Are you sure?
This action might not be possible to undo. Are you sure you want to continue?
Course 221
David Simms
School of Mathematics
Trinity College
Dublin
Typesetting
Eoin Curran.
1
2
Contents
1 Lebesgue Measure 7
1.1 Algebra of Subsets . . . . . . . . . . . . . . . . . . . . . . . . 7
1.2 The Interval Algebra . . . . . . . . . . . . . . . . . . . . . . . 9
1.3 The length measure . . . . . . . . . . . . . . . . . . . . . . . . 11
1.4 The σalgebra . . . . . . . . . . . . . . . . . . . . . . . . . . . 13
1.5 The outer measure . . . . . . . . . . . . . . . . . . . . . . . . 14
1.6 Extension of measure to σalgebra, using outer measure . . . . 16
1.7 Increasing Unions, Decreasing Intersections . . . . . . . . . . . 20
1.8 Properties of Lebesgue Measure . . . . . . . . . . . . . . . . . 22
1.9 Borel Sets . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 24
2 Integration 29
2.1 Measure Space, Measurable sets . . . . . . . . . . . . . . . . . 29
2.2 Characteristic Function . . . . . . . . . . . . . . . . . . . . . . 30
2.3 The Integral . . . . . . . . . . . . . . . . . . . . . . . . . . . . 31
2.4 Monotone Convergence Theorem . . . . . . . . . . . . . . . . 34
2.5 Existence of Monotone Increasing Simple Functions converg
ing to f . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 36
2.6 ’Almost Everywhere’ . . . . . . . . . . . . . . . . . . . . . . . 39
2.7 Integral Notation . . . . . . . . . . . . . . . . . . . . . . . . . 45
2.8 Fundamental Theorem of Calculus . . . . . . . . . . . . . . . 46
2.9 Fatou’s Lemma . . . . . . . . . . . . . . . . . . . . . . . . . . 48
2.10 Dominated Convergence Theorem . . . . . . . . . . . . . . . . 49
2.11 Diﬀerentiation under the integral sign . . . . . . . . . . . . . . 50
3 Multiple Integration 53
3.1 Product Measure . . . . . . . . . . . . . . . . . . . . . . . . . 53
3.2 Monotone Class . . . . . . . . . . . . . . . . . . . . . . . . . . 56
3.3 Ring of Subsets . . . . . . . . . . . . . . . . . . . . . . . . . . 57
3.4 Integration using Product Measure . . . . . . . . . . . . . . . 59
3.5 Tonelli’s Theorem . . . . . . . . . . . . . . . . . . . . . . . . . 63
3
3.6 Fubini’s Theorem . . . . . . . . . . . . . . . . . . . . . . . . . 64
4 Diﬀerentiation 71
4.1 Diﬀerentiation . . . . . . . . . . . . . . . . . . . . . . . . . . . 71
4.2 Normed Space . . . . . . . . . . . . . . . . . . . . . . . . . . . 72
4.3 Metric Space . . . . . . . . . . . . . . . . . . . . . . . . . . . 73
4.4 Topological space . . . . . . . . . . . . . . . . . . . . . . . . . 73
4.5 Continuous map of topological spaces . . . . . . . . . . . . . . 75
4.6 Homeomorphisms . . . . . . . . . . . . . . . . . . . . . . . . . 77
4.7 Operator Norm . . . . . . . . . . . . . . . . . . . . . . . . . . 77
4.8 Diﬀerentiation . . . . . . . . . . . . . . . . . . . . . . . . . . . 80
4.9 Notation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 83
4.10 C
r
Functions . . . . . . . . . . . . . . . . . . . . . . . . . . . 84
4.11 Chain Rule . . . . . . . . . . . . . . . . . . . . . . . . . . . . 87
5 Calculus of Complex Numbers 89
5.1 Complex Diﬀerentiation . . . . . . . . . . . . . . . . . . . . . 89
