Surface area of revolution: Question bank — Maths deep dive

Quick reminders of the vocabulary these questions use (all built in the parent note):

$d s$ = the arc-length element, the length of a tiny slanted piece of curve, $d s = 1 + (y^{'})^{2} d x$ .
$d x$ = the horizontal shadow of that piece (its projection onto the axis).
The radius = the distance from a point on the curve to the rotation axis (how far it travels in one full spin, divided by $2 π$ ).
A frustum = a cone with its tip sliced off; the thin band each curve-piece sweeps.

True or false — justify

TF1. Rotating a curve about the x-axis, the surface area integrand is

2 π y d x

False. It is

2 π y d s

; using

d x

takes only the horizontal shadow of each piece and undercounts every part of the curve that is not flat.

TF2. If a curve is perfectly horizontal (constant

y = c

) on

[a, b]

, then

d s = d x

there.

True. A flat piece has

d y = 0

, so

d s = 1 + 0 d x = d x

— the slant equals the shadow only when nothing rises or falls.

TF3. Surface area of revolution always uses

π r^{2}

somewhere, just like volume of revolution.

False. Volume stacks disks (

π r^{2}

); surface wraps rings whose contribution is circumference

2 π r

times slant

d s

. See Volume of revolution (disk & shell) for the contrast.

TF4. Rotating the segment

y = x

x \in [0, 1]

about the x-axis gives the same surface area as rotating it about the y-axis.

True here by symmetry of that particular line about

y = x

, but not in general — the radius factor changes from

y

x

, so only symmetric configurations coincide.

TF5. If you double every length in a curve (scale by 2), the surface of revolution has double the area.

False. Area scales with the square of length, so it becomes

4 \times

larger — both the radius and the

d s

each double.

TF6. The factor

1 + (y^{'})^{2}

can never be less than 1.

True. It is

1

only where

y^{'} = 0

and grows for any nonzero slope, since a square is never negative; the slant edge is always at least as long as its shadow.

TF7. Rotating

y = f (x)

about the line

y = 3

still uses radius

y

False. The radius is the distance to the axis, here

∣ y - 3∣

; using

y

silently assumes the axis is

y = 0

TF8. For a curve dipping below the x-axis, you should use

2 π y

with

y

possibly negative.

False. Distance is never negative, so the radius is

2 π ∣ y ∣

; a negative radius would give nonsensical negative area contributions.

Spot the error

SE1. "To rotate

x = g (y)

about the y-axis,

S = \int 2 π y 1 + (d x / d y)^{2} d y

The radius is wrong: distance to the y-axis is

x = g (y)

, not

y

. It should read

2 π x = 2 π g (y)

;

y

measures height, not distance from the axis.

SE2. "Since

d s = 1 + (y^{'})^{2} d x

, I'll integrate

\int 2 π y 1 + (y^{'})^{2} d y

over the

y

-limits."

The variable is mismatched: the

d x

form of

d s

must pair with

x

-limits and

d x

. If you switch to

y

-limits you must rewrite

d s = 1 + (d x / d y)^{2} d y

SE3. "A full cone of slant

L

, radius

r

has lateral area

π r^{2} L

The area is

π r L

(unroll the cone into a circular sector).

π r^{2}

would be the flat base disk, a different thing entirely — see Frustum and cone geometry.

SE4. "Frustum lateral area is

2 π r_{1} ℓ

where

r_{1}

is the smaller radius."

You must use the average radius:

π (r_{1} + r_{2}) ℓ

. Using only the small radius undercounts the flaring band; the mean circumference is what wraps.

SE5. "For

y = r^{2} - x^{2}

rotated about the x-axis, I'll integrate from

0

r

and that's the whole sphere."

That covers only the hemisphere. The full sphere needs

x \in [- r, r]

; symmetry lets you double the

[0, r]

result instead, but you cannot just relabel the limit.

SE6. "By Pappus,

S = 2 π \overset{r}{ˉ} L

where

\overset{r}{ˉ}

is the centroid of the region."

For surface area

\overset{r}{ˉ}

is the centroid of the curve (arc), not the enclosed area. Confusing the two mixes up the surface theorem with the volume version — see Pappus's theorem.

SE7. "The curve crosses the axis at one point, so the surface has a hole there."

Where the curve meets the axis the radius is

0

, so that ring shrinks to a point — the surface simply closes to a smooth tip, no hole. Think of a sphere's poles.

