Idli batter does its useful fermenting in about 15 hours: lactic acid bacteria drop the pH from 6.3 to about 4.5 and double the volume with carbon dioxide. Left for a year, it goes through a succession: it over-sours within a day or two, then yeasts and moulds colonise the surface, eat the acid and let spoilage bacteria such as Bacillus break down the protein. Open the lid after a year and you would find a dried, mouldy, ammonia-smelling crust over a sour sludge; in a truly sealed jar the gas from the first days alone could have built up several bar of pressure and blown the lid.
- About 15 hours is all the useful fermenting idli batter needs at about 30°C.
- pH 6.3 → 4.5 in a day, as Weissella and Leuconostoc multiply roughly a thousandfold.
- Two days at 30°C is when measured quality starts to fall.
- Yeasts, then moulds, then Bacillus take the batter apart over the following months.
- About 4 L of CO₂ per kilo: enough, in a perfectly sealed jar, for roughly 8 bar.
Fermentation is ecology in a bowl. Every organism that thrives in batter changes it, and in doing so makes room for the next. Food scientists have measured the first 24 hours in detail.
Nobody, as far as I can find, has published a year-long idli experiment. So beyond the first few days I lean on the closest well-studied system, silage, where farmers have kept fermented plant matter for months and measured exactly what happens when air gets in. I'll mark clearly where measurement ends and estimate begins.
What happens in the first 24 hours?
Lactic acid bacteria, mostly Weissella and Leuconostoc, multiply about a thousandfold, turn sugars into lactic acid and carbon dioxide, and raise the batter. This part is measured, not guessed.
Mandhania and colleagues at Pune's National Centre for Cell Science sampled a home-style batter (parboiled rice and black gram, 2:1) every 3 hours for 24 hours at about 30°C. Their numbers:
| Hour | pH | Volume rise (mL) | Lactic acid bacteria (×10⁶ CFU/g) |
|---|---|---|---|
| 0 | 6.3 | 0 | 0.06 |
| 6 | 6.0 | 1 | 33.6 |
| 12 | 4.6 | 34 | 37.6 |
| 18 | 4.2 | 52 | 51.1 |
| 24 | 4.5 | 62 | 77.2 |
Weissella jumped from 9% to 71% of the bacteria between hours 6 and 9, and the Leuconostoc family made up 71 to 85% of DNA reads at hours 9 and 12.
Rice-only batter did not ferment at all; black gram-only batter did. The gram carries the microbes, and the rice supplies most of the starch.
See the data as a table
| pH (pH) | |
|---|---|
| 0 h | 6.3 |
| 6 h | 6.0 |
| 12 h | 4.6 |
| 18 h | 4.2 |
| 24 h | 4.5 |
Source: Mandhania et al., Applied and Environmental Microbiology (2019), Table 2 · Home-style batter at about 30°C with no starter culture.
The team fermented batter without starter at about 30°C, sampled it every 3 hours for a day, and profiled the bacteria by 16S rRNA sequencing, qPCR and culturing 354 isolates; vendor batters from Bangalore and Pune were also tested. pH fell from 6.3 to 4.5, bulk density from 1.2 to 0.5 g/cm³, and lactic acid bacteria stayed above 85% of the community, led by Weissella. Batter kept at 4°C did not ferment, and Weissella alone could ferment sterilised batter. The limit: only 24 hours were tracked and yeasts were not counted, so the study says nothing about long storage.
Read the paper (free)Can a ferment live for a century?
Only if someone keeps feeding it. A batter left alone is a closed meal that its microbes finish in days; a starter that is refreshed every day can, in principle, run indefinitely.
San Francisco's Boudin Bakery traces its sourdough to 1849, when Isidore Boudin, from a family of Burgundy bakers, is said to have got a wild starter from a Gold Rush prospector. In 1868 the bakery committed to that mother dough even as commercial cake yeast spread.