5.2 Path Integrals . . . . . . . . . . . . . . . . . . . . . . . . . . . 91
5.3 Cauchy’s Theorem for a triangle . . . . . . . . . . . . . . . . . 95
5.4 Winding Number . . . . . . . . . . . . . . . . . . . . . . . . . 98
5.5 Cauchy’s Integral Formula . . . . . . . . . . . . . . . . . . . . 101
5.6 Termbyterm diﬀerentiation, analytic functions, Taylor series 104
6 Further Calculus 107
6.1 Mean Value Theorem for Vectorvalued functions . . . . . . . 107
6.2 Contracting Map . . . . . . . . . . . . . . . . . . . . . . . . . 108
6.3 Inverse Function Theorem . . . . . . . . . . . . . . . . . . . . 109
7 Coordinate systems and Manifolds 115
7.1 Coordinate systems . . . . . . . . . . . . . . . . . . . . . . . . 115
7.2 C
r
manifold . . . . . . . . . . . . . . . . . . . . . . . . . . . . 117
7.3 Tangent vectors and diﬀerentials . . . . . . . . . . . . . . . . . 118
7.4 Tensor Fields . . . . . . . . . . . . . . . . . . . . . . . . . . . 121
7.5 Pullback, Pushforward . . . . . . . . . . . . . . . . . . . . . 125
7.6 Implicit funciton theorem . . . . . . . . . . . . . . . . . . . . 127
7.7 Constraints . . . . . . . . . . . . . . . . . . . . . . . . . . . . 130
7.8 Lagrange Multipliers . . . . . . . . . . . . . . . . . . . . . . . 132
7.9 Tangent space and normal space . . . . . . . . . . . . . . . . . 133
7.10 ?? Missing Page . . . . . . . . . . . . . . . . . . . . . . . . . . 134
7.11 Integral of Pullback . . . . . . . . . . . . . . . . . . . . . . . 135
7.12 integral of diﬀerential forms . . . . . . . . . . . . . . . . . . . 136
4
7.13 orientation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 140
8 Complex Analysis 145
8.1 Laurent Expansion . . . . . . . . . . . . . . . . . . . . . . . . 145
8.2 Residue Theorem . . . . . . . . . . . . . . . . . . . . . . . . . 147
8.3 Uniqueness of analytic continuation . . . . . . . . . . . . . . . 151
9 General Change of Variable in a multiple integral 153
9.1 Preliminary result . . . . . . . . . . . . . . . . . . . . . . . . . 153
9.2 General change of variable in a multiple integral . . . . . . . . 156
5
6
Chapter 1
Lebesgue Measure
1.1 Algebra of Subsets
Deﬁnition Let X be a set. A collection / of subsets of X is called an
algebra of subsets of X if
1. ∅ ∈ /
2. E ∈ / =⇒ E
∈ /. E
= ¦x ∈ X : x ∈ E¦
3. E
1
, . . . , E
k
∈ / =⇒ E
1
∪ E
2
∪ ∪ E
k
∈ /. i.e., / is closed under
complements and under ﬁnite unions. Hence / is closed under ﬁnite
intersection. (because E
1
∩ ∩ E
k
= (E
1
∪ ∪ E
k
)
)
Example Let 1 be the collection of all ﬁnite unions of intervals in R of the
form:
1. (a, b] = ¦x ∈ R : a < x ≤ b¦
2. (−∞, b] = ¦x ∈ R : x ≤ b¦
3. (a, ∞) = ¦x ∈ R : a < x¦
4. (−∞, ∞) = R
1 is an algebra of subsets of R, called the Interval Algebra.
£
¢
£
¢
£
¢
7
We want to assign a ‘length’ to each element of the interval algebra 1, so
we want to allow ‘∞’ as a length. We adjoin to the real no’s the 2 symbols
∞ and −∞ to get the Extended Real Line:
Deﬁnition The Extended Real Line:
R∪ ¦−∞, ∞¦ = [−∞, ∞]
We extend ordering, addition, multiplication to [−∞, ∞] by:
1. −∞< x < ∞∀x ∈ R
2. Addition
(a) ∞+∞= x +∞ = ∞+x = ∞
(b) (−∞) + (−∞) = x + (−∞) = (−∞) +x = −∞
(Don’t deﬁne ∞+ (−∞) or (−∞) +∞.)
