The short answer
The standard freshwater dose is 1–3 g of salt per litre of actual tank water, and it treats far fewer things than its reputation suggests. Salt is genuinely excellent at one job, useful at two more, and completely useless at most of what it gets recommended for.
The single most important fact about it is this: salt does not evaporate and the filter does not remove it. It leaves only in the water you physically take out. Every top-up you make to replace evaporation adds fresh water without adding salt — but every dose you add on top of an existing dose stacks. People who "add a bit of salt" at each water change without doing the arithmetic are running a slowly rising concentration that nobody ever measured.
The dose chart
All figures are per litre of real water volume, which is not the number printed on the tank. See the volume note further down.
| Concentration | Per 100 L | Used for | Duration |
|---|---|---|---|
| 1 g/L (0.1%) | 100 g | Nitrite protection during a cycling crash; mild stress relief after transport | Days to a couple of weeks |
| 2 g/L (0.2%) | 200 g | General supportive treatment for tolerant species | Up to about two weeks |
| 3 g/L (0.3%) | 300 g | Upper end for external protozoa on salt-tolerant fish | Under ten days, then dilute out |
| 10–20 g/L (1–2%) | Separate container only | Short dip for external parasites, in a bucket — never the display tank | 5–10 minutes, supervised |
For reference, a level tablespoon of granulated salt is roughly 15 g, but the variation between salt grades is large enough that a kitchen scale is worth more than any spoon-based rule.
The one use with strong evidence: nitrite
This is the reason to keep salt in the cupboard.
Nitrite kills fish by entering the bloodstream through the gills and binding to haemoglobin, producing methaemoglobin, which cannot carry oxygen. The fish then suffocates in fully oxygenated water. Chloride ions compete with nitrite for the same uptake pathway at the gill, so raising chloride in the water sharply reduces how much nitrite gets in.
Practically: if your nitrite is above zero and you cannot fix it fast enough with water changes alone, 1 g/L of salt buys the fish protection while the bacteria colony catches up. It does not remove the nitrite and it does not speed up cycling. It protects the fish from a poison that is still present.
This is the situation to remember it for: a tank that is still cycling, a filter that was cleaned too thoroughly, or a recovery after an extended power cut. The cycling process, day by day.
The two reasonable uses
Osmotic relief. A freshwater fish is saltier than the water around it, so it constantly leaks salts out and takes water in, and it spends energy pumping the excess back out. Raising the dissolved salt slightly narrows that gradient and reduces the metabolic cost. For a fish recovering from transport, injury or a difficult water change, that saved energy is real. This is where the traditional 1 g/L "tonic" dose comes from.
Some external protozoa and slime irritations. At 2–3 g/L, salt is hostile to a number of single-celled external parasites and can help a fish shed excess slime coat. This works on the free-swimming stages of some organisms — not on anything encysted, and not on anything inside the fish.
What salt does not do
This list is longer than the one above, and this is where most of the money and most of the dead fish go.
- It does not reliably cure ich. The free-swimming stage is vulnerable; the cyst stage on the fish and the reproductive stage in the substrate are not, at any dose a fish would survive. Salt at 2–3 g/L can support a temperature-based approach, but on its own it usually produces a cycle of apparent improvement followed by reinfection. The full treatment for ich, fin rot and fungus.
- It does not touch internal bacterial infections. Dropsy, internal ulcers, septicaemia — salt in the water does not reach them.
- It does not remove ammonia, and it does not reduce ammonia toxicity. Only water changes and a functioning filter do that.
- It does not raise hardness in any meaningful way. This is a persistent confusion. Sodium chloride adds sodium and chloride; general hardness is calcium and magnesium. Adding salt to a rift lake cichlid tank does not make it a rift lake. What African cichlids actually need from UAE tap water.
- It does not make up for poor water quality. If the underlying number is wrong, salt masks a symptom while the cause continues.
- It does not evaporate. Worth saying twice.
