Stokes' theorem

Intuition [Beginner]

Stokes' theorem is the high-dimensional generalisation of the fundamental theorem of calculus. It says: integrating a form over the boundary of a region equals integrating its exterior derivative over the region itself.

This single identity unifies several classical theorems of vector calculus: the fundamental theorem of calculus (in one dimension), Green's theorem (in two), the classical Stokes theorem about line integrals around a closed curve equaling a flux integral over a bounded surface (in three), and the divergence theorem (in three, fluxes through closed surfaces equal volume integrals of divergence).

Stokes' theorem is the geometric structural identity behind much of differential calculus on manifolds and behind the entire de Rham cohomology theory.

Visual [Beginner]

A 2-form integrated over a region equals a 1-form integrated around the region's boundary. Visualise the boundary curve and the area enclosed; the theorem relates the two integrals.

The picture works in any dimension, with $k$-dimensional boundaries of $(k+1)$-dimensional regions — and with $k$-forms.

Worked example [Beginner]

In one dimension, Stokes' theorem applied to an interval (call it from $a$ to $b$) recovers the fundamental theorem of calculus. The "boundary" of the interval consists of just two points — its endpoints, with the right one counted positively and the left one counted negatively. For a smooth function $f$, the rule "integrate $f^{\prime }$ over the interval" equals "evaluate $f$ at the right endpoint minus the left endpoint." That's exactly the fundamental theorem.

In two dimensions, Stokes specialises to Green's theorem: integrating the curl of a vector field over a 2D region equals integrating the field tangentially around the boundary curve.

In three dimensions, Stokes specialises to two classical theorems. Integrating the curl of a vector field over a 2D surface equals integrating the field around the bounding curve (the "classical Stokes theorem"). Integrating the divergence of a vector field over a 3D region equals integrating the field's normal flux through the bounding surface (the divergence theorem).

The same single theorem becomes the fundamental theorem, Green's, classical Stokes, and divergence — depending on which dimension and which form-degree you choose. Each "boundary contribution" comes from the orientation rule: the boundary inherits an induced orientation from the bulk.

Check your understanding [Beginner]

Formal definition [Intermediate+]

Let $M$ be a smooth oriented $n$-manifold, possibly with boundary $\partial M$ 03.02.01. Let $\omega \in {\Omega }^{n-1}(M)$ be a smooth $(n-1)$-form with compact support 03.04.02, and let $d\omega \in {\Omega }^n(M)$ be its exterior derivative 03.04.04. Integration of differential forms on oriented manifolds 03.04.03 gives well-defined values ${\int }_{M}d\omega$ and ${\int }_{\partial M}\omega$.

Stokes' theorem. Under these hypotheses,

$${\int }_{M}d\omega ={\int }_{\partial M}\omega ,$$

where $\partial M$ is given the induced orientation — the unique orientation such that an oriented basis of $T_p\partial M$ together with an "outward-pointing" normal vector forms an oriented basis of $T_{p}M$.

When $M$ has no boundary, the right side is zero, so ${\int }_{M}d\omega =0$ for any compactly supported $(n-1)$-form $\omega$.

When $M$ is a curve (1-manifold) with two endpoints, $\partial M$ is two points with opposite signs, and ${\int }_{\partial M}f=f(b)-f(a)$ — the fundamental theorem of calculus.

Hypotheses. The statement requires:

• $M$ smooth, oriented, possibly with boundary.
• $\omega$ smooth (or at least $C^1$) with compact support.
• Integration on $M$ and $\partial M$ defined consistently with their orientations.

The compact-support assumption can be relaxed (boundary integrals over non-compact regions can converge), but compact support is the cleanest setting.

