What is an Active Sourdough Starter?

At peak, a starter has risen as high as it will go since your last feed, often to around double. The surface will be domed or slightly convex — bulging upward at the centre rather than sitting flat. You’ll see bubbles throughout: large bubbles at the top, smaller ones along the sides of a glass container, and a bubbly, web-like internal structure when you stir it. Peak is the highest rise, not a reading of how active the yeast is: the most yeast, gas and acid do not have to line up with it.

The texture changes too. A fresh starter is thick and paste-like. An active, peaked starter is lighter and airier — it holds its structure but has a loose, spongy feel when you scoop a spoon through it.

Smell is one of the most reliable indicators of starter health. An active, well-balanced starter smells pleasantly sour and yeasty — like a mild yogurt, fresh beer, or tangy apple cider vinegar, but not harsh or unpleasant.

Drop a small spoonful of starter into a glass of water. If it floats, it’s full of gas bubbles and actively fermenting — a traditional sign it’s ready to use. If it sinks, it hasn’t produced enough gas yet.

The float test is a useful quick check but not infallible. A very stiff starter can float before peak, and a 100% hydration starter can sink even when active if stirred vigorously before testing. Use it as one signal alongside volume rise and smell, not the only test.

The best time to use your starter is at or just before peak — fully risen but not yet collapsing. Two things set when that happens: temperature and feeding ratio. Holding the ratio at 1:1:1, my planning estimates move like this:

Watch your starter rather than the clock. Mark the starting level with a rubber band. When the rise slows and the dome is at its highest point, that’s peak — use it within an hour for the best leavening power. Miss it and your loaf will be more acidic with slightly less lift, so catch it on the way up. The sourdough starter manager shows the expected peak on the clock and learns your starter’s timing from the peaks you record. To work out how much to feed so you have enough active starter at peak, use the sourdough starter feeding calculator.

It depends on what you fed it, and on whether it kept rising. Doubling, the highest point and the first fall are three different moments. A doubling time means little on its own, because the ratio sets how much food the culture has to work through:

All three are for 75°F / 24°C. Planning estimates for an established, healthy 100% hydration wheat starter, not one in its first two weeks. The 1:1:1 and 1:2:2 rows bracket what King Arthur Baking saw at about 78°F, checking every 4 hours; the other cells are our extrapolations. Doubling, the highest point and the first fall are separate moments, and your own starter will vary, so mark the jar and learn its timing. Read your own starter against the row that matches your feed:

Flour matters too, and it’s the variable most people forget. Rye and whole wheat bring more nutrients and enzymes than white flour, so a starter fed 50% rye can peak noticeably earlier than the same culture on plain all-purpose. If you change flour, expect the timing to move and re-baseline before judging the culture.

For what each ratio does to timing and flavour in full, see the feeding ratio chart.

One good rise proves very little. Starters produce impressive one-off rises for reasons that have nothing to do with baking strength — a warm afternoon, a fresh bag of rye, or, in a young starter, bacteria rather than yeast. What you’re actually testing for is repeatability, and the usual bar is three consecutive feeds:

1. 1
   Fix the ratio and the time Pick one ratio and stay on it for the whole test, feeding at roughly the same time each day. Changing ratio mid-test resets the baseline and tells you nothing.
2. 2
   Watch it peak three feeds running Each of the three should double within the window for that ratio, and peak at roughly the same time as the one before it.
3. 3
   Consecutive feeds, not one day These are three feeds in a row — not three peaks crammed into a single day. Feeding once daily means the test takes three days. Feeding twice daily is fine too; it just finishes sooner.
4. 4
   Bake on the third Three in a row at consistent timing means the culture is predictable, which is the whole point. Two out of three is close — don't restart the count, just carry on and watch the next feed.

Consistency is what you’re after, not speed. A starter that reliably peaks at 9 hours is more useful than one that peaks at 4 hours today and 11 tomorrow, because you can plan a bake around the first and not the second.

A brand-new starter often rises dramatically around days 2–4, then goes quiet for the best part of a week. That first rise is real, but it’s usually driven by Leuconostoc and other bacteria that thrive briefly in fresh flour and water — not by the wild yeast that leavens bread. The quiet spell afterwards is the yeast taking over, and it’s a normal stage rather than a failure.

So before you apply the three-feed test, ask how old the culture is. On a starter under about a week old, three good rises in a row are still worth having — but treat the first bake as an experiment rather than a certainty, and keep feeding through any slump that follows it.

Beyond the basic signs, three practical checks tell you whether the culture has real leavening power:

A layer of grey liquid (hooch) on top doesn’t mean the starter is dead — it means it’s hungry. Pour off the hooch, discard most of the starter, and feed with fresh flour and water. Give it 24–48 hours of regular feeds before writing it off.

If your starter is alive but sluggish — rising slowly, not doubling, smelling off — follow this revival protocol:

1. 1
   Warm it up Temperature is often the cause. Move the jar somewhere warmer — on top of the fridge, near a radiator, or in an oven with just the light on (79–82°F / 26–28°C). A few degrees warmer can shorten the time to peak noticeably.
2. 2
   Feed small and often Switch temporarily to 1:1:1 (equal parts starter, flour, water by weight) and feed twice daily. A 1:1:1 gives less food per feed than a bigger ratio, but feeding often keeps fresh food coming and regular discarding stops old acid building up while it recovers.
3. 3
   Use strong flour Whole wheat or rye holds more wild yeast and nutrients than white. Add 10–20% of either to your feed for a few cycles to give the culture a boost.
4. 4
   Be patient and consistent Feed at the same time each day. After 3–5 days of twice-daily feeds most sluggish starters recover. Don't bake until it has passed the three-feed test above — a single good rise during recovery isn't enough to plan a bake around.
5. 5
   Discard properly Keep only 20–30g before each feed. Discarding keeps the acid load low — too much accumulated acid inhibits yeast and is a common cause of sluggishness.

A starter kept on the counter and fed daily is in continuous active fermentation mode. A starter stored in the fridge is in a maintained, dormant state — the cold slows activity dramatically but doesn’t kill the culture.

To use a fridge starter, take it out 12–24 hours before baking, give it one or two feeds at room temperature, and wait for it to peak. A week in the fridge usually needs one feed to bounce back; a month dormant may need two or three.

Record the feeding ratio, flour and temperature before comparing how quickly two cultures rise. A large feed at a cool temperature should not be judged against a small warm feed. Changing to rye, whole wheat or a stiff consistency changes the baseline again.

A practical record has four entries: time fed, temperature, time of highest rise and what the surface looked like then. Repeat the same conditions before concluding that the starter is weak. Bubbles alone show activity, but do not show whether the culture can leaven your planned bread on schedule.

The float test is optional. Hydration, flour and handling affect whether a sample retains enough gas to float; a stirred active starter can sink. Use repeatable expansion and its recent feeding response as your main evidence. If the jar has been stored, follow refreshment checkpoints before promising yourself a mixing time.