The Sourdough Starter Microbiome: A Stable Ecosystem
A mature sourdough starter (maintained for weeks or months) is a remarkably stable ecological community despite never being sterilized or inoculated with pure cultures. The stability arises from the microenvironment the organisms themselves create: LAB produce lactic and acetic acid that lower the pH to 3.5–4.5, killing most competing bacteria; yeast produce ethanol that further suppresses pathogens; and the organisms that survive have co-evolved specific tolerance to both acid and ethanol.
The Core Organisms
Despite geographic variation, most sourdough starters worldwide converge on similar core organisms. The Reese 2020 (Cell Host Microbe) study — the most comprehensive sourdough microbiome survey, analyzing starters from 595 bakers across 4 continents — found:
- Lactic acid bacteria (LAB) dominance: LAB constitute 70–90% of bacterial cells in most mature starters. The most commonly identified species globally: Lactiplantibacillus plantarum (the most widespread, tolerates wide pH and salt range), Limosilactobacillus sanfranciscensis (formerly Lactobacillus sanfranciscensis — the classic San Francisco sourdough organism, heterofermentative, acetic acid-dominant), Fructilactobacillus sanfranciscensis (reclassified 2020), and Limosilactobacillus fermentum. Geographic signal: European starters favor different LAB species than American starters, but the functional outcomes (acid production, phytase activity, flavor) are similar.
- Wild yeast: Saccharomyces cerevisiae is present in many starters, but Kazachstania humilis (formerly Candida humilis) is more commonly the dominant yeast in traditional sourdough, particularly in European starters. K. humilis is more acid-tolerant than S. cerevisiae and co-evolved with LAB in the sourdough environment. Unlike S. cerevisiae, K. humilis cannot ferment maltose — a key evolutionary partnership: LAB amylase converts starch to maltose, LAB consumes the maltose, leaving glucose for the yeast; this division of substrate means LAB and yeast compete less directly than would two glucose-consuming organisms.
Autolyse: Passive Gluten Development
Autolyse is a rest period (30–60 minutes) after mixing flour and water — before adding sourdough starter and salt. During autolyse:
- Glutenin and gliadin proteins fully hydrate and begin forming disulfide cross-links passively
- Endogenous flour amylases begin converting damaged starch to maltose (feeding LAB later)
- Endogenous flour proteases (particularly at whole-wheat ratios) partially cleave gluten proteins, improving extensibility
The result is a dough that requires substantially less mechanical mixing to develop adequate gluten — important for artisan bakers using minimal equipment. Salt added after autolyse further tightens gluten (salt ions compete with water molecules for binding sites on gluten proteins, reducing water activity and stiffening the network) and suppresses premature enzymatic activity.
| Fermentation Variable | Effect on LAB | Effect on Yeast | Effect on Flavor | Baker's Control |
|---|---|---|---|---|
| Temperature ↑ (25–28°C) | Faster LAB growth; favors homofermentative LAB (more lactic acid) | Faster CO₂ production; shorter bulk fermentation time needed | Milder, more lactic sourness; less complex (fewer heterofermentative byproducts) | Warm kitchen (summer) or proofing drawer; reduce bulk time by 20–30% |
| Temperature ↓ (4–8°C) | Slowed LAB but heterofermentative species retain activity longer; favors acetic acid accumulation | Significantly slowed; cold retard stops yeast activity more than LAB | More acetic (sharper) sourness; longer flavor complexity development; open crumb structure preserved | Overnight cold retard in refrigerator after shaping; allows flexible schedule and score-and-bake timing |
| Hydration ↑ (75–85%) | More mobile aqueous phase; faster LAB metabolism; more lactic acid production | Faster fermentation rate; CO₂ produced more quickly | Open, irregular crumb; milder flavor (faster LA dominance) | More difficult to shape; requires wet hands and bench scraper; high hydration sourdoughs need strong gluten development (more folds) |
| Hydration ↓ (60–68%) | Slower metabolism due to reduced water activity; longer fermentation needed; acetic acid-favoring conditions | Slower; stiff doughs ferment more slowly even at same temperature | Tighter crumb; more acetic sourness; longer shelf life (lower water activity inhibits mold) | Easier to shape; preferred for sandwich loaves and bagels; used deliberately for sharper San Francisco-style sourdough |
| Whole wheat % ↑ | Higher LAB activity (more nutrients, more phytase for acid activation); faster fermentation | More minerals for yeast nutrition; faster | More complex flavor from bran phenolics and amino acid diversity; more pronounced sour notes | Shorten bulk fermentation by ~15% per 10% added whole wheat; reduce proof time accordingly; consider higher hydration to compensate for bran water absorption |
Sourdough Baker's Science Checklist: Why Loaves Fail and What to Adjust
- Dense, gummy crumb (underproofed): Underproofing means bulk fermentation ended before adequate CO₂ was produced and gluten relaxed enough to hold the gas structure. Diagnosis: the poke test — a properly proofed loaf springs back slowly and incompletely when poked with a floured finger; underproofed dough springs back quickly and completely. Fix: extend bulk fermentation until the dough has increased ~75–100% in volume (not doubled — most sourdoughs don't fully double), shows visible bubbles on the surface and sides, and feels light and airy when gently shaken. The "float test" (dropping a small piece of dough in water — if it floats, it's ready) is unreliable; volume increase + surface texture + poke test are more accurate combined indicators.