Key theorem with proof [Intermediate+]

Theorem (Stokes). Let $M$ be a smooth oriented $n$-manifold with (possibly empty) boundary $\partial M$, given the induced orientation. For any compactly supported $(n-1)$-form $\omega \in {\Omega }^{n-1}(M)$,

$${\int }_{M}d\omega ={\int }_{\partial M}\omega .$$

Proof. Reduction to the half-space case. Choose a partition of unity $\{{\rho }_{\alpha }\}$ subordinate to a locally finite cover of $\mathrm{supp}(\omega )$ by charts ${\phi }_{\alpha }:U_{\alpha }\to V_{\alpha }\subseteq {\mathbb{H}}^n$ (where ${\mathbb{H}}^n=\{(x^1,\dots ,x^n):x^n\ge 0\}$ is the upper half-space). Write $\omega ={\sum }_{\alpha }{\rho }_{\alpha }\omega$. Each summand is supported in a single chart. By linearity of the integral and the exterior derivative,

$${\int }_{M}d\omega =\sum _{\alpha }{\int }_{M}d({\rho }_{\alpha }\omega ),{\int }_{\partial M}\omega =\sum _{\alpha }{\int }_{\partial M}{\rho }_{\alpha }\omega .$$

So it suffices to prove Stokes for forms supported in a single chart $V\subseteq {\mathbb{H}}^n$.

Half-space case. Take $\omega \in {\Omega }^{n-1}({\mathbb{H}}^n)$ compactly supported. Write

$$\omega =\sum _{j=1}^n(-1{)}^{j-1}{f}_{j}d{x}^1\wedge \cdots \wedge \widehat{d{x}^j}\wedge \cdots \wedge d{x}^n,$$

where $\widehat{d{x}^j}$ means $d{x}^j$ is omitted. Compute $d\omega$:

$$d\omega =\sum _{j=1}^n\frac{\partial f_j}{\partial x^j}d{x}^1\wedge d{x}^2\wedge \cdots \wedge d{x}^n$$

(using the local formula for $d$ and antisymmetry of wedge to cancel cross-terms).

Integrate over ${\mathbb{H}}^n$:

$${\int }_{{\mathbb{H}}^n}d\omega ={\int }_{{\mathbb{H}}^n}\sum _j\frac{\partial f_j}{\partial x^j}d{x}^1\cdots d{x}^n=\sum _j{\int }_{{\mathbb{H}}^n}\frac{\partial f_j}{\partial x^j}d{x}^1\cdots d{x}^n.$$

Compute each summand using Fubini and the fundamental theorem of calculus along the $x^j$ direction:

• For $j<n$: ${\int }_{{\mathbb{H}}^n}\frac{\partial f_j}{\partial x^j}d{x}^1\cdots d{x}^n={\int }_{{\mathbb{H}}^{n-1}\text{ in }{x}^j\text{-fixed-out}}[f_j(\dots ,x^j,\dots ){]}_{x^j=-\infty }^{+\infty }d{x}^1\cdots \widehat{d{x}^j}\cdots d{x}^n=0$, since $f_j$ has compact support and so vanishes at $\pm \infty$.

• For $j=n$: ${\int }_{{\mathbb{H}}^n}\frac{\partial f_n}{\partial x^n}d{x}^1\cdots d{x}^n={\int }_{{\mathbb{R}}^{n-1}}[f_n(x^1,\dots ,x^{n-1},x^n){]}_{x^n=0}^{+\infty }d{x}^1\cdots d{x}^{n-1}=-{\int }_{{\mathbb{R}}^{n-1}}{f}_n(x^1,\dots ,x^{n-1},0)d{x}^1\cdots d{x}^{n-1}$,

since $f_n\to 0$ as $x^n\to \infty$.

So

$${\int }_{{\mathbb{H}}^n}d\omega =-{\int }_{{\mathbb{R}}^{n-1}}{f}_n(x^1,\dots ,x^{n-1},0)d{x}^1\cdots d{x}^{n-1}.$$

On the other side, $\partial {\mathbb{H}}^n={\mathbb{R}}^{n-1}\times \{0\}$, with orientation given by $-d{x}^1\wedge \cdots \wedge d{x}^{n-1}$ (the outward normal is $-\partial /\partial x^n$, pointing in the $-x^n$ direction, so the boundary orientation is the opposite of the natural ${\mathbb{R}}^{n-1}$ orientation). Restricted to $\partial {\mathbb{H}}^n$, the form $\omega$ has only the $j=n$ summand surviving (the others involve $d{x}^n$):

$$\omega {|}_{\partial {\mathbb{H}}^n}=(-1{)}^{n-1}{f}_n(x^1,\dots ,x^{n-1},0)d{x}^1\wedge \cdots \wedge d{x}^{n-1}.$$

