The romantic version says you're capturing wild yeast floating through your kitchen. It's a lovely image and it's mostly wrong. The microbes that become your starter arrive on the flour — they were on the grain, they survived milling, and they're sitting in the bag right now. Your hands contribute some. The air contributes very little.
That single fact changes how you approach this. You're not fishing. You're running a selection process: creating conditions where the few organisms you want — wild yeasts and lactic acid bacteria — outcompete everything else in the bag. You do that by feeding them repeatedly and letting the environment turn acidic, which the competition can't tolerate.
Which is why the answer to almost every problem below is keep feeding it and wait. You're not waiting for luck. You're waiting for arithmetic.
Whole grain flour to begin — whole wheat or rye. Whole grain carries far more microbial life and more minerals than white flour, because the bran and germ are still there. Rye is the fastest starter of the two if you have it. White bread flour for later. A jar around 500 ml, with a lid you rest on top rather than seal. A digital scale.
If your starter stalls and you can't work out why, this is worth ruling out before you change anything else. Not every supply uses chloramine — your water utility publishes which one they use — but most large systems have switched.
| Method | Chlorine | Chloramine | Worth it? |
|---|---|---|---|
| Leave it out overnight | Works | No — takes days | Only if you have chlorine |
| Quick boil, like for tea | Works | No — clears roughly a third | No |
| Gentle boil, 20 minutes | Works | Works | Effective but tedious |
| Carbon block or catalytic filter | Works | Works | Best everyday fix |
| Basic pitcher filter | Works | Unreliable | Better than nothing |
| Pinch of vitamin C powder | Works | Works | Cheap, instant |
| Campden tablet | Works | Works | The brewer's answer |
| Bottled water | Works | Works | Simplest of all |
On the two chemical options: ascorbic acid is potent enough that 1,000 mg treats 30–50 gallons, so a few grains handles a jar — it nudges the pH down very slightly, which a starter won't mind. A campden tablet (potassium metabisulfite, from any homebrew shop) does one tablet per 20 gallons, so a quarter tablet is generous for a week of feeding; give it five minutes. Both are standard practice in brewing, where chloramine ruins a batch outright.
Warmth matters more than anything else on this list. The whole process wants 75–80°F / 24–27°C. At 65°F it can take three weeks instead of ten days. Find the warmest stable spot in your home — on top of the fridge, in the oven with only the light on, near a radiator but not on it.
-
Mix itDay 1 · 2 minutes
50 g whole grain flour + 50 g water at around 85°F / 29°C. Stir to a thick paste with no dry pockets. Scrape the sides down, rest the lid on top without sealing, and leave it somewhere warm.
Mark the level with a rubber band or a marker on the glass. You will be judging everything from here by how far it moves.
-
Look, don't feedDay 2 · 30 seconds
Probably a few bubbles. Possibly nothing. Either is fine.
Don't feed it today. A culture this young is tiny, and feeding it now just dilutes what little has established. The instinct to intervene is the enemy for the next two weeks — most starters that fail are over-managed, not neglected.
-
The lieDay 3 · 3 minutes
There's a good chance you wake up to a jar that has doubled, is full of bubbles, and smells sharp and slightly cheesy. It looks like you've done it in two days.
You haven't. This is the false rise, and almost everyone gets fooled by it. It's driven by Leuconostoc and other bacteria that thrive briefly at neutral pH, produce a lot of gas, and then die off as the mixture turns acidic. There's very little yeast involved. If you bake with this, you get a brick.
Feed it anyway: discard all but 50 g, then add 50 g flour and 50 g water. From today, switch to half whole grain, half white bread flour. Same time each day.
Why discard at all Without it you'd have kilos of starter inside a week, and the culture would sit in its own acid and waste. The discard isn't failure — it's how you keep the food-to-microbe ratio high enough for the population to keep growing. -
The quiet days — where everyone quitsDays 4–7 · 3 minutes daily
The false rise collapses and the jar goes dead. Flat. No bubbles. It may smell unpleasant — sour milk, gym socks, nail polish, occasionally something like vomit. That smell is real and it is normal.
What's actually happening is the handover: the early bacteria have acidified the environment enough to kill themselves off, and the acid-tolerant organisms you actually want are still building numbers. There's a gap between the two, and the gap looks exactly like failure.
Keep feeding once a day. Change nothing. Do not throw it out. This stage lasts three or four days and it is the single most common point of abandonment.
Still nothing after day 8? Then it's worth troubleshooting — nine times out of ten it's temperature or chloramine. Starter not rising works through the causes in order. -
It comes backDays 8–11 · 6 minutes daily
Activity returns, and it looks different this time: smaller, more uniform bubbles spread through the whole mass rather than a few big ones, a domed top, and a rise that's predictable rather than explosive. The smell turns — tangy, yogurty, a little beery. Pleasant, or at least clean.
Once it's reliably rising within about 12 hours, move to twice-daily feeding, roughly 12 hours apart, same 1:1:1 ratio. You can go fully white bread flour now if you want a milder starter, or hold at 25% whole grain for more activity and flavor.
-
ReadyDay 12 onward
It's ready when it does the same thing three days running: doubles within 4–8 hours of feeding, domes on top, then slowly flattens. Predictability is the test, not vigor. A starter that doubles once impressively and then sulks for two days isn't ready.
Now name it and go bake something.
| Smell | Usually means | Do |
|---|---|---|
| Sweet, like damp hay or raw flour | Day 1–2. Nothing has started yet | Wait |
| Sharp, cheesy, faintly like vomit | Day 3–6. The bacterial handover | Keep feeding. Normal |
| Nail polish remover, solvent | Hungry. Alcohol building up between feeds | Feed more often, or warmer |
| Yogurt, beer, clean sourness | You're there. Yeast is in charge | Move to twice daily |
| Harsh vinegar | Over-acidified, underfed | Too sour |
| Grey liquid on top | Hooch. Hungry, not dead | Hooch |
| Pink, orange, or fuzzy growth | Contamination — the one real failure | Throw it out and restart |
The float test — drop a spoonful in water, see if it floats — is treated as gospel and it's barely better than a coin flip. A stiff starter floats when it isn't ready; a slack one sinks when it is. It measures trapped gas, which correlates with readiness only loosely.
Use the jar instead. Band on the glass at feeding, then watch: doubles in 4–8 hours, domes, then flattens — three days running. That's it. That's the test.
| Symptom | Most likely cause | Go here |
|---|---|---|
| Nothing at all after 8+ days | Too cold, or chloramine in the water | Starter not rising |
| Rose once, then died | That was the false rise. Keep going | Days 4–7 above |
| Grey liquid on top | Underfed for the temperature | Hooch |
| Rises but the bread doesn't | Used past peak, or too young yet | No oven spring |
| Pink or orange streaks, fuzz | Contamination | Discard and restart |
Notes from the counter
Two weeks is a long time to maintain faith in a jar of grey paste, so here's the thing worth holding onto: the failure rate on this is very low, and almost all of it happens on days four to six.
People don't fail at making a starter. They stop. They see the day-three fireworks, believe it worked, come back on day five to a flat dead jar that smells like a locker room, conclude they've killed it, and start over — which puts them back on day one, where in three days they'll see fireworks again. I've watched people loop that three times and decide sourdough isn't for them.
So: the ugly stage is the proof it's working. That awful smell is the sound of the wrong organisms losing. Nothing is wrong. Feed it and go do something else.
The other thing — don't chase someone else's schedule too precisely. Your kitchen has a temperature and a flour and a water supply, and those three things set the pace. If the guide says day eight and you're on day twelve, you're not behind. You're just colder.