How to Cycle a Fish Tank Properly (Fishless, With Numbers)

You have a tank sitting in the corner, filled, heater on, filter humming, and nothing in it. Someone at the shop told you to wait a week. Someone on a forum said two months. Your partner keeps asking why there is a bucket in the hallway.

That gap between setup and fish is the part nobody explains well, and it is the biggest reason beginner fish die in the first month. So if you are searching how to cycle a fish tank, the honest version is this: you are not cleaning the water, you are farming bacteria, and bacteria grow on their own schedule rather than yours.

The process is measurable, which is the good news. You can watch it happen in test tube colours, tell exactly which stage you are at, and prove it is finished rather than guessing. That is what a fishless cycle gives you and what waiting a vague number of weeks does not.

What follows has numbers attached: how much ammonia to add, what the readings should look like on which days, what genuinely speeds things up, what stalls them, and how to run a fish-in cycle as gently as possible if you already bought the fish.

The short answer

Cycling a fish tank means growing two groups of bacteria in your filter: one that converts ammonia to nitrite, and one that converts nitrite to nitrate. You feed them by adding pure ammonia to an empty, heated, running tank.

Dose to about 2 ppm ammonia. Test daily with a liquid kit. Ammonia will hold, then start falling as the first bacteria establish, and nitrite will appear and climb. Nitrite spikes, sometimes off the top of the chart, then collapses as the second group catches up. Redose ammonia to 2 ppm whenever it reads zero, so the colony never starves.

The tank is cycled when a 2 ppm ammonia dose reads zero ammonia and zero nitrite 24 hours later, with nitrate showing. Expect three to six weeks at 78°F (26°C). Then do a large water change to drop the accumulated nitrate, and stock in stages.

What the nitrogen cycle actually is

Fish produce ammonia, mostly across their gills rather than in their waste, and more of it comes from uneaten food and decaying plant matter. Ammonia burns gill tissue at concentrations well under 1 ppm. Nothing in an aquarium removes it except bacteria and water changes.

The first group, ammonia-oxidising bacteria, take ammonia and produce nitrite. Textbooks name Nitrosomonas, and that is a fair shorthand, though freshwater filters host a mix that includes ammonia-oxidising archaea too. Nitrite is not an improvement from the fish’s point of view. It binds to haemoglobin and stops blood carrying oxygen properly, which is why nitrite poisoning is sometimes called brown blood disease.

The second group, nitrite-oxidising bacteria, convert nitrite to nitrate. In aquaria these are Nitrospira rather than the Nitrobacter that older books and older bottled products name. That correction came out of research published in the late 1990s by Tim Hovanec and colleagues, and it matters commercially: a bottle full of the wrong genus will not colonise your filter well. Nitrospira strains that perform both steps alone, called comammox, have since been described and reported in freshwater biofilters, so the tidy two-step diagram is a simplification.

Nitrate is the end product and far less toxic, but it accumulates. Removing it is your job, via water changes, or the plants’ job if you keep enough of them. That handover is covered in our water change guide.

One detail catches people in soft-water areas: nitrification consumes alkalinity. Roughly 7 mg of alkalinity as calcium carbonate disappears per mg of ammonia nitrogen oxidised, a figure borrowed from wastewater engineering. If your tap water sits at 2 dKH or less, a hard-running cycle strips the buffer, the pH drops into the fives, and the bacteria stop. It is the stall nobody diagnoses.

What you need before you start

Not much, and none of it is expensive.

• A liquid test kit covering ammonia, nitrite, nitrate and pH. Strips are fine for a rough pH check and close to useless for nitrite and ammonia at the low end, which is exactly where the decisions live.
• A source of pure ammonia. Either an aquarium ammonium chloride solution sold for cycling, or plain janitorial ammonia with no surfactants, dyes or perfume. Shake the bottle: if it foams, it has detergent in it, and it is not usable.
• A dechlorinator, and the knowledge of whether your supply uses chlorine or chloramine.
• A heater and a thermometer. Aim for 78 to 82°F (26 to 28°C) during the cycle.
• A KH test, ideally. It is the cheapest insurance against a stall you will otherwise spend a fortnight not understanding.
• Somewhere to write daily readings. Phone notes are fine. Memory is not.

A safety note that deserves its own line: never mix ammonia with bleach or any chlorine-based cleaner. The reaction produces chloramine gas. Use it in a ventilated room, away from children and pets, stored in its original labelled bottle. And check your own supplier’s water report rather than assuming, because whether you are dosing against chlorine or chloramine varies by utility and by season.

