For a community tropical tank: ammonia 0, nitrite 0, nitrate under 40 ppm, pH 6.5–7.5, KH 3–8 dKH, GH 4–12 dGH, 23–26 °C (73–79 °F). Other tank types are in the second table. And the one thing no other chart says: what a given ammonia reading is worth depends on your pH — the same 0.25 ppm is 85× more toxic at pH 8.5 than at 6.5.
The chart
Community tropical numbers — the default if you have not deliberately set up for something else. Print it, tape it inside the cabinet door.
| Parameter | Target | Act when | Why it matters |
|---|---|---|---|
| Ammonia (total) | 0 ppm | any reading at all | What your kit measures is NH₃ + NH₄⁺ together. How much of that is the toxic form depends on pH — see the ammonia section. Zero is the only reading that is safe at every pH. |
| Nitrite | 0 ppm | any reading; urgent over 0.5 ppm | Blocks the blood from carrying oxygen. In a cycled tank any nitrite means the filter is not keeping up — overfeeding, a dead fish, or media rinsed under the tap. |
| Nitrate | under 40 ppm | over 40 ppm | The end of the cycle and the only part you remove by hand. Its rate of climb is also the most honest measure of how heavily stocked you are. |
| pH | 6.5–7.5 | a moving reading, more than an unusual one | The parameter people worry about most and need to act on least. Almost every community fish sold is captive-bred and tolerant; what they do not tolerate is a number that swings. |
| KH (carbonate hardness) | 3–8 dKH | under 3 dKH | The buffer that stops pH moving. Nitrification consumes it continuously, so it falls on its own — this is the test nobody runs until something has already gone wrong. |
| GH (general hardness) | 4–12 dGH | outside the range for your species | Dissolved calcium and magnesium. Livebearers and shrimp genuinely need it; soft-water species genuinely need it low. Most other community fish are indifferent. |
| Temperature | 23–26 °C (73–79 °F) | drifting, or swinging day to night | Sets metabolic rate, oxygen capacity and how fast everything else happens. A steady wrong-ish temperature beats a correct one that moves. |
| Tank type | pH | GH | KH | Temp °C | °F | NO₃ max |
|---|---|---|---|---|---|---|
| Community tropicalMost tetras, rasboras, corydoras, danios, gouramis | 6.5–7.5 | 4–12 | 3–8 | 23–26 | 73–79 | 40 |
| PlantedCommunity fish in a plant-first tank, with or without CO₂ | 6.2–7.4 | 3–8 | 2–6 | 22–26 | 72–79 | 20 |
| Soft water / blackwaterDiscus, cardinal tetras, apistogramma, chocolate gourami | 5.5–6.8 | 1–5 | 0–3 | 27–30 | 81–86 | 10 |
| Rift lake cichlidMalawi mbuna, peacocks, Tanganyikan shell dwellers | 7.8–8.6 | 10–20 | 8–18 | 24–27 | 75–81 | 40 |
| LivebearerGuppies, mollies, platies, swordtails | 7.2–8.2 | 10–20 | 5–15 | 24–27 | 75–81 | 40 |
| Neocaridina shrimpCherry, blue dream, yellow — the hardy ones | 6.8–7.8 | 6–10 | 2–6 | 20–24 | 68–75 | 20 |
| Goldfish / coldwaterFancy goldfish, white cloud mountain minnows | 7.0–8.4 | 8–20 | 5–15 | 18–22 | 64–72 | 40 |
Hobby-consensus ranges, cross-checked against the water these fish come from; nitrate ceilings from this site's water-change engine. GH and KH in German degrees. Ammonia and nitrite are 0 for every row. A total ammonia reading is far more dangerous in a hard alkaline tank than a soft acidic one — at pH 8.5 even 0.16 ppm reaches the long-term exposure limit.
Targets by tank type, and why they differ
The second table above is not seven opinions — it is roughly the water these fish evolved in. Rift lake cichlids come out of hard alkaline lakes; blackwater tetras come out of soft acidic water stained with tannins. Where a captive-bred fish is many generations removed from either, the range widens and stability starts to matter more than the target:
- Community tropical. The default. Almost anything sold as a community fish is captive-bred and tolerant of a wide range — held steady.
- Planted. Injected CO₂ pulls pH down by a point or more during the photoperiod; that swing is normal and KH is what keeps it orderly.
- Soft water / blackwater. Low KH means pH moves easily — this is the one setup where you check buffering rather than assume it. Discus want the top of the temperature band.
