You dose alkalinity. It reads 7.4 dKH the next morning. You dose more. Still 7.4. You check your doser, you check your test kit, you buy a second test kit, and you start quietly wondering whether the tank is haunted. Somewhere around week three of this, someone on a forum asks you what your magnesium is, and you realise you haven't tested it since the tank was set up.
Quick answer: Magnesium is what stops calcium and carbonate from crashing out of your water as solid limestone. It sits at roughly 1285 ppm in natural seawater and should be kept somewhere in the 1250-1400 ppm range in a reef tank. When magnesium drops much below that, calcium and alkalinity become nearly impossible to hold you can pour dose into the tank and watch the numbers refuse to move.
What does magnesium actually do in a reef tank?
Here's the thing nobody explains properly when they hand you a magnesium test kit: magnesium isn't really a nutrient your corals are hungry for. It's a chemical bodyguard for the two parameters you do care about.
Reef tank water is supersaturated with calcium and carbonate. Chemically speaking, it shouldn't be. Those two ions want to find each other and precipitate out as calcium carbonate the same stuff your rock, your coral skeletons, and the crust on your heater are made of. Left alone, that reaction would strip your tank of calcium and alkalinity within days.
Magnesium is what prevents it. Magnesium ions are similar enough to calcium that they slot into the surface of a forming calcium carbonate crystal but different enough that once they're there, the crystal stops looking like calcium carbonate to the next ion drifting past. The growing surface gets clogged, the crystal stalls, and precipitation grinds to a halt. Randy Holmes-Farley describes it as magnesium effectively poisoning the crystal surface, and it's the single most useful mental image in reef chemistry.
So magnesium doesn't build coral skeletons. It keeps the ingredients for coral skeletons dissolved in your water long enough for the coral to use them.
What should my magnesium level be?
Natural seawater sits at about 1285 ppm, and that's the honest target. In practice there's a comfortable band around it, and stability inside that band matters far more than hitting an exact number.
| Level | What it means |
|---|---|
| Below 1200 ppm | Danger zone expect alkalinity and calcium to become unstable and resist dosing |
| 1250-1400 ppm | The working range for a reef tank; most reefers aim for 1300-1350 |
| ~1285 ppm | Natural seawater, and the number to steer toward if you want one |
| Above 1500 ppm | Usually harmless but pointless, and expensive to maintain |
Worth noticing: magnesium sits at roughly three times the concentration of calcium in seawater (about 1285 ppm versus about 412 ppm). If you've ever wondered why a magnesium supplement bottle empties so much faster than you expected when you're correcting a low tank, that ratio is why. There is a lot of water to fill.
Why does low magnesium make alkalinity impossible to hold?
Because you've removed the bodyguard. With magnesium down around 1100 ppm, abiotic precipitation the non-biological kind, nothing to do with your corals starts running in the background of your tank continuously.
Every millilitre of alkalinity you dose gets partially consumed by that reaction instead of going where you wanted it. So does your calcium. The frustrating part is that the two numbers fall together, in a roughly 1:1 chemical relationship, which is a useful diagnostic: if your alkalinity and calcium are both drifting down despite steady dosing, and you can't find a coral load big enough to explain it, test magnesium before you change anything else.
You'll sometimes also see it as a fine white dust on heaters, pump impellers, and powerhead housings. That's your alkalinity, precipitated onto the warmest and most turbulent surfaces in the tank.
What actually consumes magnesium?
Less than you'd think, and not what you'd expect. Corals do incorporate a little magnesium into their skeletons, but the bigger consumer in most tanks is coralline algae, which builds itself out of high-magnesium calcite. A tank with purple encrusting every wall and powerhead is pulling magnesium down noticeably faster than a bare one.
Consumption rates vary enormously by tank anywhere from almost nothing to around 5-8 ppm a day in a heavily loaded system. That range is too wide to guess at, which is exactly why this is a parameter worth logging rather than assuming. Two or three tests spaced a week apart will tell you your own tank's drawdown far more reliably than any general rule.
Do I even need to dose magnesium?
Quite possibly not. This is the honest answer that magnesium supplement manufacturers are not incentivised to give you.
Most commercial salt mixes come out of the bucket at or slightly above natural seawater magnesium. If you're doing regular water changes on a tank with modest coralline and coral demand, those water changes may replace magnesium as fast as you consume it, and your level will just sit there for months. Plenty of reefers genuinely never dose it.
You find out which camp you're in by testing monthly is plenty for a stable tank and only reaching for a supplement when the number actually moves.
How do you raise magnesium without causing a different problem?
Slowly, and with the right stuff. Two rules matter here.
First, don't raise magnesium by more than about 50 ppm in a day. Rapid swings in a major ion aren't worth the risk to livestock, and there's no prize for fixing it by Tuesday.
Second, pay attention to what the supplement is made of. Magnesium supplements are built from magnesium chloride and magnesium sulfate, and dosing pure magnesium chloride in large volumes will slowly skew your tank's chloride-to-sulfate ratio away from seawater. Commercial blends and the common DIY recipes both mix the two a roughly 5:2 ratio of chloride to sulfate by volume is the usual maintenance blend. Buy a blended product or follow a published recipe rather than improvising with one salt.
The one thing magnesium won't fix
You'll come across the bryopsis trick raising magnesium to 1600-1800 ppm with a specific brand of supplement to kill off bryopsis algae. It's real in the sense that reefers have genuinely had it work, but the effect is generally attributed to an impurity in one particular product rather than to magnesium itself, and results are inconsistent. Treat it as a last-resort experiment, not a chemistry principle.
Otherwise: test it, log it, and let it be boring. A magnesium level that hasn't moved in six months isn't a parameter you're neglecting. It's one that's doing its job.