3. Multiplication
(a) ∞∞= ∞= (−∞)(−∞)
(b) ∞(−∞) = (−∞)∞= −∞
(c) x∞ = ∞x =
∞ x > 0
0 x = 0
−∞ x < 0
(d) x(−∞) = (−∞)x =
−∞ x > 0
0 x = 0
∞ x < 0
Deﬁnition If A ⊂ [−∞, ∞] then
1. supA = ∞ if ∞∈ A or if A has no upper bound in R.
2. infA = −∞ if −∞ ∈ A or if A has no lower bound in R.
Deﬁnition Let / be an algebra of subsets of X. A function m : / −→[0, ∞]
is called a measure on / if:
1. m(∅) = 0
2. if E =
¸
∞
j=1
E
j
is countably disjoint with E
j
∈ / ∀j then m(E) =
¸
∞
j=1
m(E
j
). (m is Countably Additive.)
8
E
Figure 1.1: E is a disjoint union of sets
Note:
1. Countable means that E
1
, E
2
, . . . is either a ﬁnite sequence or can be
labelled by (1, 2, 3, . . .).
2. Disjoint means that E
i
∩ E
j
= ∅ ∀i = j
3. m(E) =
¸
∞
j=1
m(E
j
) means either:
(a)
¸
∞
j=1
m(E
j
) is a convergent series of ﬁnite no’s with the sum of
the series as m(E)
(b)
¸
∞
j=1
m(E
j
) is a divergent series with m(E) = ∞
(c) m(E
j
) = ∞ for some j, and m(E) = ∞
1.2 The Interval Algebra
Deﬁnition For each E ∈ 1, the interval algebra (See Section 1.1), write
E = E
1
∪ ∪ E
k
(disjoint) where each E
i
is of the form (a, b], (−∞, b], (a, ∞), or(−∞, ∞). The
length of E is m(E) = m(E
1
) +m(E
2
) + +m(E
k
) where m(a, b] = b −a.
m(−∞, b] = m(a, ∞) = m(−∞, ∞) = ∞
Deﬁnition a set V ⊂ R is called an Open Subset of R if:
for each a ∈ V ∃ > 0 s.t.(a −, a +) ⊂ V
£
¢
¡
£
¢
¡
a − a + a
i.e., every point is an interior point: it has no end points.
9
Theorem 1.2.1. (HeineBorelLebesgue) The closed interval is compact.
Let ¦V
i
¦
i∈I
be a family of open sets in R which cover the closed interval
[a, b]. Then ∃ a ﬁnite number of them: V
i
1
, . . . , V
i
k
(say) which cover [a, b]
(i.e. [a, b] ⊂ V
i
1
∪ ∪ V
i
k
.)
Proof. Put
K = ¦x ∈ [a, b] s.t. ∃( ﬁnite set ¦i
1
, . . . , i
r
¦ ⊂ I s.t. [a, x] ⊂ V
i
1
∪ ∪V
ir
)¦
c a b
a ∈ K =⇒ K = ∅
Let
c = sup K
Then
a ≤ c ≤ b c ∈ V
j
(say)
V
j
is open, therefore
∃ > 0 s.t. (c −, c +) ⊂ V
j
and
∃ k ∈ K s.t. c − < k ≤ c,
[a, k] ⊂ V
i
1
∪ ∪ V
ir
(say)
Then
[a, min(c +
2
, b)] ⊂ (V
i
1
∪ ∪ V
ir
∪ V
j
)
therefore
min(c +
2
, b) ∈ K
But c +
2
∈ K (as c = sup K) =⇒ b ∈ K as required.
10
1.3 The length measure
Theorem 1.3.1. The length function
m : 1 −→[0, ∞]
is a measure on the interval algebra 1
Proof. We have to check that m is countably additive, which reduces to
showing that if
(a, b] =
∞
¸
j=1
(a
j
, b
j
]
is a countable disjoint union then
∞
¸
j=1
(b
j
−a
j
)
is a convergent series with sum b −a.