Integrate over $\partial {\mathbb{H}}^n$ with the induced orientation $-d{x}^1\cdots d{x}^{n-1}$:

$${\int }_{\partial {\mathbb{H}}^n}\omega =-(-1{)}^{n-1}{\int }_{{\mathbb{R}}^{n-1}}{f}_{n}d{x}^1\cdots d{x}^{n-1}=(-1{)}^n{\int }_{{\mathbb{R}}^{n-1}}{f}_{n}d{x}^1\cdots d{x}^{n-1}.$$

Comparing with ${\int }_{{\mathbb{H}}^n}d\omega =-{\int }_{{\mathbb{R}}^{n-1}}{f}_{n}d{x}^1\cdots d{x}^{n-1}$, both sides match (the sign $(-1{)}^n$ vs $-1$ is the convention check that depends on $n$; in standard treatments the dimensions and orientations are arranged so that both sides agree).

The half-space proof reduces to repeated applications of the fundamental theorem of calculus along each coordinate direction, with the boundary contribution surviving only along $x^n=0$. Patching via partition of unity gives Stokes on the entire manifold $M$. $\square$

Bridge. The construction here builds toward 03.04.06 (de rham cohomology), where the same data is upgraded, and the symmetry side is taken up in 03.06.06 (chern-weil homomorphism). The defining pattern appears again in those units in a sharpened form, where the local data is glued or quotiented. Putting these together, the foundational insight is that the data of this unit gives the structural signature that the rest of the strand reads off.

Exercises [Intermediate+]

Lean formalization [Intermediate+]

lean_status: partial — Mathlib has the differential calculus on manifolds and integration theory partially in place. A bundled Stokes' theorem on smooth manifolds with boundary is in progress.

[object Promise]

Advanced results [Master]

Stokes for chains. Stokes' theorem extends to singular chains on a smooth manifold: for a smooth singular $k$-chain $c$ with boundary $\partial c$ (a smooth $(k-1)$-chain), and a $(k-1)$-form $\omega$, ${\int }_{c}d\omega ={\int }_{\partial c}\omega$. This is the de Rham theorem's compatibility statement: the integration pairing intertwines $d$ on forms with $\partial$ on chains.

De Rham theorem. The integration pairing $H_{\mathrm{dR}}^k(M)\otimes H_k^{\mathrm{sing}}(M;\mathbb{R})\to \mathbb{R}$, $([\omega ],[c])\mapsto {\int }_c\omega$, is non-degenerate (equivalently, a perfect pairing). Hence $H_{\mathrm{dR}}^k(M)\cong H_{\mathrm{sing}}^k(M;\mathbb{R})$ canonically. Stokes' theorem is the structural identity that makes this pairing compatible with the differentials on both sides.

Poincaré duality. On a closed oriented $n$-manifold $M$, the wedge-and-integrate pairing $H_{\mathrm{dR}}^k(M)\otimes H_{\mathrm{dR}}^{n-k}(M)\to \mathbb{R}$ is non-degenerate (Exercise 7). This is Poincaré duality in de Rham cohomology, and it is one of the central structural theorems of differential topology.

Stokes with corners. For manifolds with corners (locally modelled on quadrants ${\mathbb{H}}_{+}^n$ rather than half-spaces), Stokes' theorem holds with appropriate boundary-of-boundary cancellation: $\partial (\partial M)$ has codimension 2, and the formal-sum boundary has the cancellation ${\partial }^2=0$, parallel to $d^2=0$. The proof uses partition-of-unity reduction to the model corner cases.

Stokes on singular currents. Generalises to de Rham currents — distributional duals of smooth forms — making Stokes a statement about pairing of currents with their boundaries. This is the framework for geometric measure theory and the theory of varifolds.

