Generalized Stokes’ Theorem

Colin M. Weller
June 5 2020

Contents
1 The Essentials and Manifolds 2

2 Introduction to Differential Forms 4

3 The Wedge Product 6

4 Forms on Manifolds and Exterior Derivative 7

5 Integration of Differential Forms 8

6 Generalized Stokes’ Theorem 10

7 Conclusion 12

8 Acknowledgements 13

Abstract

We introduce and develop the necessary tools needed to generalize Stokes’ Theo-
rem. We begin our discussion by introducing manifolds and differential forms. Going
beyond our treatment of integrable functions in 3 dimensions we go through what it
means to integrate a smooth differential form on a smooth manifold with boundary.
To do this we also introduce helpful tools like the wedge product and the exterior
derivative. We do not give a proof of every theorem in this construction, but rather
give the reader a detailed guide on how those theorems would be used to prove the
main result of the paper: Stokes’ Theorem. Unless otherwise noted all references are
from [3].

1
1 The Essentials and Manifolds
The theory of Topological manifolds does not explicitly mention calculus. To un-
derstand what we mean by derivatives of functions, curves or maps we introduce a
”smooth manifold”. For this formalism we restrict our study to the subsets of Rn .
We begin by recalling the definition of a smooth map.

Definition 1.1. Let U and V be open subsets of Rn and Rm , respectively. A map

F : U → V is said to be smooth if each of the component functions of F has
continuous partial derivatives of all orders.

To extend the idea of smoothness we introduce a diffeomorphism.

Definition 1.2. If a smooth map F is in addition bijective – one-to-one and onto–

and has a smooth inverse map, it is called a diffeomorphism .

With the goal to additionally give subsets of Rn a smooth structure we must base
our construction on the calculus of maps between Euclidean spaces. To study this
calculus we introduce the following definitions.

Definition 1.3. Let M be a subset of Rn . A coordinate chart on M is a pair (U, ϕ),

where U is an open subset of M and ϕ : U → U e is a homeomorphism from U to an
open subset Ũ = ϕ(U ) ⊂ Rn .

Definition 1.4. Let M be a subset of Rn . If (U, ϕ), (V, ψ) are two charts such that
U ∩ V 6= ∅, then the composite map ψ ◦ ϕ−1 : ϕ(U ∩ V ) → ψ(U ∩ V ) is a composition
of homeomorphisms, and is therefore itself a homeomorphism. Two charts (U, ϕ)
and (V, ψ) are said to be smoothly compatible if either U ∩ V = ∅ or the transition
map ψ ◦ ϕ−1 is a diffeomorphism.

Loosely speaking, we need a way to keep track of what the charts are communi-
cating about that subset through the so-called atlas.

Definition 1.5. We define an atlas for M to be a collection of charts whose domains

cover M. An atlas A is called a smooth atlas if any two charts in A are smoothly
compatible with each other.

In practice, many different choices of an atlas would yield the same ”smooth
structure”(which we have not officially defined yet). That is, it gives us the same set
of smooth functions on the manifold. To resolve this, we have a maximal atlas.

2
Definition 1.6. A smooth atlas A on M is called maximal if it is not contained
in any strictly larger smooth atlas. This just means every chart that is smoothly
compatible with every chart in A is already in A.
With this, we let a smooth structure on a manifold in Rn be a maximal smooth
atlas.
Definition 1.7. A smooth manifold is a manifold together with a smooth structure
In the literature, the smooth structure is commonly omitted. In many cases we
will not be able to find a smooth structure in which case the manifold is not smooth
or we could possibly find a smooth structure of a given manifold. The difference is
an interesting area of study but we will, for the purposes of this paper, just refer to
manifolds as smooth or not smooth. In addition to a smooth manifold, we have the
notion of a smooth manifold with boundary. Just as the name suggests, a smooth
manifold with boundary is exactly like the smooth manifolds we have already defined
but with additional structure known as the boundary. In order to rigorously define
such an object, we introduce an analogue to Rk .
Definition 1.8. The closed upper half space {(x1 , . . . , xn ) : xn ≥ 0} is denoted Hn .
Definition 1.9. A smooth manifold with boundary is smooth manifold M in which
every point has a neighbourhood homeomorphic to an open subset of the closed
n-dimensional upper half space Hn .
This definition will be essential in generalizing Stokes’ Theorem. We also recall
from our previous 3-dimensional treatment of Stokes’ Theorem that it was important
to have an orientable surface, or to orientate the surface. How do we go about giving
a manifold in n dimensions an orientation? We recall that a pointwise orientation
is given by choice of orientation of each tangent space in the 3 dimensional case and
this definition is generalized to any number of dimensions. The same theorem applies
as well.
Theorem 1.1. A connected, in the topological sense, orientable smooth manifold
with boundary admits exactly two orientations.
A theorem that we present without proof will become useful for later in the paper.
Theorem 1.2. If M is any smooth manifold with boundary, there is a smooth
outward-pointing vector field along ∂M
To conclude, we introduce the partition of unity. First, the idea of a support and
its properties.

