Key takeaways
- Watts per gallon worked only because every fluorescent tube on the market converted electricity into light at close to the same rate.
- A mid range LED runs roughly double the efficacy of T8 fluorescent and aims almost all of that output downward instead of scattering it into a hood.
- The rule never accounted for water depth or tank footprint, so a 20 gallon long and a 20 gallon high were treated as the same lighting problem.
- Replace it with three figures: substrate PAR, rim to substrate depth, and the demand of the hardest plant you intend to keep.
- Wattage still tells you running cost, waste heat, and which driver to order when one dies.
Forty watts of T8 fluorescent and forty watts of LED, hung over the same 90 cm tank, do not grow the same plants. One of them may not grow much at all, and the rule that treats the two as equivalent is still being repeated in shops and printed in books sold today.
Watts per gallon was a good rule for about fifteen years. It worked because of an accident of manufacturing, and that accident ended somewhere around 2010.
What follows is what the rule was measuring by proxy, which physical link snapped, and the three figures I use in its place when a tank arrives and nobody can tell me whether the fixture is strong enough.
The rule, and the era that made it true
The formula was 1 to 2 watts per gallon for undemanding plants, 2 to 3 for a moderately planted tank, and 3 or more for anything difficult. On a 20 gallon (76 litre) tank that meant 40 to 60 watts of tube, which was one or two standard lamps.
It held because the entire hobby was buying one technology. A 90 cm fluorescent tube from any manufacturer turned electricity into light at close to the same rate as any other 90 cm tube, in the same shape, in the same fitting, over the same footprint.
Wattage was standing in for light output. While the conversion between those two stayed fixed across the whole market, the substitution cost you nothing.
A watt describes the electricity bill
A watt is a rate of energy consumption. It states what the fixture pulls from the wall and says nothing about what leaves the front of it.
The property that actually varies is efficacy, meaning light produced per watt consumed. Across the lamp types this hobby has used, efficacy spans a factor of roughly three, and inside the LED category alone it spans nearly two.
Once efficacy varies, wattage stops carrying information. Two fixtures both labelled 40 watts can differ by more than 100 percent in the light they land on a substrate, and nothing on either box will warn you.
The numbers that broke the rule
These figures are approximate and they shift with manufacturer and vintage. The ranking is stable, though, and the gaps are wide enough that precision is not the point.
| Lamp type | Typical efficacy | Share aimed at the water | Common in |
|---|---|---|---|
| T12 fluorescent, magnetic ballast | 50 to 70 lumens per watt | Roughly half, via a plain reflector | 1980s to late 1990s |
| T8 fluorescent, electronic ballast | 80 to 95 lumens per watt | Half to two thirds | Late 1990s to mid 2000s |
| T5 high output | 90 to 105 lumens per watt | Two thirds with a good parabolic reflector | Mid 2000s to early 2010s |
| Mid range LED aquarium bar | 100 to 150 lumens per watt | Nearly all of it, aimed by lens | 2012 onward |
Two properties changed at once, which is why the rule failed outright rather than drifting slowly out of calibration. Efficacy roughly doubled, and the direction the light travelled changed completely.
Optics, the part a tube never had
A fluorescent tube radiates in every direction. Half its output leaves the glass heading upward into the hood, and a stamped aluminium reflector sends back some fraction of that, usually with losses at every bounce.
An LED emits into a hemisphere from the start, and a lens narrows the beam further. Almost every photon a decent aquarium bar produces is already pointed at the water before it clears the housing.
That difference is worth something like one and a half to two times the delivered light, stacked on top of the efficacy gain. It also concentrates the output instead of spreading it, which creates a fresh problem examined in the trade between coverage and raw output.
What the rule ignored, even in 1995
Depth. A 20 gallon long and a 20 gallon high hold the same water and score identically on watts per gallon, but one is about 30 cm deep and the other closer to 40 cm.
Light crossing that extra 10 cm gives up a real share to distance and absorption before a leaf sees it. Same volume, same watts, different tank, and the rule cannot tell them apart.
Footprint was invisible to it as well. Thirty watts stretched across a 120 cm tank is a different proposition from 30 watts concentrated over a 45 cm cube, and volume alone will never separate those two cases.
The rule also quietly assumed a hood resting on the cover glass. Move that same fixture 30 cm up onto a hanging kit and substrate readings can fall by half, while the number on the label does not move at all.
