Japan Hojicha AssociationJapan Hojicha Association

Tea gets more interesting with roasting.

Reading what is inside, from the words on the packet

Chapter 6 · 5 of 5 sections readable now

Illustration of tea in three shallow dishes, with colours ranging from bright gold to dark brown.

Light roast and dark roast on a packet are relative terms that hold only inside the shop that printed them, defined by neither temperature nor colour.

01

Hojicha roast levels: the industry has no shared definition

Who decides what light roast and dark roast mean on a packet, and by what measure? This piece follows those words from the shop, to the label rules on the back, to the shelf.

Two packets marked dark roast, bought at two different shops, sit side by side on the kitchen table. Opened, the leaves are not quite the same colour. On the back, the name field on both says only hojicha. Light and dark appear on the front, and nowhere else.

What follows looks at who chooses those words, what the label rules say, and how the shelves are actually sorted.

The same word, pointing up and down

Compare the price tags and the darker packet is cheaper at one shop, dearer at the next. Light and dark are decided by the shop that sells the tea.

It is a little like the word upstream, which depends entirely on where the river is joined. One long-established tea house writes that the finer the tea, the lighter the roast, which puts dark roasts on the cheaper side.

Other sellers name their range by roast depth, and one large maker puts dark roast in a product name, as a mark of skill. The sources go as far as this: the same word can point in opposite directions from shop to shop. From the shop's words, the next stop is the back of the packet.

No seat in the small table

Turn a packet over and there is a table in tiny print. The industry's labelling rules settle the name of the tea, and stop there.

The rules list the names of teas, hojicha among them, and no term for roast depth. They fix what must go inside the table and what may be printed outside it. Light and dark roast belong to neither list. If the depth sits outside the table, it is not carelessness on the shop's part: there is simply no seat for it.

A roasting temperature does appear once, in an appendix to the rules. What it separates is not depth but whether a packet names a manufacturer or a processor. The sources show that no yardstick for depth is written down anywhere. After the rules, the shelf.

Opening the shelves one by one

Open the sellers' own online shops one after another and packets fill the screen. Only a small share of them use words for roast depth.

Counted up, fewer than one seller in four described the depth of the roast. The rest sorted their teas by part of the plant, leaf or stem, or by grade names. Some sorted by method instead, such as sand roasting or roasting twice.

None of them published a roasting temperature or time. The sources show that depth words are the exception on the shelf, and no more. The sellers' names and the date of the count come together at the end.

Figure 6-1-1What 18 companies divided their hojicha by (confirmed on each company's official site on 24 September 2026)

Divided by (companies)Businesses
Words for the depth of the roast (4)Nakamura Tokichi Honten (炒りが浅い, lightly roasted) / Fujita Chaen (深焙り and 浅焙り) / HOJICHA CO. (Light Roast, Gold Roast, Dark Roast) / Ito En (深炒り)
Part of the plant, leaf or stem (4)Fukujuen / Arai Seicha / Kirin / Uogashi Meicha (grades of stem hojicha)
Grade or trade name (3)Kanbayashi Shunsho Honten / Gion Kitagawa Hanbee / Chasandai
Method or process (4)Kaneju Farm / Yamamasa Koyamaen / Tsuchiyama Hitoban Hōji / Lupicia (two-stage firing)
Origin or certification, or no account given (3)Itohkyuemon / Ocha no Igeta / Ooigawa Chaen
A survey of the official sites of 18 businesses, opened and counted on 24 September 2026. Each company is placed in one group only; a company using more than one axis is placed under its main one. No company published a roasting temperature or time; every figure that does appear is about brewing. The pages of Kanbayashi Shunsho Honten, Uogashi Meicha, Itohkyuemon and Ocha no Igeta are drawn in JavaScript, so they were reopened in a browser the same day and read in full. Everything here is the sellers' own account of their products, not a statement based on measuring components.