In 1906 the great earthquake hit San Francisco, and the fire that followed destroyed the bakery's Broadway building. As the bakery tells it, Louise Boudin, who ran the business after Isidore's death in 1887, scooped the mother dough into a bucket as the ground shook and got out with it. The bakery moved to 10th Avenue, where it still is, and SFist reported at its 175th anniversary in 2024 that the starter had been fed every day since. The family account is the only source for the rescue itself, so treat the details as the bakery's tradition.
Boudin Bakery: Our history (1906 timeline entry); SFist (Jay Barmann): Boudin Bakery turns 175 (Sep 2024)The difference between Boudin's starter and a forgotten tub of batter is food and fresh flour every day. Stop feeding, and the succession below begins.
Where does the gas come from, and how much is there?
From heterolactic fermentation: Leuconostoc and Weissella split each glucose into one lactic acid, one ethanol and one carbon dioxide, so every gram of acid in the batter comes with gas.
The 1965 study by Mukherjee and colleagues first showed that Leuconostoc mesenteroides alone could both sour and leaven idli batter, the first recorded case of a lactic acid bacterium doing the job yeast does in bread. Now the worked example, for 1 kg of batter:
- Acid made. Assume the bacteria stop at about 1.5% lactic acid by mass, when the acid stalls their own growth (my estimate, typical of sour fermented foods): 15 g, or \( 15/90 = 0.167 \) mol.
- Gas and alcohol made. One mole of each per mole of acid: 0.167 mol CO₂ and \( 0.167 \times 46 \approx 7.7 \) g of ethanol, a faintly boozy batter.
- Gas volume at 30°C and 1 atm. \( V = nRT/P = 0.167 \times 8.314 \times 303 / 101{,}325 \approx 4.2 \) L, about four times the batter's own volume.
- Pressure if the lid were perfectly sealed over 0.5 L of headspace, ignoring the CO₂ that dissolves:
\[ P = \frac{nRT}{V} = \frac{0.167 \times 8.314 \times 303}{0.0005} \approx 8.4\times10^{5}\ \text{Pa} \approx 8\ \text{bar} \]
Some CO₂ dissolves in the liquid and much escapes early, so the real figure is lower; but even a tenth of it is enough to pop a snap-on plastic lid.


What happens in the first week?
The batter over-sours, loses its gas and starts to taste bitter and alcoholic, and yeasts begin to multiply. Quality falls measurably within two days at room temperature.
- Over-fermentation. Kavitake and colleagues note that traditional fermentation runs 12 to 14 hours and that going past about 15 hours brings high acidity and nutrient degradation. The bacterial community keeps shifting: 23 core genera disappeared and 12 new ones appeared after 12 hours.
- Collapse. Chelliah and colleagues found sensory scores of plain batter dropped after two days at 30°C, with texture lost as the batter stopped holding gas.
- Yeasts. In the same study yeast counts rose steadily at room temperature and held steady at 4°C. CSIR-CFTRI notes that at about 40°C batter over-ferments and turns too sour, so a warm kitchen speeds the whole sequence.
Batter of black gram and rice at 1:4 was fermented at 30 ± 2°C and sequenced at 0, 6, 12 and 15 hours, yielding 165,154 high-quality reads and 1,298 OTUs. Firmicutes, the group that includes the lactic acid bacteria, rose from 7% to 68% of the community as Proteobacteria fell from 93% to 31%. The authors note that fermentation beyond about 15 hours brings high acidity and nutrient loss. The limit: the paper's reported percentages differ slightly between sections, and it did not measure pH or storage.
Read the paperThe authors fermented batter with and without curry leaf powder and stored it at 30°C and 4°C, tracking microbes and sensory scores. Plain batter lost sensory quality after two days at 30°C, turned sour after 20 to 24 hours of fermentation, and lost gas-holding texture; yeast counts in it rose steadily at room temperature but stayed level in the fridge, while lactic acid bacteria declined slowly. Curry leaf powder slowed the decline. The limit: storage counts are reported only in figures, and the study ran for days, not months.