Take the ﬁrst n intervals and relabel them:
(a
1
, b
1
], . . . , (a
n
, b
n
]
so that:
a a
1
b
1
a
2
b
2
a
3
b
3
an bn b
then
¸
n
j=1
(b
j
−a
j
) = b
1
−a
1
+b
2
−a
2
+b
3
−a
3
+ +b
n
−a
n
≤ a
2
−a +a
3
−a
2
+a
4
−a
3
+ +b −a
n
= b −a
therefore
∞
¸
j=1
(b
j
−a
j
)
converges and has sum ≤ b −a
Let > 0 and put
0
=
2
,
1
=
4
, . . . ,
j
=
2
j+1
, . . .
so
j
> 0 and
¸
∞
j=1
j
=
Then
(a −
0
, a +
0
), (a
1
, b
1
+
1
), (a
2
, b
2
+
2
), . . .
is a family of open sets which cover [a, b]
11
By the compactness of [a, b] ∃ a ﬁnite number of these open sets which
cover [a, b], and which by renumbering and discarding some intervals if nec
essary we can take as:
(a −
0
, a +
0
), (a
1
, b
1
+
1
), . . . , (a
n
, b
n
+
n
)
so that:
a −
0
a +
0
a a
1
a
2
b
2
+
2
b
1
+
1
a
3
b
3
+
3
a
n−1
b
n−1
+
n−1
b
an bn + n
b −a = a
1
−a +a
2
−a
1
+ +a
n
−a
n−1
+b −a
n
<
0
+(b
1
+
1
−a
1
) + +(b
n−1
+
n−1
−a
n−1
) +(b
n
+
n
−a
n
)
< +
¸
∞
j=1
(b
j
−a
j
)
is true ∀ > 0. Therefore
b −a ≤
∞
¸
j=1
(b
j
−a
j
)
Therefore
∞
¸
j=1
(b
j
−a
j
) = b −a
as required.
Let m :−→[0, ∞] be a measure m on an algebra / of subsets of X.
E ⊂ F =⇒
Thenm(E) ≤ m(F)
and
m(F ∩ E
) = m(F) −m(E) if m(F) = ∞
Then
F = E ∪ (F ∩ E
)
E
F
is a disjoint union, and
m(F) = m(E) +m(F ∩ E
)
12
Theorem 1.3.2. (m is subadditive on countable unions)
m(
∞
¸
j=1
) ≤
∞
¸
i=1
m(E
i
)
Proof. We proceed as follows:
Let
F
n
= E
1
∪ E
2
∪ ∪ E
n
1
E E
2
E
3
E
n
∞
¸
j=1
E
j
= E
1
∪ (E
2
∩ F
1
) ∪ (E
3
∩ F
2
) ∪
m(
¸
∞
j=1
E
j
) = m(E
1
) +m(E
2
∩ F
1
) +m(E
3
∩ F
2
) +
≤ m(E
1
) +m(E
2
) +m(E
3
) +
hence the result:
m(
∞
¸
j=1
E
j
) ≤
∞
¸
j=1
m(E
j
)
m is subadditive on countable unions.
1.4 The σalgebra
Our aim is to extend the notion of length to a much wider class of subsets of
R. In particular to sets obtainable from 1 by a sequence of taking countable
unions and taking complements.
Deﬁnition an algebra / of subsets of X is called a σalgebra if for each
sequence
E
1
, E
2
, E
3
, . . .
of elements of /, their union
E
1
∪ E
2
∪ E
3
∪
is also an element of /.
So / is closed under countable unions, and hence also under countable
intersections.
