Your fish give off ammonia around the clock. In a lake that ammonia is diluted into millions of gallons. In a 75-gallon tank it has nowhere to go, so a film of microbes on your filter media, gravel and glass has to convert it, first to nitrite and then to nitrate. When that film is healthy, you never notice it. When it's missing, as in a new tank or after a bad cleaning, ammonia and nitrite climb within days. That conversion is the aquarium nitrogen cycle.
Most hobby guides still name the same two bacteria they did in the 1980s. When researchers sampled real aquarium filters, they mostly found other microbes, and that changes how you should cycle a tank and clean a filter.
How the aquarium nitrogen cycle works
Fish excrete most of their nitrogen as ammonia, mostly across the gills rather than in solid waste (UF/IFAS FA-16), and uneaten food and dead leaves add more as they break down. Ammonia-oxidizing microbes live on that ammonia and give off nitrite, which is less toxic but still interferes with the blood's ability to carry oxygen (SRAC 451). Nitrite-oxidizing bacteria then turn nitrite into nitrate, which is far less toxic and builds up until a water change or plants take it out.
A normal filter removes little or no nitrogen from the water. The cycle turns a very toxic compound into a mildly toxic one, and the water change finishes the job. That's why a cycled tank still needs routine water changes. For what each compound does to fish and safe levels, see our guides to ammonia, nitrite and nitrate.
Who does the work: an update on the names
The names in most aquarium books, Nitrosomonas for ammonia and Nitrobacter for nitrite, come from early work on wastewater and soil. When researchers sampled actual aquariums, they found something different.
- Nitrite is handled mainly by Nitrospira, not Nitrobacter. In a 1998 study of freshwater aquaria, the nitrite oxidizers were Nitrospira. Nitrobacter added from a commercial product did not establish itself in the tanks (Hovanec et al. 1998, Applied and Environmental Microbiology).
- Some Nitrospira do both steps. In 2015, two research groups independently found Nitrospira that oxidize ammonia all the way to nitrate on their own, called "comammox" (Daims et al. 2015, Nature; van Kessel et al. 2015, Nature).
- Ammonia is handled mostly by microbes the old books never mention. In a 2011 survey of 27 freshwater aquarium filters, ammonia-oxidizing archaea (AOA), a separate branch of life from bacteria, outnumbered ammonia-oxidizing bacteria in 23 of them, and the lower the tank's ammonia, the larger the archaea's share (Sauder et al. 2011, PLoS ONE).
That survey came before anyone knew to look for comammox. When the same lab went back in 2024 and tested the filters of 38 freshwater aquariums in homes and pet stores for all three groups, comammox Nitrospira were in every one, and by gene counts they were the most abundant ammonia oxidizer in 30. Archaea were the most abundant in 7, and classic ammonia-oxidizing bacteria were scarce except in the tank with the most ammonia (McKnight & Neufeld 2024, Applied and Environmental Microbiology). Gene counts show who is present, not who is doing the most work. These come from home and pet-store tanks in Ontario, so treat the shares as a snapshot. A freshwater recirculating fish farm showed the same pattern, with its nitrifiers made up mainly of archaea and comammox Nitrospira, Nitrosomonas far outnumbered and no Nitrobacter detected (Bartelme et al. 2017, Frontiers in Microbiology).
The practical lesson goes beyond names. The community that runs a mature tank is adapted to low, steady ammonia. Ammonia-oxidizing bacteria tend to show up where ammonia runs high, and archaea and comammox where it runs low. A new tank cycled on heavy ammonia doses may start with a different mix than the one it settles into. Nobody has measured this in a fishless-cycled tank. In a 2025 study of three fish-in home tanks, the order varied: comammox showed up first in one, while classic ammonia-oxidizing bacteria made up most of the ammonia oxidizers for the first weeks in another. All three ended up dominated by comammox (McKnight et al. 2025).
Heterotrophic bacteria
A second, much larger group of bacteria feeds on organic matter: uneaten food, mucus, fish waste and dead leaves. These heterotrophs don't take part in nitrification directly, but they break organic waste down and release ammonia as they do it. That's part of why overfeeding raises ammonia: any food the fish don't eat becomes ammonia a little later. Even when every bit gets eaten, more food still means more ammonia. Fish excrete more the more they digest, and these heterotrophs break down the extra waste they pass (UF/IFAS FA-16).
