A driver pulls up to a 1,000-gallon horizontal diesel tank, drops a gauge stick, reads 18 inches wet, and pulls the laminated chart taped inside the cab door. The chart says 612 gallons. Back at the desk somebody runs the same 18 inches through a volume formula and gets 605. Seven gallons apart, and nobody's wrong.
That gap is the most misunderstood thing about gauging a cylinder on its side. The formula answer and the chart answer come from two different definitions of what "gallons in this tank" even means, and the trades that move fuel for a living learned a long time ago to keep both around. One is geometry. The other is a measurement of a specific, dented, real piece of steel sitting in the yard.
The two definitions in play
The calculated number treats the tank as an ideal shape: a perfect cylinder, perfectly round, with flat ends and walls of zero thickness. Feed it diameter, length, and the wet height on the stick, and it returns the volume of fluid that shape would hold. It is exact for the shape it describes and it costs nothing but the arithmetic.
The strapping table, also called a tank capacity table or gauge chart, is the opposite philosophy. Somebody physically measured *that* tank and built a lookup table: at every gauge height, here are the real gallons. It does not assume the tank is round, because the tank isn't quite round. It does not assume the ends are flat, because they probably aren't. It is the answer for one tank and one tank only.
Neither is the "true" number in some absolute sense. They answer slightly different questions, and the difference between them is exactly the catalog of ways a real tank departs from an ideal one.
Where the seven gallons hide
When a strapping table disagrees with a clean formula, the gap is built out of a short list of physical realities the formula was never told about.
| Source of difference | What it is | Typical effect on a small tank |
|---|---|---|
| End-cap shape | Dished, domed, or hemispherical ends hold more than flat ends | Adds volume, often 2–6% on a short tank |
| Out-of-round shell | A used tank sags or bulges; the cross-section isn't a true circle | A few gallons, either direction |
| Shell plate thickness | The formula uses outside dimensions; fluid sits inside the steel | Removes a small, fixed slice |
| Deadwood | Internal baffles, heater coils, sumps, fittings | Removes volume at specific heights |
| Tilt | A tank that isn't dead level reads long or short on the stick | Skews the whole table |
The American Petroleum Institute writes the rules for building these tables under its Manual of Petroleum Measurement Standards, Chapter 2.2 (the same content that became ISO 12917 in 2002). Chapter 2.2E covers calibrating a horizontal tank by manual measurement; Chapter 2.2F covers doing it with an internal electro-optical distance-ranging instrument that maps the shell. Those standards exist precisely because the geometry formula isn't trusted alone for custody transfer. The published procedures even list the correction terms a calibrator has to account for: strapping-tape tension, the tape's thermal expansion, shell expansion under the weight of the liquid, plate thickness, and deadwood. Every one of those is a place a real tank quietly drifts away from the textbook cylinder.
Why the formula is still the right tool most of the time
Read all that and you might conclude you should always strap the tank. For custody transfer, where money changes hands per gallon and a regulator may audit you, that's the correct instinct. For a farmer deciding whether to order fuel this week, it's wildly out of proportion.
The flat-end circular-segment formula is accurate to a fraction of a percent *for the shape it models*, and for the rectangular-section diesel and heating-oil tanks that dominate farms and small dealers, that shape is nearly the real one. Those tanks are stamped from flat sheet with shallow ends. The "error" against a proper strapping table on a tank like that lives in the low single digits of percent, and most of that is the end caps. When you're ordering 600 gallons and the truck meter is the thing you actually pay against, a 1% modeling gap on your own gauge reading is noise.
The place the formula genuinely earns its keep is the non-linearity that no chart-free eyeball can handle. Tip a cylinder on its side and equal inches of stick stop meaning equal gallons, because the cross-section is fat in the middle and pinched top and bottom. That curve is real and steep near the ends, and it's the reason a stick reading needs *some* math behind it whether that math is a formula or a pre-computed table.
A worked check you can run in your head
Take a tank 48 inches in diameter and 96 inches long. Full capacity is roughly 752 gallons. Now compare three gauge heights and watch the curve:
- Stick at 24 inches (exactly half the diameter): about 376 gallons, a clean 50%. This is the one height where the symmetry makes "half stick" equal "half tank."
- Stick at 12 inches (a quarter of the diameter): not 188 gallons, but closer to 142, roughly 19% of capacity, well under the quarter you'd guess.
- Stick at 36 inches (three-quarters up): around 610 gallons, about 81%, not the 75% the stick height suggests.
The two off-center readings are where intuition fails and where a hand-scrawled "each inch is about 8 gallons" rule does real damage. Those numbers come straight out of the circular-segment geometry. A strapping table for an actual flat-end tank of these dimensions would land within a handful of gallons of each, and the gaps would be the end-cap and out-of-round terms from the table above.
The unit underneath all of it
There's one more place precision sneaks in, and it predates every tank in the field by three centuries. A US gallon is defined as exactly 231 cubic inches. That number isn't a rounding of anything tidy; it's the old English wine gallon fixed by a 1706 statute under Queen Anne (5 Anne c. 27), which the US Treasury adopted in 1832 and which the country has used ever since. Every gallon figure on a strapping table or out of a formula is, at bottom, cubic inches of fluid divided by 231. Litres are cleaner, since a US gallon is exactly 3.78541 litres, which is part of why the metric ag world rarely argues about this.
It matters here because a tank computed in inches and reported in gallons is doing that 231 division at the end, and a tank strapped in the field is doing it too. When the formula answer and the chart answer disagree, the gallon definition is never the culprit. The culprit is always the shape.
What to keep on hand
The honest setup for most field gauging is both numbers, used for what each is good at. Keep a fast calculation for the day-to-day question of how many gallons are in the tank and how many to order, and keep a proper strapping table for the tank if you ever need a custody-grade figure or your tank has pronounced dished ends. The calculation handles the non-linear curve correctly and instantly; the table handles the dents, the deadwood, and the end caps that the calculation can't see.
If you want the fast side of that pair, a flat-end horizontal-tank reading turned into gallons remaining and gallons to fill from a single stick measurement on the phone you already have at tank-side, Dipfill does exactly that, with the flat-end assumption stated right next to the result so you always know which of the two answers you're holding.