Solid geometry

The formula (the core equation(s), each with a one-line reading of what it says)

Prism / cylinder:   V = A·h

Reads: any solid with a constant cross-section is just that cross-section extruded — area times length, no exceptions.

Pyramid / cone:   V = (1/3)·A·h

Reads: taper the cross-section linearly down to a point and exactly one third of the extruded volume survives. Any base shape, any apex position, same third.

Sphere:   V = (4/3)·π·r³        A = 4·π·r²

Reads: the volume grows with the cube of radius, the skin with the square — and the skin is exactly four great-circle disks.

Conical frustum:   V = (h/3)·(A₁ + A₂ + √(A₁·A₂))

Reads: a cone with the tip cut off. Note the geometric-mean term — it is not the average of the two end areas times height.

Spherical cap:   V = (π·h²/3)·(3r − h)

Reads: the volume of liquid depth h in a spherical head of radius r — the workhorse of every dished-end tank calculation.

Prismatoid:   V = (h/6)·(A_bot + 4·A_mid + A_top)

Reads: sample the cross-section at the bottom, middle, and top, weight the middle by four. This is Simpson's rule wearing a hard hat, and it is exact for cones, spheres, frustums, wedges, and hoppers.

Cone slant height and lateral skin:   s = √(r² + h²)       A_lat = π·r·s

Reads: the sheet-metal length runs along the slant, not the axis; the wrap-around surface is measured on s, never on h.

Dihedral angle:   cos θ = (n₁·n₂)/(|n₁|·|n₂|)

Reads: the angle between two flat faces comes from their normals — the number a sine bar or a CMM report is actually giving you.

Euler's polyhedron formula:   V − E + F = 2

Reads: for any polyhedron with no holes through it, vertices minus edges plus faces is always two. A free integrity check on every solid model and mesh.

Where you meet it (2-4 concrete engineering situations, specific: bench, test stand, review board)

Propellant load and ullage on the test stand. A horizontal tank with dished heads is a cylinder plus two spherical caps, and the level sensor reads depth, not volume. Converting one to the other is the circular-segment area for the barrel section and the cap formula for the ends — evaluated at every depth the strapping table covers. Get the cap term wrong and the loaded mass the stand reports is wrong by whole percent.

The mass-properties estimate at the design review. Before the CAD model is trusted — or when the board wants a sanity check on it — you decompose the part into primitives: this boss is a cylinder, that gusset is a wedge, the transition is a frustum, subtract the bore. Volume times density per primitive, sum, and compare against the CAD mass readout. A 15% disagreement means somebody's model has a surface that didn't knit, and Euler's V − E + F = 2 on the mesh is the fastest way to prove it.

Flat-pattern development at the sheet-metal bench. A right cone is cut flat and rolled, and the flat pattern's radius is the slant height s = √(r² + h²), not the part height — a 400 mm tall cone of 150 mm base radius unrolls from a 427 mm slant radius, not 400. Layout off the drawing's height dimension and the rolled part comes up short every time. Two nearby parts do not obey that formula: a reducer (a frustum) unrolls as an annular sector whose inner and outer radii are slant distances measured from the virtual apex the cone would taper to, and an offset transition duct is an oblique cone with no single slant radius at all — it has to be developed by triangulation. √(r² + h²) is the right cone's rule only.

Weld prep, chamfers, and fixture angles. The included angle called out on a groove weld, the countersink spec, the compound angle where two mitered members meet — all dihedral angles. When the fit-up gap varies along a joint that should be constant, the first check is whether someone measured the angle between the faces while the print defined it between the normals: those two numbers are supplements, and 60° versus 120° has scrapped real weldments.

How it works (the real substance — behavior, gotchas, limits of validity, the mistake people make)

Two scaling facts drive most of the intuition. First, the one-third factor: taper any cross-section linearly to a point and the area shrinks with the square of the remaining distance, so the volume integral of (z/h)² delivers exactly 1/3 — for a cone, a pyramid, an off-center pyramid, anything. Second, the square–cube law: scale a part up 2× and its surface doubles twice (4×) while its volume doubles three times (8×). Heat dissipation, casting solidification time, weight versus skin friction — every "it worked at half scale" surprise lives in that gap.

The mistake that actually costs money is linear interpolation on things that aren't linear. A frustum's volume is not the average of its end areas times height: for a 2:1 taper in radius the average-area shortcut reads 7.1% high — the geometric-mean term in (h/3)(A₁ + A₂ + √(A₁A₂)) exists precisely to kill that error. Same disease on horizontal tanks: at 25% of diameter in depth, a horizontal cylinder holds 19.6% of its volume, not 25%. Any gauge, spreadsheet, or hand rule that treats depth as proportional to volume on a round or dished vessel is wrong everywhere except empty, half, and full.

