A hydrometer is a weighted glass float with a paper scale inside it. Drop it in a liquid and it sinks until it has displaced its own weight; sugary wort is denser than water, so it floats higher. That is the entire instrument — no battery, no calibration drift, nothing to go wrong except the way you read it. Which is why it is slightly absurd that the pages telling you how to read it do not agree.
What does a hydrometer actually measure?
Density, relative to water. Water reads 1.000. A wort with enough dissolved sugar to make a five-percent beer reads about 1.050 — it is 5% denser than water. As yeast turns that sugar into alcohol, which is less dense than water, the number falls. The gap between the reading before fermentation (original gravity, OG) and after (final gravity, FG) is what the ABV calculator turns into a percentage.
It measures density, not sugar — and that matters once alcohol is in the beer, because alcohol drags the reading down independently of the sugar left. A hydrometer handles that correctly and a refractometer does not, which is the whole reason refractometer readings need a second correction after fermentation starts.
How do you take a hydrometer reading?
- Sanitize whatever goes in the beer. The thief or baster, the trial jar, and the hydrometer if it is going into the fermenter rather than into a jar. Drawing into a jar is the safer habit: less time with the lid off, and one less thing to sanitize on brew day.
- Fill the jar about three-quarters full. The hydrometer displaces its own volume, so a jar filled to the brim overflows when you lower it in — and a jar too shallow lets it touch bottom, which reads whatever you like.
- Lower it in and spin it. A gentle twist sheds the bubbles that cling to the bulb, and each one is buoyancy you did not intend. Let it stop drifting before you read.
- Get your eye level with the surface. Reading down at the scale from above adds a point or two of parallax on its own.
- Read it near the temperature on the paper — most are 60 °F (15.6 °C), many metric ones 20 °C. If your sample is hotter, the correction calculator has the chart and the arithmetic.
Top or bottom of the meniscus — which is right?
Wort climbs the glass where it touches, so the surface in the jar is a shallow dish: a low flat middle and a raised rim. That curve is the meniscus, and the ranking pages send you to opposite edges of it. The American Homebrewers Association says to "carefully look where the bottom point of meniscus is and take a reading from that level". Home Brew Answers says to "note the level of the beer, not where the meniscus curves up to meet the hydrometer". DeFalco's instruction sheet says to "read the scale inside the instrument at the level where the liquid contacts the glass" — which is the top. Australian Homebrewing splits the difference with "where the surface of the sample intersects with the scale".
The instrument settles it, and none of them mention how. A hydrometer is not a ruler that you may read from either end; it is calibrated to be read at a defined place. In the hydrometry literature the rule is that hydrometers "are calibrated to give correct indications when read at the principal surface of the liquid" — the flat part, the bottom of the meniscus. Read there and you need no correction, because the manufacturer already made one for you. Read at the rim and you are reading high by a fixed amount.
About one graduation. Homebrewing DIY quotes a British Standard permitting you to read the top of the meniscus and then "correct by using the division (line) that is 2 mm below" — and the lines on a homebrew triple-scale hydrometer are printed every 0.002. So the two ways of reading differ by roughly two gravity points. Yours may be one or three: it depends on the bore of your jar and the cleanliness of the stem, and it is a constant for your equipment, which is the useful part.
Does reading it wrong actually change your beer?
Here is the part nobody computes. Suppose you read the rim every time — consistently high, on the OG and on the FG. This is what that does to the numbers you draw off a 21 L batch of 5 kg pale malt that really went 1.050 to 1.010:
| Misread | ABV | Attenuation | Efficiency |
|---|---|---|---|
| read right | 5.25% | 80.0% | 68.0% |
| +1 point | 5.25% | 78.4% | 69.4% |
| +2 points | 5.25% | 76.9% | 70.7% |
| +3 points | 5.25% | 75.5% | 72.1% |
Apparent attenuation and brewhouse efficiency, computed from this site's engine for 5 kg of 37 PPG pale malt into 21 L. The misread is applied to both readings, as a consistent habit would.
The ABV does not move at all. Not "barely" — it is 5.25% on every row, because alcohol is calculated from the difference between two readings and a constant error subtracts itself out. The number on the bottle, the number your friends ask about, is completely immune to the argument the internet is having.
Everything measured against a single reading moves. Apparent attenuation falls from 80% to 75%, which is the difference between a yeast performing as the lab said and one you would write off as lazy. And brewhouse efficiency climbs from 68% to 72% — four points, for free, from a habit. Four points is the gap people buy a new mill over.
