August 27, 2026

Which salinity tester can you trust: refractometer, hydrometer or digital?

Your swing arm hydrometer says 1.023. The refractometer you bought to check it says 1.026. The shop's digital tester read your sample at 1.028. Three instruments, three answers, spread across 0.005 SG, more than 6 ppt, wider than the whole range a reef is supposed to live in.

Here is the short version: none of the three measures salinity directly, and none can be trusted until it has been checked against a 35 ppt reference solution. A calibrated refractometer or conductivity tester will get you within 0.001 SG. A swing arm hydrometer will not. And whoever calibrated most recently is the one who is right.

What each salinity tester actually measures

Nothing on your bench weighs dissolved salt. A hydrometer measures density. A refractometer measures refractive index, how far light bends passing through the sample. A digital salinity tester is either a refractometer with a sensor in place of your eye, or a conductivity tester that measures how easily a current flows through the water. Natural seawater at 35 ppt conducts at about 53 mS/cm at 77 F (25 C).

Every conversion is temperature dependent and assumes a particular mix of salts. Natural seawater at 35 ppt is about 1.026 SG at 77 F (25 C), and 0.001 SG is a little over 1 ppt, so a 0.002 gap between two devices is nearly 3 ppt. If you are not sure what number to aim for, what is the ideal salinity for a reef tank settles the 1.025 versus 1.026 question.

Refractometer vs hydrometer: why the cheap one loses

A swing arm hydrometer is a plastic box with a pivoting needle. Fill it, the needle floats, read where it settles. It is cheap and it is where most tanks start. It has three problems, and they stack.

Bubbles. Air clings to the needle and lifts it, so the reading comes out high. Tap it, tilt it, fill it from the side, and you will still get a bubble one time in three.

Temperature. Density falls as water warms, and every hydrometer is calibrated to read correctly at one temperature: most current ones at 77 F (25 C), older designs at 60 F (16 C). Use a 60 F model on 78 F (26 C) tank water and it reads roughly 0.002 to 0.003 low.

Salt creep. Dried salt on the pivot makes the needle stick and dried salt on the needle changes its weight. Rinse after every use and soak in vinegar monthly, and it still drifts.

A swing arm being 0.002 off is therefore normal, not a defect, and that is the whole width of the 1.024 to 1.026 band most reefers aim for. The refractometer vs hydrometer argument was settled years ago. Checked against a calibrated refractometer it makes a fast daily check on top-off drift, but it is not something you set salinity by.

Optical refractometer: good, if you calibrate it properly

An optical refractometer is a metal tube with a prism at one end and an eyepiece at the other. Two or three drops on the prism, close the cover, hold it to a window, and read where the colour changes from blue to white against a scale in ppt and SG.

Most hobby refractometers are built for brine, not seawater. Even one sold with a reef on the box usually carries the scale for a plain sodium chloride solution. Seawater bends light slightly less per gram. Natural seawater at 35 ppt has the same refractive index as a 36.5 ppt sodium chloride solution, so a brine refractometer zeroed with distilled water reads seawater about 1.5 ppt, roughly 0.001 SG, high. You dial the tank in to 35 ppt on the eyepiece and the water is actually at 33.5.

Distilled or RO water only sets the zero, and zero is 35 ppt away from where you read. Calibrate with a 35 ppt reference solution instead and the offset disappears at the one point on the scale that matters. If yours did not come with a bottle, buy one.

ATC means automatic temperature compensation, and it compensates for the instrument, not the sample. Inside an ATC refractometer a bimetallic strip nudges the scale as the metal body warms and cools, typically over 50 to 86 F (10 to 30 C). That works because two drops take the temperature of the prism within about 30 seconds, and it does not work if you read immediately or if the body has just come out of a cold garage. Let the refractometer, the reference and the sample sit in the same room for 20 minutes, apply the drops, wait 30 seconds, then read.

Digital salinity tester: accurate, but the probe needs looking after

Digital refractometers use the same prism with a light sensor in place of your eye. A seawater scale and a temperature sensor are built in, so the brine and temperature problems go away. The catch is the published accuracy: the Milwaukee MA887 is specified at plus or minus 2 ppt, or 0.002 SG. No more precise than a cheap optical refractometer with a reference bottle.

Conductivity testers are what people usually mean by digital salinity tester: pen units like the Hanna HI98319, and the salinity probes on aquarium controllers. The Hanna is specified at plus or minus 1 ppt, or 0.001 SG. Conductivity moves about 2 percent per degree C, so every one has a built in temperature sensor. Give the probe 30 to 60 seconds in the sample until the number stops moving.

Three rules keep one honest:

• Calibrate it with the right fluid. Conductivity probes calibrate at 35 ppt, or 53 mS/cm, not at zero. A sodium chloride solution cannot match seawater in density, refractive index and conductivity all at once, so a refractometer reference and a conductivity standard are different solutions even when both say 35 ppt. Use the one made for the instrument, from a fresh sachet rather than an open bottle.
• Rinse the probe in RO water after every use and never let tank water dry on the electrodes. Salt deposits change the cell geometry and the reading drifts with no warning.
• Recalibrate on a schedule. Every 2 to 4 weeks for a pen tester, monthly for a controller probe, and after any reading that surprises you. Controller probes also collect bubbles and biofilm, so mount them in flow and wipe them when you calibrate.