Why questions

W1. Why is the slant length

d s

, not the flat

d x

, the correct "how long" factor?

The surface wraps along the tilted curve itself; a steep piece covers more surface than its short horizontal shadow suggests, and only

d s

measures that true tilted length.

W2. Why does the frustum formula use the average of the two radii?

Unrolling the band gives a shape whose width varies linearly from

2 π r_{1}

2 π r_{2}

, so its area is the mean circumference

2 π \frac{r _{1} + r _{2}}{2}

times the slant — exactly like averaging the two ends of a trapezoid.

W3. Why can

r_{1} \approx r_{2} \approx y

once we make the band infinitesimal?

Over a vanishingly short piece the curve barely changes height, so both edge radii collapse to the single value

y

; the tiny difference is higher-order small and drops out in the limit.

W4. Why is the radius factor "distance to the axis" and not just a coordinate?

A point sweeps a circle whose circumference is

2 π

times how far it is from the spinning axis; the coordinate only equals that distance when the axis happens to be the corresponding coordinate axis.

W5. Why does substitution appear so often in these integrals (e.g.

u = 1 + 4 x^{2}

The

1 + (y^{'})^{2}

often hides an inner function whose derivative already sits outside as the radius factor, so

u

-substitution collapses the whole thing to a clean

\int u d u

— see Integration by substitution.

W6. Why does the sphere derivation produce a constant integrand

r

The height

y = r^{2} - x^{2}

and the arc-slant

\frac{r}{r ^{2} - x ^{2}}

multiply so the awkward roots cancel, leaving

r

— a signature of the sphere's uniform curvature.

W7. Why do we treat parametric curves with

(d x / d t)^{2} + (d y / d t)^{2}

instead of

1 + (d y / d x)^{2}

When both coordinates move with

t

, arc length is Pythagoras on

d x

and

d y

jointly; the

1 + (y^{'})^{2}

form is just this after factoring out

d x

, which fails when

d x / d t = 0

(vertical tangent). See Parametric curves.

Edge cases

E1. What is the surface area contribution exactly where the curve touches the rotation axis?

Zero, because the radius is

0

there so

2 π r d s \to 0

; the ring degenerates to a single point and the surface smoothly closes to a tip.

E2. What happens to the integrand if the curve has a vertical tangent (

d y / d x \to \infty

In the

x

-form

1 + (y^{'})^{2}

blows up, signalling you should switch to integrating in

y

(use

1 + (d x / d y)^{2}

) where the slope is finite.

E3. Can the surface area be finite even if the curve stretches to infinity?

Yes — a curve on

[1, \infty)

(like Gabriel's Horn's boundary) can give a divergent surface while a related quantity stays finite; whether

S

converges depends on how fast

y d s

decays.

E4. What is the surface area of revolution of a single point?

Zero. A point has no arc length (

d s = 0

), so there is nothing to sweep — it traces at most a circle (a curve), which has area zero.

E5. If a curve lies entirely on the rotation axis, what surface does it sweep?

None with area — every point has radius

0

, so the whole thing spins into a line segment on the axis, contributing

2 π \cdot 0 \cdot d s = 0

E6. Rotating a straight vertical segment (constant

x = c > 0

y

from

0

h

) about the y-axis — what shape and area?

A cylinder of radius

c

, height

h

; here

d s = d y

and

S = \int_{0}^{h} 2 π c d y = 2 π c h

, the familiar "unrolled rectangle" lateral area.

E7. Rotating the same vertical segment about the x-axis instead — now what?

It sweeps a flat annular disk (a washer), not a lateral band; the surface-of-revolution lateral formula doesn't apply since the piece is parallel to the axis — treat it as a flat ring of area

π (x_{2}^{2} - x_{1}^{2})

E8. What is

S

when the curve is closed and symmetric, like a full circle rotated about a diameter?

You get a sphere; but you must integrate the whole boundary and watch for double-counting — the top and bottom semicircles trace the same surface, so use one of them over the full width.

Connections

Surface area of revolution — the parent note these traps drill.
Arc length — the $d s$ that makes the "slant not shadow" trap.
Volume of revolution (disk & shell) — the $π r^{2}$ vs $2 π r d s$ contrast in TF3.
Frustum and cone geometry — source of the average-radius rule (SE4, W2).
Parametric curves — the vertical-tangent edge case (E2, W7).
Integration by substitution — why $u$ -subs appear (W5).
Pappus's theorem — arc-centroid vs region-centroid (SE6).