3
Definition 1.10. The support of a function f on a smooth manifold M , denoted
supp f is the closure of the set where f is nonvanishing:
supp f = {p ∈ M : f (p) 6= 0}
Definition 1.11. Now let U = {Uα }α∈A be an arbitrary open cover of a smooth
manifold M. A partition of unity subordinate to U is a collection of smooth functions
{ϕα : M → R}α∈A , with the following properties:
(i) 0 ≤ ϕα (x) ≤ 1 for all α ∈ A and all x ∈ M
(ii) supp ϕα ⊂ Uα P
(iii) the set of supports {supp ϕα }α∈A is locally finite; and ( iv ) α∈A ϕα (x) = 1 for
all x ∈ M

2 Introduction to Differential Forms

To begin our discussion on differential forms we discuss multilinear functions – the
language of tensors. We have already been introduced to, and are quite familiar
with, some mulilinear functions: the cross product, dot product, and determinant.
All of these functions are linear in each separate argument. Formally, we introduce
the following definition.
Definition 2.1. Let k be a natural number and let V be a finite dimensional vector
space. A k-tensor on V is a real-valued function T on the cartesian product V k
where T is real-valued multilinear function of k elements of V and
T : V × ··· × V → R
| {z }
k copies

The natural number k is usually referred to as the rank of the tensor. A neat way
to make other tensors and study their properties is to perform operations on them
like multiplying two real numbers. For tensors we have the following definition.
Definition 2.2. Let V be a finite-dimensional real vector space and let S ∈ T k (V ), T ∈
T l (V ). Where T k (V ) and T l (V ) are the set of rank k and l tensors in V respectively.
Define a map
S ⊗ T : V × ··· × V → R
| {z }
k+l copies

by
S ⊗ T (X1 , . . . , Xk+l ) = S (X1 , . . . , Xk ) T (Xk+1 , . . . , Xk+l )
With both S and T both depending linearly on its arguments, the Xi , it follows that
S ⊗ T is a rank-(k + l) tensor and is called the tensor product.

4
Theorem 2.1. The tensor product is associative; T ⊗ (R ⊗ Q) = (T ⊗ R) ⊗ Q.

Proof. Let T, R, and Q be rank p, q, and s rank tensors, respectively. By the asso-
ciativity of the real numbers we have that

((T ⊗ S) ⊗ R) (X1 , . . . , Xp , Xp+1 , . . . , Xp+q , Xp+q+1 , . . . , Xp+q+s )

= [T (X1 , . . . , Xp ) · S (Xp+1 , . . . , Xp+q )] · R (Xp+q+1 , . . . , vp+q+s )
= T (X1 , . . . , Xp ) · [S (vp+1 , . . . , Xp+q ) · R (Xp+q+1 , . . . , Xp+q+s )]
= (T ⊗ (S ⊗ R)) (X1 , . . . , Xp , Xp+1 , . . . , Xp+q , Xp+q+1 , . . . , Xp+q+s )

This type of product will eventually formulate the wedge product of differential
forms. To develop this mathematical operation we need to know about a special
kind of tensor– the alternating tensor.

Definition 2.3. A covariant k-tensor T on a finite-dimensional vector space V is

said to be alternating if it has the following property:

T (X1 , . . . , Xi , . . . , Xj , . . . , Xk ) = −T (X1 , . . . , Xj , . . . , Xi , . . . , Xk )

To describe the algebra and applications of alternating tensors we recall the def-
inition of the sign of a permutation– a function that, in this context, will rearrange
the arguments of a tensor. The sign is +1 if the the permutation can be written as
a even number of compositions and −1 otherwise. Alternating tensors are interest-
ing objects of study, but is there a way to make alternating tensors out of ordinary
tensors? This process is called the alternating projection.

