A hydrometer is calibrated to read accurately at one specific temperature — commonly 60°F or 68°F, marked on the instrument itself — and a sample read at any other temperature needs correcting before you compare it to a recipe target or a previous reading. This matters more than it sounds like it should: a sample straight out of a warm kettle can be off by several points uncorrected.
The correction table (calibrated at 60°F, raw reading 1.050)
| Sample temperature | Corrected reading |
|---|---|
| 40°F | 1.049 |
| 50°F | 1.049 |
| 60°F (calibration point) | 1.050 |
| 70°F | 1.051 |
| 80°F | 1.052 |
| 90°F | 1.054 |
| 100°F | 1.056 |
| 110°F | 1.058 |
| 130°F | 1.064 |
| 150°F | 1.070 |
These figures come from a widely-published polynomial fit to the physical expansion behaviour of water and wort as temperature changes, calibrated so the correction is exactly zero at the reference temperature (60°F here) and grows in both directions away from it. Notice the correction is small and barely matters within about 10°F of calibration, but grows sharply at temperature extremes — a sample at 150°F needs a correction of a full 20.0 points (1.050 corrects to 1.0700), while the same instrument at 90°F needs a correction of about 4.1 points (1.050 corrects to 1.0541).
Why it works this way
Liquid expands as it warms, becoming less dense — a hydrometer floats lower in warmer liquid than it would in the same liquid cooled to its calibration temperature, which makes the RAW reading understate the true gravity at temperatures above the reference point, and overstate it slightly at temperatures below. The correction table (and the underlying formula) restores what the reading would show at the calibration temperature, which is the number that's actually comparable to your recipe's target gravity or your previous reading.
Cooling the sample beats correcting it
Cool your sample close to your hydrometer's marked calibration temperature before reading it whenever practical — it's simpler than correcting, and removes the correction step entirely. When that's not practical (a quick check straight from an actively-mashing kettle, for instance), note the sample temperature at the moment you read it and apply the correction before recording the number anywhere. A reading taken hot and left uncorrected, then compared later against a reading taken cold and also uncorrected, can look like a meaningful gravity change when it's actually just two different, uncorrected temperature errors pointing in different directions.
This correction is separate from — and often confused with — refractometer correction, which corrects for a completely different effect (alcohol's influence on refractive index) and only applies once fermentation has actually started. A hydrometer reading fresh wort before pitching needs temperature correction only; a refractometer reading fermenting or fermented beer needs both.