Starting a saltwater aquarium: live rock, dry rock or a deep sand bed

Updated on August 25, 2026 · ~7 min read

A marine tank is started by filling with saltwater at about 35 ppt, adding the rock and letting it mature. With live rock the cycle takes 2-4 weeks, with dry rock 6-10 because it starts from zero. Fish go in once the cycle is confirmed, corals much later — when nitrate and phosphate are low and stable.

In a marine tank the nitrogen cycle is the same as in freshwater — ammonia, nitrite, nitrate, the same bacteria and the same wait. But the start-up is longer and less forgiving, for three reasons worth understanding before you fill the tank: saltwater holds less oxygen at the same temperature, corals tolerate far smaller swings than fish do, and the rock you put in decides on its own how long the whole thing takes.

Live rock, dry rock or inert: what changes

Rock is not decoration: in a marine tank it is the filter. Its porous surface hosts most of the bacterial population — the aerobic part near the surface, the anaerobic part deep inside, where nitrate is reduced to nitrogen gas.

Live rock — taken from another mature aquarium or from the sea. It brings the grown bacteria with it, and with them the hitchhiking fauna: micro-crustaceans, sponges, worms. It is the fastest route (2-4 weeks) and it gives the biologically richest tank. The price: it costs a lot, it can introduce unwanted guests, and rock shipped by courier arrives with some of its organisms dead — the die-off — which releases ammonia for days.

Dry rock — the same rock, dried and lifeless. Cheaper, carries nothing unwanted, and easier to build into an open structure. The cycle, though, starts from zero exactly as in freshwater: 6-10 weeks. It shortens considerably by adding a single piece of live rock, or a handful of sand from a healthy tank, as an inoculum.

Inert or man-made rock — porous ceramic or cement. It behaves like dry rock: no biological risk, slow colonisation, often better weight and shape.

On the bottom, two schools:

• A thin bed (1-2 cm) or bare bottom. Easy to siphon, nothing accumulates, no surprises. Today the more common choice in reef tanks.
• A deep sand bed, 8-10 cm of fine sand. In the deep, oxygen-free zones denitrifying bacteria break nitrate down: a well-built bed keeps nitrate low by itself. The downside is that it accumulates, matures over months and must then be left alone — stirring it years later can release in one go what it has stored. It is a decision to make at the start, because it cannot be converted later.

Salt and salinity

Water is mixed outside the tank, in a bucket with a pump, dissolving synthetic salt in RO water. Not tap water: the nitrate, phosphate and silicate in tap water are the most common cause of algae and trouble in a reef's first months.

The reference is ~35 ppt, which corresponds to a specific gravity of 1.025-1.026. A fish-only tank can sit a little lower, 30-35 ppt. In practical terms, taking RO water to 35 ppt needs roughly 36 g of salt per litre — about 3.6 kg per 100 L, or roughly half a pound per gallon. Let it dissolve completely before testing, and always test at tank temperature: both specific gravity and conductivity shift with heat, and a refractometer calibrated at 20 °C read on water at 26 °C gives the wrong number.

Maturation, week by week

Weeks 1-2. Tank filled, rock in, pumps and skimmer running, lights off or nearly. With shipped live rock the die-off produces a clear ammonia spike: test every two days. With dry rock nothing happens and you have to supply an ammonia source, exactly as in a fishless cycle.

Weeks 2-5. Nitrite appears and falls. Keep testing, add nothing. Lights stay short: at this stage they only feed algae.

Weeks 3-6: the algae stages. They arrive in order, and they are a phase, not a fault. First diatoms, the brown film: they eat the silicate in new water and finish on their own (the reef reference is under 0.1 ppm of silicate, ideally undetectable). Then often cyanobacteria, the slimy red-brown sheet, which points to poor flow and excess nutrients. Finally green filamentous algae. This is when the first "cleaners" go in — a few hermits, a few snails.

Weeks 4-10: cycle confirmed. Ammonia and nitrite at 0 ppm, nitrate present. Water change, then the first fish — a few, one or two at a time.

From here on: corals. Not yet. Corals demand the stability a two-month-old tank does not have: nitrate low (~1-10 ppm) and steady, phosphate 0.02-0.1 ppm, KH 7-9 dKH that does not swing, calcium 400-450 ppm, magnesium 1250-1400 ppm (about three times the calcium). Start with the most tolerant species after the third or fourth month, and wait longer still for SPS.

A marine tank's parameters, and when each starts to matter

Parameter	Reef reference	When it starts to matter
Salinity	~35 ppt (1.025-1.026)	day one
Temperature	24-26 °C (75-79 °F)	day one
Ammonia / nitrite	always 0 ppm	during the cycle
Nitrate	~1-10 ppm	before corals
Phosphate	0.02-0.1 ppm	before corals
KH (alkalinity)	7-9 dKH, stable	with the first corals
Calcium	400-450 ppm	with the first stony corals
Magnesium	1250-1400 ppm	if calcium and KH will not hold
Silicate	under 0.1 ppm	first months (diatoms)

A fast swing in KH does more damage than a slightly off value: correct in small steps, no more than ~1 dKH a day. And if calcium or KH refuse to come into range despite dosing, the first suspect is magnesium, not the dose.

Mistakes that cost a start-up

• Tap water. Even just for top-offs. It is the fastest way to buy yourself months of algae.
• Rushing corals. A coral in a six-week-old tank is a coral you lose, and not from a dosing error: from instability.
• Salt dissolved directly in the tank. Mix outside, always.
• Full lighting during the cycle. It achieves nothing and feeds algae.
• Too much rock. An open structure with room for flow filters better than a solid wall: dead spots are what create problems.
• Measuring salinity with a swing-arm hydrometer. The least reliable instrument on the shelf. A refractometer is cheap and calibrates with distilled water.

How Gurglee does it

Related guides