Calibrated monthly, it is as trustworthy as any salinity tester a hobbyist can buy. Never calibrated, it is a random number generator with a battery.

Best salinity tester for a reef tank

For most tanks: an optical refractometer plus a 35 ppt reference solution. Cheap, no batteries or probes to fail, and calibrated properly it reads within 0.001 SG. If you would rather not squint at a boundary line, a conductivity tester is the upgrade: faster, reads to 0.1 ppt, and the natural choice if you already run a controller. Budget for calibration fluid, because without it the accuracy on the box means nothing.

A swing arm hydrometer belongs in the cupboard as a backup with its offset written on the side in marker. Whichever you buy, the reference solution is the real instrument. The tester is just a way of comparing tank water to it.

How to calibrate a refractometer

Five minutes, before every batch of new saltwater.

1. Get a 35 ppt reference. Buy a bottle labelled for refractometers, or make one: dissolve 3.65 g of pure sodium chloride (no iodine, no anti caking agent) in 96.35 g of RO water, weighed on a scale that reads to 0.01 g. That solution optically matches 35 ppt seawater. Keep it sealed so it does not evaporate and drift upward.
2. Equalise temperatures. Leave the refractometer, the reference and your tank sample together for 20 minutes at room temperature, roughly 68 to 77 F (20 to 25 C).
3. Apply and wait. Two or three drops of reference on the prism, close the cover with no bubbles under it, wait 30 seconds.
4. Adjust. If the line is not on 35 ppt (1.026 SG), turn the calibration screw until it is. Rinse, dry, and confirm with a second drop.
5. Test the tank. Rinse, dry, drops of tank water, wait 30 seconds, read. Rinse and dry again before it goes back in the box, because dried salt scratches the prism and a scratched prism gives a fuzzy line.

Repeat the check monthly, after any knock, and whenever a reading would make you change something. Never adjust salinity on a reading you have not verified against the reference that day. The same applies to your mixing bin, since salt mixes differ enough between brands to be worth testing on every new bucket.

Your reading disagrees with the store: who is right?

You read 1.024 at home. The store reads your sample at 1.027. That is 4 ppt, and the fish do not care whose fault it is.

Neither of you is right by default. The store owns a device that gets used forty times a day and calibrated far less often, frequently zeroed with RO water on a brine scale, which makes it read high. Ask two questions: what did you measure it with, and when did you last calibrate it? "This morning, with 35 ppt solution" wins the argument. "It is digital, it does not need it" loses it.

Then use the tie breaker. Put your reference solution on your refractometer. If it reads 35 ppt, your tank reading stands. Take the bottle to the store and ask them to read it on theirs. Whichever device does not read the reference at 35 is the one that is wrong.

Rule out the sample itself too. A sample that sat in a car at 50 F (10 C) and went straight onto the prism reads high, so let it warm up. An open cup loses freshwater to the air and creeps saltier, so carry a sealed bottle filled to the top.

If you were the one who was wrong, do not fix it in an afternoon. Move salinity no faster than about 0.001 SG (roughly 1 ppt) per day, because fish and corals tolerate slow drift and punish step changes, so a 0.003 correction is a three day job. To raise it, mix the next water change or your top-off water slightly above target. To lower it, top off with RO water in small amounts through the day: on a 200 litre (50 gallon) tank, 1 ppt is roughly 6 litres (1.5 gallons) of RO beyond normal top-off, so 500 mL at a time is the right pace. Never add dry salt to the display, and never pour a correction batch in without flow to disperse it. If corals sulk afterwards, corals not opening after a water change covers the salinity mismatch case.

Log the tester, not just the number

Salinity is a weekly test plus every batch of new saltwater. A line that steps 0.002 on the day you switched from the hydrometer to the refractometer is a calibration event, not a tank event.

Log the reading in one unit, which device you used, and when it was last calibrated, in the notes column beside the reading. When salinity moves on the chart you can then see whether the water changed or the instrument did. The reef tank testing schedule shows where salinity sits among the other parameters, and the reef tank parameters chart covers which number to fix first if more than one is out.

Get the spreadsheet
Make a copy in Google Sheets (needs a Google account)
Download the Excel version (no account needed)

Free, no email, no attribution. Readings turn green inside your target band and red outside it, and the charts draw your target band as dashed lines.

Keeping track of which device read what, and when each was last calibrated, is the sort of bookkeeping that gets skipped until a salinity line steps and nobody knows why. Recif is the reef companion app we built to handle this: it logs every test, charts each parameter against your own target ranges so a drifting salinity reading stands out from a calibration jump, reminds you when a parameter has gone too long untested, and rolls the tank into one Stability Score. It is available on both the App Store and Google Play.