The compositional fact
Caffeine is a purine alkaloid produced by the tea plant and present in every true tea. It is highly water-soluble, extracts readily and early, and in isolation is intensely bitter — which is the whole of its relevance to this page. It is not the only bitter component of tea, since the catechins are bitter as well, and a cup’s bitterness is a combination rather than a single compound’s signature. Related alkaloids, theobromine and theophylline, occur in much smaller quantities. This catalogue records caffeine as a compositional property of the leaf, expressed as a range and required to carry a five-part basis stating leaf quantity, water volume, temperature, steep time and infusion number, because a caffeine figure without its basis is not a fact but a decoration. Its bitterness is also perceptually distinct from the bitterness of the polyphenol fraction, which experienced tasters can sometimes separate: one reads as a clean, sharp, quickly fading bitterness and the other as a broader one arriving with a drying grip. The distinction is difficult to demonstrate outside a laboratory and it is worth knowing that a cup’s bitterness has at least two sources, because it explains why lowering the temperature helps with some bitter cups and not with others.
Where it sits in the plant
Not evenly. The bud and the youngest leaves generally carry more than mature leaf, which is why a bud-heavy plucking is associated with more caffeine in the finished tea than a coarse one from the same bush — a fact that runs directly against the widespread belief that white tea is low in caffeine, since silver needle is buds only. Stems and mature leaf carry less, which is the compositional basis of kukicha and bancha being described as lighter. Cultivar differences are substantial. And shading is associated with a higher level in the leaf, which places gyokuro and matcha at the higher end rather than the lower. Each of these is a statement about what is in the leaf, and each of them has a straightforward flavour consequence in that the same factors shift the bitter fraction. Season and growing conditions move it as well, and in directions that are not intuitive. Rapid growth under warm conditions, the same conditions that produce a brisker summer leaf, is associated with different levels from the slow spring growth that follows dormancy. The general point for a buyer is that the leaf-level figure is set by a combination of variety, plucking, season and cultivation, none of which is visible from a category name — which is exactly why category-level claims about caffeine are so unreliable.
What brewing does
Caffeine extracts quickly and early, so a short steep collects a substantial share of what will ever come out, and extending the steep adds proportionally more of the later polyphenol fraction than of caffeine. Higher temperatures raise extraction, as they do for everything, and a higher leaf-to-water ratio raises it in direct proportion. Two consequences follow for flavour. The commonly repeated instruction to discard a short first steep in order to remove most of the caffeine does not work as described, because although caffeine does come out early it is far from completely removed, and the rinse also discards a good deal of the aromatic and sweet fraction. And a second infusion is genuinely different in balance rather than simply weaker, since the fast fraction has already largely gone. Grind and leaf size matter as much here as anywhere. A broken or fannings grade releases everything faster, so a bagged tea brewed for two minutes and a whole-leaf tea brewed for two minutes are not comparable preparations. And matcha is the categorical exception, since the leaf is consumed rather than infused and nothing is left behind — which places it apart from every infused tea in any comparison and is one of the reasons such comparisons are so frequently misleading.
Caffeine and the other compounds
Its behaviour in the cup is not independent. Caffeine forms complexes with the larger oxidation products of black tea, which is what produces the haze known as tea cream when a strong black liquor cools, and those complexes affect both the appearance and the texture of the drink. Its bitterness is perceptually suppressed by sweetness and by the amino acid fraction, which is one reason a shaded green tea can be higher in caffeine than a black tea and taste less bitter. And it is perceptually reinforced by astringency, since both arrive together and the mouth does not separate them naturally. Any account that models caffeine as a fixed quantity of bitterness added to a cup will mispredict, because what is perceived depends on everything else present. The interaction with the amino acid fraction is worth stating plainly because it inverts a common assumption. A shaded green tea can be higher in caffeine at the leaf level than a black tea and read as far less bitter in the cup, because what is perceived depends on the balance rather than on any single component. Anyone using perceived bitterness as a proxy for caffeine content is therefore using a measure that runs the wrong way for exactly the teas where the question is asked most often.
Why comparisons between teas are so unreliable
Because almost every published comparison varies several things at once. Two teas brewed at different ratios, temperatures and times cannot be compared for caffeine; nor can a cup of tea and a cup of coffee, since the preparations differ in every respect. Leaf-level content and cup-level content are different measurements and are routinely conflated. Cultivar, plucking standard, season, shading and leaf age all move the leaf figure substantially, so category-level generalisations — that green tea has less than black tea, that white tea is lowest — are unreliable at the level of any particular pair of teas. TeaHQ records caffeine estimates as wide ranges with an explicit basis and marks them as inferred, and there is no predicate in this catalogue asserting that one tea is lower in caffeine than another, precisely because the evidence does not support it. Published figures compound the problem in a further way. Much of the widely quoted data comes from a small number of analyses of particular teas under particular conditions and has been copied between sources for decades without the conditions travelling with it. A figure without its basis cannot be checked, cannot be reproduced and cannot be compared with another figure, which is precisely why the schema in this catalogue refuses to accept one that does not carry its five parts.
The boundary of this page
Everything above is about the leaf and the cup: what caffeine is, where it sits in the plant, how it extracts, and what it contributes to bitterness and texture. What a caffeinated drink does to a person is a different subject, it is governed elsewhere in this catalogue under a different standard of evidence, and nothing on this page should be read as addressing it. The distinction is not a formality. A compositional fact about a leaf and a claim about a physiological outcome are different kinds of statement, they require different support, and the habit of running them together is why so much writing about caffeine in tea is simultaneously confident and unreliable. There is one further reason the boundary is drawn where it is. The relevant claims are governed here by a health surface with its own evidence standards, its own required framings and its own review requirements, and duplicating them casually on a flavour page would put material outside that governance while looking exactly like the rest of the catalogue. Keeping the compositional discussion clean is therefore a structural safeguard as much as an editorial preference.