Calculus › Section P · Vector analysis › Topic P.4
Green's theorem
A closed curve's circulation and its interior's spin are two views of one number. Green's theorem converts between them — and its flux face turns the boundary walk into an area measurement.
From a curve to its interior
Green's theorem trades a boundary integral for a region integral — a bargain in both directions. If \(C\) is a positively oriented, simple, closed, piecewise-smooth curve enclosing a region \(R\), and \(M\) and \(N\) have continuous partial derivatives on all of \(R\), then
Positive orientation means counter-clockwise: walk the boundary and the region stays on your left. The density \(\frac{\partial N}{\partial x} - \frac{\partial M}{\partial y}\) is the planar spin density — the z-component of \(\operatorname{curl}\mathbf{F}\) from the vector-fields topic — so the theorem reads: circulation around the boundary equals total spin inside. Every closed-loop line integral is now double-integral material, and vice versa.
Worked example: the swirl again
The line-integrals topic computed by hand that \(\mathbf{F} = \langle -y, x\rangle\) earns \(2\pi\) per lap of the unit circle. Green's theorem reproduces it in one line: with \(M = -y\) and \(N = x\), the spin density is \(1 - (-1) = 2\), so
The spin density is a constant 2 — the field rotates uniformly everywhere in the disk — and integrating it over the disk counts exactly what the parametrised dot product counted before. One theorem, two routes, zero trigonometry.
Worked example: a square's work
Compute \(\oint_C 2y\,dx + 3x\,dy\) around the counter-clockwise boundary of the square \(0 \le x \le 2\), \(0 \le y \le 2\). Green: the density is \(\frac{\partial(3x)}{\partial x} - \frac{\partial(2y)}{\partial y} = 3 - 2 = 1\), so the work is \(\iint_R 1\,dA\), the area, 4. Second route, straight perimeter arithmetic: the bottom edge (\(y = 0\), \(dy = 0\)) contributes 0; the right edge (\(x = 2\), \(dx = 0\)) contributes \(\int_0^2 6\,dy = 12\); the top edge (\(y = 2\), walked right-to-left) contributes \(\int_2^0 4\,dx = -8\); the left edge (\(x = 0\)) contributes 0. The edges total \(0 + 12 - 8 + 0 = 4\). The perimeter took four integrals; Green took one.
Area from the boundary
Choose \(M\) and \(N\) so that \(\frac{\partial N}{\partial x} - \frac{\partial M}{\partial y} = 1\) and Green's theorem reads 'boundary integral = area'. Three standard choices: \(M = 0, N = x\) gives \(\text{area} = \oint_C x\,dy\); \(M = -y, N = 0\) gives \(\text{area} = \oint_C (-y)\,dx\); and their average \(\text{area} = \tfrac{1}{2}\oint_C (x\,dy - y\,dx)\) is the symmetric favourite. Any of them turns a boundary walk into an area measurement. The drafting tool called a planimeter works on exactly this principle: its wheel registers the tangential distance it rolls while the stylus traces an outline, and Green's theorem converts that boundary data into the enclosed area — assuming uniform, friction-free rolling, which is the tool's engineering assumption.
Worked example: an ellipse's area
Test drive the area formula on the ellipse \(\frac{x^2}{a^2} + \frac{y^2}{b^2} = 1\), parametrised \(x = a\cos t\), \(y = b\sin t\) for \(0 \le t \le 2\pi\). Compute \(x\,dy - y\,dx\): \(x\,dy = ab\cos^2 t\,dt\) and \(y\,dx = -ab\sin^2 t\,dt\), so \(x\,dy - y\,dx = ab(\cos^2 t + \sin^2 t)\,dt = ab\,dt\). The area is \(\tfrac{1}{2}\int_0^{2\pi} ab\,dt = \pi ab\). Second route, pure double integral: the substitution \(x = au\), \(y = bv\) is a linear stretch with Jacobian \(ab\), so the ellipse's area is \(ab\) times the unit disk's area, \(ab\pi\). Two routes, one famous formula — and for \(a = b = 1\) the boundary integral is worth \(2\pi\), the flux computation below in disguise.
