What the research says about sourdough fermentation temperature
- Your starter is two organisms with two different optima. The bacteria peak at 32 to 33°C (90 to 91°F). The yeast peaks at 27°C (81°F). They do not want the same thing.
- 26°C (79°F) is the line. Below it, both respond to temperature almost identically. Above it, the balance tilts toward acid and away from gas.
- At 37°C the yeast can disappear entirely. In one controlled study the yeast count fell below detection while the bacteria held steady at a thousand times the population.
- The rule that fermentation doubles every 10°F has no published source. We looked. It is a workable kitchen heuristic and it is not a measurement.
- Warm dough goes slack because of pH, not heat. The proteases doing the damage belong to the flour, and acidity switches them on.
The yeast and the bacteria want different temperatures
A sourdough starter is not one organism with one preferred temperature. It is a partnership between lactic acid bacteria and wild yeast, and the two have measurably different growth optima.
Gänzle, Ehmann and Hammes (1998) modeled the growth of both partners across the full range a sourdough encounters. Their cardinal temperatures are the most useful numbers in this entire subject and they are almost never quoted.
| Organism | Optimum | Minimum | Maximum |
|---|---|---|---|
| L. sanfranciscensis LTH2581 the bacterium | 32°C (90°F) | 3.0°C | 41.0°C |
| L. sanfranciscensis LTH1729 the bacterium | 33°C (91°F) | 4.1°C | 41.0°C |
| C. milleri LTH H198 the yeast | 27°C (81°F) | 8.0°C | 35.9°C |
Read the last row against the first two. The yeast stops growing at 35.9°C, a temperature at which the bacteria are still five degrees short of their own ceiling. The yeast is the fragile partner, and the yeast is the one that makes your bread rise.
That asymmetry is the single fact this whole article rests on. Everything below is a consequence of it.
Why 26°C is the temperature that actually matters
Below 26°C (79°F), the yeast and the bacteria respond to temperature almost identically. That is the finding, stated plainly in the 1998 paper, and it is the reason the two organisms coexist stably in traditional sourdoughs maintained for decades.
The practical consequence is a clean rule. Warming a dough from 20°C to 25°C speeds up the whole culture without changing its character. Both partners accelerate together, so you get the same bread sooner. That is why our bulk fermentation time table works as a straight time-for-temperature trade in the cool range.
Above 26°C the trade stops being clean. The bacteria keep climbing toward their 32 to 33°C optimum while the yeast has already peaked at 27°C and begun to decline. Every additional degree now buys you more acid production relative to gas production. The dough is not simply fermenting faster. It is fermenting differently.
This is the microbiological explanation for the most common complaint in summer baking: the starter looks active, doubles quickly, smells increasingly sharp, and the loaves get flatter. The activity is real. It is bacterial activity, and bacteria do not inflate bread.
Two honest caveats on these numbers. Gänzle's figures come from two specific bacterial strains and one yeast strain grown in controlled media, not from a jar of flour and water. And modern taxonomy has renamed both organisms: Lactobacillus sanfranciscensis is now Fructilactobacillus sanfranciscensis, and Candida milleri is now Kazachstania humilis. The optima have not moved. The names have.
What happens to a sourdough starter held at 37°C
The yeast population can collapse entirely while the bacteria carry on, which produces a starter that sours without rising.
Vrancken and colleagues (2011) propagated wheat sourdough at 23°C, 30°C and 37°C for ten days and counted both populations. At 23°C and 30°C the yeasts held steady at roughly log 8 CFU per gram. In the 37°C fermentation refreshed every 24 hours, no yeasts were found at all for most of the run, appearing only from day 8 to day 10 and never exceeding log 5 CFU per gram. The bacteria meanwhile sat between log 8 and log 9 throughout.
The species composition shifted too. At 23°C, Leuconostoc citreum dominated. At 30°C and 37°C, Lactobacillus fermentum took over. Final pH landed between 3.3 and 3.4 in most runs, but the 23°C sourdough refreshed daily finished at about 3.7.
So a hot starter is not a fast version of a cool starter. It is a different culture with different residents. If you have been keeping a starter on top of the refrigerator or near an oven vent and it has turned sharp and sluggish, the population you are feeding may no longer be the one you began with.
Fermentation does not reliably double every 10°F
This rule appears on nearly every sourdough site, including in the temperature table on our own dense-crumb article. It has no published source that we could find.