Definition 2.4. Let T be a k-tensor belonging to a finite-dimensional vector space

V we define
1 X
Alt T = (sgn σ)T σ
k! σ∈S
k

where Sk is the set of permutations of the arguments of T . And T σ denotes the

permuted arguments of T so
1 X

( Alt T ) (X1 , . . . , Xk ) = (sgn σ)T Xσ(1) , . . . , Xσ(k)
k! σ∈S
k

The following theorem is from [2]:

Theorem 2.2. Let T be a k-tensor. Then Alt T is an alternating tensor

5
Proof.
" #σ
1 X 1 X 1 X
[Alt(T)]σ = (−1)π T π = (−1)π◦σ◦σ (T π )σ = (−1)σ (−1)π◦σ T π◦σ
k! π∈S k! π∈S k! π∈S
k k k

Now let τ = π ◦ σ. This yields

" #
1 X
[Alt(T)]σ = (−1)σ (−1)τ T τ = (−1)σ [Alt(T)]
k! τ ∈S
k

Now having the ability to make alternating tensors out of ordinary ones and being
able to multiply the results together we can introduce the wedge product.

3 The Wedge Product

With our preliminary work on tensors now complete we are ready to give a useful
mathematical operation: the wedge product.
Definition 3.1. Let S and T be alternating tensors of rank k and l, respectively.
The wedge product S ∧ T is defined as
(k + l)!
S∧T = Alt(S ⊗ T )
k!l!
With this product we summarize some of the most important properties of the
wedge product with the following lemma and subsequent theorem. First a definition:
Definition 3.2. Let V be an n -dimensional vector space, and suppose (ε1 , . . . , εn )
is any basis for V ∗ . We will define a collection of alternating tensors on V that
generalize the determinant function on Rn . For each multi-index I = (i1 , . . . , ik ) of
length k such that 1 ≤ i1 , . . . , ik ≤ n, define a covariant k -tensor εI by
 
εi1 (X1 ) . . . εi1 (Xk )
εI (X1 , . . . , Xk ) = det 
 .. .. 
. . 
εik (X1 ) . . . εik (Xk )
 
X1i1 . . . Xki1
= det  ... .. 

. 
X1ik . . . Xkik

6
Lemma 3.1. For any multi-indices I = (i1 , . . . , ik ) and J = (j1 , . . . , jl )
εI ∧ εJ = εIJ
where IJ is the multi-index (i1 , . . . , ik , j1 , . . . , jl ) obtained by concatenating I and J
Theorem 3.2. Let ω, η, ξ be alternating tensors. Then the wedge product is asso-
ciative :
ω ∧ (η ∧ ξ) = (ω ∧ η) ∧ ξ.
The product is also anticommunative so if ω is a k-alternating-tensor and η is an
l-alternating-tensor then
ω ∧ η = (−1)kl η ∧ ω
Proof. To prove associativity, note that Lemma 3.1 gives
εI ∧ εJ ∧ εK = εIJ ∧ εK = εIJK = εI ∧ εJK = εI ∧ εJ ∧ εK

The general case follows from bilinearity. Similarly, using Lemma 3.1 again, we get
εI ∧ εJ = εIJ = (sgn τ )εJI = (sgn τ )εJ ∧ εI
where τ is the permutation that sends IJ to JI. It is easy to check that sgn τ =
(−1)kl , because τ can be decomposed as a composition of kl transpositions (each in-
dex of I must be moved past each of the indices of J) Anticommutativity then fol-
lows from bilinearity.

4 Forms on Manifolds and Exterior Derivative

Returning our attention to smooth manifolds we consider the subset of alternating
k-tensors on a smooth n-dimensional manifold M . We take the following definition
from [2].
Definition 4.1. Let M be a smooth manifold. A k-form on M is a function ω that
assigns to each point x ∈ M an alternating k-tensor ω(x) on the tangent space of X
at x.
By choice of coordinate chart(see Definition 1.4) we can write a k-form ω locally
as
0
X
ω= ωI dxi1 ∧ · · · ∧ dxik
I
What kind of operations can we make on differential forms? We have already seen
how we can take the wedge product but there is also the exterior derivative. This
operation generalizes the differential of a function.

7
Definition 4.2. Let ω be some smooth k-form on an open subset of Rk . The exterior
derivative of ω is then defined as the following (k + 1) form:
0
! 0
X X
I
d ωI dx = dωI ∧ dxI
I I

or equivalently
0
! 0 X
X X ∂ωI
d ωI dxi1 ∧ · · · ∧ dxik = dxi ∧ dxi1 ∧ · · · ∧ dxik
I I i
∂xi

We summarize the most important properties of the exterior derivative in the

following theorem but do not give proof of each property.

