The Stem Concentrates Floral Compounds the Leaf Does Not
A 2025 tissue-level study found oolong's stem holds more of the compounds behind its floral scent than the leaf beside it. A 2022 study found adding stem to the blend lowers one of those same compounds in the finished tea. Neither team was answering the other's question, and nobody has reconciled the two.
Oolong's stem is not dead weight kept in for shape. A 2025 analysis of the raw tissue itself found the stem holds more of the specific floral terpenes an oolong is prized for than the leaf blade beside it. A 2022 trial that instead rolled the stem into the finished tea and tasted the result found the opposite: adding stem to the blend lowers the exact compound the tissue work says the stem concentrates, while raising a different one. Neither team was answering the other's question, and nobody has reconciled the two directions since.
Pluck an oolong shoot and you take three or four open leaves and a length of stem with them, a coarser standard than the bud-only pull a fine green tea wants (Red Blossom Tea Company5). What happens to that stem once the shoot goes into the roller is where the traditions split. Anxi's Tieguanyin makers roll the shoot into its tight pellet and then pick the stem back out of the dried tea; Taiwanese gaoshan makers commonly leave it in, so the finished pellet is the whole shoot, stem included (Jiemian News3).
Whether that stem belongs in the cup is an old argument inside the trade itself. Competition lots are usually trimmed for a cleaner-reading liquor. Other tasters hold that a retained stem adds real complexity, with the honest caveat that bare stem carrying no leaf just turns the cup bitter (Red Blossom Tea Company5). The two chemistry papers behind the opening finding measured two different things, and neither by itself settles which side of that argument is right.
What the trade already believed about the stem
Chinese processing lore has an answer ready before any lab gets involved: active compounds in the stem travel into the leaf through the shoot's own vascular bundles during rolling and resting, and the reddened stem that forms alongside the reddened leaf edge during shaking is treated as a necessary condition for the tea's high aroma, not a cosmetic side effect (Rednet4). On this account the stem is a working part of the plant's plumbing during manufacture, whether or not it survives into the final dried lot.
What a 2022 study found when it added the stem back in
A team working on oolong compared tea rolled with leaf and stem together against tea rolled from leaf alone, holding the cultivar and handling otherwise constant (Lin et al., Foods, 2022). The stem-inclusive tea tested higher in umami, both by electronic tongue and by direct measurement: theanine and six other free amino acids all rose significantly. Its aroma profile shifted too, but not in one direction across the board. Nerolidol, a fruity-floral terpene, came in at 2.598 micrograms per gram in the stem-inclusive tea against 1.523 in the leaf-only version, a significant rise. Linalool, the terpene most associated with a tea-like floral character, ran the other way: 13.513 micrograms per gram with stem included against 17.257 without, a significant fall. Catechins, caffeine, theaflavins, and soluble sugars did not move either way, and a general tasting panel found no significant overall preference between the two versions. Stem changed the chemistry without producing a tea that testers simply liked more or less.
What a 2025 study found when it looked inside the tissue itself
A second team asked a narrower question: not what happens when stem is added to a blend, but what the stem tissue itself actually contains, tracked separately from the leaf through the same Tieguanyin manufacture (Ming et al., Food Chemistry, 2025). Running gas chromatography and gene-expression analysis on stem, first-leaf, and second-leaf tissue in parallel, they found the floral and fruity compounds, among them damascone, beta-ionone, geraniol, and linalool, concentrated more heavily in the stem than in the leaf blades around it. The leaf tissue, in turn, carried more nerolidol and alpha-farnesene, and it held onto more of the raw green, grassy-smelling compounds. Tissue for tissue, the stem was the more floral half of the shoot, not an inert stick riding along.
The two findings point in opposite directions on the same two compounds
Line the two papers up and a real problem appears. The 2025 tissue study says the stem itself runs higher in linalool than the leaf does; add more stem to a batch, and naive arithmetic says the resulting cup should carry more linalool, not less. The 2022 blend study measured the opposite: tea rolled with stem included came out lower in linalool than the leaf-only version. Nerolidol runs the same contradiction in reverse. The 2025 study found the leaf, not the stem, is where nerolidol concentrates; the 2022 study found adding stem to the blend raised the finished tea's nerolidol.
The gap is not a contradiction so much as two separate instruments pointed at two separate moments. One measured raw tissue composition at a point in processing; the other measured a finished, fully manufactured tea. A compound's starting concentration in stem or leaf does not have to survive intact into the cup, since withering, shaking, and roasting run their own enzymatic conversions on whatever precursors each tissue supplies, and nothing published so far has tracked the same shoot through both measurements at once. The old trade account, that stem compounds migrate into the leaf through the vascular bundles during rolling, would explain a stem-poor liquor gaining nerolidol only if the transfer moves compounds selectively rather than just diluting the mix, which is a plausible reading of the mechanism, not a claim either paper tested directly.
What this settles, and what it does not
It settles the question of whether the stem is inert. It is not. Both studies, measuring different things, agree the stem is chemically active rather than decorative packaging, which backs the half of the trade that defends leaving it in. It does not settle which finished cup, more stem or less, actually carries more of the classically floral note in the cup itself. That is a comparison nobody has run directly yet: one study on the same batches, tracking stem-tissue composition and finished-tea chemistry together, rather than two teams working the question from opposite ends five years apart.
For a drinker choosing between a trimmed Tieguanyin pellet and a stem-intact Taiwanese one, the honest position is that both makers are working with real chemistry, not folklore, and that the debate over which style tastes better is not one the lab has closed. The trade's own name for stem left deliberately bare of leaf, useless and merely bitter, still stands. Somewhere between none and too much is a ratio no published study has quantified for oolong specifically, and every maker who leaves the stem in is making that call by hand, one basket at a time.