A 55 gallon tank that lost 170 watts and gained plants
A customer brought me a 208 litre (55 gallon) tank running four 54 watt T5 high output tubes. That is 216 watts over 55 gallons, just under 4 watts per gallon, and by the old arithmetic it should have grown anything in the catalogue.
The tubes were four years old and had never been changed. Fluorescent output falls steadily with hours run, and past roughly 12,000 to 15,000 hours a T5 tube is often down to 70 to 80 percent of its new output. Substrate PAR measured 21.
We fitted a single 45 watt LED bar 20 cm above the rim. Substrate PAR came back at 46, the tank shed 171 watts of draw and roughly the same again in waste heat, and the Cryptocoryne that had sulked for two years filled the midground inside a season.
The three figures that replace it
Only the first of these is genuinely non negotiable. The other two turn it into a shopping decision.
- PAR at the substrate, measured or read off a published chart at your depth
- Water depth from rim to substrate, in centimetres
- The light demand of the hardest plant you intend to keep
PAR is the honest currency here, and what that number counts and where it has to be read are worked through in what the intensity figure is actually counting. Depth is what converts a fixture rating into a substrate reading, which is the whole subject of matching output to the height of your water column.
Plant demand closes the loop. Species by species figures sit in the demand table growers work from, and buying the fixture before choosing the plants is simply the wrong order of operations.
If nobody publishes a PAR chart
Plenty of budget fixtures publish nothing you can use. There is a crude fallback, and I want to be blunt that it is a starting guess and not a rule.
For a mid range white LED bar over water 30 to 40 cm deep, somewhere around 8 to 12 watts of LED per 30 cm of tank length puts you in medium light territory. Halve that for a shallow nano and double it past 50 cm of depth.
The heuristic only holds inside one technology, over one depth band, with broadly similar optics. Stretch it beyond those conditions and you have quietly reinvented the mistake this article is about.
Where wattage still earns its keep
Running cost first. A 45 watt fixture on eight hours a day draws about 0.36 kWh daily and near 131 kWh across a year, which at typical residential rates lands somewhere around $15 to $30 annually.
Heat second. Every watt a fixture consumes turns into heat somewhere, and on a closed hood over a small tank that can lift water temperature by 1 to 2 degrees Celsius (2 to 4 degrees Fahrenheit) through a summer afternoon.
Third, wattage sizes the power supply. If you are replacing a dead driver, the wattage rating is exactly the figure you need, and the US Department of Energy publishes a plain overview of LED efficiency and driver behaviour on its Energy Saver pages.
Never size a replacement LED by matching the wattage of the fluorescent hood it is replacing. On a shallow tank that arithmetic routinely lands you at two or three times the light you had before, and the algae shows up in about ten days.
Why the rule refuses to die
Because it is easy, and because nothing that replaced it fits on a shelf label. Watts per gallon needs no meter, no depth measurement, and no decision about which plants you want.
PAR needs either a quantum sensor at $400 to $500, such as the Apogee MQ-510 whose specifications sit at Apogee Instruments, or a manufacturer honest enough to publish readings at more than one distance. Neither is as convenient as division.
The convenient answer stopped being the correct one the year your fixture changed technology. If your tank genuinely still runs tubes, the old rule is not badly wrong, and where these figures sit alongside colour and hours is set out in the wider account of how a fixture is specified.
Questions readers ask about this
Only on a tank still lit by fluorescent tubes of one type, at a standard depth, with the hood sitting on the cover glass. Inside those boundaries 2 to 3 watts per gallon of T8 still lands roughly where it always did. Apply it to any LED fixture and the answer becomes noise.
Ask about depth instead of volume. A standard 55 gallon is around 40 cm from rim to substrate, and a mid range LED bar of 35 to 55 watts across its 120 cm length usually reaches medium light there. A 55 gallon in a taller shape needs considerably more.
Because efficacy roughly doubled and the optics now aim the output straight down. A 40 watt LED bar frequently delivers more substrate PAR than the 96 watt fluorescent hood it replaced. Cut the photoperiod by an hour, drop the output slider to 60 percent, and reassess after two weeks.
Slightly better, and still poor. Lumens weight output to human eye sensitivity, which peaks in the green band that leaves reflect rather than absorb. A fixture heavy in green will post strong lumen figures and grow plants no faster. Ask for PAR at a stated distance instead.
Every 9 to 12 months on a planted tank running eight hours a day, which works out near 3,000 hours. The tube keeps lighting the room long after it has stopped growing plants, and output loss of 20 to 30 percent is invisible to your eye but obvious on a meter.