The fishless cycle, step by step

1. Set the tank up completely. Substrate, hardscape, plants, heater, filter, lid. Fill it and dechlorinate. Do this before you start dosing so nothing disturbs the colony later.
2. Bring the temperature to 78 to 82°F (26 to 28°C) and let it stabilise for a day. Warmth roughly doubles bacterial growth rate compared to a room-temperature tank, and it costs you pennies.
3. Test your starting water. Note pH, KH and nitrate from the tap. Some supplies arrive with 20 to 40 ppm nitrate already, which changes what “finished” looks like later.
4. Calibrate your ammonia dose. Add a small measured amount, run the filter for an hour, then test. Scale from the result. Household ammonia strength is not standardised, so calculating from the label is guesswork and testing is not.
5. Dose to 2 ppm and record the volume you used. Write it on the bottle with a marker. You will use that number a dozen more times.
6. Test ammonia and nitrite daily, same time of day, in decent natural light. Colour cards are hard to read under warm bulbs.
7. Wait for ammonia to start falling. Typically days 7 to 12. Nitrite will appear at the same time, which is your confirmation that the first group has established.
8. Redose to 2 ppm whenever ammonia hits zero. Do not let the colony sit hungry, and do not stack doses on top of an existing reading.
9. Ride out the nitrite spike. It often goes above the top of the chart. Dilute a sample 1:1 with distilled or RO water, test that, and double the result to know where you really are.
10. Watch nitrite fall to zero. This is the slow half. Usually days 18 to 35.
11. Run the confirmation test. Dose 2 ppm, wait 24 hours, test. Zero ammonia and zero nitrite means cycled. Repeat once more the next day before you believe it.
12. Do a large water change, 50 to 80 percent, temperature matched and dechlorinated, to clear the accumulated nitrate. Then add fish within a day or two.

Step 12 is where people get nervous, and the nerves are misplaced. The bacteria live on filter media, substrate and every hard surface. Almost none of them are floating in the water column, so a big change costs you very little colony and removes a lot of nitrate.

What the readings should look like, day by day

The exact days shift with temperature, pH and whether you seeded the filter. The shape does not. This is a typical fishless cycle in a 20 to 40 gallon (75 to 150 litre) tank at 78°F (26°C), pH 7.6, KH 6.

Two things about that table. The nitrite plateau around days 14 to 20 is where most people give up or start fiddling, and fiddling is what extends it. And the nitrate column climbs faster than beginners expect, because every 2 ppm of ammonia you feed the tank ends up as roughly 7 to 9 ppm of nitrate, depending on whether your kit expresses ammonia as total ammonia or as nitrogen.

A worked example: sizing the dose for a 29 gallon

Say Malik has a 29 gallon tank, which the box calls 110 litres. It has an inch and a half of sand, a piece of spiderwood and a couple of rocks, so his real water volume is closer to 25 gallons (95 litres). That distinction matters here, because he is dosing to a concentration.

He has a bottle of clear janitorial ammonia. The label says 10 percent, but he does not trust it, and he is right not to. Instead:

1. He adds 2 mL with a syringe, runs the filter for an hour, and tests. Reading: 1.0 ppm.
2. Simple proportion. 2 mL gives 1.0 ppm in 95 litres, so 2 ppm needs 4 mL.
3. He adds another 2 mL, waits an hour, tests again. Reading: 2.0 ppm. Confirmed.
4. He writes “4 mL = 2 ppm” on masking tape on the bottle.

Twelve days later ammonia reads zero and nitrite reads 4 ppm. He adds his 4 mL. On day 15 nitrite is off the chart, so he takes 2.5 mL of tank water, adds 2.5 mL of distilled water from the supermarket, tests that, and gets 3 ppm. Real value: about 6 ppm. High, but not a reason to intervene.

On day 27 he doses 4 mL at 8pm. At 8pm the next evening: ammonia zero, nitrite zero, nitrate somewhere north of 60 ppm. He repeats it the following night with the same result. Cycled, on day 28.

He then changes 70 percent of the water. 25 gallons at 70 percent is 17.5 gallons (66 litres), matched to 78°F (26°C) at the tap by thermometer, dechlorinated in the bucket, poured in slowly. Post-change nitrate: about 20 ppm. Fish go in the next morning, six of them rather than twenty.

To check that 70 percent figure against your own volume, or work out where a 40 percent change lands your nitrate, the percentage calculator beats mental arithmetic over a bucket. And if your tank is sold in litres while your test kit talks gallons, the unit converter saves a conversion error that would otherwise double your dose.