What are the three scales on a triple-scale hydrometer?
Specific gravity, Brix (or Balling — near enough the same thing at brewing strengths), and potential alcohol. The first two are two ways of saying how much sugar is dissolved, and the gravity converter moves between them properly. The third is a trap.
| Specific gravity | Brix / Balling | Potential alcohol |
|---|---|---|
| 1.010 | 2.6° | 1.3% |
| 1.020 | 5.1° | 2.6% |
| 1.030 | 7.6° | 3.9% |
| 1.040 | 10.0° | 5.3% |
| 1.050 | 12.4° | 6.6% |
| 1.060 | 14.7° | 8.0% |
| 1.070 | 17.1° | 9.3% |
| 1.080 | 19.3° | 10.7% |
| 1.090 | 21.6° | 12.1% |
| 1.100 | 23.8° | 13.6% |
Brix from the ASBC polynomial. Potential alcohol from the winemaking divisor that reproduces DeFalco's published table to within a quarter point of ABV above 1.040; below that it runs high, and the guide says so rather than printing it as fact.
The potential-alcohol scale answers a winemaker's question, not a brewer's: what this liquid would reach if it fermented all the way dry. Wine usually does. Beer never does — it stops around 1.010 with unfermentable dextrins left, by design. So the scale is right about a question you are not asking.
Our 1.050 wort sits at 6.6% on the potential-alcohol scale. The beer it becomes is 5.25% — the scale reads 26% high, because it assumed a finish at 1.000 that beer does not reach. Read it at both ends and subtract, as DeFalco's sheet tells you to, and you get 5.3% — close, but still not the same number, because the scale's divisor shifts with gravity and the ABV formula's does not. Use the SG scale and let the calculator do it.
Should you pour the sample back in?
The AHA is blunt: "do not return your sample to the fermenter, as it could cause contamination." That is the safe answer and it is the right default, especially for the OG reading, where the wort is a sterile, unprotected sugar solution and anything you drop in has a clear run at it.
It is worth knowing what the caution costs, though, because nobody prints it. A 150 mL trial jar, 4 readings across a brew — OG, one mid-fermentation, and two days apart at the end to confirm — is 600 mL, or 2.9% of a 21 L batch. Run it through the packaging ledger and it is the difference between 36 and 34 half-liter bottles. Two bottles is the price of knowing when your beer is finished, which is a bargain — but pour the samples into a glass and taste them rather than down the sink. Flat, warm, young beer is the best early warning of an infection you will get.
A hydrometer is thin glass with lead shot in the bottom, and it breaks when it hits a hard tap or the bottom of a bucket. If one breaks in your beer, dump the batch. Fine glass does not settle out reliably, you cannot filter it at home with any confidence, and there is no batch worth the risk. Read into a jar, not into the fermenter, and this never comes up.
My hydrometer reads 1.002 in water — is it broken?
No, and the fix is free. The correction calculator covers running the check in distilled water at the paper's temperature; what matters is what you do with the result. As the BYO piece on MoreBeer puts it, "if it reads 1.002, for example, 0.002 becomes your correction factor for use with all future readings." Subtract it from every reading, forever. Write it on the tube in permanent marker so the next person to pick it up knows too.
And this is the real answer to the meniscus argument. If you read that distilled water the same way you read your wort — same jar, same eye level, same edge of the curve — then whatever the meniscus was worth is already inside the offset you just wrote down. Correcting by it removes your reading habit along with the factory's error. The brewers arguing about which edge to read are arguing about a number that a two-minute check makes irrelevant, and none of the pages having the argument mentions that.
It reads 1.020 after two weeks — is it stuck or done?
Almost certainly done, and higher than the recipe promised. The test is not the number, it is whether the number is still moving: take two readings three days apart. If they match, fermentation is over whatever the value, and the question becomes why it stopped there — usually a warm mash leaving unfermentable sugar, plenty of crystal malt, or an under-pitch that quit early. The attenuation calculator compares where you landed against the range your yeast strain is published at, which is the comparison that actually tells you something.
Before you conclude anything, apply your two corrections. A reading taken at 78 °F on a 60 °F hydrometer is about two points low, and an instrument that reads 1.002 in water is two points high. Two points is also roughly what the meniscus is worth. None of these is large; all three at once is the difference between "stuck" and "finished" — and Brew Day will hold the readings, the efficiency they imply and the ABV they add up to on one sheet.