13
1.5 The outer measure
Deﬁnition We deﬁne the outer measure ˆ m associated with m to be the
function:
ˆ m : ¦all subsets of X¦ −→ [0, ∞]
given by:
ˆ m(E) = inf
∞
¸
j=1
m(E
j
)
where the inf is taken over all sequences E
j
of elements of / such that:
E ⊂
∞
¸
j=1
E
j
Theorem 1.5.1. ˆ m(E) = m(E) ∀E ∈ / (i.e. ˆ m agrees with m on /)
Proof. 1. if
E ∈ /
and
E ⊂
∞
¸
j=1
E
j
with E
j
∈ /
then:
m(E) ≤
∞
¸
j=1
m(E
j
)
therefore
m(E) ≤ ˆ m(E)
2. if
E ∈ /
then
E, ∅, ∅, . . .
is a sequence of elements of / whose union contains E. Therefore:
ˆ m(E) ≤ m(E) + 0 + 0 +
therefore
ˆ m(E) ≤ m(E)
14
Theorem 1.5.2. E ⊂ F =⇒ ˆ m(E) ≤ ˆ m(F)
Proof. if E ⊂ F then each sequence in / which covers F will also cover E.
Therefore:
ˆ m(E) ≤ ˆ m(F)
Theorem 1.5.3. ˆ m is subadditive on countable unions
ˆ m(
∞
¸
j=1
E
j
) ≤
∞
¸
j=1
ˆ m(E
j
)
for all sequences E
1
, E
2
, . . . of subsets of X.
Proof. Let > 0. Choose:
0
,
1
,
2
, . . . > 0
such that
∞
¸
j=0
j
=
choose B
ij
∈ / s.t.
E
i
⊂ B
i1
∪ B
i2
∪ ∪ B
ij
∪
and s.t.
∞
¸
j=1
m(B
ij
) ≤ ˆ m(E
i
) +
i
Then
∞
¸
i=1
E
i
⊂
∞
¸
i=1
∞
¸
j=1
B
ij
and
∞
¸
i=1
∞
¸
j=1
m(B
ij
) ≤
∞
¸
i=1
ˆ m(E
i
) +
therefore
ˆ m(
∞
¸
j=1
E
j
) ≤
∞
¸
j=1
ˆ m(E
j
)
as required.
15
Deﬁnition We call a subset E ⊂ X measurable w.r.t m if:
ˆ m(A) = ˆ m(A∩ E) + ˆ m(A∩ E
)
for all A ⊂ X
(E splits every set A into two pieces whose outer measures add up.)
E
A
1
2
1.6 Extension of measure to σalgebra, using
outer measure
We can now prove the central:
Theorem 1.6.1. Let m be a measure on an algebra / of subsets of X, ˆ m
the associated outer measure, and M the collection of all subsets of X which
are measurable with respect to m.
Then M is a σalgebra containing / and ˆ m is a (countably additive)
measure on M.
Corollary 1.6.2. the measure m on the algebra / can be extended to a
measure (also denoted by m) on the σalgebra M by deﬁning:
m(E) = ˆ m(E) ∀E ∈ M
in particular:
Corollary 1.6.3. the length measure m on the interval algebra 1 can be
extended to a measure (also denoted by m) on the σalgebra M of measurable
sets w.r.t. m. We call the elements of M the Lebesgue Measureable sets,
and the extended measure the (onedimensional) Lebesgue Measure.
1 ⊂ ¦Lebesgue Measurable Sets¦ ⊂ all subsets of R
length measure m Lebesgue measure outer measure ˆ m
(countably additive) (not countably additive
=⇒ not a measure)
` ↓
[0, ∞]
16
Proof. 1. / ⊂ M
Let E ∈ /, let A ⊂ X
Need to show:
ˆ m(A) = ˆ m(A∩ E) + ˆ m(A ∩ E
)
For > 0. Let
F
1
, F
2
, . . .
be a sequence in / s.t.
A ⊂
∞
¸
j=1
F
j
and s.t.
