No fruit is involved, and that is the point
The question that confuses people most about fruit notes is whether something has been added, and for an unflavoured tea the answer is no. The compounds a leaf produces overlap with those in the reference foods, and a tasting note names the nearest familiar thing rather than reporting an ingredient. The ingredient list settles the question definitively — a flavoured tea declares a flavouring and a plain one does not — and the cup usually settles it too, because an added flavouring is strongest in the dry leaf, front-loads heavily into the first infusion and is largely gone by the second, while a leaf’s own aromatics persist across infusions in step with the tea. Knowing this removes both the suspicion and the opposite error, which is reading a note as a claim about content. It also explains why fruit descriptors vary so much between writers: they are pointers to a resemblance, and a pointer only works if the reader has encountered the thing pointed at. Blackcurrant, damson and greengage assume a British reference set; longan, lychee and jujube assume another; and neither travels well.
Carotenoid breakdown, the largest single route
The growing leaf contains carotenoid pigments, and their oxidative breakdown during withering, oxidation and firing produces a family of compounds consistently associated with fruity, floral-fruity and woody-violet character. The ionones are the most frequently named, associated with violet and with stone fruit, and damascenone appears in the same discussions, associated with a stewed-fruit and rose-adjacent quality and detectable at extraordinarily low levels. This is a well-established route rather than a convention, and it is the reason fruit character is so much more prominent in oxidised teas than in green ones: the pigments are being broken down by the same processing that darkens the leaf. Two useful consequences follow. Fruit notes in a black or an oolong tea are usually a signature of how far and how gently oxidation was allowed to run, which is a maker’s decision. And the same pigments are responsible for part of the leaf’s colour, so the aromatic gain and the visual change are two faces of one process — which is one more reason that colour and aroma in tea are correlated without either causing the other.
Stone fruit, and the muscatel case
Stone fruit — peach, apricot, nectarine — is one of the more reliable fruit descriptors and TeaHQ keeps it as an aroma facet. It appears in Darjeeling, in Oriental Beauty, in several Taiwanese oolongs, in some Chinese black teas and in aged whites. The most-discussed instance is muscatel, the character of Darjeeling second flush, and it is worth setting out because the mechanism is unusual: the accepted account is that a leafhopper feeds on the young shoots in early summer, the leaf is damaged by that feeding, and the plant’s response to the damage generates aromatic compounds that are not otherwise present. The same mechanism with a related insect is the accepted account of Taiwanese Oriental Beauty, where the honeyed character depends on the leaves having been bitten. Two things make this the most instructive fruit note in tea. The character cannot be produced on demand, because the insect population varies with conditions and the fields cannot be sprayed, which is why these teas vary so much year to year. And muscatel is a convention as well as a chemistry — the word names a recognised trade quality that a whole industry agrees on, which is why it is reliable as communication.
Citrus, and the two ways it arrives
Citrus in tea comes from three sources and telling them apart is straightforward. It can be added, which is the Earl Grey case: bergamot is an oil applied to finished leaf, so every Earl Grey is a flavoured tea and the ingredient list says so. It can be a second ingredient present in the packet, as in the citrus-peel blends and the aged tangerine-peel traditions of southern China, where the material is visible. Or it can be generated by the leaf, which is the case in a great many unflavoured teas described as lemony, bergamot-like or grapefruit-adjacent, and the compounds discussed in that connection include limonene, citral and the linalool oxides, several of which are the same compounds citrus fruits themselves rely on. TeaHQ keeps `citrus` as a flavour family separate from `fruity` because the two behave differently: citrus notes are among the brightest and most volatile in the leaf and are lost first in storage and driven off first by water that is too hot, which makes them a fair indicator of both freshness and brewing temperature.
Berry, tropical and the descriptors that are least reliable
These are the sub-families where a reader should hold a note most lightly. Berry — raspberry, blackcurrant, cranberry — is used of some Kenyan and African orthodox teas, of purple-leaf cultivars and of certain Nepali and Darjeeling lots, and the notes recur across independent descriptions often enough to be worth taking seriously while being much less stable than stone fruit. Tropical — mango, passion fruit, lychee — appears in descriptions of some Taiwanese, Vietnamese and Thai oolongs and of certain high-bud black teas, and it is heavily influenced by the reference set of whoever is writing. Both are also the descriptors most likely to be reaching for a resemblance in a tea that is genuinely fruity in some direction the writer cannot place. The honest treatment is the one this catalogue applies throughout: a note that appears repeatedly across independent sources is evidence, a single vivid note is worth enjoying and not relying on, and the presence of an unusual fruit descriptor in a listing is a reason to try a small quantity rather than a prediction of what you will find.
Dried fruit is a different family
Worth separating explicitly, because the vocabulary runs them together and the chemistry does not. Fresh-fruit notes belong largely to the carotenoid-breakdown and terpene routes described above and are characteristic of relatively fresh, aromatically intact teas. Dried-fruit notes — raisin, date, dried longan, prune, fig — belong to aged and heavily processed teas, arise through a different set of routes including the reactions between sugars and amino acids under heat and the slow changes of long storage, and are among the most persistent rather than the most fragile compounds in the leaf. That is why an aged white tea, an old sheng and a heavily roasted oolong converge towards a shared dried-fruit vocabulary from quite different starting points, and why a fresh tea and an aged one described with the same word are not describing the same thing. The practical use is diagnostic: a tea whose fruit character has shifted from fresh towards dried has moved along the storage sequence, which for a shou mei is the intention and for a Darjeeling first flush is a report on the cupboard.