- Rift lake cichlid. The one common setup where high pH is correct — and the one where an ammonia reading is most dangerous, because almost all of it is in the toxic form.
- Livebearer. Hard water is not optional for these — soft-water guppy tanks are behind a great deal of unexplained livebearer mortality.
- Neocaridina shrimp. GH is the parameter that matters here — they build shells from it, and molting failures trace back to it more often than anything else. Copper is lethal: check every medication and plant fertilizer.
- Goldfish / coldwater. No heater needed in most homes. Goldfish produce far more waste than their length suggests, so the nitrate ceiling is the binding one.
Why "ammonia: 0" is right advice for the wrong reason
Every chart in this hobby prints a zero next to ammonia and moves on. The zero is good advice. The reasoning behind it is missing, and the missing part is what tells you whether a reading is an inconvenience or an emergency.
Your test kit measures total ammonia. That total is a mixture of two things: NH₃, un-ionized ammonia, which crosses the gill membrane and is the toxic one, and NH₄⁺, ammonium, which largely does not and is far less harmful. Which way the mixture sits is set by pH, and it moves by about an order of magnitude per pH point:
| pH | Of the total, this much is toxic NH₃ | A 0.25 ppm reading, as NH₃-N | Total ammonia at the exposure limit |
|---|---|---|---|
| 6.0 | 0.06 % | 0.0001 mg/L | 42.91 ppm |
| 6.5 | 0.18 % | 0.0004 mg/L | 13.59 ppm |
| 7.0 | 0.56 % | 0.0012 mg/L | 4.31 ppm |
| 7.5 | 1.8 % | 0.0036 mg/L | 1.38 ppm |
| 8.0 | 5.4 % | 0.011 mg/L | 0.45 ppm |
| 8.5 | 15.2 % | 0.0313 mg/L | 0.16 ppm |
Computed by this site's engine at 25 °C from the ammonia dissociation constant (Emerson et al. 1975): pKa = 0.09018 + 2729.92/T, fraction NH₃ = 1/(1 + 10^(pKa − pH)). The last column is the total-ammonia reading at which un-ionized NH₃ reaches the 0.02 mg/L NH₃-N long-term exposure guideline used in aquaculture.
Read the last two columns together and the chart's zero stops being dogma. At pH 6.5, a 0.25 ppm reading is 0.0004 mg/L of actual toxin — a rounding error, and a kit could read 13.6 ppm before you crossed the long-term limit. At pH 8.5, the same 0.25 ppm is 0.031 mg/L — already over the 0.02 mg/L long-term limit — your whole budget is 0.16 ppm, and 0.4 ppm on the same kit would put you into acute harm.
Across the pH range this very chart covers, one ammonia number spans a factor of 85 in actual toxicity. Temperature adds to it: the same water at 30 °C holds 2× as much of the toxic form as at 20 °C. This is why an ammonia reading in a Rift lake tank is a drop-everything event and the identical reading in a blackwater tank is something to watch — and why nobody should be reassured by a chart that says only "0".
First: if you get an ammonia reading, do not chase it by raising pH — you would be converting harmless ammonium into the toxic form. Change water instead. Second: this is the mechanism behind the advice never to do a large water change into a tank whose pH has crashed while ammonia is present. The fresh water brings the pH back up, and the ammonia that was sitting there safely as ammonium converts as it rises. Fix the ammonia first, in small changes, then the pH. It is also why a shipped bag wants less acclimation, not more.
Which of these actually matter
A chart gives every row equal weight, and they do not have equal weight. Honestly ranked:
- Ammonia and nitrite. Non-negotiable, and the only two that do damage in hours. Both zero, always, and the nitrogen cycle is how you get there.
- Temperature. Cheap to control, easy to get wrong, and it multiplies everything else — including how much of your ammonia is in the toxic form.
- Nitrate. Slow, cumulative, and entirely in your hands. Also the number that tells you whether you are overstocked, which is a more useful question than any inch-per-gallon rule.
- KH. Under-watched. It is not a target so much as a floor: below about 3 dKH your pH is no longer held by anything.
- GH. Matters intensely for shrimp and livebearers, barely at all for most community fish.
- pH. Last, deliberately. It is the most tested, most worried-about and least often actionable number on the chart.
Stable beats perfect
The single most common way people harm fish with a parameters chart is by treating it as a set of targets to hit rather than a range to stay inside. A tank sitting contentedly at pH 8.0 does not need correcting to 7.2, and correcting it is how fish die — not from the pH, from the journey.