∞
¸
j=1
m(F
j
) ≤ ˆ m(A) +
E
A
F
i
Then
ˆ m(A) ≤ ˆ m(A∩ E) +ˆ m(A∩ E
) since ˆ m is subadditive
≤ ˆ m(
¸
∞
j=1
F
j
∩ E) +ˆ m(
¸
∞
j=1
F
j
∩ E
)
≤
¸
∞
j=1
ˆ m(F
j
∩ E) +
¸
∞
j=1
ˆ m(F
j
∩ E
)
=
¸
∞
j=1
m(F
j
∩ E) +
¸
∞
j=1
m(F
j
∩ E
) since F
j
∩ E, F
j
∩ E
∈ /
and ˆ m = m on /
=
¸
∞
j=1
m(F
j
) since m is additive
≤ ˆ m(A) +
17
therefore,
ˆ m(A) ≤ ˆ m(A ∩ E) + ˆ m(A∩ E
) ≤ ˆ m(A) +
∀ > 0, and therefore
ˆ m(A) = ˆ m(A∩ E) + ˆ m(A∩ E
)
as required.
2. M is an algebra
let E, F ∈ M; Let A ⊂ X
4
A
F
E
1
2
3
Then
ˆ m(A) = ˆ m(1 + 2) +ˆ m(3 + 4) since E ∈ M
= ˆ m(1 + 2) +ˆ m(3) +ˆ m(4) since F ∈ M
= ˆ m(1 + 2 + 3) +ˆ m(4) since E ∈ M
therefore
E ∪ F ∈ M
and M closed under ﬁnite unions.
Also, M closed under complements by symmetry in the deﬁnition.
3. M is a σalgebra
Let E =
¸
∞
i=1
E
i
be a countable disjoint union with E
i
∈ M. We need
to show that E ∈ M. Put
F
n
= E
1
∪ E
2
∪ ∪ E
n
18
n
1
A
E E
n 2
F
E
F
n
∈ M since M is an algebra. Let A ⊂ X. Then
ˆ m(A) = ˆ m(A∩ F
n
) +ˆ m(A∩ F
n
) since F
n
∈ M
=
¸
n
k=1
ˆ m(A∩ E
k
) +ˆ m(A∩ F
n
) since E
1
, . . . , E
n
∈ M,
and are disjoint
≥
¸
n
k=1
ˆ m(A ∩ E
k
) +ˆ m(A∩ E
) since E
⊂ F
n
is true ∀n. Therefore:
ˆ m(A) ≥
¸
∞
k=1
ˆ m(A ∩ E
k
) +ˆ m(A∩ E
)
≥ ˆ m(
¸
∞
k=1
A∩ E
k
) +ˆ m(A∩ E
) since ˆ m is countably subadditive
(*)
= ˆ m(A∩ E) +ˆ m(A∩ E
)
≥ ˆ m(A) since ˆ m is subadditive
All the above are equalities:
ˆ m(A) = ˆ m(A∩ E) + ˆ m(A ∩ E
)
therefore
E ∈ M
and M is closed under countable disjoint unions. But any countable
union:
E
1
∪ E
2
∪ E
3
∪
19
can be written as a countable disjoint union
E
1
∪ (E
2
∩ F
1
) ∪ (E
3
∩ F
2
) ∪
where
F
n
= E
1
∪ ∪ E
n
therefore M is closed under countable unions as required.
4. ˆ m is countably additive on M
Put A = E in (*) to get
∞
¸
k=1
ˆ m(E
k
) = ˆ m
∞
¸
k=1
E
k
as required.
1.7 Increasing Unions, Decreasing Intersec
tions
Deﬁnition We use the notation E
j
↑ E to denote that
2
1
E
E
E
E
1
⊂ E
2
⊂ ⊂ E
j
⊂
is an increasing sequence of sets such that
∞
¸
j=1
E
j
= E
Deﬁnition We use the notation E
j
↓ E to denote that
1
E
E
2
E
20
E
1
⊃ E
2
⊃ ⊃ E
j
⊃
is a decreasing sequence of sets such that
∞
¸
j=1
E
j
= E
Theorem 1.7.1. E
j
↑ E =⇒ lim
j→∞
m(E
j
) = m(E)
Proof. 1. if m(E
j
) = ∞ for some j then the result holds
2. if m(E
j
) is ﬁnite ∀j then
E = E
1
∪ (E
2
∩ E
1
) ∪ (E
3
∩ E
2
) ∪
is a countable disjoint union and
m(E) = m(E
1
) + [m(E
2
) −m(E
1
)] + [m(E
3
) −m(E
2
)] +
= lim
n→∞
m(E
1
) + [m(E
2
) −m(E
1
)] + + [m(E
n
) −m(E
n−1
)]
= lim
n→∞
m(E
n
)
as required.
