[Intermediate] — surface area and volume, and the one principle that delivers the cone and the sphere without calculus.

Before you start

  • You can find the area of plane shapes, including sectors — plane shapes.
  • You met the limiting argument that gave — same lesson, section 4.
  • You know Pythagoras, for slant heights — triangles and polygons.

What you will be able to do after this lesson:

  1. State Cavalieri’s principle and use it to derive the volume of a sphere from a cylinder and a cone.
  2. Derive the curved surface area of a cone by unrolling it into a sector.
  3. Compute surface area and volume for prisms, cylinders, cones, pyramids and spheres, and choose the right one for a composite solid.
  4. Apply the scaling laws — length, area and volume scale as , , — and explain a physical consequence.

Study route: read 1–5, attempt the prediction in section 4, then run the lab. Block 1 is the derivation that matters.


1. Why this exists

A cylindrical water tank is to be built. Its volume determines what it holds and what you are paid for; its surface area determines how much steel it costs and how fast it loses heat. These are different numbers with different units, and they do not scale together — which is the point of section 5 and the reason large animals are not simply scaled-up small ones.

There is also a mathematical reason. Areas of straight-sided plane shapes came from rearrangement. Volumes of straight-sided solids do not always: no finite cutting turns a regular tetrahedron into a cube of equal volume. That was Hilbert’s third problem, settled by Dehn in 1900 — the year it was posed. So even the pyramid’s volume needs a limiting argument, and the sphere certainly does.

2. Terminology

TermPlain-English definitionNote
PrismA solid with the same cross-section all the way alongA cylinder is a prism with a circular cross-section
Cross-sectionThe shape you get by slicingConstant for a prism, shrinking for a cone
Perpendicular heightHeight measured at right angles to the baseNot the slant
Slant height ()Distance from apex to base edge along the surfaceFor a cone,
Curved surface areaThe area of the curved part onlyExcludes the flat base(s)
Total surface areaEverything, including basesSay which you mean
Cavalieri’s principleSolids with equal cross-sectional areas at every height have equal volumeThe workhorse of this lesson

3. Prisms and cylinders

For any solid with a constant cross-section of area and perpendicular height :

For a cylinder, , so .

Curved surface of a cylinder — by unrolling. Cut the cylinder down its side and flatten it. You get a rectangle whose width is the base circumference and whose height is :

Unrolling is the key move, and it works because a cylinder is developable — it can be flattened without stretching. Section 4 uses it again for the cone. A sphere is not developable, which is exactly why every flat map of the Earth distorts something.

4. Cavalieri’s principle, and the cone and sphere

Cavalieri’s principle. If two solids have the same height, and at every level their cross-sections have equal area, then they have equal volume.

The intuition is a stack of coins: shear the stack sideways and it still contains the same amount of metal, because no slice changed size.

The cone

The can be seen directly: a cube can be dissected into three identical square-based pyramids, each with the cube’s base and height. So each is a third of the cube — and Cavalieri then extends the result from that special pyramid to every pyramid and cone of the same base area and height, because their cross-sections match level by level.

The cross-section of a cone at height is a circle of radius , shrinking linearly, so its area falls as the square of the remaining height. That is why the answer is rather than .

Curved surface of a cone — by unrolling. Cut along a slant line and flatten. You get a sector of a circle of radius (the slant height), whose arc is the base circumference . Using sector area :

with by Pythagoras.

The sphere — Archimedes’ argument

Take a hemisphere of radius . Alongside it, take a cylinder of radius and height , with a cone of the same radius and height removed from it, apex down.

Slice both at height above the base:

  • Hemisphere: the slice is a circle. By Pythagoras its radius is , so its area is .
  • Cylinder minus cone: the cylinder’s slice is a circle of radius , area . The cone’s slice at height has radius , area . The remaining ring has area .

Identical, at every height. So by Cavalieri the two solids have the same volume:

and therefore:

No calculus, and no measurement — just one principle and a Pythagorean slice. Block 1 of the lab checks the equality height by height.

Surface area of a sphere: , which Archimedes also proved. It is exactly the curved surface of the circumscribing cylinder (), a coincidence striking enough that he asked for a sphere inscribed in a cylinder to be carved on his tombstone.

TIP

Predict before running the lab. A sphere sits snugly inside a cylinder of the same radius and height . What fraction of the cylinder’s volume does the sphere occupy? And what fraction of its total surface area? Work both out before opening the answers.

