A visual essay on calculus, from circle peeling to the fundamental theorem
Previously: It's Just Comparing — a visual essay on trigonometry
Here is a circle. You know its area is $\pi r^2$. But why?
Not "because a textbook said so." Why does the area of a circle — this perfectly round thing with no straight edges — boil down to a formula involving the square of its radius?
Let me show you something. Take that circle and slice it into thin concentric rings, like a target at a shooting range. Now peel each ring open and lay it flat. The outermost ring is the longest — its length is the full circumference. The innermost ring is tiny. And when you lay them all out, they form a shape you already know how to measure.
That shape is a triangle. Its base is $2\pi r$ (the circumference), its height is $r$ (the radius), and its area is $\frac{1}{2} \cdot 2\pi r \cdot r = \pi r^2$.
Try it yourself. Use the slider below to increase the number of rings, and watch the staircase of unrolled rings converge to a perfect triangle.
What did the rings become when you unrolled them? Rectangles. And what did the collection of rectangles become? A triangle. Each ring was too curved to measure directly. But thin enough, and it stopped being curved.
That is the core move of calculus: slice something curved into pieces so thin that each piece becomes simple. A ring so thin it is basically a rectangle. A curve so zoomed-in it is basically a line.
The number of rings matters. With 4 rings, the staircase is rough — you can see the gaps between the steps and the triangle. With 20, it is close. With 100, the staircase is the triangle, for all practical purposes.
Slice a circle into concentric rings, unroll them, and they form a triangle with area $\frac{1}{2} \cdot 2\pi r \cdot r = \pi r^2$.
This trick — replacing a curved shape with many thin straight pieces — is older than the word "calculus" itself. Archimedes used it to find the area of a circle over two thousand years ago. But the insight is timeless: thin slices reveal hidden simplicity.
Before we move on, notice something odd. The area of a circle is $\pi r^2$. Its circumference is $2\pi r$. If you know a little calculus — or even if you don't — look at what happens when you ask "how fast does the area grow as you increase $r$?" The answer is $2\pi r$. The circumference. The rate at which a circle's area grows equals its boundary. Is that a coincidence, or a clue to something deeper? Hold that thought. We will come back to it.
And that raises a question. Why does zooming in make curved things look straight? What is it about curves that changes when you look closely enough?
The Earth looks flat from your backyard. That is not an illusion — it is calculus.
Imagine zooming into the Earth at any point. At 1x, you see mountains and coastlines. At 10x, a hillside. At 100x, a patch of grass. At 1000x, it is a plane. The curvature is still there, but at that scale it has become indistinguishable from a flat surface.
The same thing happens with any smooth curve. Pick a point on it. Zoom in. Keep zooming. The curve bends less and less relative to your field of view until, eventually, it looks perfectly straight.
Look at the curve below. How many times do you think you need to zoom in (2x each time) before it looks indistinguishable from a straight line?
It takes surprisingly few zooms. Even a wildly curvy function straightens out after just 4 or 5 doublings of magnification. The curve does not slowly become straight — it snaps to straight remarkably fast.
Drag the slider to zoom into the marked point. The curve straightens into a line — that line is the tangent.
This is the foundational observation of calculus. Curved things look straight when you zoom in far enough. Not approximately straight. Not "sort of" straight. At sufficient magnification, the curve and its tangent line become visually identical.
And if a curve looks like a line when you zoom in, then at that tiny scale, it behaves like a line. Lines have a slope. So curves, at each point, have a slope too.
But how do you measure the slope of something that is, in fact, curved? You cannot just "rise over run" a curve the way you can a straight line. Or can you?
Your speedometer tells you your speed right now. How is that even possible?
Speed is distance divided by time. But "right now" is a single instant — zero time passes, zero distance is covered. Zero divided by zero is meaningless. And yet your speedometer gives you a perfectly definite number. Sixty miles per hour. Right now.
The trick is exactly what you just saw. Your position over time traces a curve. Your speed at any instant is the slope of that curve at that point. And you find that slope by zooming in until the curve looks like a line, then measuring the line's slope.
More precisely: pick two nearby times. Compute the average speed between them (distance traveled divided by time elapsed). Now bring the two times closer together. The average speed changes slightly. Keep bringing them closer. The average speeds converge to a single number. That number is the instantaneous speed — the derivative.
The tangent line on the left shows the slope at each point. The right panel traces the derivative — the slope as a function of position.
