Kodo ko jaanr (finger millet beer)

A drink that is brewed before it is served

Kodo ko jaanr is a traditional fermented finger-millet product of the eastern Himalayas. Tongba is the distinctive way this fermented grain is served: placed inside a tall vessel, covered with warm or hot water, and drunk through a filtering straw.

The distinction is important.

The grain ferments before the drinker receives it. Adding water to the serving vessel extracts alcohol, acids, sugars and other soluble compounds already present in the fermented millet. It does not begin the entire fermentation process at the table.

In Nepali, kodo means finger millet, while jaanr denotes a fermented grain preparation. Tongba spelling differs between sources.

Finger millet is Eleusine coracana (L.) Gaertn., a small-seeded cereal particularly useful in upland agriculture.

The fermented grain is inoculated with marcha, a dried mixed microbial starter. Organisms in the starter convert cooked grain starch into sugars and then convert part of those sugars into alcohol.

The resulting product is mildly sweet, acidic and alcoholic.

For tongba service, the fermented millet is held in a cylindrical wooden, bamboo or metal vessel. A straw, traditionally made from bamboo, allows the liquid to pass while retaining most of the grain.

After drinking the first extraction, more water can be added.

This creates a serving experience in which one preparation yields several successive drinks.

Communities and regional terminology

The 1996 comparative survey records Mandokpenaa thee (Limbu), Sampicha ummaak (Rai), Naarr paa (Gurung), Saangla chi (Tamang), Chirs Shyaabu (Sunwar), Paadaare haan (Magar), Gyaar chyyaang (Sherpa), Minchaa chhyaang (Bhutia), and Mong Chee (Lepcha) as names for kodo ko jaanr in its surveyed communities. These spellings are transcriptions from that source, not standardized spellings across dialects and not evidence that every named community follows one method. The survey covered Darjeeling hills and Sikkim, not Nepal; broader Eastern Himalayan distribution does not justify projecting the Sikkim/Darjeeling household process onto every Nepalese community. Rice bhaate jaanr, maize makai ko jaanr and distilled raksi remain distinct products.

Historical foundations

A 1996 regional survey documents kodo ko jaanr in Darjeeling hills and Sikkim and cites older accounts of regional fermented-drink culture. These references do not establish that every historical drink called marwa or chyang was the same product, or that its current starter, vessel or method continued unchanged.

Why finger millet matters

Finger millet is adapted to environments where other cereal crops may be more difficult to grow.

It is cultivated in the Himalayan foothills and uplands, where its small seeds can be dried and stored after harvest.

The mature grain supplies substantial starch, along with protein, fiber, minerals and other components.

Starch is the principal substrate for alcohol production, but it is not immediately accessible to ordinary brewing yeasts.

The grain therefore requires cooking and subsequent saccharification.

Different local finger-millet varieties can differ in seed color, starch characteristics and processing behavior.

Traditional descriptions from Sikkim record local varieties including mudke, nangrey, fyakre and nangkatwa.

These names are regionally documented rather than a complete botanical classification.

The flavor and texture of the resulting beverage can change with the millet variety, degree of cooking and microbial starter.

Cleaning and cooking the grain

The 1996 Namchi/Algarah survey records finger millet cleaned, washed and cooked for about 30 minutes in an open cooker. Excess water was drained and the cooked seeds cooled on a woven bamboo mat called mandro.

The account gives no batch weight, cooking-water quantity, measured cooling endpoint or validated cooked-grain holding control. The roughly 30-minute cooking time is an observation from this survey, not a universal endpoint.

Adding marcha

The 1996 survey reports 0.5–2.0% powdered marcha sprinkled over cooked seeds and mixed before the mass was packed under local leaves. The account does not specify a standardized starter culture or prove that the percentage is a safety limit.

A 2022 review describes 1–2% powdered starter for several cereal beverages. This broader range does not replace the named field process or establish a home dose.

Saccharification: Making sugar without malting

Marcha contains microorganisms able to break down starch. The 1996 survey describes a two-to-four-day first stage under fern, banana or canna leaves. Sweetness and aroma were observed, but neither is a validated test of microbial safety.

The main fermentation

The 1996 field account describes transfer to an earthen pot or bamboo septu made airtight and a later fermentation of three to four days in summer or five to seven days in winter.

It gives no pressure-management method or measured pathogen-control endpoint. Its vessel description and seasonal timing are historical observations, not home operating instructions. The fermented grain and the later tongba extraction are separate stages.

Measured microbiology

A 2004 study examined 40 samples of kodo ko jaanr.

The investigated samples contained yeast populations around 7.1 log CFU/g and lactic acid bacterial populations around 5.9 log CFU/g.

