How Many Fish Can a Tank Hold? The Inch-per-Gallon Rule Is Wrong

Three rules, and the calculator ships all of them

The rules of thumb are old and they are regional. Wikipedia's aquarium article lists three side by side: one inch of adult fish length per US gallon, three centimetres of adult fish per four litres, and one centimetre of adult fish per thirty square centimetres of surface area. They are not translations of each other. One inch per gallon works out at about 1.06 cm per litre; the metric rule most often quoted in Europe is 1 cm per litre, and the cautious version halves it. None of the three is anybody's standard. They are descriptions of what shops and clubs told people, written down.

The tool ships five of them in a dropdown and lets you pick, which is more honest than most calculators — but it also means the tool has no opinion. Running the 20-gallon preset through all five gives capacities of 12.1, 12.2, 16.2, 24.2 and 24.4 inches of fish, and small-fish counts of 6, 6, 8, 12 and 12. That is the same tank, twice over, depending on a dropdown nobody explains. The tool's own footnote says so: stocking rules are rough guidelines that differ by country and school of thought.

Waste follows mass, and mass follows the cube of length

This is where the length rule breaks, and it breaks arithmetically rather than as a matter of opinion. Two fish of the same shape, one twice as long as the other: the big one is twice as long, twice as tall and twice as wide, so it has eight times the volume and eight times the mass. Metabolic rate does not follow mass one for one — it follows mass to about the three-quarter power, which is Kleiber's law, stated for animals generally. Eight to the power of 0.75 is 4.76. The big fish therefore burns, eats and excretes about 4.8 times what the small one does, and a length rule hands it exactly twice the water. It is under-provisioned by a factor of nearly two and a half.

Push it further and the rule collapses outright. Ten 2-inch tetras and one 20-inch fish both read as twenty inches of fish, so both get twenty gallons. But the single fish weighs about a thousand times one tetra, and a thousand to the power of 0.75 is 178 — against 10 for ten tetras. The one big fish loads the filter roughly eighteen times as hard as the school it is supposedly equivalent to. This is why a goldfish in a bowl is the classic disaster and a shoal of neons in the same water is not: the rule that equates them measures the wrong thing.

One honest caveat: Kleiber's three-quarter exponent is stated for animals in general, and the exponent measured for individual fish species varies with temperature, feeding state and how it is measured. Treat 4.8 as the order of the effect, not a specification. The direction is not in doubt — mass rises faster than length, and waste rises with mass — and the direction alone is enough to kill the rule.

The rated volume is not the water in the tank

A tank sold as 20 gallons is 20 gallons of glass box, measured to the rim. Nobody fills to the rim: the water line sits an inch or so below it, because a lid needs clearance and a hang-on filter needs a return path. Substrate takes the bottom two inches. Rocks and wood take more. The stocking tool is unusual in modelling this — it asks for the tank height, the water level and the substrate depth separately — and on its own 20-gallon preset it returns 19.9 gallons gross against 16.2 gallons of water. Nineteen per cent of the tank you paid for is not water, and every stocking rule is a rule about water.

The companion tool, the aquarium volume calculator, does not model it at all. It multiplies length by width by height and stops, so its answer is the glass box brim-full. Two translation keys for a water level exist in all six of its language files and are never rendered on screen — the field was designed and never wired up. On a 100 × 40 × 50 cm tank it answers 200 litres; allow 5 cm of freeboard and 5 cm of gravel and the real figure is 160, twenty per cent lower. Its surface-area card has the same problem in reverse: it reports 22 000 cm², the whole outside of the glass box, when the number a surface rule wants is the water surface, 4 000 cm². Use the volume tool for weight and use the stocking tool for water.

One thing the stocking tool used to get wrong here, and no longer does. It removed the full substrate depth from the water column, as though gravel were solid rock. Gravel is roughly a quarter to two fifths pore space and the pores are full of water, so the real displacement is smaller. It now subtracts sixty-five per cent of the bed depth: on the 20-gallon preset that takes out about 1.6 gallons where it used to take 2.5. The old behaviour erred toward less water and therefore fewer fish, which is the safe direction — but a number that is wrong in a safe direction is still wrong, and it made the tool disagree with the substrate calculator sitting next to it.

Surface area, filtration and the things a volume cannot see

Oxygen enters a tank across the water surface and carbon dioxide leaves the same way. That is why the surface-area rules exist and why they are not simply a metric restatement of the volume rules: a tall, narrow column and a low, wide breeder can hold identical water and offer very different surfaces. Wikipedia puts the aquarists' version plainly — some hold that a deeper aquarium takes no more fish than a shallower one of the same surface area. The tool computes the footprint for every shape it offers and uses it directly for its two surface methods, so switching to one of those is how you make the tool notice the difference between a column and a breeder.