Read the paper (free)What takes over after the bacteria are done?
Air-loving yeasts, then moulds, then spore-forming bacteria, in a succession that silage scientists have mapped in detail. Without that analogy, the months after week one would be guesswork.
Silage is grass or maize fermented by lactic acid bacteria in the absence of air, preserved by the same low pH that sours batter. When air reaches it, Woolford and colleagues found that yeasts are the main cause of the breakdown: they raise the pH, heat the mass and burn off dry matter.
Every organism that thrives in batter changes it, and in doing so makes room for the next.
Applied to an open bowl of batter, the sequence I would expect is:
- Weeks 1 to 3: the yeast film. Acid-tolerant yeasts form a creamy skin on top and use oxygen to burn the lactic acid as fuel. Less acid means a rising pH at the surface.
- Weeks 2 to 8: moulds. Once the surface is less acidic, moulds root into it, forming fuzzy white, green or black patches, and begin drying it into a crust.
- Months 2 to 6: protein breakdown. Spore-forming bacteria such as Bacillus, which survive almost anything, use proteases to break down the black gram protein. Ammonia from amino acids pushes the pH up further and brings the smell of rot. The same Bacillus subtilis is useful in fermented foods like natto and a spoilage agent in bread.
- Months 6 to 12: a dry cake. Water evaporates through the crust. At an indoor evaporation rate of a millimetre or two a day (my estimate), the half-kilo of water in a kilo of batter can leave within months, and with it most microbial activity. What remains is a hard, discoloured, mould-veined cake over whatever damp sludge the crust protected.
Stages after the first week are my estimates built on silage science and general food microbiology; I found no long-term measurements on idli batter itself.
The authors exposed fermented whole-crop barley and wheat silages to air and used selective antimicrobials to see which organisms drove the spoilage. In barley silage, deterioration was essentially caused by yeasts, which were linked to the rise in pH, the temperature increase and the loss of dry matter; filamentous fungi and streptomycetes seemed not to contribute at first, with lactobacilli and protein-digesting bacteria possibly active in the final stages. Wheat silage stayed stable, which the authors put down to its butyric acid and higher dry matter. The limit: the abstract reports no numbers, and silage is not batter, so I use it for the order of events, not the timing.
Read the paperWould a sealed jar be any different?
Yes, very. Without air, the yeasts and moulds that undo the acid cannot get started, so a truly airtight jar behaves like a silo: sour, alcoholic and fairly stable, apart from the gas.
That is how silage keeps for a season. The catch is the first day's carbon dioxide. As worked out above, a sealed litre of batter could in principle build several bar, far more than a kitchen lid survives. So in practice the lid leaks or blows, air gets in, and you are back in the open-bowl timeline.
If the lid held
- a hiss as you open it;
- a sharp vinegary and alcoholic smell;
- a grey, watery batter separated into layers, sour enough at around pH 4 that most food-poisoning bacteria struggle;
- useless for idli, because the starch and protein have been worked over and the gas is gone.
Gas in a sealed batter jar
What changes if the assumptions move?
| Change | What you open the lid to | Why |
|---|---|---|
| Fridge at 4°C | Sour but recognisable batter for about a week, slow decline after | Fermentation stops at 4°C (Mandhania); yeasts stay level (Chelliah) |
| Hot kitchen, 35 to 40°C | Same sequence, compressed into weeks | CFTRI notes batter over-ferments at about 40°C |
| Only half the acid, 0.75% | Half the gas, about 4 bar if sealed | Gas is proportional to acid |
| Airtight, pressure-rated jar | Sour, alcoholic, separated batter, little mould | No oxygen for yeasts and moulds (silage logic) |
| Batter dried right after fermenting | A shelf-stable powder | No water, no microbes: CSIR-CFTRI makes dehydrated fermented mixes this way |
As usable idli batter, a year is impossible: the useful fermentation is over in about 15 hours, quality falls within two days at 30°C, and air lets yeasts, moulds and Bacillus spend the following months undoing the acid and digesting the protein. A sealed jar avoids most of that but makes about 4 L of CO₂ per kilo, enough in principle for around 8 bar. Change the method and the year becomes possible: ferment for 15 hours, then dry the batter, as CSIR-CFTRI's dehydrated ready mixes do, and the fermented flavour keeps on a shelf.