Identified yeasts included Saccharomyces cerevisiae, Pichia anomala, Saccharomycopsis fibuligera and Candida glabrata.

Lactic acid bacteria included Pediococcus pentosaceus and organisms identified as Lactobacillus bifermentans.

The species names reflect the taxonomy used in the study. Subsequent taxonomic revisions may alter how particular organisms are classified.

These findings demonstrate that kodo ko jaanr is not a pure Saccharomyces fermentation.

The diverse microorganisms contribute to starch breakdown, ethanol production, acidification and flavor.

Their proportions are not fixed between households.

What happens during a measured fermentation

During the 2006 experimental time course, saccharification and liquefaction activities were strongest around the second day, while reducing sugars rose until roughly the third day and then declined. Yeast and lactic acid bacterial counts changed during the sequence. Alcohol in that experimental material rose from 0.1% to 6.9% by day six. This is a measurement of that fermented grain under the experiment's conditions, not an alcohol value for another batch or for a tongba serving after water extraction.

Tongba: The serving apparatus

The fermented grain is made before the serving step. The 1996 source names toongbaa or dhungro as a sipping vessel made from wood, bamboo or sometimes earthenware; wooden examples may be decorated or lidded. A narrow bamboo pipsing straw has a side hole near its lower end to keep grits out. The existing Tongba vessel equipment record is separate from this recipe and linked as equipment-for: it is service/extraction equipment, not the required primary fermenter.

Preparing the drink for serving

The 1996 account describes placing about 500–600 g fermented millet in a toongbaa vessel, adding warm water, waiting 10–15 minutes and drinking through a pipsing straw. Additional water could be added for later extractions.

This is a serving description, not a fermentation step or sterilisation stage. The water volume, extraction yield and alcohol strength of each serving were not measured.

Why repeated infusions change the drink

Warm water extracts soluble material from the fermented grain. Later infusions can differ as the grain is extracted and diluted.

Neither taste nor perceived warmth establishes alcohol strength. The 2006 figure of 6.9% was measured in a fermentation experiment and cannot be assigned to a tongba serving without measuring the extracted liquid.

Flavor and accompaniments

The 1996 survey records slightly acidic, sweet and mildly alcoholic flavour as preferred in its sampled communities, and mentions fried meat or pickles with toongbaa. These observations do not establish a safe batch endpoint or apply unchanged to every Himalayan community.

Recognizing faults

The 1996 survey reports that pronounced sourness and unpleasant flavour were disliked in the sampled regional product and that prolonged fermentation was associated with bitterness and greater alcohol. These are historical sensory observations, not validated safety tests or controls. They do not provide a measurable endpoint or determine whether a particular batch is safe.

Safety, alcohol and contemporary production

The cited microbiological and chemical analyses describe specific samples and experimental batches. They do not establish a defined starter, validated pathogen controls, water-quality requirements, a safe household endpoint, alcohol strength for each serving infusion or safe storage duration. Warm water poured into a tongba/toongbaa vessel extracts and dilutes fermented material; it is not a validated kill step. Batch size, water volume, yield and process limits remain unestablished, and no sensory cue or elapsed time is presented as proof of safety.

Separate fermentation, product stages and serving

A 1996 field description from Namchi in South Sikkim and Algarah in Kalimpong describes cleaned and washed finger millet cooked for about 30 minutes, drained and cooled on the woven bamboo mandro mat, then mixed with powdered marcha. The mass is held with fern, banana or canna leaves for two to four days of saccharification before transfer to an earthen pot or bamboo septu for room-temperature fermentation, described as three to four days in summer or five to seven in winter. A separate 2006 laboratory time-course used 2% marcha, 100 g in 250 ml bottles, two days at 28°C followed by eight days at 28°C. That experiment is not a household schedule. The toongbaa/tongba is the later drinking vessel, not the necessary primary fermenter.

A 2024 sample distinguishes related kodo drinks

A 2024 Singalila ridge study labels Tongba (also written Tumba in its discussion) as a finger-millet fermented drink and sampled it at Kalipokhri. It separately sampled chhyang and nigar at Gurdum from a related kodo fermentation. These names identify separate sampled products or stages in that study; they are not evidence that every local household makes the same products in the same way. Its metabolite and in-vitro bioactivity assays describe samples and do not validate a home process or safety controls.

Geographic scope of older and broader sources

The 1996 field survey concerns Darjeeling hills and Sikkim. The 2004 characterization article and a 2022 regional review place finger-millet jaanr/tongba within the wider eastern Himalaya, including Nepal, but do not make the detailed Sikkim/Darjeeling household sequence a universal Nepalese recipe. The 2022 review explicitly describes historical evidence as sporadic; a single origin and continuous recipe are not established.