5. The scaling laws

Scale every length of a solid by a factor :

QuantityScales asExample:
Lengthdoubles
Area
Volume

Because area and volume scale differently, the ratio of surface area to volume falls as . A large object has proportionally less surface than a small one.

That single fact explains a great deal: why a mouse must eat constantly and an elephant need not (heat is lost through surface, generated through volume); why fine gravel dissolves faster than a boulder; why cells are microscopic; and why a scaled-up insect could not breathe, since gas exchange is a surface process supporting a volume of tissue.

Worked example — runnable

Runnable example: save as solids.py in any empty directory and run python3 solids.py. Standard library only; writes no files.

"""Volumes and surface areas, including Cavalieri's derivation of the sphere."""
import math
 
EPS = 1e-9
 
 
def cylinder_volume(r, h):
    return math.pi * r * r * h
 
 
def cone_volume(r, h):
    return math.pi * r * r * h / 3
 
 
def sphere_volume(r):
    return 4 * math.pi * r ** 3 / 3
 
 
def cone_slant(r, h):
    return math.hypot(r, h)
 
 
def cone_curved_area(r, h):
    return math.pi * r * cone_slant(r, h)
 
 
def sphere_area(r):
    return 4 * math.pi * r * r
 
 
def slice_area_hemisphere(r, y):
    """Cross-section of a hemisphere at height y: a circle of radius sqrt(r^2-y^2)."""
    return math.pi * (r * r - y * y)
 
 
def slice_area_cylinder_minus_cone(r, y):
    """Cross-section of (cylinder radius r, height r) with a cone removed."""
    return math.pi * r * r - math.pi * y * y
 
 
def volume_by_slicing(area_at_height, lo, hi, n=200000):
    """Riemann midpoint sum of cross-sectional areas - Cavalieri, numerically."""
    dy = (hi - lo) / n
    return sum(area_at_height(lo + (i + 0.5) * dy) for i in range(n)) * dy
 
 
if __name__ == "__main__":
    R = 3.0
 
    print("Block 1 - Cavalieri: hemisphere vs (cylinder minus cone), slice by slice")
    print("      y    hemisphere   cyl - cone    difference")
    for i in range(6):
        y = R * i / 5
        a1 = slice_area_hemisphere(R, y)
        a2 = slice_area_cylinder_minus_cone(R, y)
        print(f"  {y:5.2f}  {a1:11.6f}  {a1:11.6f}  {a1 - a2:.1e}")
        assert abs(a1 - a2) < EPS
    print("  equal at every height, so the two solids have equal volume")
 
    hemi = cylinder_volume(R, R) - cone_volume(R, R)
    print(f"  cylinder(r={R}, h={R}) - cone(r={R}, h={R}) = {hemi:.6f}")
    print(f"  2/3 pi r^3                                  = {2*math.pi*R**3/3:.6f}")
    print(f"  so the whole sphere is                       {2*hemi:.6f}")
    print(f"  and 4/3 pi r^3 is                            {sphere_volume(R):.6f}")
    assert abs(2 * hemi - sphere_volume(R)) < EPS
 
    print()
    print("Block 2 - the same volumes by numerical slicing")
    hemi_numeric = volume_by_slicing(lambda y: slice_area_hemisphere(R, y), 0, R)
    print(f"  hemisphere by Riemann sum = {hemi_numeric:.6f}")
    print(f"  closed form               = {sphere_volume(R)/2:.6f}")
    assert abs(hemi_numeric - sphere_volume(R) / 2) < 1e-6
    cone_numeric = volume_by_slicing(
        lambda y: math.pi * (R * (1 - y / 6.0)) ** 2, 0, 6.0)
    print(f"  cone (r={R}, h=6) by Riemann sum = {cone_numeric:.6f}")
    print(f"  1/3 pi r^2 h                     = {cone_volume(R, 6.0):.6f}")
    assert abs(cone_numeric - cone_volume(R, 6.0)) < 1e-6
 
    print()
    print("Block 3 - unrolling a cone gives a sector")
    H = 4.0
    l = cone_slant(R, H)
    base_circumference = 2 * math.pi * R
    sector_angle = base_circumference / l           # radians: arc = radius * angle
    sector_area = 0.5 * l * l * sector_angle        # 1/2 r^2 theta
    print(f"  r={R}, h={H} -> slant l = {l:.6f}")
    print(f"  unrolled: a sector of radius {l:.4f}, arc {base_circumference:.6f}")
    print(f"  sector angle {sector_angle:.6f} rad ({math.degrees(sector_angle):.4f} deg)")
    print(f"  sector area  {sector_area:.6f}   pi r l = {cone_curved_area(R, H):.6f}")
    assert abs(sector_area - cone_curved_area(R, H)) < EPS
    assert sector_angle < 2 * math.pi, "the sector never closes into a full circle"
 