The full science of the normal overnight ferment is in idli and dosa batter fermentation, the method in how to make idli and dosa batter, and the cold-weather problem in why idli batter won't ferment in winter. More from the Taste Lab is on the Rasa world page.
What's new
- Kerala names Bacillus subtilis, the hardy fermenter and bread spoiler, its state microbe Onmanorama, Feb 2026
- iD Fresh adds protein, ragi, multigrain and mallige idli batters in tubs FoodTechBiz, Jan 2025
What if next?
- What if you cooked biryani with a lightning bolt? From a year of slow chemistry to a few microseconds of very fast physics.
- What if you ate only one taste for a week? A year-old batter is pure sour. A week of one taste is its own experiment.
- What if you drank 10 cups of filter coffee a day? The idli's partner, taken to its own extreme.
Questions people ask
How long can idli batter be kept?
At room temperature, quality falls within about two days at 30°C. In the fridge, fermentation stops and yeasts stop multiplying, so it keeps noticeably longer, though it still slowly sours. Curry leaf powder has been shown to slow the decline.
Why does idli batter become too sour?
Lactic acid bacteria keep making acid after the batter has risen. Past about 15 hours at 30°C, acidity climbs and nutrients degrade, and heat speeds it up.
Is it safe to use batter with a layer of liquid on top?
A thin clear liquid on a day-old batter is usually just separation from over-fermentation. Fuzzy growth, pink, green or black spots, or an ammonia smell mean it should be thrown away.
What bacteria ferment idli batter?
Mainly heterofermentative lactic acid bacteria: Weissella, which reached up to 71% of the community in one study, and Leuconostoc, followed by Enterococcus and Streptococcus after about 12 hours.
Sources
- Mandhania MH et al.: Diversity and succession of microbiota during fermentation of the traditional Indian food idli, Applied and Environmental Microbiology (2019) (pH, volume, density and bacterial counts over 24 h)
- Kavitake D et al.: Bacterial diversity of traditional fermented food, idli, by high throughput sequencing, Journal of Food Science and Technology (2022) (over-fermentation after 15 h)
- Chelliah R et al.: Bio-fortification and shelf-life extension of idli batter using curry leaves, Journal of Food Science and Technology (2016) (quality loss after 2 days at 30°C, rising yeasts)
- Mukherjee SK et al.: Role of Leuconostoc mesenteroides in leavening the batter of idli, Applied Microbiology (1965) (lactic acid bacterium as the leavening agent)
- Woolford MK, Bolsen KK, Peart LA: Studies on the aerobic deterioration of whole-crop cereal silages, Journal of Agricultural Science (1982) (yeasts drive aerobic spoilage and pH rise)
- India Science and Technology (CSIR-CFTRI): Fermented dehydrated ready mixes for dosa and idli batter (15–18 h fermentation, over-fermentation at about 40°C)
- Onmanorama: Kerala's new state microbe has been in our food all along (2026)
- FoodTechBiz: iD Fresh expands flagship batter range (2025)
- Boudin Bakery: Our history (the bakery's account of 1849, 1868 and the 1906 rescue of the mother dough; the story)
- SFist (Jay Barmann): SF's oldest continually operating business, Boudin Bakery, turns 175 (2024) (Isidore Boudin, the move to 10th Avenue, starter fed daily)