The tea plant

一One species, two main varieties#

All the teas in this database come from one species, Camellia sinensis (L.) O. Kuntze. Two botanical varieties matter in practice. Var. sinensis, the China type, is described in the PNAS genome paper as a slower-growing shrub with small leaves that tolerates cold. Var. assamica, the Assam type, grows fast, has large leaves and is sensitive to cold, so it is grown mainly in warm, tropical areas. Wuyi cultivars such as Shuixian, Rougui and Qidan are var. sinensis. Pu'er is made from Yunnan large-leaf tea, 云南大叶种 (Yúnnán dàyèzhǒng), which is var. assamica.

The Chinese trade defines "large leaf" by size: a leaf area of roughly 28–50 cm², estimated as length × width × 0.7. Yunnan large-leaf plants are mostly 乔木型 (qiáomù xíng, tree type), around 5–6 m when left unpruned and occasionally over 20 m. GB/T 22111-2008 defines Yunnan large-leaf tea as the tree and small-tree large-leaf varieties grown in Yunnan, and makes it the only permitted raw material for pu'er.

Both genomes have been sequenced: the assamica cultivar 云抗10号 (Yúnkàng 10) in 2017 (3.02 Gb, about 81% repeats) and the sinensis cultivar 舒茶早 (Shūcházǎo) in 2018 (genome estimated at about 2.98 Gb, assembly 3.1 Gb). These are reference points for research, not tools a buyer can use.

二Where cultivated tea came from#

Genetic studies give a layered picture. Using nuclear microsatellites on 392 and then 402 samples, Meegahakumbura and colleagues found three genetic groups: China type, Chinese Assam type and Indian Assam type. They read this as three separate domestications, with Chinese Assam tea arising in southern and western Yunnan. About half of the Chinese Assam landraces they sampled showed admixture with Camellia taliensis, a wild relative.

Dates depend on the method and should be read as model estimates. The 2018 PNAS genome paper puts the split between var. sinensis and var. assamica at about 0.38–1.54 million years ago. The marker-based study puts it at about 22,000 years ago, and the split between Chinese and Indian Assam tea at about 2,770 years ago. These point estimates should not be compared without their models and uncertainty. The microsatellite study assumes a 12-year generation time and reports a 95% highest posterior density interval of 2,598–1,043,593 years for the first split, and 1,200–235,932 years for the second. Its first interval overlaps part of the genome estimate. This does not reconcile the studies, but it makes a simple “one must be wrong” verdict unjustified. Neither estimate directly dates the first cultivation of tea.

C. taliensis, 大理茶 (Dàlǐ chá), is a separate species in section Thea. It grows at 1,300–2,700 m in south-western Yunnan and neighbouring Myanmar and Thailand, and local people use it in place of ordinary tea. Min Tianlu's widely used revision counts 12 species and 6 varieties in the section. This matters for "wild tea" (野生茶, yěshēngchá) sold as pu'er: leaf from wild Camellia that is not Yunnan large-leaf C. sinensis falls outside the GB/T 22111 definition of pu'er (our inference from the definition) and can taste and act differently.

三Landraces and clonal cultivars#

Chinese horticulture divides tea plants by how they are propagated. 有性系 (yǒuxìngxì, seed-propagated) varieties, popularly 群体种 (qúntǐzhǒng, population varieties), grow from seed, so every plant is a little different. 无性系 (wúxìngxì, clonal) cultivars come from one selected plant multiplied by cuttings (扦插, qiānchā) or layering, so every bush is genetically the same. By the end of 2017 clonal cultivars covered 60.94% of China's tea area.

The distinction maps onto the teas here. Wuyi's old landrace is 菜茶 (càichá); the published GB/T 18745-2026, effective from 1 April 2027, defines Qizhong 奇种 as yancha made from the seed-propagated Wuyi caicha, and Da Hong Pao as a blend of several clonal cultivars. Rougui, Shuixian, Qidan and Beidou are clones. The Da Hong Pao mother bushes at Jiulongke were propagated by cuttings by 1982, and in 1988 experts judged the cutting-grown plants to match the originals.

Yunnan's great large-leaf "varieties" are different. 勐海大叶种 and 勐库大叶种 (Mengku large-leaf) are seed populations that were recognised as national cultivars in the first national approval batch. Old pu'er gardens are therefore genetically diverse, and "cultivar" on a pu'er wrapper usually means a population or a local type such as purple bud, not a clone.