The prismatoid formula is the honest general-purpose weapon. Because it is Simpson's rule applied to the cross-section area function, it is exact whenever A(z) is a polynomial up to cubic — which covers every solid in this entry, including the sphere (check: A_mid = πr², ends zero, V = (2r/6)·4πr² = 4πr³/3). For a lofted or free-form solid it is no longer exact, but three sections still beat two, and for a fabricated hopper or foundation excavation it is usually within the concrete truck's rounding.

Angle bookkeeping has one standing trap. The dihedral angle from face normals and the angle you'd measure laying a protractor inside the joint are supplements of each other; which one the software reports depends on which way its normals point. A regular tetrahedron's faces meet at acos(1/3) ≈ 70.53° — if your fixture math says 109.47° for the same joint, nothing is wrong except a sign convention, but only one of those numbers cuts correctly.

Limits of validity worth stating out loud: V = A·h demands the cross-section be constant and h be measured perpendicular to it — a skewed prism still obeys it, but only with the perpendicular height. The cap and segment formulas assume truly spherical heads; real tank heads are often ASME torispherical or 2:1 elliptical, which need their own coefficients, and applying the spherical cap to an elliptical head is a built-in few-percent error. And V − E + F = 2 holds only for solids topologically like a sphere: put a hole through the part — a handle, a torus, a hollow waveguide loop — and the count drops to 0. That failure is a feature: a mesh that returns the wrong Euler count has told you its genus, or its defect, before any downstream analysis runs.

History (who derived it and when, told as a short story with inline [n] citations)

The volume rules are older than proof itself. Democritus (c. 460–c. 370 BC) — the atomist — asserted that a cone holds one third of its cylinder and a pyramid one third of its prism, apparently by imagining solids sliced into thin layers; Archimedes later credited him with stating both results without proving them ^[1]^[2]. The proof came from Eudoxus of Cnidus (408–355 BC), whose method of exhaustion squeezed a curved or tapered solid between inscribed and circumscribed stacks until no room for error remained ^[1]. Euclid packaged that machinery around 300 BC: Books XI through XIII of the Elements build up parallelepipeds, prove the cone and pyramid thirds by exhaustion in Book XII, and close the whole work by constructing the five regular polyhedra and proving there are no others ^[3]^[4].

Archimedes pushed past the straight-edged solids. In On the Sphere and Cylinder he proved that a sphere has exactly two thirds the volume and two thirds the surface of its circumscribing cylinder, and he rated it his best result — he asked for the figure to be carved on his tomb ^[5]^[6]. He got his wish: in 75 BC, some 137 years after Archimedes died in the sack of Syracuse, Cicero found the neglected grave by hunting for the sphere-and-cylinder marker ^[5]^[6]. Half a world away and seven centuries later, Zu Geng (c. 450–520 AD), working with his father Zu Chongzhi, derived the sphere's volume from the principle that solids with matching cross-section areas at every height have matching volumes ^[9]^[10].

The Renaissance rediscovered that principle through a wine barrel. At his 1613 wedding, Kepler watched a merchant gauge barrels by poking a rod diagonally through the bung-hole and wondered why that worked; the result was Nova stereometria doliorum vinariorum (1615), which computed volumes of revolution by resolution into indivisibles ^[7]^[8]. Cavalieri systematized the idea in his Geometria indivisibilibus of 1635, and "Cavalieri's principle" — the same statement Zu Geng had used — became the standard bridge between slices and volumes until calculus made it a theorem ^[9]^[7].

The last classical surprise came by mail. In 1750 Euler wrote to Goldbach with the observation that vertices minus edges plus faces equals two for the solids he could count, publishing papers on it in 1752 — the first admitting he could not prove it ^[11]^[12]. Archimedes and Descartes had both written extensively on polyhedra and missed it, likely because it is a fact about connection rather than measurement ^[11]. Rigorous proofs arrived in the nineteenth century, and the formula became the seed of topology — which is why a 2,000-year-old subject about cones and pyramids ends up validating finite-element meshes.

Related tools (bullet list of HE calculator slugs that use or neighbor this topic, as /tools/ links)

/tools/tank-volume-horizontal — the circular segment plus spherical-cap machinery, depth in, gallons out
/tools/tank-volume-vertical — cylinder-plus-heads volume for the upright case
/tools/concrete-slab-volume — V = A·h with a waste factor, priced by the yard
/tools/concrete-column-volume — cylinders and rectangular prisms, batched
/tools/convert-area-volume — m³, ft³, gallons, liters; the units half of every volume problem
/tools/moment-of-inertia-shapes — the next integral up from volume over the same primitive solids
/tools/convert-angle — dihedral results arrive in degrees, radians, or mils depending on who fitted the planes

Sources

Written by HE in our own words from the cited sources — engineering judgment included, your stamp still required. All entries →