Where beneficial bacteria live
Nitrifiers live attached to surfaces, in the slimy biofilm on filter media, gravel, decorations and glass, rather than floating free in the water (SRAC 451). Filter media matters because it packs a lot of surface into a small space with constant water flow carrying ammonia and oxygen past it.
Do water changes remove beneficial bacteria? Very few. Since the colony is on surfaces, draining and refilling water barely touches it. The water you add can hurt it, though. Untreated chlorine or chloramine kills bacteria, so treat new water before it goes in. Our chloramine guide explains why chloraminated water also releases a little ammonia once it's treated.
Can you clean the filter? Yes, gently. Rinse sponges and media in a bucket of old tank water and don't scrub them clean or run them under the tap. Avoid replacing all of the media at once. Swap half, let the new half colonize for a few weeks, then do the other half.
What about a power outage? Nitrifiers need oxygen. Keepers commonly report that media in a stopped canister starts running short of oxygen within a few hours. When the power comes back, the first water to flush out can carry decaying waste. On a long outage, take the media out and keep it in a bucket of tank water with an air stone if you have battery air.
What slows the nitrogen cycle down
- Low pH and alkalinity. Nitrification is an acid-producing process. It uses up alkalinity, and activity drops when pH falls below about 7 (SRAC 451). Very soft water with almost no KH can see its pH slide over time, and the cycle slows with it. Our pH, GH and KH guide covers how to keep enough buffer in soft water.
- Cold water. Nitrifiers work more slowly in cooler water, so a tank cycled at 80°F and run at 72°F needs time to catch up.
- Low oxygen. Clogged filters and stopped flow starve the colony.
- Medications. Some disease treatments, especially antibiotics, can harm the biofilter (UF/IFAS FA-16). Test ammonia and nitrite during and after any treatment.
- Disinfectant in new water. Covered above. Always use a conditioner rated for chloramine.
How long does it take to cycle a tank?
There's no fixed number. University of Florida extension puts it at six to eight weeks for a new biofilter (UF/IFAS FA-16). In a 2025 study that followed three newly set-up home aquariums stocked with fish from the start, ammonia and nitrite dropped to undetectable by week 3 in the two tanks with live plants and by week 8 in the third, which had artificial plants and was dosed with a bottled bacteria product (McKnight et al. 2025). Three lightly stocked tanks is a small sample, and an easier test than clearing a 2 ppm fishless dose in a day. Expect several weeks, and up to the six to eight UF/IFAS gives, without a head start. Warm water, enough KH and seeded media can shorten that. Cool or soft, acidic water can stretch it. Go by the test results, not the calendar or how clear the water looks. A tank is cycled when it converts its ammonia load to nitrate within about a day, with ammonia and nitrite both reading zero.
Fishless cycling, step by step
Fishless cycling is the cleanest way to start a tank, because no fish has to sit in rising ammonia while the colony grows. The doses below are common hobby practice, not a published standard:
- Set up the tank with the filter, heater and treated water running at the temperature you plan to keep.
- Add an ammonia source. Dose pure ammonia (the label should list only ammonia and water, with no surfactants, scents or dyes) or ammonium chloride to about 2 ppm on your test kit. Anything from 2 to 3 ppm is fine, but don't go above 5 ppm. Keepers often report that the bigger 4 to 5 ppm doses in older guides drag out the nitrite stage. That fits wastewater research showing free ammonia, the share of total ammonia that rises with pH and temperature, starts inhibiting nitrite oxidizers at roughly 0.1 to 1 mg/L, a level big doses can reach in harder, higher-pH water (Anthonisen et al. 1976), but nobody has measured it in aquariums. The more common cause of a stalled cycle is pH sliding below about 7 as nitrification uses up KH (see step 5). Fish food works but is messier and harder to control.
- Test every day or two for ammonia, nitrite and nitrate. Ammonia falls first, nitrite spikes, then nitrite falls as nitrate rises.
- Redose ammonia when it reaches zero, but once nitrite shows up, cut the dose to half or less, and skip a dose if nitrite reads above about 5 ppm. Piling ammonia onto high nitrite can stall the nitrite-eating microbes. A day or two without ammonia won't starve the colony.
