August 27, 2026

What is the ideal salinity for a reef tank: 1.025 or 1.026?

Search "ideal salinity for reef tank" and the first page gives you 1.025, 1.026, 35 ppt and 1.023, each presented as the answer. Your refractometer reads 1.024, the salt bucket says 35 ppt, and they cannot all be right.

Here is the short version. Natural seawater over most coral reefs is 35 ppt, which is a specific gravity of 1.0264 at 25 C (77 F). So 1.026 is the closer match to what your corals evolved in, 1.025 is inside the range of real reefs, and anything from 1.025 to 1.026 is a good target. Pick one number in that band and hold it. A tank that reads 1.025 every week of the year is in better shape than one that averages 1.026 while wandering between 1.023 and 1.028 on the way.

35 ppt, 1.026 and 53 mS/cm are the same water

Salinity is the amount of dissolved salt in the water. Nobody in the hobby measures that directly. Every tester measures a proxy, and each proxy has its own units.

ppt (parts per thousand) is closest to the real thing: grams of salt per kilogram of seawater. It does not change with temperature. The open ocean averages 35 ppt and most coral reefs sit between 34 and 36 ppt.

Specific gravity (SG) is a density ratio: how heavy your water is compared with pure water. Seawater at 35 ppt is 1.0264 times as dense as pure water when both are at 25 C (77 F). That is the whole "35 ppt specific gravity 1.026" pairing. Hydrometers measure density directly, and hobby refractometers print an SG scale next to the ppt one.

Conductivity is what probes and controllers measure. Dissolved salt carries current, and 35 ppt seawater is about 53 mS/cm at 25 C; the probe converts that to ppt or SG for you.

The conversions you actually need, all at 25 C (77 F):

The step to remember is that 0.001 SG is about 1.3 ppt. The whole reef tank salinity 1.025 or 1.026 debate is a disagreement over 1.3 ppt, roughly 4% of the salt in the tank, and real reefs vary by more than that between seasons. It is worth getting right. It is not what kills corals. Sitting at 1.023, or bouncing around, is.

Why the same water reads differently at different temperatures

Water expands as it warms. A litre of 35 ppt seawater weighs about 1.023 kg at 25 C (77 F) and about 1.026 kg at 15.6 C (60 F). The salt has not changed. The density has, and every instrument assumes a temperature.

Hydrometers are calibrated at one reference temperature. Older glass and swing arm types use 60 F (15.6 C); some newer ones, Tropic Marin's for example, use 77 F (25 C). Put a 60 F hydrometer into 77 F tank water and it reads roughly 0.002 to 0.003 low, because the warm water is less dense than what it was designed for. A reading of 1.024 on that instrument is a real 1.026 to 1.027, and the gap grows as the tank gets warmer. Plenty of tanks "running 1.023" in the old days were sitting at natural seawater.

Refractometers measure how far the water bends light, which also shifts with temperature. Most have automatic temperature compensation, but it corrects for the prism, not the drop you just put on it, so give the sample 30 seconds. Most hobby refractometers are also built for sodium chloride brine, and one zeroed on RO water reads real seawater about 1.5 ppt high. Calibrate with a 35 ppt reference solution instead. Which salinity tester can you trust compares all three.

Conductivity probes compensate automatically. Seawater conductivity rises about 2% per degree C, so they have to.

The practical rule: know what temperature your instrument assumes, test at the tank's normal running temperature, and read every target chart as meaning 25 C (77 F) unless it says otherwise. Most reefs run slightly above that, 25 to 27 C (77 to 81 F), and what temperature a reef tank should be is a question of its own.

Why fish-only tanks run lower, and why that is wrong for corals

Older books, and most of the results for "salinity for saltwater tank", put fish-only systems at 1.020 to 1.023. There is a logic to it.

A marine fish is less salty than the sea around it. It loses water through its gills constantly, so it drinks seawater and pumps the salt back out. That costs energy. Lower the salinity and the fish does less osmotic work, and the common parasites, Cryptocaryon (marine ich) in particular, cope worse than the fish do. Taken to the extreme, hyposalinity treatment holds fish at 1.009 (about 12 ppt) for 30 days or more to break the ich life cycle. It works because the fish can regulate and the parasite cannot.

Corals, anemones, shrimp, snails and clams cannot regulate. Their tissue sits at the same salinity as the water, so drop the tank to 1.020 and their tissue drops with it. Hyposalinity kills invertebrates outright, which is why it is only ever done in a bare quarantine tank.

There is a second, quieter problem. Every ion in a salt mix scales with salinity. A salt that gives 420 ppm calcium, 1350 ppm magnesium and 8 dKH at 35 ppt gives about 365 ppm, 1175 ppm and 7 dKH when mixed to 1.023 (30.5 ppt). Nothing is wrong with the salt. The tank is 13% short of everything at once, and you end up dosing calcium and alkalinity to patch a salinity problem. Does your reef salt mix change your parameters has the numbers for the common brands.