四How cultivars are approved and registered#

Mainland China set up a national tea cultivar approval committee (全国茶树良种审定委员会) in 1981. The first batch of national cultivars is dated 1985 in the Tea Research Institute's review of Chinese tea breeding; some popular sources give 1984, probably the year the committee recognised them. It had 30 cultivars, 17 seed-propagated populations and 13 clones, including 福鼎大白茶 (Fuding Dabai), 铁观音 (Tieguanyin), 福建水仙, 凤凰水仙, 勐海大叶种, 勐库大叶种 and 凤庆大叶种. Five batches between 1985 and 2002 brought the national total to 95. Compulsory approval ended with the 2000 Seed Law, voluntary national identification followed from 2003, and by 2014 ten batches had produced 134 national cultivars, 117 clonal and 17 seed-propagated, alongside more than 200 approved by provinces. (A figure of 124 national cultivars for 1985–2002, found in one provincial source, does not match the published batch lists.)

Under the current Seed Law, crops other than the five major ones (rice, wheat, maize, cotton and soybean) are registered rather than approved. Under the Ministry of Agriculture's 2017 measures, in force 1 May 2017, a variety on the list must be registered before it is promoted; an unregistered variety may not be advertised or sold under a registered variety's name. The tea plant is item 28 of 29 on the first list (MOA Announcement 2510). Registration numbers take the form GPD + crop + (year) + two-digit province code + four-digit serial. Plant variety rights (植物新品种权) are a separate protection, held by more than 40 tea cultivars by the end of 2018.

None of this certifies the plant in a given garden. A cultivar name on a tea is the producer's statement about planting material, and no routine market test checks it.

五Plant-age categories in pu'er#

Pu'er is priced as much by the age and growth habit of the plant as by place. The trade uses a ladder of terms. 台地茶 (táidìchá) is terrace-plantation tea from densely planted, pruned bushes. 生态茶 (shēngtàichá, "ecological tea") is a loose label for less intensively managed bushes. 荒地茶 (huāngdìchá) and 野放 (yěfàng) mean plantation bushes left untended until they grew tree-like. 大树 (dàshù) means tall, unpruned trees, very roughly several decades old. 古树 (gǔshù) means old trees, conventionally a century or more. See plantation tea, eco tea, abandoned-plantation tea, large-tree and old-tree.

Only the top rung has a legal definition, and that definition was written for conservation. Yunnan's regulations on protecting old tea trees, adopted on 30 November 2022 and in force from 1 March 2023, cover wild and cultivated tea trees of 100 years or more. The 2023 survey rules count a plant as an old tea tree if its age is shown to be 100 years or more, if records show 100 years of cultivation, or if the trunk measures at least 15 cm across, 10 cm above the ground on the uphill side. A draft Jingmai group standard counts all wild and "transitional" types regardless of age. Chaozhou, in Guangdong, also sets 100 years and adds protection tiers at 300 and 500 years.

六Why age claims are so loose#

Three gaps make old-tree claims hard to rely on. First, the definitions above govern which trees are protected; none of them governs what a wrapper may say. The Yunnan regulation encourages traceability for old-tree products but contains no labelling rule. Second, age in practice is often inferred from trunk girth rather than proven, and girth depends on site and pruning history. Third, there is no accepted test that a buyer or regulator can apply to a cake to read the age of the trees it came from.

Volume shows how far claims run ahead of supply. Official, village and press estimates for Lao Banzhang range from about 60 to 117 t a year, and China Newsweek reported ¥9.9 cakes sold under the name. Core Bingdao Laozhai produces about 8 t of old-tree spring tea, while "Bingdao" sales run to hundreds of tonnes or more.

Yancha has the same problem in a smaller form. 老枞 (lǎocōng, old bush; see lao cong) has no unified age standard. The six Da Hong Pao mother bushes have not been picked since 2006, so "mother-tree Da Hong Pao" for sale is not what it says. Claims that those bushes are thousands of years old are unsupported. Treat every age figure on a listing as the seller's claim, and record it as one.

七Green and white tea: plant names are not product names#

Longjing 43 makes the difference between a population and a clone concrete. The Chinese Academy of Agricultural Sciences describes selecting one plant from the Longjing population, multiplying it by cuttings and shortening its selection number, Longjing 6043, to Longjing 43. The cultivar is therefore related to Longjing’s seed population, but the two names do not describe the same planting material.

Baiye 1 offers a different naming lesson. Its pale spring shoots supply Anji Baicha, whose manufacture includes heat fixation. The “white” in that tea name refers to the plant’s pale growth, while its processing places the product in green tea. It is not the withered-and-dried white-tea category.

Conversely, Baihao Yinzhen is a white-tea product name: clause 3.1 of GB/T 22291-2017 specifies single buds from Dabai or Shuixian tea cultivars. It is not the name of one clone. Record plant identity, picked material and finished product separately; each explains a different part of what reaches the cup.