Descriptions of hojicha talk about how strong the fire is. Over a strong flame. High heat, all at once. Few of them give a number. The numbers are in the research, in a report that roasted stem tea at three temperatures.

What follows looks in turn at what was established, where the temperature was taken, the sample and conditions, the taste compounds, and how such numbers can be compared.

The hotter, the toastier

Inside the sound of a turning drum, the leaf shifts colour little by little. Within the range tested, the higher the heat, the more toasty compounds formed, and the colour moved from yellow-green towards yellow-red.

The toasty compounds all came out higher at the higher temperatures. The colour drifted from the green side towards the red. What was not established is how hot the leaf itself became. Where that temperature was taken is the next question.

What was hot was the inside of the drum

The inside of a turning drum gets too hot to hold a hand near. The report's temperature belongs to the inner surface of the drum, not to the leaf.

The researchers read the drum's inner surface with an infrared thermometer. Other reports measured in other places, and some do not say where at all. How the measuring points differ is set out in this section's table.

The sources tell us which part of the machine was measured, and that is where they stop. Next, the stems that were roasted and how.

Figure 5-4-1Roasting and hiire temperatures are measured in different places from source to source

Source (year)TemperatureWhat the temperature was measured onTime
Shiono 1974, tray type65°Cexhaust temperature5 hours
Japan Tea Central Association, Green tea labelling standard, appendixhiire 70–130°C / roasting 140°C or abovenot statednot stated
Shiono 1974, air-through type90–110°Cinternal temperature20–25 minutes
Takahashi 2015around 100°C (80°C for higher grades)not stated30–40 minutes
Kochi Prefecture (Table 1, 2013)exhaust temperature 100°C at the second roasting (the same for light, medium and dark)exhaust temperaturenot stated
NARO 2015100 / 115 / 130°Cleaf surface temperaturenot stated
Kawakami 1999150 / 130 / 100°Cnot stated12 min / 10 min / 15 min
Shiono 1974, drum-over-fire type140–150°Cexhaust temperature40–60 minutes
Hara 1970160–180°Cdrum inner wall temperaturenot stated
Sasaki et al. 2020180 / 200 / 220°Cdrum inner surface temperature (infrared thermometer)40–420 seconds
GJTA (an industry body)200°Cnot stated10 minutes or more
Temperature, place of measurement and time for roasting and for hiire, source by source, from the lower temperatures to the higher. "Not stated" means the source does not say; nothing has been estimated to fill it. The time column carries only the values written in the sources, with no conversions. The labelling standard row is not a definition of tea types but a worked list for sorting out who is responsible for the labelling. The Kochi Prefecture row has no time; the report records only exhaust temperature and rotation speed (first-flush tea was separated by drum speeds of 17 / 12 / 8 rpm; the first roasting was at an exhaust temperature of 90°C and 22 rpm). GJTA (Global Japanese Tea Association) is an industry body's account. Sources: Shiono T. (1974); Takahashi J. (2015); Tea Research Station, Kochi Prefectural Agricultural Technology Center (data for 2013); NARO National Institute of Vegetable and Tea Science (2015); Hara T. and Kubota E. (1970); Kawakami M. and Yamanishi T. (1999); Sasaki et al. (2020); Japan Tea Central Association, Green tea labelling standard (last amended 2026); GJTA.