    print()
    print("Block 4 - Archimedes' tombstone: sphere inside its snug cylinder")
    cyl_v = cylinder_volume(R, 2 * R)
    cyl_total_area = 2 * math.pi * R * (2 * R) + 2 * math.pi * R * R
    print(f"  sphere volume {sphere_volume(R):.6f},  cylinder volume {cyl_v:.6f}")
    print(f"  ratio = {sphere_volume(R) / cyl_v:.10f}")
    print(f"  sphere area   {sphere_area(R):.6f},  cylinder total area {cyl_total_area:.6f}")
    print(f"  ratio = {sphere_area(R) / cyl_total_area:.10f}")
    assert abs(sphere_volume(R) / cyl_v - 2 / 3) < EPS
    assert abs(sphere_area(R) / cyl_total_area - 2 / 3) < EPS
    print("  both ratios are exactly 2/3 - which is what he wanted on his grave")
    # the curved part of the cylinder alone equals the sphere's whole surface
    assert abs(2 * math.pi * R * (2 * R) - sphere_area(R)) < EPS
 
    print()
    print("Block 5 - scaling: length k, area k^2, volume k^3")
    print("      k    volume x    area x    area/volume x")
    base_v, base_a = sphere_volume(1.0), sphere_area(1.0)
    for k in (1, 2, 3, 10):
        v, a = sphere_volume(k), sphere_area(k)
        print(f"  {k:5}  {v/base_v:9.1f}  {a/base_a:8.1f}  {(a/v)/(base_a/base_v):13.4f}")
        assert abs(v / base_v - k ** 3) < 1e-6
        assert abs(a / base_a - k ** 2) < 1e-6
    print("  surface-to-volume falls as 1/k: big things have proportionally less skin")
 
    print()
    print("solids: passed")

Expected output:

Block 1 - Cavalieri: hemisphere vs (cylinder minus cone), slice by slice
      y    hemisphere   cyl - cone    difference
   0.00    28.274334    28.274334  0.0e+00
   0.60    27.143361    27.143361  3.6e-15
   1.20    23.750440    23.750440  0.0e+00
   1.80    18.095574    18.095574  3.6e-15
   2.40    10.178760    10.178760  1.8e-15
   3.00     0.000000     0.000000  0.0e+00
  equal at every height, so the two solids have equal volume
  cylinder(r=3.0, h=3.0) - cone(r=3.0, h=3.0) = 56.548668
  2/3 pi r^3                                  = 56.548668
  so the whole sphere is                       113.097336
  and 4/3 pi r^3 is                            113.097336

Block 2 - the same volumes by numerical slicing
  hemisphere by Riemann sum = 56.548668
  closed form               = 56.548668
  cone (r=3.0, h=6) by Riemann sum = 56.548668
  1/3 pi r^2 h                     = 56.548668

Block 3 - unrolling a cone gives a sector
  r=3.0, h=4.0 -> slant l = 5.000000
  unrolled: a sector of radius 5.0000, arc 18.849556
  sector angle 3.769911 rad (216.0000 deg)
  sector area  47.123890   pi r l = 47.123890

Block 4 - Archimedes' tombstone: sphere inside its snug cylinder
  sphere volume 113.097336,  cylinder volume 169.646003
  ratio = 0.6666666667
  sphere area   113.097336,  cylinder total area 169.646003
  ratio = 0.6666666667
  both ratios are exactly 2/3 - which is what he wanted on his grave

Block 5 - scaling: length k, area k^2, volume k^3
      k    volume x    area x    area/volume x
      1        1.0       1.0         1.0000
      2        8.0       4.0         0.5000
      3       27.0       9.0         0.3333
     10     1000.0     100.0         0.1000
  surface-to-volume falls as 1/k: big things have proportionally less skin

solids: passed

Block 1 is Archimedes’ argument, executed. The two solids look nothing alike, yet every slice matches — and that is enough to force equal volumes.

NOTE

In block 4 the sphere’s volume and surface area print as the same number, . That is a coincidence of choosing : and only when . They are still a volume and an area, in different units, and at any other radius the numbers differ.