Watch what happens at the peaks and valleys of the curve. The tangent line goes flat — horizontal — and the derivative crosses zero. At a peak, the curve stops rising and starts falling. At a valley, it stops falling and starts rising. In both cases, for one instant, it is doing neither: the slope is zero.
This is not a coincidence. It is the definition of a peak: a point where the slope changes from positive to negative. The derivative captures this transition precisely.
We write the derivative of $f(x)$ as $f'(x)$, and it means exactly what you just saw: the slope of $f$ after zooming in at $x$ until the curve is straight.
So the derivative is just the slope after zooming in. But slopes have patterns. When you zoom into $x^2$ at different points, the slopes follow a rule. Can you see it?
If the derivative is just the slope after zooming in, then the "rules of calculus" are just descriptions of what happens to slopes when you do arithmetic.
Start with the simplest curves: powers of $x$. Below, you can zoom into $x^2$ at any point and read off the slope. Try a few points and look for a pattern.
Click points on the curve to record their slopes. Can you spot the pattern?
At $x = 1$, the slope of $x^2$ is $2$. At $x = 2$, it is $4$. At $x = 3$, it is $6$. The slope of $x^2$ at any point $x$ is $2x$.
Switch to $x^3$. At $x = 1$, the slope is $3$. At $x = 2$, it is $12$. The slope of $x^3$ at $x$ is $3x^2$.
The pattern: the derivative of $x^n$ is $nx^{n-1}$. Pull the exponent down in front, reduce it by one. This is the power rule, and it is not arbitrary — it follows from what happens when you zoom in on $x^n$ and measure the slope.
$$\frac{d}{dx} x^n = n \cdot x^{n-1}$$
The derivative is the limit of the slope between two nearby points. For $x^2$:
$$f'(x) = \lim_{h \to 0} \frac{(x+h)^2 - x^2}{h}$$Expand $(x+h)^2 = x^2 + 2xh + h^2$ and cancel:
$$= \lim_{h \to 0} \frac{x^2 + 2xh + h^2 - x^2}{h} = \lim_{h \to 0} \frac{2xh + h^2}{h} = \lim_{h \to 0} (2x + h) = 2x$$The same approach works for any power. For $x^3$, expanding $(x+h)^3$ and canceling gives $3x^2$. The pattern holds: the binomial expansion of $(x+h)^n$ always leaves $nx^{n-1}$ as the surviving term.
The other rules of calculus follow the same logic. The product rule, the chain rule — each one describes how zooming in interacts with a particular arithmetic operation. They are not formulas to memorize. They are consequences of one idea: curved things look straight when you zoom in.
We have been measuring slopes — how fast things change at a point. But what about the opposite question? Instead of zooming in to find a slope, what if you want to add up a bunch of zoomed-in pieces to find a total?
You want to know the area under a curve. The curve is not a rectangle, so you cannot just multiply base times height. But you know the move by now: slice it into thin pieces.
Each thin vertical slice is almost a rectangle. Its width is some small number $\Delta x$, and its height is the value of the curve, $f(x)$. Multiply width times height to get the area of that slice. Add up all the slices, and you get an approximation to the total area.
The approximation gets better as the slices get thinner. With 4 rectangles, it is rough. With 40, it is close. With 400, you cannot see the error.
The exact area under $x^2$ from $0$ to $1$ turns out to be a surprisingly clean number. How many equal-width rectangles do you think it takes to get within 1% of it?
Just 18 rectangles gets you within 1% of the true area. The convergence is much faster than most people expect — the rectangles do not need to be that thin before the sum is remarkably accurate.
Slide to add rectangles. The sum of their areas converges to the true area under the curve: $\int_0^1 x^2\, dx = \frac{1}{3}$.
This process of adding up thin slices is called integration. The symbol $\int$ (a stretched "S" for "sum") and the notation $\int_a^b f(x)\,dx$ mean exactly what you just did: slice the region from $a$ to $b$ into thin rectangles of width $dx$, each with height $f(x)$, and add up their areas.
$$\int_a^b f(x)\,dx = \lim_{n \to \infty} \sum_{i=1}^{n} f(x_i) \cdot \Delta x$$
So we have two operations. The derivative zooms in to find a slope. The integral zooms in to find an area (by making each slice so thin it becomes a rectangle). They seem like entirely different things — one is about rates, the other about totals.
And yet they are connected. Deeply, beautifully, inevitably connected.
Watch both panels below. On the left is a curve $f(x)$. On the right, the shaded area under $f$ accumulates as you sweep right — this accumulated area is a new function, $F(x)$.