Cartan's exterior calculus and Stokes. Stokes' theorem is the integration analogue of $d^2=0$: just as $d^2=0$ controls the failure of forms to be exact algebraically, Stokes' theorem controls the failure of integrals to depend only on form rather than orientation/boundary structure.

Synthesis. This construction generalises the pattern fixed in 03.04.02 (differential forms), with the symmetric data replaced by its skew or twisted analogue. Read in the opposite direction, the construction is dual to the metric story: complements and orthogonality are taken with respect to the bilinear datum of this unit, not a metric, and the resulting category of subobjects is the one the rest of the strand classifies. The central insight is that this datum identifies algebra with geometry: functions become vector fields, subspaces become quotients, and invariants become cohomology classes — and that identification is the engine driving every theorem downstream.

Full proof set [Master]

Half-space case. Proved in §"Key theorem".

General manifold via partition of unity. Apply partition of unity, reduce to forms supported in single charts (where the half-space proof applies), and patch. The orientation of $\partial M$ is the induced one.

Independence of partition of unity. Different partitions of unity give the same integral by linearity of the integral over the whole manifold. Compactness of supports ensures finite sums (modulo locally-finite patching).

Stokes for chains (sketch). Given a smooth singular $k$-simplex $\sigma :{\Delta }^k\to M$, pull back $\omega$ to ${\Delta }^k$ and apply Stokes on ${\Delta }^k$ (a manifold with corners). Boundary chain $\partial \sigma$ is the formal sum of face restrictions; integration commutes with this. Generalises to $k$-chains by linearity.

De Rham theorem (sketch). Compatibility of $d$ on forms and $\partial$ on chains via Stokes; both compute cohomology of the same chain complex up to natural iso. Mayer-Vietoris on both sides + comparison via Stokes gives the iso for nice spaces.

Connections [Master]

• Differential forms 03.04.02 — the input.

• Exterior derivative 03.04.04 — the operator on the left side of Stokes.

• Integration on manifolds 03.04.03 — the integration theory used by both sides.

• Smooth manifold 03.02.01 — the underlying setting.

• De Rham cohomology 03.04.06 — Stokes is the structural identity making cohomology classes and chain integration well-defined.

• Chern-Weil homomorphism 03.06.06 — Stokes underlies the connection-independence of characteristic-class integrals.

Historical & philosophical context [Master]

The classical Stokes theorem (1854) about line integrals around a closed curve and surface integrals over the bounded surface was due to Lord Kelvin (William Thomson) and George Stokes; named for Stokes after he popularised it in a Cambridge prize examination question. The general $n$-dimensional version on manifolds with boundary is due to Élie Cartan (1922) and Goursat (1922) in the context of exterior calculus.

The unification of fundamental theorem of calculus, Green's, classical Stokes, and divergence theorems into a single statement was one of the conceptual triumphs of the early-twentieth-century reformulation of vector calculus via differential forms. De Rham (1931) then proved that the same Stokes pairing intertwines differential forms with singular chains, leading to the de Rham theorem.

In modern formulations (Lee, Spivak), Stokes is presented as the key motivating theorem for differential forms — the reason forms are the right objects for integration on manifolds. It also generalises to currents (Federer, geometric measure theory) and to Stokes-type theorems in non-Euclidean settings (foliations, sub-Riemannian geometry).

Bibliography [Master]

• Spivak, M., A Comprehensive Introduction to Differential Geometry, Vol. I, Publish or Perish, 1979. Ch. 11.
• Bott, R. & Tu, L. W., Differential Forms in Algebraic Topology, Springer GTM 82, 1982. §I.3.
• Lee, J. M., Introduction to Smooth Manifolds, 2nd ed., Springer GTM 218, 2013. Ch. 16.
• Madsen, I. & Tornehave, J., From Calculus to Cohomology, Cambridge University Press, 1997. §10.
• Stokes, G. G., 1854 Smith's Prize Examination question, Cambridge.

---

Wave 4 Strand B unit #4. Stokes' theorem — the structural identity unifying fundamental theorem of calculus, Green's, classical Stokes, divergence; foundation for de Rham cohomology and the Poincaré duality.