The nitrogen cycle in plain language

The nitrogen cycle is how bacteria turn toxic fish waste into a relatively harmless plant nutrient, running in two steps: ammonia → nitrite → nitrate. One group of bacteria converts ammonia to nitrite, and a second, slower group converts nitrite to nitrate; a tank is only "cycled" once both colonies hold ammonia and nitrite at 0 mg/L. That build-up typically takes four to six weeks, and the nitrate left at the end is comparatively safe, fine up to roughly 25 mg/L in a planted community tank.

Most aquarium books open with the nitrogen cycle and then fail you in one of two ways. Some hand you the chemistry with no sense of why it matters at your tank. Others give you reassuring metaphors and skip the equations that explain when fish actually die. I want to do both here, the way I'd explain it standing next to your tank with a test kit in hand.

So here is the whole thing in one breath. Ammonium turns into nitrite, nitrite turns into nitrate. Each step is run by its own bacteria, each with its own clock and its own ways of breaking. And the reason "cycled" is a word worth caring about comes down to one quiet, nasty fact: the second step is more poisonous than the first and takes longer to switch on. A tank halfway through cycling can be deadlier than one that has not started.

What's actually in the water

Fish, shrimp, snails, leftover food, and rotting plant bits all dump nitrogen into the water in the same starting form: ammonia. Once dissolved, it splits between two versions that keep converting back and forth, the uncharged NH₃ and the protonated NH₄⁺, balanced on an equilibrium that pH controls:

NH₃ + H₂O ⇌ NH₄⁺ + OH⁻

Below pH 7, almost everything sits as NH₄⁺. That form is relatively gentle. Push above pH 8 and the balance swings hard toward NH₃, the uncharged version that slips across fish gill membranes and does the killing.

So the same reading means very different things in different tanks. "0.25 mg/L total ammonia" in a soft, acidic Amazon biotope is close to a non-event. The identical number in a hard, alkaline African cichlid tank can be a death sentence, because so much more of that total is sitting as toxic NH₃. Your test kit can't separate the two for you. It reports the sum, partly because the two forms interconvert during the test itself, so reading it correctly depends on you knowing your pH. The NH₄ reference page lays out the math at a few common pH values.

In a tank with no working biological filter, ammonia climbs in a straight line with the bioload. Nothing is eating it. Once a filter is established, ammonia gets consumed faster than the fish produce it, and your reading sits at zero. That zero is the whole game.

The first conversion

The first step belongs mostly to bacteria in the genus Nitrosomonas, with a hand from a few ammonia-oxidising archaea. They run this:

NH₄⁺ + 1.5 O₂ → NO₂⁻ + H₂O + 2 H⁺

Read that reaction closely and it tells you most of what you need to know about managing it.

It is hungry for oxygen. A colony stuck in an oxygen-starved pocket simply doesn't work. Deep dead substrate, a clogged filter chamber, the stagnant corner behind a big rock: ammonia oxidation stalls in all of them. Good flow through the filter and the occasional poke at the substrate aren't fussiness, they are what keeps the reaction fed.

It makes acid. Two hydrogen ions for every ammonium consumed. One molecule is nothing, but months of a busy filter is a real, slow drain on your carbonate buffer, and a big part of why a low-KH tank drifts toward pH crashes when water changes don't keep topping the buffer up.

Its product is nastier than its input. Nitrite crosses into the blood and binds haemoglobin, leaving it unable to carry oxygen around the body. Hold a sensitive freshwater fish at around 0.5 mg/L and you will usually see stress; push past 1 mg/L and deaths can follow within hours, sooner in soft water with little chloride to compete at the gill.

This is where beginners get caught. The first colony almost always establishes faster than the second, which opens a window where the tank is happily converting ammonia (NH₄ reads zero) while nitrite quietly piles up (NO₂ climbs). Watching ammonia fall to zero, you would swear the tank had turned the corner, when in fact the water is now more toxic than where it started. This is the phase that kills fish during fish-in cycling. The NO₂ page is the one to watch through this stretch, because it is still telling you the truth when NH₄ has gone quiet.

The second conversion

The second step, in freshwater, is mostly the work of Nitrospira. Older books credit Nitrobacter, and you'll still see that name repeated everywhere, but in freshwater it's Nitrospira. The reaction is simple:

NO₂⁻ + 0.5 O₂ → NO₃⁻

This is the step that closes the loop into something harmless, and it is also the slow, temperamental one. Nitrospira grows more slowly than Nitrosomonas, starts from a smaller seed population, and bruises more easily when conditions shift. In a fishless ammonia cycle, this second colony usually trails the first by about two weeks. That two-week lag is the dangerous window above, stretched out long enough to matter.

