Hydrometer vs Refractometer: Wort & Must Gravity (2026)
A glass triple-scale hydrometer lists at $8.79 at MoreBeer; a dual-scale ATC refractometer runs $15–$25 on Amazon and $49.99 at Midwest Supplies. On original gravity the two land within a point or two of each other. They diverge the moment yeast starts working. A hydrometer measures density, which is exactly what fermentation changes. A refractometer measures refractive index, and ethanol (nD 1.3611) bends light more than water (nD 1.3330) while making the liquid less dense — so a fermenting sample reads high, and the error grows as ABV climbs. Stack on the fact that refractometers are calibrated against pure sucrose while wort is not sucrose (the 1.02–1.06 wort correction factor, 1.04 typical), and every post-pitch refractometer reading carries two separate corrections.
| Factor | Hydrometer | Refractometer |
|---|---|---|
| 2026 price (USD) | $7–$15 for a glass triple-scale — MoreBeer lists one at $8.79. A NIST-traceable precision finishing hydrometer graduated in 0.001 runs $25–$35. Budget another $8–$12 for a 250 mL test jar, which is not optional. | $15–$25 for a dual-scale ATC unit on Amazon; homebrew shops still list them near $50 (Midwest Supplies: $49.99). Price includes a hard case, pipette, microfiber cloth, and calibration screwdriver. Before ~2010 these sold for $50–$70. |
| Sample size needed | 100–250 mL per reading — the float needs depth to swing free without touching bottom or wall; 250 mL is the standard test-jar size. Four readings off a 5-gallon (18.9 L) batch at 200 mL each is 800 mL, about 4.2% of the beer poured away, plus four openings of the fermenter. | 2–3 drops, roughly 0.1–0.15 mL. Ten readings total under 2 mL — about 0.01% of a 5-gallon batch. You can read a sample off the mash paddle and lose nothing. |
| What it physically measures | Density, by buoyancy (Archimedes). The float sits higher in a denser liquid; the scale reports specific gravity relative to water at the printed calibration temperature. Ethanol is less dense than water, so the reading drops as fermentation proceeds — a genuine density change, not an artifact. | Refractive index: how sharply light bends crossing the prism/sample boundary, read as a shadow line and converted to Brix on a pure-sucrose calibration. Ethanol’s nD of 1.3611 exceeds water’s 1.3330, so alcohol bends light more while lowering density. The two instruments err in opposite directions. |
| Accuracy before fermentation (OG) | ±0.001 SG once temperature-corrected. A typical triple-scale is graduated in 0.002 SG steps (0.982–1.160), so you interpolate between marks; dedicated finishing hydrometers are graduated in 0.001. Main error source is reading the meniscus wrong or trapped CO2 bubbles lifting the float. | ±0.2 Brix, graduated in 0.2 Brix steps over 0–32% — about ±0.0008 SG. Matches the hydrometer on unfermented wort or must provided you divide the display by your wort correction factor. For OG work the two instruments agree within 0.001–0.002 SG. |
| Accuracy AFTER fermentation starts | Unchanged. Sugar converts to ethanol and CO2, density falls, and density is precisely what the instrument reads. No formula, no correction factor, no dependence on the OG reading. This is why it stays the reference instrument. | Wrong without correction, and progressively worse as ABV rises. Ethanol lifts the refractive index, so the shadow line stalls high. A beer that truly finished at 1.010 from OG 1.050 still displays about 5.8 Brix, which is 5.58 true Brix after dividing by a WCF of 1.04. Recovering the real FG requires Terrill’s cubic using BOTH readings, each first divided by your wort correction factor: FG = 1.0000 − 0.0044993·RIi + 0.011774·RIf + 0.00027581·RIi² − 0.0012717·RIf² − 0.0000072800·RIi³ + 0.000063293·RIf³, where RIi is the WCF-corrected OG Brix and RIf the WCF-corrected current Brix. Do not feed raw display readings into this formula. |
| Temperature sensitivity & correction | High, and manual. On a 60 °F-calibrated hydrometer a sample at 80 °F reads about 0.0024 low and at 100 °F about 0.0059 low. Either cool the sample to the printed reference (60 °F/15.6 °C or 68 °F/20 °C — check the paper scale, they differ) or apply the standard polynomial, good to 0.001 SG from 32–212 °F. | Automatic — within limits. ATC compensates only from 50–86 °F (10–30 °C). A 212 °F boil sample is far outside that band, but 2–3 drops spread on a metal prism reach room temperature in roughly 30 seconds, which is what makes the instrument practical on brew day. Wait the 30 seconds; don’t read hot. |