Common pitfalls and traps

  • Using slant height where perpendicular height belongs, or the reverse. Volume of a cone uses ; curved surface uses . They are different numbers related by , and swapping them is the classic error.
  • Forgetting which surfaces are included. “Curved surface area” of a cone excludes the base. A closed cylinder has two circular ends; an open pipe has none. Read the question.
  • Assuming a cone is half a cylinder. It is a third. The cross-section shrinks linearly, so its area shrinks quadratically.
  • Applying Cavalieri without checking every height. The principle requires equal cross-sections at every level, not just at a few. Matching at three heights proves nothing.
  • Expecting the sphere to unroll. It cannot — a sphere has intrinsic curvature and no flat development exists. This is a theorem, not a lack of ingenuity, and it is why world maps must distort area, angle, or both.
  • Scaling area and volume by the same factor. Doubling the radius of a tank multiplies steel by and capacity by . Costing a scaled design as though both grow together is a real and expensive mistake.

Check your understanding

  1. A cylinder has radius cm and height cm. Find its volume and total surface area. Use .
  2. A cone has radius and perpendicular height . Find its slant height, curved surface area and volume.
  3. A sphere and a cone have the same radius , and the cone has height . Which has the greater volume, and by what ratio?
  4. A hemisphere of radius sits on top of a cylinder of the same radius and height . Give the total volume and the total exterior surface area.
  5. If a model is built at scale, what fraction of the original’s surface area and volume does it have?
Answers — open only after an attempt
  1. cm³. Total surface cm².
  2. . Curved surface . Volume .
  3. Sphere ; cone . The sphere is greater, by a factor of exactly .
  4. Volume . Exterior surface cylinder’s curved part hemisphere’s curved part the cylinder’s base — the top of the cylinder is covered by the hemisphere and is not exterior. Total .
  5. Surface area ; volume .

And the prediction from section 4: the cylinder has volume , and the sphere — a ratio of exactly . The cylinder’s total surface is , and the sphere’s is — again exactly . Both ratios being the same number is Archimedes’ result, and block 4 confirms both.

Practice — independent task

Implement volume_of_revolution(radius_at_height, lo, hi, n): the volume of the solid formed by rotating a curve about an axis, computed by slicing.

  1. Use the disc method: a slice at height is a circle of radius radius_at_height(y), so its area is . Sum over slices.
  2. Reproduce every solid in this lesson: cylinder (constant radius), cone (linear), sphere (sqrt(r^2 - y^2)), and check each against its closed form to within 1e-6.
  3. Compare midpoint, trapezoidal and left-endpoint sums at the same n. Report which converges fastest, and by how much — with numbers you produced.
  4. Now do something with no elementary closed form: rotate about the -axis for . Report your value and estimate its error by comparing and .
  5. Investigate the sphere near its poles. The radius function has infinite slope at , so the discs are a poor fit there. Measure how the error behaves as grows — does it improve like , , or something worse? State what you observed.

Edge cases: lo == hi; a radius function returning zero throughout; a negative radius (decide whether to raise or square it away, and justify).

Done when: all four closed forms are reproduced, your step-3 comparison is backed by measured numbers, and your answer to step 5 identifies the convergence rate you actually observed rather than the one you expected.

Tradeoffs, limits and extensions

Cavalieri is integration without the notation. “Equal cross-sections at every height means equal volume” is the statement that depends only on . The lab’s volume_by_slicing is a Riemann sum, and applications of integration formalises it. The advantage of Cavalieri’s form is that it gives the sphere exactly, with no limit taken.

Some solids genuinely need calculus. A torus, an ellipsoid, or the solid of revolution in step 4 above have no elementary dissection argument. The slicing method handles all of them uniformly, which is why it eventually replaced the geometric arguments entirely.

Numerical accuracy has a shape. A midpoint Riemann sum converges as for a smooth integrand, but the sphere’s radius function has unbounded derivative at the poles, so the observed rate there is worse. Step 5 asks you to measure this rather than assume it — the gap between the theoretical rate and the observed one is exactly where numerical analysis lives.

Before moving on

You are done with this lesson when you can:

  • State Cavalieri’s principle and use it to derive the sphere’s volume from a cylinder and a cone.
  • Derive the cone’s curved surface area by unrolling it into a sector.
  • Choose correctly between perpendicular height and slant height.
  • Apply the , , scaling laws and give a physical consequence.

Recap for later lookup: prism/cylinder ; cylinder curved surface ; cone , curved surface with ; sphere , surface ; sphere : circumscribing cylinder in both volume and surface area; lengths scale as , areas , volumes .

Next: Latitude and Longitude — the sphere again, but now measuring on its surface, where none of the plane formulas apply.

↑ Back to Geometry and Trigonometry