- Watch pH and KH. Nitrification uses up KH. Every 2 ppm dose converted to nitrate eats roughly 0.65 to 0.8 dKH (about 12 to 14 mg/L as CaCO₃), depending on whether your kit reads ammonia as nitrogen or as total ammonia. In soft water, a few days of redosing can use up all the KH you have, and the cycle slows as pH falls below about 7. Don't let KH drop below about 1 to 2 dKH. UF/IFAS notes nitrifiers stop working effectively below 20 mg/L total alkalinity, about 1.1 dKH (UF/IFAS FA-16). Top it up with a water change or a little baking soda (see our GH vs KH guide).
- Call it done when a measured 2 ppm ammonia dose reads zero ammonia and zero nitrite 24 hours later.
- Do a large water change to bring nitrate down, then add fish within a day or two so the colony doesn't go hungry.
Shortcuts that work, and ones that might
Seeding with established media is the most reliable shortcut. A handful of media or a sponge from a healthy, disease-free tank carries a working colony. Treat any media from another tank as a possible carrier of disease, the same way you'd treat a new fish.
Bottled bacteria are a mixed bag. The two products tested in the 2011 Sauder study contained ammonia-oxidizing bacteria but no archaea (Sauder et al. 2011). In the Hovanec study, Nitrobacter from a commercial additive never took hold (Hovanec et al. 1998). As of 2024, researchers described store-bought supplements as containing mainly ammonia- and nitrite-oxidizing bacteria, rarely archaea, and noted that none had yet been designed around comammox Nitrospira (McKnight & Neufeld 2024). Products that list Nitrospira as the nitrite oxidizer at least match what lives in real tanks, and they may give you a head start. Independent tests of specific products are scarce, though, so treat any bottle as a head start, not a finished cycle, and keep testing. One caution: some of this research has industry ties. The 1998 study came partly from Aquaria Inc., Marineland's parent company at the time. Its lead author later founded a company selling Nitrospira starters, and the Waterloo lab later took a grant with a bacteria-supplement maker. The independent studies agree, but no one has independently tested the products head to head.
What beneficial bacteria don't do
They don't control algae. Nitrifiers change nitrogen's form. They don't remove it. Algae and plants can use ammonium as well as nitrate, so a biofilter neither starves algae nor feeds it extra. What decides algae is how much nitrogen and phosphorus go in versus come out (feeding, stocking, water changes, plants) and how much light the tank gets. Houseplants with their roots in the tank are one way to take nitrogen out (see our guide to houseplants in aquariums).
Nitrifiers don't clear cloudy water either. A milky haze in a new tank is usually a bloom of free-floating heterotrophic bacteria feeding on excess organics. It isn't the nitrifying colony. It typically clears on its own once the excess food and waste are under control. A green haze is different. That's free-floating algae, and it's a light-and-nutrient problem.
A cycled tank still needs water changes, because nitrate keeps building up.
Before new fish arrive
A shipped fish has spent the night in its bag, and it goes into your tank water as soon as it's acclimated. That water should be cycled, temperature-matched and treated for chloramine before the box shows up. Adding several fish at once raises the ammonia load all at once. A filter sized for a few small fish needs a few days to catch up to a big new order. Feed lightly for the first several days and test ammonia and nitrite daily.
Our shipping article covers what happens chemically in the bag and why the first hour after opening it matters.
FAQ
How do I know my tank is cycled?
It converts its normal ammonia load to nitrate within about a day, with ammonia and nitrite both reading zero and nitrate rising over time. In fishless cycling, a measured 2 ppm ammonia dose should read zero ammonia and zero nitrite 24 hours later. Clear water alone doesn't tell you.
Do water changes remove beneficial bacteria?
Hardly any. Nitrifiers live in biofilm on surfaces, not in the water column. Treating the new water for chlorine and chloramine is what protects them.
Should I clean my filter?
Yes, but rinse the media in old tank water, not tap water, and don't replace all of it at once.
Do bottled bacteria products work?
Some may give you a head start. But the two products tested in a 2011 study contained no archaea, and as of 2024 none had been designed around comammox Nitrospira, which were the most abundant ammonia oxidizer in 30 of the 38 freshwater filters in that 2024 survey. Test ammonia and nitrite instead of trusting the label.
Why did my cycled tank spike after I added fish?
The colony was sized for the old load. A big addition of fish, or heavy feeding, can outrun it for a few days. Feed lightly and test daily until it catches up.
Cover photo (cropped): Christine Loew / iNaturalist, CC BY 4.0
