Why the distinction is worth the effort
A reader who cannot separate these words will read a tea description and take away nothing checkable. Told that a tea is made from Tie Guan Yin, they cannot tell whether that is a statement about the plant in the field, about the manufacturing style, or about the market category. Told that a Longjing is qun ti zhong, they cannot tell that this is a claim about how the field was planted rather than a grade. Told that a Darjeeling is clonal, they cannot tell that this says nothing about which plant. Each of the five terms answers a different question, and once they are separated the same sentence carries several distinct pieces of information instead of one blur. What follows is not pedantry for its own sake. Nearly every recurring confusion in the tea vocabulary — famous-name substitution, wild and ancient claims, cultivar names on labels that cannot be verified, the belief that a tea style must come from the plant of the same name — is a collapse of two of these five into each other.
Variety: a rank in botany
A botanical variety sits below species, and in tea there are two accepted ones — var. sinensis and var. assamica — plus a contested third. This is a taxonomic statement about a naturally occurring population, made by botanists, and it is the only one of the five that has a formal nomenclatural code behind it. It is written in the botanical convention: italicised species name, then var., then the variety epithet. Two cautions. English usage of the word variety almost never means this, and a packet describing its variety is usually naming a cultivar or a product type. And a botanical variety is not a quality tier: assamica is not a lesser sinensis or the reverse, they are two forms of one plant adapted to different conditions. Where TeaHQ needs the botanical sense it says variety group and constrains the value to a fixed list, precisely so the loose sense cannot creep in.
Cultivar: a plant somebody selected and named
A cultivar is a cultivated variety: a specific plant, chosen by a person for its properties, given a name, and maintained in cultivation so that the name continues to refer to the same thing. Yabukita, Jin Xuan, TRI 3055 and TRFK 6/8 are cultivars. The essential feature is identity over time — a cultivar name is a promise that the plant in this field is the same plant as the one in that field — and in tea that promise can only be kept by propagating clonally, because tea seed does not breed true. A cultivar name is therefore a claim about a propagation chain reaching back to a selection event. It is worth noting what a cultivar name does not assert. It says nothing about where the plant is grown, how it was managed, or how the tea was made, and two teas from the same cultivar in different countries can be entirely unalike. It is one variable, held constant.
Clone: a fact about propagation, not a plant
A clone is a set of plants propagated vegetatively from one individual, so that they are genetically identical. Every clonally propagated cultivar is a clone; not every clone is a cultivar. An estate that selected a good bush in its own fields and multiplied it without ever registering or naming it has a clone and arguably not a cultivar, and a great deal of Indian, Sri Lankan and East African planting material is exactly that — numbered estate selections in local circulation. When a Darjeeling estate says a field is clonal it is making a statement about propagation method, distinguishing it from the old seed-grown sections. It is not naming a plant, and it is emphatically not a quality claim, although the market has come to read it as one. The two words are used loosely enough that it is worth asking which is meant. Clonal usually means not-seedling. Cultivar usually means we can tell you which one.
Landrace: a population, adapted to a place
A landrace is a seed-grown population that has been cultivated in one place long enough to have adapted to it, selected by growers keeping seed from plants that did well and by the site killing the ones that did not. Qun ti zhong in China, zairai in Japan, the Kangra jat and the old Darjeeling China stands in India, Trung Du in Vietnam, the Turkish seedling fields: these are landraces. They are not cultivars, not wild, and not uniform. Their defining property is internal variation — every plant is a different individual — and that variation is what makes them valuable as breeding material and awkward as a crop. The single most common error in this area is treating landrace, heritage, ancient and wild as one idea. A landrace is a specific agricultural thing with a definition, and it can be twenty years old or four hundred; a wild population is not a landrace; and an old individual tree is not a population at all.
Style: what was done to the leaf, and what it is sold as
The fifth term is not botanical at all. A commercial tea style is a manufacturing method plus a market category — Tie Guan Yin as a rolled, lightly oxidised, sometimes roasted Anxi-type oolong; Da Hong Pao as a roasted Wuyi rock oolong; Longjing as a flat pan-fired green tea. Styles have names, and in tea those names very often coincide with the name of a plant, because a style frequently began with one and then outgrew it. A style name tells you what to expect in the cup and nothing reliable about the field. Kept apart, the five terms let a single tea be described precisely: a var. sinensis plant, of the cultivar Qi Dan, propagated as a clone, grown in a field replanted from a Wuyi landrace, made in the Da Hong Pao style. Collapse them and all you can say is Da Hong Pao, which is what most labels say.