The numbers in detail

Lighter for finer tea
Nakamura Tokichi Honten, in Uji, Kyoto, writes that the higher the price, the lighter the roast, and so the paler the brown.
Depth as a name
HOJICHA CO., which sells online only, offers Light Roast, Gold Roast and Dark Roast. Ito En uses dark roast (深炒り) in a product name.
Labelling rules
The Japan Tea Central Association's green tea labelling standard lists 17 names, hojicha among them. No roast-level term is included.
Label items
The standard sets 11 mandatory items and 7 optional ones, the optional ones printed outside the labelling table. Roast level is in neither.
Appendix temperature
A roasting temperature appears once, in the standard's process chart. It separates the manufacturer and processor categories, not roast depth.
The 18-site count
The official sites of 18 businesses were opened and counted on 24 September 2026. Four used words for roast depth.
Words in use
The four used lightly roasted, dark roasted (two spellings) and Dark Roast.
The other 14
They sorted their products by part of the plant, grade or method. None published a roasting temperature or time.

Sources

Updated 17 October 2026This section on its own page

02Members only

Light roast vs dark roast hojicha: can the levels be matched across shops?

If the packet gives no figures, the research surely does. This piece asks whether the stages named in studies can be lined up with a shop's light and dark, by looking at how each study measured heat.

An illustration of a wooden desk with two plain paper bags set well apart, a small dish of differently roasted tea leaf in front of each, and a blank notebook, a pencil and a ruler between them.

Leaf through the studies on roasting and the stages are all there: light, medium, dark. A small table on paper looks far more certain than the words on a packet. Put the two together, and it seems possible to say what temperature dark roast means.

This piece looks at whether that single table can be made, starting from how each study built its stages.

Only the names line up

Lay several reports side by side and each has its light, medium and dark. Even so, no table can match those stages to roast levels across shops.

It is like measuring a mountain from a different valley each time. What can be built is the order of light and dark within one shop's range. On the research side, too, each study defines its stages differently.

The sources go as far as this: what lines up is the names of the stages. After the names comes the question of where the heat was measured.

Measuring heat in different places

Around a roasting machine the heat gathers unevenly, and the reading depends on where the probe sits. Temperatures cannot be lined up from one study to the next.

One prefectural station measured the air leaving the machine. A national institute measured the surface of the leaf, and another prefecture's study the inside of the drum. The same number of degrees points to a different place each time.

The sources show that the figures cannot simply be compared as higher or lower. After where the heat was read comes how the stages were made.

The three sections that remain look at a trial that made its roast stages by slowing the drum rather than changing the heat, and at studies that set no stages at all. Temperatures and drum speeds close the piece.

The rest of this article is for members.

Membership is available in Japanese. It takes an email address and a password, and it is free.

Open the Japanese membership page

Sources

Updated 17 October 2026This section on its own page

03

Roasting hojicha at home: three levels, and what makes them three

To roast tea leaves to three depths in a home kitchen, what should stay the same and what should change? This piece looks at dividing by time and keeping a record of colour.

An illustration of a frying pan and a wooden spatula on the stove, with three small heaps of tea leaf on white paper in front, lightest on the left, and a kitchen timer and a damp cloth beside them.

Without a clay roaster, a frying pan will roast tea leaves. Move them with a wooden spatula over a low flame and the colour shifts, then the smell of the room. The hard part is knowing when to take the pan off the heat. Three stages are the aim, but no kitchen comes with a yardstick for them.

This piece looks at what to divide the three stages by, and how to keep a record of the colour.

Change only when it stops

The leaves begin to rustle at the tip of the spatula. The three stages are divided by the time before the pan comes off the heat.

The pan, the amount of leaf and the flame stay the same all three times. Only the length of time before stopping changes. A prefectural station's trial was built the same way, holding the temperature and changing the time in the machine.

The sources show that the same design works at home. Once the stages are set by time, the colour is next.

Colour can be the yardstick

Set the three side by side and the colours run from a light brown to a deep one. That colour is a fair guide has been checked.

It is like a hillside turning in autumn, from the foot upward. A report from another prefecture found that whatever the temperature, the leaf's colour moved along the same path. The same report advised watching lightness and yellowness rather than redness.

Lightness rises briefly at the start of roasting and then falls. Judged by lightness alone, early leaf and well-roasted leaf can be mistaken for each other. The sources show that laying all three on the same paper is enough. A photo under the same light keeps the colours as they are.

Numbers move with the method

It is tempting to turn the colour into a number, but no kitchen has the instrument. Even if it did, the number would not travel beyond the house.

The prefectural station ground the tea to powder before measuring; the other report measured whole leaf. The instruments came from the same maker, yet grinding alone moves the figures. No line can be drawn at the number where dark roast begins.

The sources show that a colour reading belongs to the way it was taken. The figures and conditions come together at the end.

