Visual walkthrough — Volume of revolution — shell method

---

Step 0 — What are we even measuring?

WHAT. We have a flat region in the plane — a shape you could cut out of paper. We spin it around a straight line (the axis of rotation). As it spins, it sweeps out a 3D solid, like clay on a potter's wheel. We want that solid's volume — the amount of space it fills.

WHY start here. Every symbol later ($r$, $h$, $x$, $a$, $b$) is a measurement on this picture. If the picture isn't clear, the symbols are just marks.

PICTURE. The pale-yellow shaded region below is our paper shape. The chalk-blue dashed line is the axis. Follow one point on the region: as it spins it traces a circle around the axis.

---

Step 1 — Slice the region into thin vertical strips

WHAT. We chop the region into many tall, thin vertical rectangles. Pick one strip at horizontal position $x$. It has a tiny width we call $dx$ (read: "a little bit of $x$") and a height reaching from the $x$-axis up to the curve, so its height is $f(x)$.

WHY vertical, and why parallel to the axis? The axis is the $y$-axis — a vertical line. A strip that also stands vertical will, when spun, sweep out a clean tube wrapped neatly around the axis. If we'd sliced horizontally, one strip would sweep out a flat washer instead — that's the other method (disk/washer). Choosing strip direction is the whole art; see Choosing dx vs dy strips.

PICTURE. One highlighted strip in pink, width $dx$, height $f(x)$, sitting at distance $x$ from the axis.

$$\underset{\text{a thin rectangle}}{\underbrace{\text{strip}}}:\underset{\text{how far from axis}}{\underbrace{x}},\underset{\text{how tall}}{\underbrace{f(x)}},\underset{\text{how wide}}{\underbrace{dx}}$$

---

Step 2 — Spin ONE strip: it becomes a shell

WHAT. Take the single pink strip and spin it a full turn around the axis. It sweeps out a thin hollow tube — a cylindrical shell, like a tin can with no top or bottom.

WHY one strip at a time? Volume is additive: if we can find the volume of one shell, we just add them all up. Divide-and-conquer. This is the same trick as areas — measure one strip, then sum (Area between two curves).

PICTURE. The flat pink strip and, beside it, the tube it becomes. Label its three measurements:

• radius $r=x$ — the strip was at distance $x$ from the axis, so the tube's wall is that far out.
• height $h=f(x)$ — the strip's height is unchanged by spinning.
• thickness $dx$ — the strip's width is unchanged too.

---

Step 3 — Unroll the shell into a flat sheet

WHAT. Slit the tube straight down one side and flatten it out. It becomes a thin rectangular slab — a flat sheet.

WHY does this help? We don't have a volume formula for a "tube" memorised, but we absolutely have one for a box: length × height × thickness. Unrolling converts an unfamiliar shape into a box we can measure. This is the key idea of the whole method.

PICTURE. The tube slit open and laid flat. Watch what each measurement becomes:

• the tube's height $h$ → the sheet's height (unchanged),
• the tube's thickness $dx$ → the sheet's thickness (unchanged),
• the tube's circumference → the sheet's width (this is the new part!).

WHY the width is a circumference. The sheet's width is "how far you'd walk going once around the tube." Walking once around a circle of radius $r$ is exactly its circumference.

---

Step 4 — Where does $2\pi r$ come from?

WHAT. The width of the flattened sheet is the distance once around the circle of radius $r$. That distance is $2\pi r$.

WHY $2\pi r$ and not $\pi r$ or $\pi r^2$?

• $\pi r^2$ is an area (a filled disk). We don't want the filled disk — we're walking the rim, not paving the floor.
• $\pi r$ is half the way around (a semicircle). We spin the strip a full turn, so we go all the way: $2\pi r$.

Think of $\pi$ as "how many diameters fit around a circle" — but a diameter is $2r$, so once around is $\pi \times (2r)=2\pi r$.

PICTURE. A circle of radius $r$ with its rim "unbent" into a straight segment of length $2\pi r$.

---

Step 5 — Is unrolling exactly right, or only approximately?

WHAT. A worry: when we flatten a tube, the inner wall (radius $r$) is a bit shorter around than the outer wall (radius $r+dx$). So which circumference did we use? Let's check the flattening isn't cheating.

WHY this matters. If unrolling introduced an error that didn't vanish, our formula would be wrong. We must prove the error disappears.

HOW — exact volume of the shell. A real shell is the space between two solid cylinders: an outer one of radius $r+dx$ and an inner one of radius $r$, both of height $h$. Cylinder volume is $\pi (\text{radius}{)}^2\times \text{height}$, so:

$$dV=\underset{\text{outer solid cylinder}}{\underbrace{\pi (r+dx{)}^{2}h}}-\underset{\text{inner solid cylinder}}{\underbrace{\pi r^{2}h}}$$

Expand $(r+dx{)}^2=r^2+2rdx+d{x}^2$:

$$dV=\pi h(r^2+2rdx+d{x}^2-r^2)=\pi h(2rdx+d{x}^2)=\underset{\text{our formula}}{\underbrace{2\pi rhdx}}+\underset{\text{the leftover}}{\underbrace{\pi hd{x}^2}}$$

PICTURE. The two term sizes shown as areas: $2rdx$ is a long thin strip; $d{x}^2$ is a tiny square. As $dx$ shrinks, the tiny square shrinks far faster (its side shrinks in both directions).

WHY the leftover vanishes. The term $\pi hd{x}^2$ has $dx$ squared. If $dx=0.01$, then $d{x}^2=0.0001$ — a hundred times smaller. As $dx\to 0$ the leftover dies far faster than the term we keep. So in the limit: $dV=2\pi rhdx✓\text{ (exact)}$ This is why unrolling is legitimate: the "inner vs outer circumference" difference lives entirely in the $d{x}^2$ term, which we throw away.

