアの古代乳製品の再現と同定 (Reproduction and identification of ancient dairy products in East Asia based on 'Seiminyoujyutu'), Milk Science 59(1):9–22 (full text read on the Japanese agriknowledge repository)paper
  4. [oki2014] Oki K., Dugersuren J., Demberel S., Watanabe K. 2014, Pyrosequencing analysis of the microbial diversity of airag, khoormog and tarag, traditional fermented dairy products of Mongolia, Bioscience of Microbiota, Food and Health 33(2):53–64paper
  5. [nirvana-ws] Mahāparinirvāṇa Sūtra (大般涅槃經), 36-juan text, juan 13; Chinese text on Wikisource, read 2026-10archive