Chapter 1

Euclidean Spaces
1.1 Smooth Functions on a Euclidean Space
1.1.1 C

versus Analytic Functions

Write coordinate on R
n
as x
1
, x
2
, , x
n
and let p = (p
1
, p
2
, , p
n
) be a point
in an open set U in R
n
. The indices on coordinates are superscripts.
Denition 1. Let k be a non negative integer. A real valued function f : U R
is C
k
at p if its partial derivatives of all orders less than or equal k exist and
are continuous at p. We similarly say that a vector-valued function is C
k
at a
point if all of its component functions are. We say that a function is C

if it
is C
k
for all k. A similar denition holds for functions on open sets.
We treat the terms C

and smooth as the same. A C

0
function is a contin-
uous function.
A neighborhood of a point is an open set containing the point. The function
is real-analytic at p if in some neighborhood of p it is equal to its Taylor series
at p, and is necessarily C

. A C

function need not be real-analytic (example

p.5: function very at at 0).
1.2 Taylors Theorem with Remainder
A smooth function need not equal to its Taylor series, but there is a Taylors
theorem with remainder for smooth functions.
We say that a subset S of R
n
is star-shaped with respect to a point p in S
if for every x in S the line segment from p to x lies in S.
Lemma 1 (Taylors Theorem with Remainder). Let f be a C

function on an
open subset U of R
n
star-shaped with respect to a point p in U. Then there are
functions g
i
(x) C

(U) such that

f(x) = f(p) +
n

i=1
(x
i
p
i
)g
i
(x), g
i
(p) =
f
x
i
(p).
1
2 CHAPTER 1. EUCLIDEAN SPACES
1.3 Tangent Vectors in R
n
as Derivatives
1.3.1 The Directional Derivative
The elements of the tangent space T
p
(R
n
) are called tangent vectors at p. The
directional derivative of f (which is C

in a neighborhood of p) in the direction

v at p is dened to be
D
v
f = lim
t0
f(p +tv) f(p)
t
=
d
dt
|
t=0
f(p +tv).
We write
D
v
=

v
i

x
i
|
p
for the map that sends a function f to the number D
v
f.
We have
D
v
(f) =

v
i
f
x
i
.
1.3.2 Germs of Functions
Consider the set of all pairs (f, U), where U is a neighborhood of p and f : U
R is a smooth function. We say that (f, U) and (g, V ) are equivalent if there is
an open set W U V containing p such that f = g when restricted to W. The
equivalence class of (f, U) is called the germ of f at p. We write C

p
(R
n
) for
the set of all germs of C

function on at p. The addition and multiplication of

functions induce corresponding operations on germs, making it into an algebra
over the reals.
1.3.3 Derivations at a Point
A linear map D : C

p
R satisfying the Leibniz rule
D(fg) = (Df)g(p) +f(p)Dg
is called a derivation at p or a point-derivation of C

p
. The set of all derivations
at p is denoted D
p
(R
n
). This set is a real vector space.
Directional derivatives at p are derivations at p, so there is a linear map
: T
p
(R
n
) D
p
(R
n
),
where v D
v
.
Lemma 2. If D is a point-derivation of C

p
, then D(c) = 0 for any constant
function c.
Proof. Obvious.
Theorem 3. The linear map : T
p
(R
n
) D
p
(R
n
) is an isomorphism of vector
spaces.
Proof. Injectivity follows from the fact that D
v
(x
i
) = v
i
. Surjectivity follows
from applying a general derivation D on Taylors theorem with remainder and
realizing that D = D
(Dx
1
, ,Dx
n
)
.
1.4. THE EXTERIOR ALGEBRA OF MULTICOVECTORS 3
This shows that we can identify the tangent vectors at p with the deriva-
tions at p. Under the vector space isomorphism between them, the standard
basis e
1
, , e
n
for T
p
corresponds to the set /x
1
|
p
, , /x
n
|
p
of partial
derivatives. We identify a tangent vector v =

v
i
e
i
as
v =

v
i

x
i
|
p
.
1.3.4 Vector Fields
A vector eld X on an open subset U of R
n
is a function that assigns to each
point p in U a tangent vector X
p
in T
p
(R
n
). The vector X
p
is a linear combi-
nation
X
p
=

a
i
(p)

x
i
|
p
, p U, a
i
(p) R.
We say that the vector eld X is C

on U if the coecient functions a

i
are all
C

on U. The ring of C

functions on an open set U is commonly denoted

by C

(U) or F(U). Multiplication of vector elds by functions on U is dened

pointwise:
(fX)
p
= f(p)X
p
, p U.
Clearly, if X is a C

vector eld and f is a C

function on U, then fX is a
C

module on U, denoted X(U).

