Green's Theorem

Why boundaries and interiors are connected

Think of water flowing in a river. If you dropped a little paddlewheel at some point, the flow would spin it - faster in some places, slower in others. That local spinning is what we'll call the scalar curl.

Now imagine a closed loop of fence posts in the river. The total flow pushing along that fence - the line integral around the loop - should somehow reflect all the little spinnings happening inside the loop. That's the core intuition here. Let's make it precise.

Picture the region $R$ chopped into a grid of tiny squares. Around each tiny square, compute the line integral - it's approximately $(\text{curl at the center}) \times (\text{area of the square})$. Now add them all up.

Here's the beautiful part: on every interior edge, two adjacent squares share a boundary but traverse it in opposite directions. These contributions cancel perfectly. The only edges that survive are the ones on the outer boundary.

Click through the views: tiny circulations → interior edges cancel → only the outer boundary $C$ survives.

So the sum of all tiny circulations = the line integral around the boundary. And the sum of (curl $\times$ area) over all tiny squares = the double integral of the curl over $R$. Since these are the same sum, we get:

That's Green's theorem. The left side is a line integral around the boundary $C$ (counterclockwise). The right side is a double integral of the scalar curl $\frac{\partial Q}{\partial x} - \frac{\partial P}{\partial y}$ over the enclosed region $R$. The boundary tells you everything about what's happening inside.

The scalar curl: rotation you can see

The quantity $\frac{\partial Q}{\partial x} - \frac{\partial P}{\partial y}$ shows up everywhere in Green's theorem, so let's build real intuition for what it measures. For a field $\mathbf{u} = P\,\mathbf{i} + Q\,\mathbf{j}$, the scalar curl tells you: how much is the field spinning at this point?

Let's look at three fields on the unit circle. For each one, we compute the line integral around $C$ and the double integral of the curl over $R$. They should always match.

$\mathbf{u} = (-y, x)$: Pure rotation. Every point spins the same way, curl $= 2$ everywhere. Total circulation $= 2 \times \pi(1)^2 = 2\pi$.

$\mathbf{u} = (x, y)$: Pure outward push, no spinning at all. Curl $= 0$ everywhere. Circulation $= 0$, no matter the curve.

$\mathbf{u} = (x^2, xy)$: Now it gets interesting - curl $= y$. The field spins counterclockwise where $y > 0$ and clockwise where $y \lt 0$. These partly cancel, and the double integral tells us exactly how much net circulation remains.

Area as a line integral

Here's a genuinely surprising consequence of Green's theorem. Choose $P = -\tfrac{y}{2}$ and $Q = \tfrac{x}{2}$. Then:

$$\frac{\partial Q}{\partial x} - \frac{\partial P}{\partial y} = \frac{1}{2} - \left(-\frac{1}{2}\right) = 1$$

So the double integral side becomes $\iint_R 1\,dA$ - which is just the area of $R$. That gives us:

Let that sink in. You can compute the area of any shape just by walking around its boundary. You never need to look inside. This is exactly how a planimeter works - a mechanical device that measures area by tracing a curve. And it's the idea behind the shoelace formula for polygon areas: just sum up $\frac{1}{2}\sum (x_i y_{i+1} - x_{i+1} y_i)$ around the vertices.

Pick a shape and watch the boundary integral compute its area:

When Green's theorem breaks

Green's theorem has three requirements, and violating any of them can give you wrong answers. Let's name them clearly.

Requirement 1: closed curve. The theorem relates a line integral around a closed loop to a double integral over the enclosed region. No closed curve, no enclosed region, no theorem.

Requirement 2: counterclockwise orientation. The region $R$ must sit to the left as you walk along $C$. If you traverse clockwise, you get a minus sign. Easy fix - just check your orientation - but easy to forget.

Requirement 3: $P$ and $Q$ must be smooth inside $R$. This is where things get genuinely tricky. Consider the field:

$$\mathbf{u} = \frac{-y\,\mathbf{i} + x\,\mathbf{j}}{x^2 + y^2}$$

Compute its curl: $\frac{\partial Q}{\partial x} - \frac{\partial P}{\partial y} = 0$ everywhere - except at the origin, where the field blows up. So if you naively apply Green's theorem on the unit circle, you'd predict circulation $= \iint_R 0\,dA = 0$.

But the actual line integral around the unit circle gives $2\pi$. The answer is not zero.

What went wrong? The origin is a singularity - a point where $P$ and $Q$ aren't defined, let alone smooth. Green's theorem never promised to work when the field has a hole in it.

The fix for multiply connected domains (regions with holes): include all boundary curves. The outer boundary goes counterclockwise, but inner boundaries around holes go clockwise. The contributions from the inner boundaries account for the missing pieces. Kaplan covers this in detail in §5.7.