How long it really takes, and what changes it

Three to six weeks is the honest range for an unseeded fishless cycle in warm water. I have seen people finish in 16 days and I have seen a tank take nine weeks, and both were doing everything correctly.

What actually moves the needle:

Seeded media is the cheat code, and it is not really a cheat. A squeeze of sponge from a running filter, a used cartridge, a cup of established substrate, or a handful of gravel in a mesh bag will carry a live colony straight into your filter. Ask at a local shop or a local club. The risk is that you may also import snails, planaria or a pathogen from a tank you have not seen, so take it from a source you would happily buy fish from.

Plants help in a quiet way. A heavily planted tank with fast growers like Hygrophila polysperma or floating Salvinia strips ammonia directly, since plants take up ammonium preferentially over nitrate. That can make ammonia readings fall early and suggest the bacteria are further along than they are. Worth knowing before you conclude you cycled in nine days.

Bottled bacteria: what I would and would not claim

This is where the hobby gets shouty, so let me be careful.

The concept is sound. Nitrifying bacteria can be cultured, bottled and shipped, and if the bottle contains living organisms of the right genus in sufficient numbers, they will colonise your filter and shorten the wait. Products built around Nitrospira have a better reputation than older formulas built around Nitrobacter, for the reason described earlier: Nitrobacter is a high-nitrite, soil-and-sewage organism that does not compete well in an aquarium filter.

The problem is everything between the culture vessel and your tank. These are living products with a shelf life, sitting on warm shop shelves, in hot delivery vans, in warehouses for months. Refrigerated brands exist for a reason. Two bottles of the same product bought six months apart may behave nothing alike, which is exactly the pattern you see in forum threads: one person swears by a brand, the next had no effect at all, and both are telling the truth about their own bottle.

My position, and it is a position rather than a fact: use one if it is cheap and available, keep it cold, buy the freshest bottle you can, and treat any speed-up as a bonus. Plenty of these labels promise fish-ready water on day one, and I would not stock a tank on that promise without the 24-hour ammonia test backing it up.

There is one situation where I would spend the money without hesitating: you have fish already, you cannot get seeded media, and you are trying to keep them alive through a fish-in cycle. In that case even a partial effect is worth it.

The parts people get wrong

Dosing too much ammonia. The classic. Someone reads an old guide saying 4 or 5 ppm, or panics that 2 ppm looks low, and doses to 8. High free ammonia and high nitrous acid both inhibit the nitrite-oxidising stage, so the cycle sticks at “ammonia clears fine, nitrite never moves” for weeks. The fix is a 50 to 70 percent water change and a return to 2 ppm dosing.

Cleaning the filter mid-cycle is the second one. You are growing a biofilm and then washing it down the sink. Leave the media completely alone until the cycle is done, even if it looks grubby. Grubby is the point.

Inconsistent dechlorination causes a stall that looks like nothing at all. Chlorine and chloramine kill nitrifying bacteria as effectively as they kill anything else, so topping up evaporation with a jug of straight tap water three times during the cycle means periodically bleaching your colony. Dechlorinate every drop.

Then there is the KH crash. Soft water areas, particularly parts of the UK, Ireland and the US Pacific Northwest, often have tap KH of 1 to 3 dKH. The cycle burns through that buffer, pH falls into the low sixes, and nitrification slows to almost nothing. Readings freeze. People assume the bacteria died. They did not, they are just working in acid. Bicarbonate of soda, or crushed coral in a media bag, brings KH back to 4 to 6 dKH and things restart within days.

Trusting strips for nitrite is a quieter failure. Many strips read nitrite in bands so wide that 0 and 0.5 ppm look similar, and some degrade badly once the tub has been opened in a humid bathroom. If you take one piece of kit advice from this post, make it the liquid kit. There is a fuller comparison of test methods in what your test results actually mean.

Last, the nitrate test that reads zero when it should not. Most bottle-based nitrate tests have a reagent that settles into a stubborn precipitate, and the instruction to bang and shake that bottle for a full minute exists for a reason. Skip it and you get a false low, panic, and start over for nothing.

Fish-in cycling, as humanely as it can be done

Sometimes the fish arrived first. Shop advice, a gift tank, a child’s birthday. There is no point moralising about it, so here is how to reduce the damage.

The goal is simple: keep ammonia and nitrite each below 0.25 ppm at all times, while the bacteria establish anyway. That means testing daily without fail, and changing water whenever either reading crosses that line. In practice, a lightly stocked tank in week one often needs a 25 to 50 percent change every day or two.

Feed lightly. Every other day is plenty for most community fish, and a lean fish produces far less ammonia than a well-fed one.