Theorem 4.1. The exterior differentiation operator defined for forms on arbitrary
manifolds with boundary has the following properties:
1. d (ω1 + ω2 ) = dω1 + dω2
2. d(ω1 ∧ ω2 ) = (dω1 ) ∧ ω2 + (−1)k ω1 ∧ dω2 , where ω1 is a k-form.
3. d(dω) = 0

With these properties and definition we know, for example, the exterior derivative
of a real-valued function f is just its differential df . We emphasize this result because
of its importance later in the paper.

5 Integration of Differential Forms

Our study of the integration of differential forms will just consider forms on the
subsets of Rn . We begin by defining what we mean by a smooth n-form on a compact
set. We know from our study of complex functions this quarter that a smooth form
ω on a set C has a smooth extension to some open set containing C.

Definition 5.1. Let C be a compact set in Rn . Let ω be a smooth n-form on C.

Z Z
1 n
f dx ∧ · · · ∧ dx = f dx1 · · · dxn
C C

Recalling Folland’s change of variable theorem for integrable functions of n vari-

ables we have from [1]:

8
Theorem 5.1. Let A be an invertible n × n matrix, and let G(u) = Au be the
corresponding linear transformation of Rn . Suppose S is a measurable region in Rn
and f is an integrable function on S. Then G−1 (S) = {A−1 x : x ∈ S} is measurable
and f ◦ G is integrable on G−1 (S), and
Z Z Z Z
n
· · · f (x)d x = | det A| · · · f (Au)dn u
S G−1 (S)

To prove the following theorem would require additional definitions and ideas
that don’t help the reader understand the main result. We use Theorem 5.2 to prove
the more ”useful” Corollary that follows.

Theorem 5.2. Let D and E be domains of integration in Rn , and let ω be an n -form

on E. If G : D → E is a smooth map whose restriction to Int D is an orientation-
preserving(det(dG) > 0) or orientation-reversing(det(dG) < 0)) diffeomorphism onto
Int E, then (R
G∗ ω
Z
DR
if G is orientation-preserving,
ω=
E − D G∗ ω if G is orientation-reversing.

Corollary 5.2.1. Suppose U, V are open subsets of Rn , G : U → V is an orientation-

preserving diffeomorphism, and ω is a compactly supported n-form on V. Then
Z Z
ω= G∗ ω
V U

Proof. Let E ⊂ V be a domain of integration containing supp ω(just the support

of ω). Since smooth maps take sets of measure zero to sets of measure zero, D =
G−1 (E) ⊂ U is a domain of integration containing supp G∗ ω. Therefore, the result
follows from the preceding theorem.
To make sense of integrating a form over a whole manifold we have to consider a
partition of unity(see Definition 1.11 ) and arrive at the following:

Definition 5.2. Let {(Ui , φi )} be a finite cover of the support of ω and let {ψi } be
a subordinate partition of unity. Then the integral of ω over M is given by
Z XZ
ω= ψi ω
M i M

R
Theorem 5.3. The definition of M ω given above does not depend on the choice of
oriented charts or partition of unity.

9
6 Generalized Stokes’ Theorem
All the previous sections now come together to give a beautiful and general result.

Theorem 6.1. Let M be an oriented n dimensional manifold with boundary, and let
ω be a compactly supported (n − 1)-form on M. Then
Z Z
dω = ω
M ∂M

Proof. We begin by considering a very special case: Suppose M is the upper half
space Hn itself. Then the fact that ω has compact support means that there is a
number R > 0 such that supp ω is contained in the rectangle A = [−R, R] × · · · ×
[−R, R] × [0, R]. We can write ω in standard coordinates as
n
X
ω= ωi dx1 ∧ · · · ∧ dx
ci ∧ · · · ∧ dxn
i=1

where the hat means that dxi is omitted. Therefore,

n
X
dω = dωi ∧ dx1 ∧ · · · ∧ dx
ci ∧ · · · ∧ dxn
i=1
n
X ∂ωi j
= j
dx ∧ dx1 ∧ · · · ∧ dx
ci ∧ · · · ∧ dxn
i,j=1
∂x
n
X ∂ωi
= (−1)i−1 i dx1 ∧ · · · ∧ dxn
i=1
∂x

Thus we compute
Z n Z
X
i−1 ∂ωi 1
dω = (−1) i
dx ∧ · · · ∧ dxn
Hn
i=1 A ∂x
n Z RZ R Z R
X
i−1 ∂ωi
= (−1) ··· i
(x)dx1 · · · dxn
i=1 0 −R −R ∂x

We can rearrange the order of integration in each term so as to do the xi integration

