Flour plus water plus patience equals bread — the oldest recipe on earth. But the sourdough revival taught millions of locked-down bakers a stranger truth: you’re not following a recipe, you’re founding an ecosystem. The thesis from a decade of microbiome research: a starter is a wild, competitive microbial jungle that assembles itself by ecological succession, and understanding its rules makes you a better baker than any schedule.

Who lives in there

A mature starter holds roughly a hundred bacterial cells for every yeast cell. The yeasts (often Saccharomyces cerevisiae alongside wild species like Kazachstania humilis) ferment sugars into CO₂ and ethanol — the lift. Lactic acid bacteria, rod-shaped and tiny, convert the same sugars into lactic and acetic acid — the tang, the keeping quality, the mold resistance (Serious Eats’ starter science gives the friendliest full tour; the PMC ecology review gives the rigorous one).

Why don’t they outcompete each other? Elegant niche partitioning: yeasts prefer glucose and fructose while key lactobacilli specialize on maltose — they eat at different tables. Each side also poisons the commons just enough: ethanol from yeasts, acid from bacteria, both tolerated by the incumbents and lethal to invaders. Low pH (around 3.5–4) is why starters rarely harbor pathogens and why sourdough keeps longer than yeasted bread.

Succession: your starter grows up

Starters mature like forests. Research describes three phases: a chaotic influx of flour microbes, a rise of sourdough-adapted genera, then dominance by acid-tolerant specialists. Early days favor mild homofermentative bacteria (creamy, yogurty notes); mature starters shift to heterofermentative species producing acetic acid — which is why a two-month-old starter tastes sharper than a two-week-old one. Temperature steers the outcome: warmth (30–35 °C) favors lactic softness, cooler room temperatures (15–22 °C) favor acetic tang. Your feeding ratio, flour choice, and fridge schedule are selection pressures, not just chores.

The scale of diversity startled scientists. The eLife study of 500 starters found a median of three bacterial and one yeast species each — but across starters, over 70 yeast types, with acetic acid bacteria emerging as underappreciated drivers of aroma and rise speed. And in a striking experiment (Reese et al., mSphere), 18 professional bakers on two continents made starters from identical flour and recipe — yet each starter diverged, flavored measurably differently, seeded partly from the flour and partly from the bakers’ own hands.

The counterpoint: don’t romanticize the wild

Ecology cuts both ways. The same research shows starter age effects are inconsistent, geography matters less than folklore claims (“San Francisco mystique” is mostly L. sanfranciscensis thriving in many places, not terroir magic), and acetic acid bacteria can slow a rise when overabundant. A starter is resilient, not precious: it survives fridge neglect, recovers from hooch, and rarely truly dies. The failures beginners fear are usually just impatience during the chaotic middle phase, when smells turn alarming right before stability arrives.

Takeaway

Treat feedings as ecosystem management: consistent flour, consistent timing, and warmth tuned to the flavor you want. Taste and smell are your instruments — creamy and mild means young or warm-kept; sharp and vinegary means mature or cool-kept. You’re not baking with an ingredient. You’re collaborating with a jungle, and once it settles, it bakes with you for decades.