We looked specifically. There is no peer-reviewed Q10 value for sourdough or for yeasted dough fermentation. The only printed doubling figure we located anywhere is on Weekend Bakery, which offers it uncited as a rule of thumb and says every 5°C, not every 10°F. Those two versions of the rule differ from each other by roughly a factor of two.
What does exist is the growth-rate data. Gänzle's maximum specific growth rates at the optimum are 0.71 and 0.67 per hour for the two bacterial strains and 0.42 per hour for the yeast. Those are single-point measurements at each organism's own optimum, not a curve you can convert into a kitchen rule.
Keep using the heuristic. It is directionally right, it is easy to hold in your head, and doubling per 10°F errs toward giving the dough more time in a cool kitchen, which is the safer error. Just do not repeat it as though somebody measured it, because as far as we can establish, for sourdough specifically, nobody has.
Warm fermentation slackens dough through acid, not heat
Long or warm bulk fermentation produces slack, spreading dough because the flour's own enzymes get switched on by falling pH. Heat is the accelerator, not the mechanism.
Thiele, Grassl and Gänzle (2004) tested this directly and the design is what makes it convincing. They compared sourdoughs against sterile doughs acidified chemically to the same pH, and against sterile doughs held at neutral pH. Gluten depolymerization occurred in the sourdoughs and in the acid sterile doughs, but not in the neutral ones. Their conclusion: gluten breakdown in sourdough is caused mainly by pH-dependent activation of the cereal's own enzymes.
The microbes are not eating your gluten. The acid they produce is unlocking proteases that were sitting dormant in the flour all along.
Temperature enters this story only indirectly, and it matters a great deal anyway: a warmer dough reaches a low pH sooner, so it spends more of its bulk fermentation inside the window where those enzymes are active. The fix is not to fear warmth. It is to shorten bulk when the dough is warm, which is exactly what the percentage-rise targets in why sourdough spreads flat instead of rising are built to do.
Dough temperature and room temperature are different numbers
The temperature that governs fermentation is the temperature inside the dough, and it is routinely several degrees away from the reading on your wall thermometer.
Professional bakers manage this with desired dough temperature, or DDT, which works backward from a target to tell you how warm your water should be. Both major published versions agree on the target and disagree on the arithmetic.
King Arthur Baking multiplies the target by 3 for a straight dough and by 4 when a preferment is involved, then subtracts room temperature, flour temperature, the preferment temperature where applicable, and a friction factor. The Perfect Loaf always multiplies by 4, since a levain is always in the formula. Both name 75 to 78°F (24 to 26°C) as the target.
Where they actually conflict is the friction factor for hand mixing. King Arthur puts it at 6 to 8°F for eight minutes of hand kneading. The Perfect Loaf sets it to zero, stating so explicitly. Apply the two formulas to the same kitchen and you get water temperatures 6 to 8°F apart. For stand mixers the gap is wider still: King Arthur measured 22 to 24°F on their own machine and cites Jeffrey Hamelman at 24 to 28°F, while The Perfect Loaf offers 20 to 40°F.
Nobody is wrong here. Friction factor is equipment-specific and it is meant to be calibrated by measuring your own dough. Take that as the instruction: a $12 instant-read thermometer in the dough beats any published constant.
What the research does not settle
Two questions bakers ask constantly have no good published answer, and pretending otherwise is how bad numbers spread.
- How much sourer a warm ferment actually gets. The direction is settled. Brandt, Hammes and Gänzle (2004) found acetate formation decreased at 35°C while lactate and ethanol were unaffected, which means the ratio shifts toward lactic acid. But no study we could open measures the lactic-to-acetic ratio across a temperature series, so any specific number you see attached to this is not coming from the literature. The field does not even agree on a target ratio: De Vuyst's group works from about 2.5, while a 2024 paper in Foods cites 4:1 to 10:1 as typical.
- What cold retarding does chemically. Every baker knows an overnight fridge proof changes flavor. We could not verify a single measured acid or aroma-compound figure for wheat sourdough retarding from any source we were able to open. Even America's Test Kitchen's cold fermentation page reports only a balloon-inflation observation, with no chemical measurement behind its flavor claim.
Retard your dough anyway. It works, it makes scoring easier, and it fits a normal schedule. Just be aware that the explanation attached to it is craft consensus rather than published measurement.
