The Sourdough Float Test: How to Do It and Why It Fails

The float test is probably the best-known way to check a sourdough starter: drop a spoonful into water, and if it floats, it’s ready. It’s quick and fun, and it’s right often enough to stay popular. It’s also wrong often enough to cause a lot of confusion.

Here’s how to do it properly, what it really tells you and what to use instead.

How to do the float test

1. Fill a glass with room-temperature water.
2. Don’t stir the starter. Stirring knocks out the gas you’re trying to measure.
3. Scoop about a teaspoon from the top of the jar, where the starter is most aerated.
4. Gently drop it onto the surface of the water.
5. Floats: the spoonful holds enough gas to be less dense than water. Sinks: it doesn’t, at least for that spoonful.

What the float test actually measures

Floating answers one narrow question: is this spoonful holding gas? Readiness is a broader question: is the starter strong enough, and at the right point in its rise, to leaven a loaf? The two are related, but they aren’t the same thing.

Why the float test gives false results

• Consistency changes everything. A stiff starter, around 50–60% hydration, traps gas well and can float long after its peak. A runny, high-hydration starter lets gas escape as you scoop it and can sink even at its best.
• Rye and whole-grain starters are denser and pastier. A ready rye starter may sink.
• Scooping and stirring knock gas out. Scoop from the bottom, or after stirring, and the result changes.
• Past-peak starter can still float. A starter that’s beginning to collapse still contains plenty of gas.
• Very young starters can float early on from a burst of bacterial activity, before they’re strong enough to raise bread.

In other words, both a float and a sink can be wrong.

Four better ways to check your starter

1. Rise. Mark the level with a rubber band right after feeding. A typical 100% hydration starter, mostly white flour, at least doubles at peak. Rye and whole-grain starters may rise only 1.5–2x.
2. Dome. At peak the surface is domed or just starting to flatten. A dipped or collapsed top with streaks on the glass means the peak has passed.
3. Bubbles throughout. Look through the side of the jar: bubbles should run all the way through, not only on the surface.
4. Timing. A healthy starter peaks at a predictable time after each feed at a given ratio and temperature. Roughly, a 1:1:1 feed peaks in about 4–6 hours at 24 °C (75 °F) and 5–8 hours at 21 °C (70 °F).

If all four line up, your starter is ready, whatever the glass of water says. The full checklist is in how to tell when your sourdough starter is ready.

When the float test is still useful

• As a tiebreaker. If the starter looks nearly ready but you’re unsure, a float is one more vote in favor.
• For a consistent starter you know well. If your 100% hydration starter reliably floats at peak, a sink on a given day is worth investigating.
• Not as the only test, and especially not for runny, rye or stiff starters.

My starter sinks. What now?

• Check the time. Is it before or after the expected peak? Wait an hour and look at the rise again.
• Check the temperature. A cold kitchen slows everything; see why your sourdough starter isn’t rising.
• Check the consistency. Weigh flour and water instead of using cups. If your starter is very runny, reduce the water slightly for a few feeds. The hydration calculator helps with the math.
• Feed it consistently for a few days at the same ratio and temperature, and watch whether the rise becomes predictable.

How Boulie helps

Boulie replaces the guesswork with a record. Each feed you log, with ratio, flour and kitchen temperature, teaches it your starter’s rhythm, so it can predict the next peak. The photo rise check reads the jar and describes the rise and bubbles it sees, and you can confirm or correct the reading.

Key takeaways

• Floating means the spoonful holds gas, nothing more.
• Runny, rye and whole-grain starters can sink at peak; stiff and past-peak starters can float.
• Rise, dome, bubbles and timing are more reliable signals.
• Use the float test as a tiebreaker, never as the deciding vote.