Stokes' Theorem — Physica

Equivalent forms

\oint_C \mathbf{F}\cdot d\mathbf{r} = \int_S (\nabla\times\mathbf{F})\cdot\hat{\mathbf{n}}\, dA

\Gamma = \int_S \boldsymbol{\omega}\cdot d\mathbf{A}\quad(\text{circulation} = \text{flux of vorticity})

Green's, Gauss's, and the gradient theorem are all shadows of one statement — ∫_Ω dω = ∫_∂Ω ω — the generalized Stokes theorem of differential forms.

Symbol	Name	SI unit	Dimension	Range
$\Gamma$	Circulationoutput Line integral of F around the boundary loop	m^2/s	L^2 T^-1	-50 – 50
$\omega$	Curl strength Uniform (∇×F)·n̂ over the surface	1/s	T^-1	-3 – 3
$A$	Surface area Area of the bounded surface	m^2	L^2	0.1 – 30

Unit systems

Symbol	SI	CGS	Imperial
$A$	m^2	cm^2	ft^2
$\omega$	1/s	1/s	1/s
$\Gamma$	m^2/s	cm^2/s	ft^2/s

Where it holds

F continuously differentiable on an oriented, piecewise-smooth surface S with boundary curve

\partial S

.

Dimensional analysis

\oint F\cdot dr \to [F]\cdot [L]

\int (\nabla \times F)\cdot dA \to ([F]/[L])\cdot [L^{2}] = [F]\cdot [L]

Both sides carry field-units times length. With a velocity-like field,

\Gamma has

units

m^{2}/s

.

\checkmark

Discovery

Lord Kelvin / George Gabriel Stokes · 1854

Kelvin stated the result in an 1850 letter to Stokes, who then set it as a Cambridge exam problem in 1854 — which is how it came to bear Stokes's name. It generalizes Green's theorem into three dimensions and is the curl half of the fundamental theorems of vector calculus.

Try this

Walk the fence around a field and add up the wind pushing you along. That single lap tells you the total swirl of every gust inside — without stepping foot in the field.

Drive the curl across a surface and watch the circulation around its boundary loop rise to match it, exactly.

Research status: stable

Real-world applications

Ampère's circuital law in magnetostatics
Circulation and lift in aerodynamics (Kutta–Joukowski)
Vorticity transport in fluid dynamics
Berry phase and holonomy in quantum systems

Common misconceptions

The surface must be flat — any surface bounded by the loop works; the result is the same for all of them.
Orientation doesn't matter — the right-hand rule between loop direction and n̂ fixes the sign.
It needs the loop to be a circle — any closed boundary curve qualifies.

Experimental verification

Ampè

re's law \oint B\cdot dr = \mu _{0}I_enc

is Stokes' theorem applied

to \nabla \times B

; every solenoid and current-loop measurement confirms it. In aerodynamics, the Kutta–Joukowski lift theorem ties circulation to measured lift.

Derivation

Tile the surface with tiny loops.

Around each, circulation

= (\nabla \times F)\cdot n

̂ dA by the definition of curl.

Edges shared by adjacent tiles are traversed in opposite directions and cancel, leaving only the outer boundary.

Summing yields

\oint _\partial S F\cdot dr = \int _S (\nabla \times F)\cdot dA

.

Limiting cases

\nabla \times F = 0

(conservative)⟶

\Gamma = 0

Curl-free fields do no net work around any loop — the path doesn't matter.

Shrinking loop⟶

\Gamma

/

A \to (\nabla \times F)\cdot n

̂Recovers the definition of curl as circulation per unit area.

Closed surface (no boundary)⟶

\int (\nabla \times F)\cdot dA = 0

A boundaryless surface has no loop, so the total curl flux vanishes — consistent with

\nabla \cdot (\nabla \times F) = 0

.

What if…

What if two surfaces share the same boundary?

Both give the same circulation — the curl flux through any capping surface is identical.

What if the loop encircles a region where F isn't smooth?

Stokes can fail; the missing curl shows up as a singular contribution (e.g. a vortex line or a wire carrying current).

1

Circulation of a swirl

Given ·

F:: $(-y$ , x, 0)
loop:: circle radius $R = 2$

Find ·

\Gamma

Steps

$\nabla \times F = 2$ ẑ
Flux of curl $= 2\cdot$ (area $) = 2\cdot \pi R^{2}$
$= 2\cdot \pi \cdot 4 = 8\pi \approx 25.1\,\mathrm{m}^{2}/s$

Result ·

\Gamma = 2\cdot \pi R^{2} = 8\pi \approx 25.1

2

Direct line integral check

Given ·

F:: $(-y$ , x, 0)
loop:: circle radius $R = 2$

Find ·

\oint F\cdot dr

Steps

On the circle $F\cdot dr = R^{2}d\theta$
$\oint$ from 0 to $2\pi$ gives $2\pi R^{2} = 8\pi \checkmark$
Matches the curl-flux computation

Result ·

\oint F\cdot dr = \oint R^{2} d\theta = 2\pi R^{2}

...

= 8\pi

Curl→Green's Theorem→Divergence (Gauss) Theorem→Gradient→