Fermentation time is the other half of this subject, and it is the variable with the best-measured consequences of any we cover. The nutritional claims people attach to long fermentation are examined in is sourdough easier to digest.
Sources
| Claim | Source |
|---|---|
| Cardinal temperatures: optima 32 and 33°C for L. sanfranciscensis, 27°C for C. milleri; yeast maximum 35.9°C; similar response below 26°C. Confirmed from full text. | Gänzle, M. G., Ehmann, M., & Hammes, W. P. (1998). Applied and Environmental Microbiology, 64(7), 2616-2623. doi:10.1128/AEM.64.7.2616-2623.1998 |
| At 37°C with 24-hour backslopping, no yeasts detected except days 8 to 10 below log 5 CFU/g, while LAB held log 8 to 9. Species shift from Leuconostoc citreum at 23°C to L. fermentum at 30 and 37°C. Confirmed from full text. | Vrancken, G., Rimaux, T., Weckx, S., Leroy, F., & De Vuyst, L. (2011). Applied and Environmental Microbiology, 77(8), 2615-2622. doi:10.1128/AEM.02470-10 |
| Acetate formation by L. sanfranciscensis decreased at 35°C while lactate and ethanol formation were unaffected. Confirmed from abstract only. | Brandt, M. J., Hammes, W. P., & Gänzle, M. G. (2004). European Food Research and Technology, 218(4), 333-338. doi:10.1007/s00217-003-0867-0 |
| Gluten depolymerization in sourdough is caused mainly by pH-dependent activation of cereal enzymes, demonstrated against acid and neutral aseptic controls. Confirmed from abstract only. | Thiele, C., Grassl, S., & Gänzle, M. (2004). Journal of Agricultural and Food Chemistry, 52(5), 1307-1314. doi:10.1021/jf034470z |
| Desired dough temperature formulas, 75 to 78°F target, friction factors of 6 to 8°F by hand and 22 to 24°F for a stand mixer. | King Arthur Baking, Dough temperature reference |
| Alternative DDT formula with a hand-mixing friction factor of 0 and 20 to 40°F for stand mixers. | The Perfect Loaf, The importance of dough temperature in baking |
Related reading
- Why is my sourdough dense? The bulk fermentation time table this article's microbiology sits underneath.
- Why is my sourdough starter not rising? The troubleshooting path when temperature is the cause.
- Sourdough starter smells like acetone? What the 27°C yeast optimum has to do with a solvent smell.
FAQ
What is the best temperature for sourdough bulk fermentation?
For a balanced loaf, hold the dough between 24 and 26°C (75 to 78°F). Both King Arthur Baking and The Perfect Loaf publish 75 to 78°F as their target dough temperature. The microbiological reason to stay under 26°C is that below that point the sourdough yeast and the lactic acid bacteria respond to temperature almost identically, so warming the dough speeds everything up without changing the balance between rise and sourness.
Does a proofing box or bread proofer help sourdough?
It helps if you set it low. The common instinct is to run a proofer at 30°C or higher because the dough moves faster, and that is past the 27°C yeast optimum measured by Gänzle and colleagues. A proofer set to 24 to 26°C gives you a stable, repeatable dough temperature, which is the actual benefit. Speed is not the benefit.
Is it better to ferment sourdough fast and warm or slow and cool?
Cool is more forgiving and warm is faster, and the window you have to catch the dough differs by hours. Neither is better in a general sense. What the research supports is that the two are not equivalent in flavor: above roughly 27°C you preferentially favor the acid-producing bacteria over the gas-producing yeast, so a warm ferment is not simply a compressed version of a cool one.
Why does my starter double fast in summer but the bread still tastes flat?
A fast rise is not the same as a healthy balance. In warm conditions the lactic acid bacteria keep accelerating past the point where the yeast has peaked, so a starter can look vigorous while its acid load climbs. The dough then over-acidifies during bulk, which softens the gluten and costs you oven spring.
Does a warmer starter make sourdough more sour?
It makes it more acidic overall, but not in the way most guides describe. Brandt and colleagues found that acetate formation by Lactobacillus sanfranciscensis actually decreased at 35°C while lactate and ethanol formation were unaffected. Acetic acid is the sharp, vinegary note. So very warm fermentation tends toward a rounder lactic sourness rather than a sharper one.