"pH down" products are the clearest case. They work by overwhelming your KH, the buffer that was holding the pH steady. Once the KH is spent the pH does not sit at the new number; it swings, and then it rebounds when the next water change brings fresh carbonate in. The tank ends up moving between two numbers instead of resting at one, which is worse than either.
If you genuinely need different water — and soft-water species and shrimp are genuine cases — change the source water, not the tank. RO water remineralized to a chosen GH and KH gives you a number you can reproduce every week, which is the part that matters. The one exception to "leave pH alone" is a pH that has fallen because KH ran out: that is not a preference, it is a tank running out of buffer, and it is the same arithmetic the cycling calculator works out for a tank being cycled.
How often to test
| Tank | Test | How often |
|---|---|---|
| Cycling | Ammonia, nitrite | Daily; KH weekly |
| Newly stocked | Ammonia, nitrite, nitrate | Every 2–3 days for a month |
| Established | Nitrate | Monthly — or weekly, to track stocking |
| Established | KH, GH, pH | Every month or two |
| Anything changed | Ammonia, nitrite | Next day, and two days later |
"Anything changed" means new fish, a filter clean, a dead fish, a medication, a power cut, or a holiday with someone else feeding.
Testing an established, stable tank every day is a habit worth breaking. It costs reagent, it produces a stream of identical numbers, and it trains you to react to the noise in your own test kit. What earns its keep is a reading right after a water change and another before the next one, because the difference between them is the only number on this page that describes your particular tank rather than tanks in general.
When a number is off
- Ammonia or nitrite above zero, fish in. Water change now, and again tomorrow. Stop feeding for a day or two. Check the free NH₃ table above against your pH before deciding how alarmed to be — and do not add an ammonia-binding conditioner in place of changing water, or you lose the ability to read your own tests.
- Nitrate over the ceiling. Arithmetic, not a judgment — the water change calculator sizes it from your reading and your tap water.
- KH under 3 dKH. A little baking soda, or a water change if your tap is harder than your tank. Fix this before you worry about the pH it is causing.
- pH outside the range but steady. Usually leave it, and check it is actually outside the range for your species rather than for the community default.
- pH moving. This one is real. Test KH first — a drifting pH is nearly always a buffer that has run out.
- Temperature drifting. Suspect the heater before the thermometer, and never trust a stick-on strip to two degrees.
What you'll be told, and what's true
- "You need to get the pH to 7.0." There is nothing special about 7. It is the neutral point of water, not a target for fish, and almost every community species in the trade is content anywhere from about 6.5 to 8.0 as long as the number holds still.
- "Any ammonia reading is an emergency." It depends entirely on your pH, which is the point of the table above. At pH 6.5, 0.25 ppm is 0.0004 mg/L of toxin and you have time to act calmly. At pH 8.5 the same reading is 85× that, over the long-term limit, and a reading of 0.4 ppm would be doing acute damage. Both need fixing; only one is an emergency.
- "pH-down products fix a high pH." They spend your KH to move the number, and once the buffer is gone the pH swings instead of settling. If you need different water, make it outside the tank and put it in gradually.
- "Nitrate is harmless." Far less harmful than the other two — by a factor of roughly a hundred — which is not the same as harmless. Chronic exposure suppresses growth and breeding, and a high reading is usually telling you something about maintenance rather than about nitrate itself.
Common questions
What is the most important water parameter?
Ammonia and nitrite, and it is not close — both should read zero, and both do real damage within hours. After that, temperature stability, then nitrate. pH is the parameter people worry about most and the one that is least often the actual problem.
My pH is 8.2 — do I need to lower it?
Almost certainly not. Nearly every community fish in the trade is captive-bred and settles happily anywhere from about 6.5 to 8.0, and a stable 8.2 is better for them than a 7.0 you have to keep correcting. The exceptions are real but specific: soft-water species like discus and cardinal tetras, and shrimp. If you are keeping those, change the source water rather than the tank water.
How often should I test an established tank?
Nitrate monthly, or weekly if you want it to tell you how heavily stocked you are. KH every month or two, because it falls on its own. Ammonia and nitrite only when something has changed — new fish, a filter clean, a death, a medication — or when a fish looks wrong. Daily testing belongs to a cycling tank, not a running one.
Are test strips good enough?
For GH, KH and a rough pH, yes — they are quick and the trend is what you want anyway. For ammonia and nitrite, no. Those two matter most in the range between "zero" and "nearly zero", which is exactly where strips cannot separate one color from the next. Keep strips for the routine and a liquid kit for the nitrogen trio.