What salt harms
Before dosing the display tank, check the guest list. Some of the most popular fish in this hobby are the least tolerant.
| Inhabitant | Tolerance |
|---|---|
| Live plants | Poor. Most stem plants and mosses show damage well below 2 g/L, and many will melt. Treat plants and salt as incompatible. |
| Shrimp | Poor at treatment doses. Move them or treat elsewhere. |
| Snails | Poor. Nerites and mystery snails react badly to sustained salt. |
| Corydoras, loaches, plecos and other scaleless or thin-scaled fish | Low. Some keepers use half doses successfully; many do not. Not the tank to experiment in. |
| Tetras and most soft-water species | Low to moderate. Tolerable briefly at 1 g/L; unhappy above it. |
| Livebearers, goldfish, most cichlids | Good. These handle 2–3 g/L without difficulty. |
| Filter bacteria | Fine at these doses. Salt is not a filter killer, which is one of its advantages over most medications. |
Note that a planted community tank with corydoras and shrimp — an extremely common setup — fails on three rows of that table at once. In that tank, salt belongs in a separate container or not at all. More on corydoras and the bottom layer.
The volume mistake
Dosing is calculated on water, and a tank never contains the volume printed on the box.
Substrate, rock, wood, equipment and the empty space between the waterline and the rim typically account for 10–20% of the stated capacity. A tank sold as 100 litres commonly holds 80–85 litres of actual water. Dose it as 100 and you have overdosed by roughly a fifth.
The reliable method: measure the internal length, width and water depth in centimetres, multiply them, divide by 1,000. That gives litres of water above the substrate. Subtract a rough allowance for hardscape. Write the number on a label and stick it to the cabinet, because you will need it for every dose you ever calculate. Dimensions, litres and filled weight.
How to dose without shocking the fish
- Work out real water volume as above.
- Weigh the salt. Do not use spoons for treatment doses.
- Dissolve it fully in a jug of tank water first. Undissolved crystals landing on a resting fish or a shrimp cause chemical burns.
- Add it in thirds, spaced two to three hours apart. A sudden change in salinity is its own stressor — which defeats the point if you are dosing to reduce stress.
- When topping up evaporation, add plain dechlorinated water only. No extra salt. Evaporation removes water and leaves the salt behind, so the concentration is already rising on its own.
- To remove it, do plain water changes. Each 25% change removes roughly 25% of the salt. Three or four changes over a fortnight brings a treatment dose back to negligible.
Which salt
Plain sodium chloride with nothing else in it. That means no anti-caking agents, no added minerals, no flavourings. Iodine at the levels found in iodised table salt is not the real risk; the additives used to keep table salt free-flowing are the reason to avoid it when a clean alternative is available.
What it is not: marine or reef salt, which is a full synthetic seawater mix containing calcium, magnesium, carbonates and trace elements, and which will change your water chemistry in ways you did not intend. The two are not interchangeable despite both being called salt.
A note on UAE tap water
Local supply is hard and alkaline, commonly around pH 7.5–8.2, and much of it originates from desalination. That gives people two false impressions worth correcting.
The first is that the water is "already salty", so salt is redundant. Desalinated water is not salty — removing salt is the entire point of the process — and the hardness in local supply is largely calcium and magnesium from post-treatment remineralisation, not sodium chloride.
The second is that because the water is hard, soft-water fish will benefit from salt. They will not. Salt does not soften water and it does not lower pH; a soft-water species in hard water stays mismatched. What is actually in Dubai tap water.
Common questions
Should I add salt to every water change as a preventative? No. Continuous low-dose salting keeps salt-tolerant organisms selected for, provides no benefit to a healthy fish in clean water, and makes the tank incompatible with plants, shrimp and several popular bottom-dwellers. Salt is a treatment, not a maintenance additive.
How long can I leave a treatment dose in? Ten days to a fortnight at 2–3 g/L, then dilute it out with plain water changes. At the 1 g/L nitrite-protection dose, until nitrite has read zero for several consecutive days.
Can I use salt and a medication together? Sometimes, but check the specific product — several treatments interact badly with raised salinity, and some become more toxic. Combining two things you do not fully understand on an already sick fish is how tanks crash.
My fish looks worse an hour after I added salt. Either the dose went in too fast, it went in undissolved, or the species is not salt-tolerant. Do a 30–50% water change with dechlorinated, temperature-matched water and reassess before adding anything else.
The shop told me salt cures everything. It cures very little. It protects against nitrite, it reduces osmotic workload, and it inconveniences some external protozoa. Everything beyond that is folklore, and treating a water-quality problem with salt lets the actual cause run for another week. Start from the diagnostic order instead.