| Scale read (SG vs Brix) | Reads specific gravity directly, typically 0.982–1.160, usually printed alongside Brix (−4 to 35) and a potential-alcohol scale (0–20%). No conversion factor, no arithmetic. Read the bottom of the meniscus at eye level. | Reads Brix on a sucrose scale, so wort needs a wort correction factor of 1.02–1.06 (1.04 default) — divide the display by the WCF to get true Brix. A wort at true 12.4 Brix (SG 1.050) displays about 12.9 Brix. Find your own WCF at real wort strength: dissolve about 3.5 oz DME in 32 oz water to land at SG 1.040–1.050, rest 15 minutes, then read the same sample on both instruments. Convert the hydrometer reading with the real cubic, Brix = 143.254·SG³ − 648.670·SG² + 1125.805·SG − 620.389, and divide the refractometer display by that value. Sanity check: if your computed WCF falls below 1.00, you have made an arithmetic error — wort always reads high on a sucrose scale, so the WCF is always greater than 1. Average across several worts. |
| Durability & breakage | A thin borosilicate tube about 12–14 in long, weighted with shot. It breaks when dropped, when it strikes the bottom of a too-narrow jar, and from thermal shock if you drop it into hot wort. Buying two at once is standard practice for a reason. Plastic versions survive falls but read less precisely. | Metal body, rubber grip, hard storage case — survives a brew-day drop onto concrete. Failure modes are a scratched prism and calibration drift, not shattering. Re-zero on distilled water at room temperature with the included screwdriver before each brew day. |
| Speed of use | 5–15 minutes per reading: draw sample, fill the jar, cool to reference temperature or measure and correct, spin the float to shed bubbles, read, then sanitize the jar, thief, and hydrometer. The sample is normally discarded rather than returned to the fermenter. | Under 30 seconds start to finish: two drops on the prism, close the cover plate, hold to light, read the shadow line, wipe. Fast enough to check every runnings during a sparge and cut off the instant pre-boil gravity hits target. |
| Best use case | Final gravity, ABV for a label or competition entry, confirming a genuinely stuck fermentation, and any number you intend to publish. It is also the reference standard you calibrate the refractometer’s WCF against — the hydrometer never needs the refractometer, but the refractometer needs the hydrometer. | Everything before pitch: first runnings, sparge cutoff, pre-boil gravity, end-of-boil gravity, and blending or diluting must to a target sugar level. After pitch it is a trend tool — good for confirming gravity is still falling day over day, not for producing a number. |
Our Verdict
Own both. Together they cost under $60, and each covers the other’s blind spot. Use the refractometer wherever no alcohol exists yet — first runnings, sparge cutoff, pre-boil, end of boil — where 2 drops and 30 seconds beat cooling a 250 mL sample, and where saving 800 mL over four readings is real beer. The moment yeast goes in, the refractometer becomes a direction indicator only: read it daily to confirm gravity is still moving, but do not convert the number. Take final gravity with the hydrometer. If you must use the refractometer for FG, run both the original and the final Brix through a proper correction such as Terrill’s cubic — never the SG scale printed on the eyepiece, which is nothing but a sucrose conversion. Skipping that step is the single most common homebrew measurement blunder. A beer that actually finished at 1.010 still shows about 5.8 Brix, and 5.8 Brix converts straight to 1.023 on the printed scale. From OG 1.050 that yields (1.050 − 1.023) × 131.25 = 3.5% ABV instead of the true (1.050 − 1.010) × 131.25 = 5.25% ABV, and an apparent attenuation of 54% instead of 80%. The brewer then diagnoses a stuck fermentation that finished a week earlier — repitching yeast, raising the temperature, rousing the cake, and in the worst case bottling a beer they believe has 23 points of sugar left. The hydrometer answers that question in one reading with no formula at all.