As you sweep, pay attention to the slope of $F(x)$. Something controls how steep it is at each point. What is it?
As you sweep right, what determines how steep $F(x)$ is at each point?
It is the height of $f(x)$. Where $f$ is tall, the area accumulates quickly and $F$ is steep. Where $f$ is small, the area grows slowly and $F$ is nearly flat. The slope of the area function at any point equals the height of the curve at that point.
The slope of $F(x)$ always equals the height of $f(x)$. Derivatives and integrals are inverses.
You just watched the slope of $F(x)$ track the height of $f(x)$ in real time. Where $f$ is tall, $F$ is steep. Where $f$ is small, $F$ is flat. The slope of the accumulated-area function at any point equals the height of the original curve at that same point.
$$F'(x) = f(x)$$
The derivative of the integral gives you back the original function. Differentiation and integration are inverse operations. This is the Fundamental Theorem of Calculus.
Why does it work? Because when you add a thin slice of area of width $dx$ and height $f(x)$, the total area $F(x)$ increases by $f(x) \cdot dx$. So the rate of increase of $F$ — its derivative — is $\frac{f(x) \cdot dx}{dx} = f(x)$.
That is the whole proof. The area grows at a rate equal to the height of the curve. The "hard" theorem of calculus is, in the end, almost obvious once you see it from the right angle.
So calculus gives you two tools: zoom in to find a slope, zoom in to find an area, and the two tools are inverses. But what happens when instead of computing a slope, someone tells you the slope at every point and asks you to find the curve?
Physics gives you the slopes. The universe traces the curve.
A falling object accelerates at $9.8$ meters per second per second. That is a statement about the derivative of velocity — which is itself the derivative of position. A population grows at a rate proportional to its size. That is a statement about the derivative of population.
In each case, you are not given the function. You are given a rule about its slope at every point. An equation involving a function and its derivatives is called a differential equation, and solving it means finding the function whose slopes match the given rule.
The interactive below shows a slope field: at every point in the plane, a short line segment shows the slope that any solution must have at that point. Click anywhere to drop a starting point and watch the solution curve trace itself, following the arrows like a ball rolling downhill.
Each tiny arrow shows the slope prescribed by the equation. Click to drop a starting point and watch the solution follow the arrows.
Notice how different starting points can lead to wildly different curves, even though they all follow the same slope rule. The differential equation constrains the direction at every point, but the starting position determines which curve you trace. One equation, infinitely many solutions — each determined by where you begin.
This is how most of physics works. Newton's laws, Maxwell's equations, the Schrodinger equation — they are all differential equations. They do not tell you the state of the world directly. They tell you the slopes. And from those slopes, the world's trajectory emerges.
One equation, infinitely many solutions. But there is a question we left unanswered at the very beginning — about a circle, its rings, and a coincidence involving the circumference.
Let's go back to where we started. The circle, the rings, the triangle.
When you peeled the circle into concentric rings and unrolled them, you were integrating. Each ring at radius $r$ has circumference $2\pi r$ and thickness $dr$, so its area is $2\pi r \cdot dr$. Adding up all the rings from $r = 0$ to $r = R$:
$$\int_0^R 2\pi r\, dr = \pi R^2$$
That is the area of a circle, derived by integration — by zooming in until each ring is a rectangle, computing its area, and summing. The visual argument you saw in Section 1 was calculus the whole time. You just did not need the notation to see it.
And notice the integrand: $2\pi r$ is the circumference of a circle of radius $r$. The derivative of $\pi r^2$ (the area) with respect to $r$ is $2\pi r$ (the circumference). The rate at which a circle's area grows as you expand its radius equals the circumference at that radius. This is the Fundamental Theorem in action: the derivative of the accumulated area gives back the "height" of each ring.
This is calculus. Not a collection of formulas. Not a set of rules to memorize. One idea, applied twice: curved things look straight when you zoom in. The derivative is the slope after zooming in. The integral is the area computed by zooming in until each piece is a rectangle. And the Fundamental Theorem says these two operations — finding slopes and finding areas — are inverses of each other.
The circumference is the derivative of the area. The surface area of a sphere is the derivative of its volume. Is that a coincidence? Or is zooming in revealing something about how shapes grow?
Every extension of calculus — into higher dimensions, across surfaces and volumes, into the language of linear algebra — rests on the same foundation. You zoom in. The curved thing becomes straight. And then you measure.