Filtration is the other cap, and no volume rule can see it. A tank does not fail because it ran out of water; it fails because ammonia arrives faster than the bacterial colony in the filter can oxidise it to nitrite and then to nitrate. That colony grows to match the food it is given, which takes weeks, and it collapses within hours if the filter is cleaned in tap water or left unpowered. Two identical tanks with the same fish and different filters are not the same tank. Nothing in a stocking calculator asks what filter you own, which is exactly why the answer it gives is a ceiling and not a plan.

Species behaviour beats arithmetic every time

The arithmetic treats fish as interchangeable units of length, and they are not. A schooling tetra or rasbora is not optional in a group: kept as two or three it hides, colours down and nips, and the fix is more of them, not fewer. So the honest lower bound for a shoaling species is a number — six, often eight or ten — and if the tank cannot hold that number the answer is a different species, not a smaller school. A territorial cichlid or a bottom-dwelling loach measures its world in floor area and lines of sight, so a 24-inch tank and a 48-inch tank with the same water are entirely different propositions to it. And a fish that is 2 inches in the shop and 12 inches at three years old must be counted at 12.

One number in the tool is worth knowing about here. The "approximate small community fish" figure is not a species judgement: it is the capacity in inches divided by two and rounded down, so it silently assumes every fish is exactly 2 inches, about 5 centimetres, long. That is a reasonable stand-in for a neon or a rasbora and useless for anything else, and the assumption is not stated anywhere on the page. The capacity in length is the tool's real output; the fish count is a convenience built on one hidden number.

How to use the number without being misled by it

Run the tank once on the conservative rule and once on the surface rule and treat the smaller of the two as the ceiling. Then subtract for the things the calculator cannot see: an unfiltered or under-filtered tank, a species that grows past 4 inches, anything that produces a lot of waste for its size, and the first three months of a new tank while the filter is still building its colony. Stock in stages rather than all at once, so the bacteria have something to grow on before the full load arrives.

And check the tank height against the water level before you trust the volume. The tool applies no sanity check between the two: type 40 inches of water into a 16-inch tank and it will happily report 47.4 gallons of water in a tank whose gross volume is 19.9, with no warning that the two numbers cannot both be true. It is an easy slip to make when you type a tank's outside height into the water field.

Frequently asked questions

Is the inch-per-gallon rule ever right?

It is roughly defensible in one narrow case: small, slim, well-filtered community fish under about 3 inches, in a tank with a large surface area, stocked gradually. That is the population it was derived from. Outside that case it degrades fast, and it degrades in the dangerous direction — it always overstates capacity for a big or a chunky fish, never understates it. A goldfish, a plecostomus and an oscar all break it badly, and all three are sold to beginners as small fish.

Does the gallon in the rule mean the tank's rating or the water in it?

The water, and this is where most people lose a fifth of their margin without noticing. The rule was always about water; the rating on the box is the glass. The stocking tool works on the water: it takes the height, the water level and the substrate depth separately and returns both figures, so on its 20-gallon preset you see 19.9 gallons gross and 16.2 gallons of water side by side. If your calculator only asks for three dimensions, it is giving you the glass, and you should knock roughly 15 to 25 per cent off before applying any rule.

Why do two of the five methods give double what two others give?

Because two of them are deliberately the cautious halves of the other two. The metric volume rule allows 1 cm of fish per litre; the conservative variant allows 1 cm per 2 litres, exactly half. The tropical surface rule allows 1 cm per 30 cm² of footprint; the coldwater rule allows 1 cm per 60 cm², exactly half, because coldwater fish are usually bigger, need more oxygen per unit of water and grow for years. So the range you see is not disagreement about facts — it is the same rule with a safety factor of two applied or not.

Does a bigger filter let me keep more fish?

Up to a point, and then no. A bigger filter holds more media, so it hosts a bigger bacterial colony and can oxidise more ammonia per hour — that genuinely raises the ceiling. What it does not change is the water surface available for gas exchange, the floor area a territorial fish needs, the nitrate that accumulates between water changes, or the fact that a 12-inch fish is 12 inches long in whatever water it is swimming in. Filtration removes one constraint out of four. When people say a heavily filtered tank can be stocked harder they are right about the chemistry and usually wrong about the space.

How do I count a schooling fish against a stocking limit?

Count the whole school as one indivisible entry, not as individual fish you can trim. Six rasboras at 1.5 inches each are 9 inches of fish and there is no such thing as three of them. The practical consequence is that a shoaling species sets a minimum tank rather than fitting into whatever tank you have: if six will not fit inside your ceiling, the species does not go in that tank. This is where the arithmetic and the welfare answer diverge most sharply, and the welfare answer wins — an under-schooled tetra is a stressed tetra, and a stressed fish is a sick fish sooner than a well-kept one.