Use a conditioner that detoxifies ammonia rather than a plain dechlorinator. Several products claim to convert free ammonia to a bound, less toxic form for around 24 to 48 hours, which buys time between water changes. Two honest caveats: the ammonia is still there and will still register on most test kits, and the exact chemistry is not fully public, so the hobby argues about how well it works and for how long. It is a bridge, not a solution.

Keep the temperature at the lower end of your species’ comfortable range. Warmer water speeds bacteria up but also raises fish metabolism and holds less dissolved oxygen, and gill-damaged fish need oxygen. Add an air stone.

Expect it to take longer than a fishless cycle, because a few small fish produce far less ammonia than a 2 ppm dose. Four to eight weeks is realistic. Be clear-eyed about the outcome too: fish that survive a fish-in cycle can carry gill scarring you will never see, which shows up later as reduced tolerance to everything else.

Signs your tank is actually cycled

One test settles it. Dose to 2 ppm ammonia, wait 24 hours, test ammonia and nitrite. Both zero, nitrate present, twice in a row on consecutive days. That is it.

Not evidence: clear water, four weeks elapsed, the shop said it looked fine, you added bacteria in a bottle, nitrate appeared once. A tank can show nitrate from tap water alone.

Once you have passed the test, change water heavily and stock in stages. A cycled filter carries a colony sized to whatever it has been fed, so a tank fed 2 ppm a day supports a modest stocking immediately, not a full one. Add a small group, wait two weeks, test, add the next. This is where the Aquarium Stocking Calculator earns its keep: it works out what your real water volume can carry by bioload rather than by the inch-per-gallon myth, and it runs entirely in your browser with nothing uploaded and no account to make.

Saltwater runs on the same nitrogen chemistry but differs in the detail, and dosing there is better served by the Reef Dosing Calculator.

Frequently asked questions

How long does it take to cycle a fish tank?

Three to six weeks is typical for a fishless cycle held around 78°F (26°C) with a pH above 7. Colder water, soft water with low KH, or a pH under 6.5 can push it past two months. Seeded media from an established filter can cut it to under a week, because you are transplanting the bacteria rather than growing them from scratch.

Can I speed up cycling with bottled bacteria?

Sometimes. Products containing Nitrospira-type nitrite oxidisers, kept refrigerated and used well inside their date, have a reasonable track record. Older formulas built on Nitrobacter tend to disappoint. Results genuinely vary between bottles and batches, the hobby argues hard about which brands work, and no product removes the need to test daily until your readings prove the tank is processing ammonia.

How much ammonia should I add to a fishless cycle?

Dose to about 2 ppm and no higher. Going to 4 or 5 ppm does not build more bacteria faster, and concentrations above roughly 5 ppm can inhibit the nitrite-oxidising bacteria you are waiting on, which is the single most common way people stall a cycle. Redose back to 2 ppm each time the reading hits zero.

How do I know when my tank is cycled?

Dose to 2 ppm ammonia, wait 24 hours, and test. Cycled means ammonia reads zero and nitrite reads zero in that window, with nitrate present. Do it twice on consecutive days before you trust it. Clear water, a set number of weeks elapsed, or a nitrate reading on its own are not proof of anything.

Do I need to do a water change before adding fish?

Usually yes. By the end of a fishless cycle nitrate has accumulated, often to 40 to 80 ppm or higher, and pH may have drifted down as alkalinity was consumed. A large change, 50 to 80 percent, with temperature-matched and dechlorinated water brings nitrate down before your fish arrive. The bacteria live in the filter and substrate rather than the water, so you are not undoing the cycle.

What happens if my ammonia stops dropping partway through?

A stall usually means one of four things: ammonia dosed too high, pH dropped below about 6.5, KH ran out so the pH is unstable, or the temperature fell. Test KH and pH first. Do a 50 percent water change to dilute ammonia and nitrite, top the KH back up, hold the temperature near 78°F (26°C), and resume dosing at 2 ppm.

Where to go from here

Set the tank up, get it warm, dose to 2 ppm, and then mostly leave it alone while writing down two numbers a day. That is the entire job. The discipline is in not intervening during the nitrite plateau, which is the fortnight where every instinct tells you to change water, clean the filter, or buy something.

Keep the daily readings somewhere you can look back at, because the shape of your own curve tells you whether a future problem is new or an old one returning. A countdown timer set for 24 hours after each ammonia dose stops the confirmation test drifting, and the days between dates calculator gives an honest count when your memory says six weeks and the calendar says nineteen days.