The other face: flux
Green's theorem has a second face. For \(\mathbf{F} = \langle M, N\rangle\) and \(\mathbf{n}\) the outward unit normal to \(C\),
Outflow through the boundary equals total divergence inside. The link is the normal element: along a counter-clockwise \(\mathbf{r}(t) = \langle x(t), y(t)\rangle\), rotating the tangent clockwise gives \(\mathbf{n}\,ds = \langle y'(t), -x'(t)\rangle\,dt\), so \(\mathbf{F} \cdot \mathbf{n}\,ds = M\,dy - N\,dx\), and applying the circulation form to \(\langle -N, M\rangle\) delivers the flux form. Mini check: \(\mathbf{F} = \langle x, y\rangle\) on the unit circle has flux \(\oint (x\,dy - y\,dx) = 2\pi\), matching \(\iint_R 2\,dA = 2\pi\). This is the divergence theorem in two dimensions; its 3D sibling ends the chapter.
Regions with holes
A region with a hole — an annulus, say — has a boundary in two pieces, and Green's theorem wants both: the outer curve walked counter-clockwise and the inner curve walked clockwise, each keeping the region on the left. With that convention the theorem holds unchanged. Test it on the punctured-plane field \(\langle\frac{-y}{x^2+y^2}, \frac{x}{x^2+y^2}\rangle\), whose spin density is zero everywhere it is defined: around any circle of radius \(r\) its tangent push integrates to \(2\pi\), so the outer boundary of the annulus contributes \(2\pi\) and the inner boundary, walked clockwise, contributes \(-2\pi\) — netting zero, exactly the spin integral over the annulus. What fails is applying Green to the full unit disk: the field is not smooth at the origin, so the disk is not a legal region — the same fine print the conservative-fields topic met at the same hole.
Common mistakes to avoid
First, orientation: a clockwise boundary negates the boundary integral, so reverse-walk it or negate before invoking the theorem — the region must sit on the left. Second, wrong density: the work form wants \(\frac{\partial N}{\partial x} - \frac{\partial M}{\partial y}\); \(\frac{\partial M}{\partial x}\) and \(\frac{\partial N}{\partial y}\) belong to the flux form, not this one. Third, applying Green where the field misbehaves inside: the punctured-plane field is smooth on the annulus but not on the disk, and the theorem notices the difference. Fourth, half a boundary: an annulus walked only on its outer circle is missing its inner piece, clockwise. Fifth, sign slips in perimeter checks: top and left edges run against their natural variable direction, so their \(dx\) or \(dy\) is negative — the classic spot where direct verification goes wrong.
Key takeaways
Green's theorem: for a positively oriented simple closed curve \(C\) enclosing \(R\), \(\oint_C M\,dx + N\,dy = \iint_R (N_x - M_y)\,dA\) — boundary circulation equals total interior spin, provided the field is smooth on all of \(R\). It collapses perimeter work into single double integrals — the square's work for \(\langle 2y, 3x\rangle\) is just \((3 - 2) \times 4\) — reproduces the swirl's \(2\pi\) as spin density 2 times area, and measures area from the boundary: \(\text{area} = \tfrac{1}{2}\oint (x\,dy - y\,dx)\), giving \(\pi ab\) for the ellipse. Its second face, \(\oint \mathbf{F} \cdot \mathbf{n}\,ds = \iint \operatorname{div}\mathbf{F}\,dA\), is the 2D divergence theorem; hole-y regions need every boundary piece, outer counter-clockwise and inner clockwise. The chapter's remaining theorems are this one, stretched into three dimensions.
Flashcards
Flashcard
∮_C M dx + N dy = ∬_R (N_x − M_y) dA
Test your understanding
Question 1 of 5
Green's theorem converts the closed-loop work integral ∮_C M dx + N dy into which double integral?
Question 2 of 5
By Green's theorem, what is the work of F = ⟨−y, x⟩ around the unit circle, counter-clockwise?
Question 3 of 5
Using area = (1/2)∮(x dy − y dx), what is the area enclosed by the ellipse x = 2 cos t, y = sin t, 0 ≤ t ≤ 2π?
Question 4 of 5
What is ∮_C 2y dx + 3x dy around the counter-clockwise boundary of the square 0 ≤ x ≤ 2, 0 ≤ y ≤ 2?
Question 5 of 5
Which orientation does Green's theorem require of the boundary curve C?
Standards: Beyond AP