The numbers in detail

Trial design
Holding the temperature and varying the time follows the trial by the Tea Experiment Station of the Kochi Agricultural Technology Center.
Stem tea colour
Ishikawa's 2020 report roasted stem tea at 180°C, 200°C and 220°C and measured the colour of the leaf itself.
Colour path
At every temperature, the three L*a*b* values followed the same path.
Values to watch
The same report advised watching lightness L* and yellowness b* rather than redness a*.
Colour values
The Kochi trial ground the tea to powder before measuring. First-flush a* was 1.34 for light and 5.11 for dark (Table 4, 2014 data).
Separate samples
These colour samples were made apart from the three-stage 2013 samples, in two stages, with the dark roast at 10 revolutions.

Sources

Updated 17 October 2026This section on its own page

04Members only

Does a darker roast mean more toasted aroma?

Does roasting deeper make hojicha more fragrant? This piece follows two trials, one holding the temperature and one holding the depth, to see what decides the roasty aroma.

An illustration of three identical white teacups in a row, the tea in them shading from pale amber on the left to deep red-brown, with a teapot and a steaming iron kettle behind.

On days that call for something roasty, the hand goes to the darker bag. A little way into roasting, the kitchen smells stronger than it did at the start. Deeper means more fragrant fits what the kitchen seems to say.

This piece looks at whether the studies bear that out, through two trials.

Darker is not always more fragrant

Lean over the cup and the roasty smell rises with the steam. That aroma does not always grow as the roast goes deeper.

When a prefectural station roasted further at the same temperature, the compounds used as its marker for roasty aroma came out higher in the light roast. Depth alone does not decide the aroma. What decides it is the pairing of temperature and how far the roast went.

For one picking season, depth made no difference at all. The sources show that within this trial the aroma did not increase. Next come two results that seem to point the other way.

Two results that look reversed

Open another report and this time the roasty compounds rise as the temperature climbs. The two look opposed, but they are answering different questions.

Climb a peak from the east and then from the west, and the view is not the same. One trial held the temperature and varied the depth; the other held the depth and varied the temperature. At one temperature, roasting further meant less; at one depth, hotter meant more.

Quote either one alone and the conclusion turns upside down. The sources show that the questions need reading separately. Beyond this lies something that happens only at the dark end.

The two sections that remain show how, at a dark roast, differences in method fade from the compounds, so that they survive mostly on the light side. Trial conditions and compound counts close the piece.

The rest of this article is for members.

Membership is available in Japanese. It takes an email address and a password, and it is free.

Open the Japanese membership page

Sources

Updated 17 October 2026This section on its own page

05Members only

How to read a hojicha label before you buy

In the seconds before buying, what can the words on a packet tell, and what can they not? This piece works through the back label, the roast words, colour and aroma, and what the shop sells itself on.

Illustration of two plain tea bags with no lettering, and two small dishes of light and dark roasted leaf.

In the few seconds before taking a packet to the till, there is not much it can tell. It cannot be opened, and the leaves are often out of sight. Even when they show, the colour of the brewed tea does not. All there is to go on is the print.

This piece sorts what the words on a packet can tell from what they cannot.

Tracing the print with a fingertip

Pick up a packet and the front carries the name and the selling lines in large type. It can tell leaf or stem, what the shop stakes itself on, and where the tea sits in that shop's range.

The temperature and time of roasting, and the colour in the cup, cannot be read from it. And packets with roast words at all are in the minority. In a separate count of shop shelves, fewer than one seller in four used them.

The sources show that what a packet will answer is narrow. After the front, the small table on the back.

The back answers tea type and origin

Turn it over and fine ruled lines hold a table of small print. That table answers the type of tea and where it came from, and no more.

The industry's rules set what must be printed and what may be. Neither list has a place for roast depth. The fields that come close to the contents are about three: the name, the ingredients and the origin.

The sources show that depth sits outside the back table. From the back, the next thing is how to read the roast words on the front.

The four sections that remain follow how roast words pair with leaf or stem, why colour and aroma stay behind the seal, and how what a shop sells on is the one thing a bag reliably tells. The counts close the piece.

The rest of this article is for members.

Membership is available in Japanese. It takes an email address and a password, and it is free.

Open the Japanese membership page

Sources

Updated 17 October 2026This section on its own page

Back to the contents