- Dense crumb with gummy streaks (overproofed): Overproofed dough has exhausted its gluten structure's ability to hold CO₂ — the gas cells coalesce and collapse. On cutting, overproofed bread shows gummy layers where the crumb structure failed to set. The proteases (which are continuously active during fermentation) eventually degrade the gluten network too much if fermentation continues past the optimal window. Fix: reduce bulk fermentation time by 15–20% (mark your container with a rubber band to track volume more precisely); if using cold retard, shorten overnight proof. Overproofed dough cannot be saved — it must be baked immediately as-is (result will be acceptable flavor with compromised crumb) or repurposed for focaccia (flat bread tolerates overproofing better because structure is less critical).
- Flat loaf (spreading instead of rising): Caused by: (1) weak starter — starter should be used at peak (doubled in volume, domed, showing vigorous bubble activity, 8–12 hours after feeding at room temperature); a starter past peak has depleted sugars and the yeast are less vigorous; (2) insufficient gluten development — inadequate folding during bulk fermentation leaves gas cells not supported; solution is stretch-and-fold every 30 minutes for 3–4 sets in the first 2 hours of bulk; (3) insufficient dough tension during shaping — tension is created by dragging the dough toward you on an unfloured surface, creating surface friction that builds the outer gluten membrane that holds shape during proofing.
- Ear and bloom: the scoring science: The "ear" (raised flap along the score line) develops because: (1) the score creates a weak point where expanding CO₂ and steam preferentially exit; (2) the angled blade (30–45° to the dough surface) creates a flap that lifts as the interior expands. Blade angle matters: a perpendicular cut creates a bulge without an ear; a 30–45° cut creates lift. Cold dough scores more cleanly than room-temperature dough because the surface is firmer and resists tearing. Scoring depth should reach gluten network (5–7mm for most loaves) — too shallow and the score seals before the loaf finishes expanding; too deep and structural support is compromised, causing loaf to split laterally.
- Baking with steam — the crust science: Steam in the first 15 minutes keeps the crust surface moist and extensible, allowing maximum oven spring (the rapid expansion of CO₂ as the cold dough reaches baking temperature — CO₂ volume increases ~12-fold from 4°C to 230°C). Without steam, the crust sets and hardens before the loaf has fully expanded — creating a smaller, denser loaf with thick crust. A Dutch oven (preheated covered baking vessel) traps the dough's own steam during initial baking, then is uncovered at 15–20 minutes to allow crust drying, Maillard browning, and the characteristic sourdough crust color and snap. Internal temperature of 96–98°C confirms the crumb is fully set (starch fully gelatinized, proteins denatured) — use a probe thermometer to verify rather than relying solely on baking time.
The three most impactful investments for sourdough quality: (1) Dutch oven (enameled cast iron, 5–6 quart) — replicates a steam-injected deck oven for crust and oven spring; any Lodge or Le Creuset-style vessel works. (2) Banneton (rattan proofing basket) — provides structure during cold retard, wicks surface moisture for better scoring, creates the signature spiral pattern; linen liner optional. (3) Bread lame (thin curved blade on a stick) — the curved blade allows the 30–45° angle necessary for creating an ear; a razor blade taped to a chopstick works but the curved lame gives more control. A digital probe thermometer to verify 97°C internal temperature is the fourth essential — baking by color alone leads to undercooked gummy crumb.