Week 6: Wild fermentation, and how to read it

What lives in a starter, which acid each organism makes, and why the end of bulk fermentation is a thing you look at rather than a time you set

29 March 2027

Open the slides

Instant yeast is one strain, freeze-dried, counted into the packet. It does the same thing on the same timetable in every kitchen in the country, which is why we started there. What you have in a jar is a population. It changes with the weather, it is a week old, and it is not the same organism as the one two benches over. Nothing it does can be looked up; all of it can be measured.

What is in the jar#

Two kinds of organism, in wildly uneven numbers. Lactic acid bacteria outnumber the yeasts by around 100 to 1 — a mature rye or wheat starter runs to roughly 10⁹ bacteria and 10⁷ yeasts per gram. The yeasts, typically Kazachstania humilis alongside Saccharomyces cerevisiae, ferment sugars to carbon dioxide and ethanol, and the carbon dioxide is what lifts the loaf. The bacteria, species like Fructilactobacillus sanfranciscensis, produce acid, and the acid is what makes the bread taste of something.

The acid also explains why the culture is stable. A ripe starter sits between pH 3.5 and 4.2, which is hostile to almost everything else that lands in the jar, and the yeasts in it are acid-tolerant in a way that most wild yeasts are not. You are not fighting contamination. You are maintaining the conditions under which the organisms you want outcompete everyone else.

Two acids#

Lactic acid bacteria split into two groups by what they do with sugar. Homofermentative species produce lactic acid and essentially nothing else. Heterofermentative species produce lactic acid plus acetic acid, carbon dioxide and ethanol. Both groups live in your starter; which one dominates decides the flavour.

Lactic acid is mild, rounded, yoghurty — the taste of a soft milky crumb. Acetic acid is sharp and vinegary, the flavour most people mean when they call a loaf sour. Three levers move the balance, and they are all things you already control.

A warm, wet, frequently fed starter gives you a mild loaf. The same jar kept stiff and cool on a 24-hour cycle gives you a sharp one. If your bread is sourer than you want, that is a setting, not a fact about your culture.

Feed by ratio, not by habit#

Write feeds as three numbers: starter, flour, water. A 1:1:1 feed is 25 g of starter with 25 g of flour and 25 g of water, which is what you have been giving yours, and at 24 °C it doubles in four to six hours and peaks an hour or two after that. A 1:5:5 feed is 20 g of starter into 100 g and 100 g, and it takes two or three times as long, because the same population starts five times more dilute with five times the food. Big ratios are how you hold a jar overnight without it starving.

Either is fine. What is not fine is a schedule you have stopped looking at. The thing to watch is the curve: the starter climbs, domes, and peaks, and shortly after that it collapses — the surface sinks, the dome caves, the smell turns from sweet-sour towards solvent and acetone as the yeasts run out of sugar. Feed at or just before the peak and the culture stays vigorous. Feed hours past collapse, day after day, and the population spends most of its life starving in acid — which is what a sluggish starter usually is, rather than a dead one. The same jar needs a different schedule at 19 °C than at 26 °C, and the schedule is the thing you change.

So measure. Mark the jar level at a feed and time the double. That number, plus the temperature you took it at, is the only honest statement you can make about how strong your starter is today.

A levain is built for one bake#

The starter is the culture you keep alive. A levain is a portion of it, fed once, in the quantity and at the hydration a particular loaf needs — for a 1 kg flour dough, something like 40 g starter with 120 g flour and 120 g water, built 8 to 12 hours ahead. Building one lets you scale up to what the bake needs without ever keeping a large jar, and lets you tune that single build warm or cool depending on which way you want the flavour to go.

Its flour and water are part of the dough, not extra to it. That 120 g of flour is 12% of the formula’s flour prefermented, and the 120 g of water counts in the hydration — write it any other way and the percentages on your page describe a dough nobody mixed.

Use it at peak, or very slightly before. A levain that has already sunk has spent its gas and gone acid, and it will lift the dough more slowly and degrade the gluten faster than one caught on the way up.

Bulk fermentation is judged by the dough#

The recipe said four hours, so the dough came out of the tub at four hours. That is the single commonest way a sourdough goes wrong, and it is worth seeing why the number was never transferable: the clock does not know your kitchen is 18 °C, that your levain was two hours past peak, or that you mixed with water straight from a cold tap.

Read three things instead. Volume: 30 to 50% above the level you marked at the start, which for a wheat sourdough is enough — double is a yeasted-dough habit and here it overshoots. Surface: domed and smooth, with a wobble like set custard when you shake the tub. Edges: bubbles visible where the dough meets the side of the tub, and some dome on the ones you can see through the wall.

End bulk early and the dough finishes fermenting in the oven instead of the tub, and it opens where your shaping was weakest rather than where you scored. Run bulk long and the acid-activated proteases cut the gluten faster than gas inflates it, which leaves a dough too slack to hold a shape at all.

The cold retard and the covered bake#

Shaped, seam up in a floured banneton, the loaf goes into the fridge overnight at 4 °C. Two things come of that. Yeast activity falls off far more sharply in the cold than bacterial acid production does, so the loaf keeps developing flavour without inflating past what the gluten can hold. And a cold dough is a firm dough — it takes a blade cleanly, where a warm slack one drags and snags.

Bake it straight from the fridge in a cast-iron pot preheated to 250 °C. The loaf is around 40% water by weight; with the lid on, the steam it gives off is trapped in a few litres of air around it, and the crust stays moist and extensible through the first ten minutes, which is when oven spring happens. Leave the lid on for twenty, then take it off and let dry heat colour the crust. No lid, no trapped steam, and the surface sets in the first few minutes with the loaf still trying to expand underneath it.

Outline#

• yeasts and lactic acid bacteria at roughly 1:100, and what each contributes
• pH 3.5 to 4.2, and why the culture resists everything else
• homofermentative and heterofermentative LAB; lactic mild, acetic sharp
• temperature, hydration and feeding ratio as the three levers on flavour
• 1:1:1 and 1:5:5, the peak, the collapse, and timing the double
• building a levain for one bake, using it at peak, counting its flour in
• ending bulk on 30 to 50% rise, a domed jiggly surface and bubbles at the wall
• the overnight retard at 4 °C, and the covered bake at 250 °C