The payoff is nitrate, which is comparatively benign. Plants take it up directly as a macronutrient, so in a planted tank it's less a waste product than a fertiliser you didn't have to buy. The NO₃ page covers what high and low look like once a tank is stocked.

One more group is worth knowing, even though the textbook two-step model skips it. In 2015, researchers described comammox bacteria (complete ammonia oxidisers), single organisms in the Nitrospira lineage that do both steps inside one cell, ammonia straight through to nitrate. In some freshwater systems they're the dominant nitrifier. None of this changes how you manage your tank. The visible chemistry still runs NH₄ → NO₂ → NO₃ on roughly the same schedule, but it does explain why a cycle occasionally finishes faster than the tidy two-colony story predicts.

Where the bacteria actually live

People say the bacteria live in your filter media, which is true and also half the story. They will colonise anywhere with flow, oxygen, and a surface to grip: filter sponges, the biofilm on rocks and glass, the substrate, plant leaves, even the inside of the heater housing.

The filter wins because of surface area, and the gap is enormous. A canister stuffed with ceramic media or bioballs can pack something like ten square metres of colonisable surface into a one-litre canister. Every other surface in the tank combined might add one more square metre. That ratio is why rinsing a sponge under the tap is so costly: chlorinated water kills the bacteria on it, and you've just deleted most of your colony in a single wash. What's left on the rocks and glass can't shoulder the bioload alone.

It is also why a fistful of mature media from an established tank beats any bottled product as a cycling accelerator. What you are moving across is not a few stray bacteria but a fully colonised surface that took months to build, ready to work the day it goes in.

What "cycled" actually means

A tank is cycled when both colonies are large enough to process the standing bioload faster than it produces ammonia. You confirm that with one test:

1. Dose ammonia to roughly 2 mg/L (or feed an equivalent ammonia source).
2. Wait 24 hours.
3. Test NH₄ and NO₂.
4. Both should read zero.

If NH₄ comes back zero but NO₂ doesn't, the first colony is working and the second hasn't caught up. If both read above zero, neither colony is sized for the bioload yet, and you wait. This is the checkpoint the first 30 days guide builds toward. It's worth repeating on an established tank too, especially before adding a serious new bioload, just to confirm the filter still has the headroom you think it does. If you'd rather see the arc laid out against a calendar, the nitrogen cycle timeline tool sketches the expected shape of each curve so you can tell whether yours is on track or stalled.

How an established cycle falls apart

Mature tanks lose their cycle more often than people expect, and the ways they do it are predictable. Most are maintenance habits picked up from advice that quietly assumed the tank was already stable.

• Rinsing the filter in tap water. Chlorine and chloramine are antibacterial on purpose, that's why they're in your drinking water. One rinse under the tap strips most of the colony. Wash filter media in old tank water siphoned out during a water change instead.
• Antibiotics in the display tank. Anything antibacterial aimed at fish disease (Furan-2, Maracyn, methylene blue at treatment strength) can't tell a pathogen from your filter bacteria. Expect a mini-cycle for about two weeks after treatment ends, and test daily through it. Better to treat sick fish in a separate hospital tank and leave the display's biology alone.
• A substrate deep-clean. A real share of your colony lives in the substrate biofilm. A full gravel-vac that churns the lower layers does two bad things at once: it releases a pulse of trapped ammonia and removes a chunk of the bacteria that would have handled it. Surface vacuuming is fine. Deep digging is not.
• A sudden bioload spike. Drop ten fish into a six-fish tank and you've roughly doubled ammonia production overnight. The colony can't double on the same timetable. Until it catches up over a few days, ammonia ticks up off zero and the tank is briefly under-filtered. Add new stock in fractions instead, maybe a quarter at a time with a week between batches.
• A pH crash. When KH falls far enough that the daily CO₂ swing is no longer buffered, pH can drop below 6.0. That's beneath the range where nitrifying bacteria work well. The cycle doesn't die so much as seize up, and ammonia starts creeping upward while nothing else has obviously changed.

Every one of these comes back to the same blind spot. Your filter is doing far more biological work than its quiet appearance suggests, and on a thinner safety margin than a settled tank lets on. Knock the colony down, whether deliberately by cleaning or medicating or accidentally by letting the chemistry slide, and you have signed up for a few days of recovery. The fix is to expect it rather than to be surprised by it.

Once you stop reading NH₄, NO₂ and NO₃ as three unrelated numbers and start reading them as one arc with a direction, most new-tank panic settles into patience. Patience happens to be Manfred's entire personality, and a cycle rewards it more than almost anything else you can do.