---

Step 6 — Add up all the shells: the integral

WHAT. We have one shell's volume $dV=2\pi xf(x)dx$. Now stack every shell from the innermost ($x=a$) to the outermost ($x=b$) and add them. "Add up infinitely many infinitely-thin pieces" is precisely what $\int$ means.

WHY an integral and not ordinary addition? Each shell is infinitely thin, so there are infinitely many of them. Plain "+" can't sum infinitely many terms; the integral is the machine built exactly for "sum of shrinking pieces" — see Definite integral as a limit of Riemann sums.

PICTURE. Many nested shells filling the solid, radius growing from $a$ (thin inner tube) to $b$ (wide outer tube).

Sanity number. Region under $y=x^2$ from $x=0$ to $x=2$, about the $y$-axis: $V={\int }_0^{2}2\pi x\cdot x^{2}dx=2\pi {\int }_0^2{x}^{3}dx=2\pi {\left[\frac{x^4}{4}\right]}_0^2=2\pi \cdot 4=8\pi .$

---

Step 7 — Edge cases: don't get ambushed

The pretty first picture always uses $r=x$, $h=f(x)$, with the axis on the left and the region on the right. Real problems break each of those. Here is every case.

Case A — the axis is a different vertical line $x=c$. Then the radius is not $x$ — it's the distance from the strip to that line, $|x-c|$. The bars $|\cdot |$ ("absolute value") mean "make it positive": a distance can't be negative.

• Axis on the right of the region ($c>x$): radius $=c-x$.
• Axis on the left ($c<x$): radius $=x-c$.

Case B — the region is between two curves. The height is no longer $f(x)$; it's top minus bottom, $y_{\text{top}}-y_{\text{bot}}$. Same "top − bottom" idea as Area between two curves.

Case C — a degenerate strip: the axis passes through the region. If the axis cuts the region, strips on one side give shells that overlap shells from the other side — you'd double-count. The safe rule: split the integral at the axis and handle each side separately, each with its own positive radius $|x-c|$.

Case D — a zero-height or zero-radius shell.

• At a point where $f(x)=0$, the strip has height $0$, so $dV=0$: it contributes nothing. Fine.
• At $x=c$ (a shell on the axis), radius $=0$, so $dV=0$: a shell of zero radius is just a line — no volume. Also fine. The formula handles these automatically; no special code needed.

PICTURE. Four small boards, one per case, each showing the strip, the axis, and the correct radius/height labelled.

---

The one-picture summary

This single board is the whole derivation: strip → spin → unroll → box volume → sum. If you can redraw this from memory, you own the shell method.

$$\underset{x,f(x),dx}{\underbrace{\text{strip}}}\stackrel{\text{spin}}{\to }\underset{r,h,dx}{\underbrace{\text{tube}}}\stackrel{\text{unroll}}{\to }\underset{2\pi r,h,dx}{\underbrace{\text{sheet}}}\stackrel{\text{box vol}}{\to }dV=2\pi rhdx\stackrel{\int }{\to }V={\int }_a^{b}2\pi xf(x)dx$$

Recall Feynman retelling — the walkthrough in plain words

I take my flat shape and cut it into loads of skinny vertical ribbons. I grab one ribbon sitting a distance $x$ out from the spin-line; it's $f(x)$ tall and a hair $dx$ wide. I spin it once around and it turns into a thin tin-can wall. I slit the can down the side and flatten it — now it's a flat sheet. How wide is the sheet? It's how far I'd walk going once around the can, which for a circle of radius $x$ is $2\pi x$. So the sheet is $2\pi x$ wide, $f(x)$ tall, and $dx$ thick — a box — so its volume is $2\pi x\cdot f(x)\cdot dx$. I do this for every ribbon, from the innermost ($x=a$) to the outermost ($x=b$), and add every box up. "Add infinitely many thin things" is the integral, so the total is ${\int }_a^{b}2\pi xf(x)dx$. The only sneaky bit is proving flattening is fair: the exact can-volume has one extra $d{x}^2$ crumb, but squaring a tiny number makes it vanish, so we can ignore it. Done.

---

Recall checkpoint

Recall Test yourself

Where does the $2\pi$ in the shell formula come from? ::: Unrolling the tube gives a flat sheet whose width is the shell's circumference, $2\pi r$ (once fully around a circle of radius $r$). Why can we drop the $d{x}^2$ term in the exact derivation? ::: It shrinks far faster than $dx$ as $dx\to 0$ (squaring a small number makes it tiny), so it vanishes in the limit. What is the radius when rotating about $x=c$? ::: $|x-c|$, the distance from the strip to the axis — not $x$. Why slice parallel to the axis for shells? ::: A strip parallel to the axis sweeps out a clean tube when spun; a perpendicular strip would sweep a washer (the disk method instead). Height for a region between two curves? ::: $y_{\text{top}}-y_{\text{bot}}$. Volume of $y=x^2$, $0\le x\le 2$, about the $y$-axis? ::: $8\pi$.

---

Connections

• Volume of revolution — shell method (Hinglish) — the parent topic, all worked examples.
• Volume of revolution — disk and washer method — the perpendicular-slice cousin.
• Definite integral as a limit of Riemann sums — why "sum infinitely many shells" is legal.
• Area between two curves — the "top − bottom" height reused.
• Choosing dx vs dy strips — the strip-orientation decision.
• Arc length and surfaces of revolution — revolve a curve to get a surface, not a solid.