1.3.5 Vector Fields as Derivations
If X is a C

vector eld on an open subset U of R

n
and f is a C

function on
U, we dene a new function Xf on U by
(Xf)(p) = X
p
f for any p U.
That is (Xf)(p) =

a
i
(p)
f
x
i
, which shows that Xf is a C

function on U.
Thus, a vector eld X gives rise to an R-libear map C

f Xf.
Proposition 4 (Leibniz rule for a vector eld). If X is a C

vector eld and

f and g are C

functions on an open subset U of R

n
, then X(fg) satises the
product (Leibniz rule):
X(fg) = (Xf)g +fXg.
Proof. Obvious.
1.4 The Exterior Algebra of Multicovectors
Linear algebra enters into manifold theory as every manifold can be locally
approximated by its tangent space at a point.
4 CHAPTER 1. EUCLIDEAN SPACES
1.4.1 Dual Space
If V and W are real vector spaces, we denote by HOM(V, W) the vector space
of all linear maps f : V W. Dene the dual space V

to be the vector space

of all real-valued linear functions on V :
V

= HOM(V, R).
The elements of V

are called covectors or 1-covectors on V .

Corollary 5. The dual space V

of a nite-dimensional vector space V has the

same dimension as V .
1.4.2 Permutations
1.4.3 Multilinear Functions
A function f : V
k
R is k-linear if it is linear in each of its k arguments. A
k-linear function on V is also called a k-tensor, and we call k the degree of f.
We deonte the vector space of all k tensors on V by L
k
(V ). For k = 0, we dene
a 0-covector to be a constant.
Denition 2. A k-function f : V
k
R is called symmetric if
f(v
(1)
, v
(2)
, , v
(k)
) = f(v
1
, v
2
, , v
k
)
for all S
k
; and is called alternating if
f(v
(1)
, v
(2)
, , v
(k)
) = (sgn)f(v
1
, v
2
, , v
k
)
for all S
k
. We deonte the space of all alternating tensors of degree k by
A
k
(V ).
1.4.4 The Permutation Actions on Mulitlinear Functions
If f is a k-linear function on V and S
k
, we dene
(f)(v
1
, v
2
, , v
k
) = f(v
(1)
, v
(2)
, , v
(k)
).
Lemma 6. If , S
k
and f is a k-linear function, then (f) = ()f.
1.4.5 The Symmetrizing and Alternating Operators
Given a k-linear function f we can make a symmetric function out of it:
Sf =

S
k
f,
and similarly we can make an alternating function out of it:
Af =

S
k
(sgn)f.
Proposition 7. If f is a k-linear function on V , then Sf is a symmetric k-
linear function, and Af is an alternating k-linear function on V .
Lemma 8. If f is an alternating k-linear function on V , then Af = (k!)f.
1.4. THE EXTERIOR ALGEBRA OF MULTICOVECTORS 5
1.4.6 The Tensor Product
If f and g are k- and l-linear functions on a vector space V , then their tensor
product f g is the k +l-linear function on V dened by
(f g)(v
1
, , v
k+l
) = f(v
1
, , v
k
)g(v
k+1
, , v
k+l
).
1.4.7 The Wedge Product
Given two alternating linear functions f A
k
(V ) and g A
l
(V ), their wedge
product, also called their exterior product, is the alternating k+l-linear function
dened by
f g =
1
k!l!
A(f g).
The factorial terms are there to compensate for repetition in the sum.
For c R and g A
k
(V ), we have c g = cg. It follows directly from the
denitino that the wedge product f g is bilinear in both f and g.
1.4.8 Anticommutativity of the Wedge Product
Proposition 9. The wedge product is anticommutative: if f A
k
(V ) and
g A
l
(V ), then
f g = (1)
kl
g f.
Corollary 10. If f is a mutlicovector of odd order, then f f = 0.
1.4.9 Associativity of the Wedge Product
Lemma 11. Let f and g be k- and l-linear functions on the vector space V .
Then
1. A(A(f) g) = k!A(f g), and
2. A(f A(g)) = l!A(f g).
Proposition 12 (Associativity of the wedge product). Let V be a real vector
space, and f, g, and h be alternating multilinear functions on V of degrees k, l,
and m respectively. Then
(f g) h = f (g h).
6 CHAPTER 1. EUCLIDEAN SPACES
Chapter 2
Manifolds
Manifolds are locally Euclidean in the sense that each point has a neighborhood,
called a chart, that is homeomorphic to an open subset of R
n
.
2.1 Manifolds
There are many kinds of manifolds, and we start by introducing topological
manifolds (which are Hausdor, second countable, locally Euclidean spaces).
2.1.1 Topological Manifolds
Denition 3. A topological space M is locally Euclidean of dimension n if
every point p of M has a neighborhood U such that there is a homeomophism
from U onto an open subset of R
n
. We call the pair (U, ) a chart, U a
coordinate neighborhood or a coordinate open set, and a coordinate map or
coordinate system on U. We say that a chart (U, ) is centered at p U if
(p) = 0.
Denition 4. A topological manifold (of dimension n) is a Hausdor, second
countable, locally Euclidean space (of dimension n).
If n = m, and open subset of R
n
cannot be homeomorphic to an open subset
of R
m
; this is called invariance of dimension.
2.1.2 Compatible Charts
Denition 5. Two charts (U, ), (V, ) of a topological manifold are C

-
compatible if the two maps

: (U V ) (U V ),
and

: (U V ) (U V )
are C

. These two maps are called the transition functions between the charts.
Denition 6. A (C

) atlas on a topological manifold M is a collection U =

{(U

)} of pairwise C

compatible charts that cover M.

7