A light roast and a dark roast stand side by side on a shop shelf. Through the window in each packet the leaf looks like a different tea altogether. Turn them over, though, and the brewing figures have not moved.
What follows is a search for guidance that brews by roast depth, and for the axis on which the figures really change: the shelf, the sellers' words, then the research.
One line on the back
Turn two packets over and the small print of their brewing notes wears the same face. Guidance that sets a different water or time by roast depth does not turn up among what is published.
Merchants and public bodies that publish brewing figures were checked, and none of them writes separate figures by depth. A single guide does publish two sets, and its axis is not the roast but the season of the leaf. Ordinary hojicha gets boiling water and a short wait; hojicha made from first-flush leaf gets cooler water and a longer one. The reason given is that second- and third-flush leaf carries less of the savoury components.
That one case is a single tea farm's published figures, not a rule of the trade. The record supports only that the season of the leaf, rather than the words light and dark, is what figures follow today. The split is laid out in the table below. After the figures come the words on packets sharing a shelf.
Figure 3-2-1The words on the packet, and the axis the figures actually divide on
| Axis | Division | Published brewing conditions |
|---|---|---|
| Words on the front of the packet | Light roast | None of the 27 with figures sets them by roast (across guides: 80–100°C, 25 seconds to about 4 minutes) |
| Words on the front of the packet | Dark roast | None of the 27 with figures sets them by roast (across guides: 80–100°C, 25 seconds to about 4 minutes) |
| The axis the figures move on (Fujita Chaen, 1 guide) | Second- and third-flush leaf = hojicha in general | 3–4 g, boiling water, 30 seconds |
| The axis the figures move on (Fujita Chaen, 1 guide) | First-flush leaf = hojicha from first-flush tea | 2–3 g, 80–90°C, 1 minute |
One way to brew, one shelf
A light-roast hojicha and a hojicha of karigane stems sit on one shop's shelf. Sold side by side, they still share one set of brewing notes.
The shop's instructions are a single set used across its site. A few guides do recommend cooler water, and none of them gives roast depth as the reason. The difference in taste, meanwhile, is put into words. One merchant calls the light roast a refined toastiness that keeps green tea's savour, and the dark roast a toastiness with a faint sweetness.
Yet neither that article nor the product page names a water temperature or a time. The record supports only that taste is told apart in words while the water is not. After the shelf come two packets that are often cited.
Pale stems and a direct-fire dark roast
Pale stem tea, and leaf roasted nearly black. The two are often brought out to compare roast depths. They cannot stand in for each other.
The lightly roasted stem tea is to be brewed short with boiling water, the direct-fire dark roast a little longer with cooler water. The depths run opposite, but the raw material and the amount of leaf differ too: they are separate products. That the words light and dark have no shared standard across shops is taken up in Chapter 6.
The record supports only that the two are brewed differently, not that depth is the reason. After the products comes a cup on the research side.
What research moves is the clock
In the laboratory, identical cups stand in a row and only the hands of the clock move. What gets varied there is the time, not the depth.
One report took several teas including hojicha, changed only the wait, and repeated the infusion. The longer it steeped, the more dissolved, until the amount levelled off.
The water for hojicha was held at the boil; its heat was never moved. The record supports only how time works. After the clock comes a comparison under matched conditions.
Matched pots, only the depth changed
The same pot, the same amount of leaf, water at the same heat. On a table set like that, only the difference in depth should reach the cup.
So the comparison begins with a light roast and a dark roast made from the same raw material, under matched conditions. Only the depth and the wait are moved. Aroma, astringency and sweetness are ranked with the labels hidden, and the colour is noted too.
The results are not written here yet. The record supports only the shape of the comparison. Its exact conditions are set out below.
The numbers in detail
- The 29 guides checked
- 29 guides were collected from 18 tea merchants and manufacturers and 11 public bodies and organisations, and the 27 that give figures were compared, leaving out two with no brewing page. None writes a different temperature or time by roast depth. Verified 2026-09-22.
- Split by season
- Fujita Chaen's “Ocha Biyori” is the one case. Ordinary hojicha is 3–4 g, boiling water, 30 seconds; hojicha from first-flush leaf is 2–3 g, 80–90°C, 1 minute. It says the former is made from second- or third-flush leaf, with fewer savoury components.
- One set for one shelf
- Maiko no Cha Honpo sells a light-roast hojicha and a karigane hojicha side by side. Its brewing instruction is one set for the whole site: 15 g, 650 mL of boiling water, 30 seconds.
- The two at 80°C
- two of the 27 with figures recommend a temperature in the 80s, and neither gives roast depth as the reason.
- Taste in words
- the descriptions are Ryuhoen's. The light roast has “a refined toasty aroma that makes the most of green tea's savour”; the dark roast “the toastiness of a strong roast with a faint sweetness”. Neither the article nor the product page gives a temperature or a time.
- The stem tea
- Maruhachi Seichajo's Kenjō Kaga Bōcha, a lightly roasted stem tea. 6 g with 260 mL of boiling water for 25 seconds.
- The dark roast
- Minokamo Chaho's direct-fire dark-roast hojicha. About 5 g with 180 mL at about 80°C for 1 minute.
- Research on time
- Yoneda and Katō's 1994 report. Six kinds of tea including hojicha, infusion times from 0.5 to 4 minutes, three infusions each. The amounts dissolved rose with time and levelled off at 3–4 minutes. The water for hojicha was held at the boil.
- The planned comparison
- 3 g of leaf, 150 mL of water, 90°C, the same pot. Waits of 30 and 60 seconds, three runs of each condition, twelve brews in all. Aroma, astringency and sweetness ranked blind by three people, recorded with the colour.