The purple leaf that was never rolled
A 2025 Indonesian study made white tea from two purple-leaf cultivars and one green-leaf clone, withered and dried only, then measured what each one carried.
White tea is defined by what is withheld. The leaf is not rolled, not fixed, not fired to redirect it. What that leaves behind is the cultivar, largely undisguised. A 2025 study from Indonesia made white tea from two purple-leaf cultivars and one green-leaf clone, put all three through the same two steps, and measured what each one still carried.
What the paper was actually asking
The cultivar comparison is not the paper's main axis, and we will not pretend otherwise. Prawira-Atmaja and Puangpraphant1 set out to test extraction method. Ultrasonic-assisted extraction, bath against probe: a bath at 35 kHz and 320 W, a probe at 250 kHz and 750 W, each run at 5, 10 and 15 minutes. The finding on that axis is a split decision. Probe extraction pulled out higher concentrations of the bioactive compounds. Bath extraction left more of the antioxidant activity intact.
The three cultivars are the material that axis was tested on. All three were processed as white tea, withered and dried, nothing else. That is what makes the study readable from here.
The two directions the chemistry went
The abstract-level result is stated in one sentence, and it points both ways. Purple-leaf cultivars showed significantly higher levels of anthocyanins, polyphenols and flavonoids (p < 0.05). The green-leaf cultivar showed higher carotenoid content.
No multiplier is given for the anthocyanin gap, so none is given here. The direction is significant. The size is not on record.
One number is on record, and it belongs to the whole panel rather than to any single cultivar. Total phenolic content across every cultivar, method and extraction time ran from 318.52 to 352.51 mg GAE/g by the bath method, and from 327.53 to 376.45 mg GAE/g by the probe. That is the window all twelve extractions landed inside. It is a range, not a verdict, and it should be read as one.
A principal component analysis explained 70.82% of the total variance. The samples clustered by cultivar type and by extraction method, both.
Where GMB-7 comes from
The green-leaf clone is GMB-7. It is an Assam-variety clone, Camellia sinensis var. assamica, bred at the Research Institute for Tea and Cinchona at Gambung in West Java. GMB is Gambung. It is not an experimental line. It is planted widely and commercially in Indonesia for yield, and the lead author's own affiliation is that same institute, which is a plain enough explanation for why it is the green-leaf comparator here rather than something else.
The purple-leaf cultivars are named SS-2 and SS-3. Their pedigree is where the record thins, and we will not fill it in. The paper identifies them as purple-leaf cultivar material tested alongside GMB-7 in the same Indonesian work. It does not give their breeding history. They were grown and tested in that context, by that author. We do not know it, so we will not write it as though we did.
A different purple, somewhere else
The purple leaf most readers will have met is Zijuan, and it is worth keeping at arm's length from this study. Zijuan is a somatic mutant of the Yunnan Daye cultivar, also var. assamica, from Yunnan in China, and it is the purple cultivar the wider tea trade actually knows by name.
Zijuan is a different plant from a different country, tested in different work, and its numbers do not belong here. What does carry over is the processing habit: Zijuan is almost always made into green, black or pu-erh style tea. It is rarely put through the two steps this piece is about.
What is left when nothing is done
The compounds counted here were counted in laboratory extracts, by two extraction methods neither of which resembles a teapot. What they do in a person is not something this paper measured, and it is not something we will claim on its behalf.
What the study does establish is narrower than that. White tea is a process, not a leaf colour. Run a purple-leaf cultivar through the wither and the dry, and the anthocyanins, polyphenols and flavonoids come through higher. Run GMB-7 through the same two steps, and the carotenoids come through higher. Neither leaf wins. They arrive carrying different things, and the process is quiet enough that the difference survives to be measured.
Almost every white tea on a shelf still comes from green-leaf stock, Fuding Da Bai and its siblings among them: this register has never before had a purple-leaf entry to compare against. It is one reason to ask, the next time a purple-leaf white turns up for sale, which cultivar is actually in the cup, not only which grade or garden.