So the ideal salinity for a reef tank is 1.025 to 1.026, 33 to 35 ppt. A fish-only tank can run 1.023 to 1.026, though many people keep it at reef salinity so a coral can go in later without a fortnight of correction.

Stability first, then the number

Corals handle a steady 1.025 and a steady 1.026 equally well. What they handle badly is movement.

A salinity swing shows up as retracted polyps, soft coral tissue that looks loose or over inflated, and in SPS as pale or receding tissue at the base. Water changes are the usual cause. A 10% change with new water at 1.028 into a tank at 1.025 nudges the display by 0.0003, which is harmless. A 30% change with the same mismatch is 0.001 in an hour. A 50% emergency change with unmatched water is 0.0015 in an hour, more than a day's safe movement delivered in the time it takes to drain a bucket, and a classic reason corals stay closed after a water change.

So match new water to the tank within 0.001 before it goes in, match the temperature too, and test every mixed batch, because buckets of the same brand vary.

Salinity creep: how the salinity of a reef tank drifts on its own

Three things move salinity without you touching anything, and they pull in different directions.

Evaporation raises it. Only the water leaves; the salt stays. A 200 litre (about 50 gallon) tank with an open top and a cooling fan can lose 1 to 2 litres a day. At 1.5 litres a day that is a rise of about 0.0002 SG per day, small enough to miss, and it compounds. A week without top off takes 35 ppt to 37 ppt, or 1.026 to 1.028. A fortnight away and the tank greets you at 1.029 or higher.

An auto top off (ATO) is the fix, and it also makes salinity the parameter you stop checking, which is where its two failure modes hide.

• Stuck off, or the reservoir empty. Salinity climbs at the evaporation rate above. You usually notice when a coral sulks for no visible reason.
• Stuck on. A float switch fouls and the pump runs until the reservoir is dry. A 20 litre (5 gallon) reservoir dumped into a 200 litre system takes it from 1.026 to about 1.024 in an hour, the fast crash that actually kills things. Size the reservoir so a complete dump cannot move the tank more than about 0.001, or fit a second float or a timed cut off.

Salt creep lowers it. The white crust on the rim and the skimmer lid is salt that has left the water, and so is every cup of skimmate you pour away. The ATO replaces all of it with fresh water, a few tenths of a ppt a month, and the tank that was 1.026 in the autumn is 1.024 by spring.

Water changes reset it toward whatever you mixed. Mix to 1.024 out of habit and weekly 10% changes will walk the tank there in a couple of months.

None of this is dramatic, which is exactly the problem. Test salinity weekly, and every mixed batch before it goes in, even with an ATO. The reef tank testing schedule has it at that cadence.

Correcting it: no faster than 0.001 SG per day

Whichever way it drifted, the rule is the same: move salinity no faster than about 0.001 SG (1.3 ppt) per day, and slower still if you keep SPS. A 0.003 jump in an afternoon is the water change swing above, delivered on purpose. Slow drip, never a jump.

If the tank is too low, the safest fix for a small gap is to do nothing: pause the ATO and let evaporation raise it at about 0.0002 SG a day while you watch the sump level. That closes 0.001 in about five days with no chance of overshoot.

For a bigger gap you need to add salt. Never put dry salt mix into the display or the sump; undissolved salt is concentrated brine wherever it lands, and it burns whatever it lands on. You need about 1.3 g of salt mix per litre of system water for every 0.001 SG, so roughly 260 g for a 200 litre (50 gallon) tank, or 5 g per gallon. Dissolve it completely in 5 to 10 litres of tank water, then drip that brine into a high flow area of the sump over several hours, never straight onto a coral. Test the next day before you add more; the reading, not the arithmetic, decides the second day's amount.

If the tank is too high, swap saltwater for RO/DI. Removing about 3.5% to 4% of system volume, 7.5 litres (2 gallons) on a 200 litre tank, and letting the ATO refill it with fresh water drops salinity by roughly 0.001. Test the next day, repeat. Never pour RO straight in, and never near a coral.

Fish tolerate a move downward better than upward, and you will read that lowering can go faster. Corals get no such concession, so in a reef 0.001 SG per day applies in both directions. A tank that came back from holiday at 1.031 takes five days to reach 1.026. That feels slow. It is the right speed.

Write the number down

Salinity is the parameter people test least often and remember least accurately. "It is always 1.026" usually means "it was 1.026 the last time I looked, in May." Writing the reading down turns a pile of them into a trend, and the trend shows the ATO losing to evaporation months before a coral has to tell you. Salinity is one row on the reef tank parameters chart, and the tracker below has a column for it.

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.

Salinity creep is slow enough that a notebook full of "1.026, fine" entries hides it until it is not fine. Recif is the reef companion app we built to make that visible: it logs every test, charts each parameter against your own target ranges so a slow slide out of the band shows as a slope instead of a surprise, works out consumption trends, reminds you when salinity has gone too long untested, and rolls the whole tank into one Stability Score. It is available on both the App Store and Google Play.