10
first. By the fundamental theorem of calculus, the terms for which i 6= n reduce to
n−1 Z R Z R Z R
X
i−1 ∂ωi
( − 1) ···
i
(x)dx1 · · · dxn
i=1 0 −R −R ∂x
n−1 Z R Z R Z R
X
i−1 ∂ωi i 1 c i · · · dxn
= (−1) ··· i
(x)dx dx · · · dx
i=1 0 −R −R ∂x
xi =R
n−1 Z RZ R Z R
c i · · · dxn
X 
= (−1)i−1 ··· ωi (x) dx1 · · · dx

i=1 0 −R −R i
x =−R
=0

because we have chosen R large enough that ω = 0 when xi = ±R. The only term
that might not be zero is the one for which i = n. For that term we have
Z Z R Z RZ R
n−1 ∂ωn
dω = (−1) ··· n
(x)dxn dx1 · · · dxn−1
Hn −R −R 0 ∂x
Z R Z R xn =R

= (−1)n−1 ··· ωi (x) dx1 · · · dxn−1
−R −R xn =0
Z R Z R
= (−1)n ωi x1 , . . . , xn−1 , 0 dx1 · · · dxn−1


···
−R −R

because ωn = 0 when xn = R. To compare this to the other side of the above

equation we compute as follows:
Z XZ
ωi x1 , . . . , xn−1 , 0 dx1 ∧ · · · ∧ dx
ci ∧ · · · ∧ dxn


ω=
∂Hn i A∩∂Hn

Because xn vanishes on ∂Hn , the restriction of dxn to the boundary is identically

zero. Thus, the only term above that is nonzero is the one for which i = n, which
becomes Z Z
ωn x1 , . . . , xn−1 , 0 dx1 ∧ · · · ∧ dxn−1


ω=
∂Hn A∩∂Hn

Taking into account the fact that the coordinates (x1 , . . . , xn−1 ) are positively ori-
ented for ∂Hn when n is even and negatively oriented when n is odd, this becomes
Z Z R Z R
n
ωn x1 , . . . , xn−1 , 0 dx1 · · · dxn−1 .


ω = (−1) ···
∂Hn −R −R

11
Next, let M be an arbitrary manifold with boundary, but consider an (n − 1)-form
ω that is compactly supported in the domain of a single chart (U, ϕ). Assuming
without loss of generality that ϕ is an oriented chart, the definition(with the right
interpretation of charts in this case) yields
Z Z Z
−1 ∗

∗

d ϕ−1 ω


dω = ϕ dω =
M Hn Hn
∗
since (ϕ−1 ) dω is compactly supported on Hn . By the computation above, this is
equal to Z

∗
ϕ−1 ω
∂Hn
n
where ∂H is given the induced orientation. since ϕ∗ takes outward-pointing vectors
on ∂M to outward-pointing vectors on Hn (by Theorem 1.2 ), it follows that ϕ|U ∩∂M
n
R an orientation-preserving diffeomorphism onto ϕ(U ) ∩ ∂H , and thus is equal to
is
∂M
ω. This proves the theorem in this case.
Finally, let ω be an arbitrary compactly supported (n − 1)-form. Choosing a
cover of supp ω by finitely many oriented coordinate charts {(Ui , ϕi )} , and choosing
a subordinate partition of unity {ψi } , we can apply the preceding argument to ψi ω
for each i and obtain
Z XZ
ω= ψi ω
∂M i ∂M
XZ
= d (ψi ω)
i M
XZ
= dψi ∧ ω + ψi dω
i M
Z ! Z !
X X
= d ψi ∧ω+ ψi dω
M i M i
Z
=0+ dω
M
P
because i ψi ≡ 1. This proves the result.

7 Conclusion
With the presentation of this general result we immediately obtain the higher dimen-
sional analogue of the Fundamental Theorem of Line Integrals, Green’s Theorem and

12
the Divergence Theorem. This result has important implications in the study of co-
homology of forms, complex analysis, and other areas of differential geometry. The
generalized Stokes’ Theorem also has rich applications in active areas of Physics
research.

8 Acknowledgements
I’d like to thank Professor Morrow and Jordan Brown for helping deliver this sequence
this year, particularly this quarter. I’d also like to thank my classmates for their
helpful comments and words of encouragement throughout the year.

References

[1]. Gerald Folland. Advanced Calculus. Prentice-Hall, Inc. 2002

[2]. Victor Guillemin and Alan Pollack. Differential Topology. Prentice-Hall, Inc. 1974
[3]. John M. Lee. Introduction to Smooth Manifolds. Springer-Verlag New York, Inc. 2003

13