Theorem 1.7.2.
E
j
↓ E
m(E
1
) = ∞
=⇒ lim
j→∞
m(E
j
) = m(E)
Proof.
(E
1
∩ E
j
) ↑ (E
1
∩ E
)
therefore:
lim
j→∞
m(E
1
∩ E
j
) = m(E
1
∩ E
)
and
lim
j→∞
[m(E
1
) −m(E
j
)] = m(E
1
) −m(E)
so
lim
j→∞
m(E
j
) = m(E)
as required.
21
Example
(a −
1
n
, b] ↓ [a, b]
£
¢
a −
1
n
a b
for Lebesgue measure m:
m[a, b] = lim
n→∞
m(a −
1
n
, b]
= lim
n→∞
(b −a +
1
n
)
= b −a
as expected.
1.8 Properties of Lebesgue Measure
We now show that the Lebesgue measure is the only way of extending the
‘length’ measure on the interval algebra 1 to a measure on the Lebesgue
measurable sets.
We need:
Deﬁnition a measure m on an algebra / of subsets of X is called σﬁnite
if ∃ a sequence X
i
in / such that
X
i
↑ X
and m(X
i
) is ﬁnite for all i.
Example the Lebesgue measure m is σﬁnite because:
(−n, n] ↑ R
and m(−n, n] = 2n is ﬁnite
£
¢
−n O n
22
Theorem 1.8.1. (Uniqueness of Extension) Let m be a σﬁnite measure on
an algebra / of subsets of X.
Let M be the collection of measurable sets w.r.t. m.
Let l be any measure on M which agrees with m on /.
Then l(E) = ˆ m(E) ∀E ∈ M
Proof. Let E ∈ M. Then for each sequence ¦A
i
¦, A
i
∈ / covering E:
E ⊂
∞
¸
j=1
A
i
we have
l(E) ≤
¸
∞
i=1
l(A
i
) =
¸
∞
i=1
m(A
i
) since l = m on /
Therefore,
(∗) l(E) ≤ ˆ m(E) by deﬁnition of the outer measure ˆ m
Now let X
i
↑ X and m(X
i
) ﬁnite, X
i
∈ /. Consider
E
i
X
l(X
i
∩ E) +l(X
i
∩ E
) = l(X
i
) = m(X
i
) = ˆ m(X
i
∩ E) + ˆ m(X
i
∩ E)
By (*) it follows that
l(X
i
∩ E) = ˆ m(X
i
∩ E)
Take lim
i→∞
to get l(E) = ˆ m as required.
As a consequence we have:
Lemma 1.8.2. Let m be the Lebesgue measure on R. Then for each mea
surable set E and each c ∈ R we have:
1. m(E +c) = m(E). m is translation invariant
2. m(cE) = [c[m(E)
Proof. Let M be the collection of measurable sets
23
1. deﬁne a measure m
c
on M by
m
c
(E) = m(E +c)
then
m
c
(a, b] = m(a +c, b +c] = (b +c) −(a +c) = b −a = m(a, b]
therefore m
c
agrees with m on the interval algebra 1. Therefore m
c
agrees with m on M.
2. deﬁne a measure m
c
on M by
m
c
(E) = m(cE)
Then
m
c
(a, b] = m(c(a, b])
=
m(ca, cb] c > 0
m¦0¦ c = 0
m[cb, ca) c < 0
=
cb −ca
0
ca −cb
= [c[(b −a)
= [c[m(a, b]
Therefore m
c
agrees with [c[m on the interval algebra. Therefore m
c
agrees with [c[m on M.
1.9 Borel Sets
Deﬁnition if 1 is any collection of subsets of X, we denote by G(1) the
intersection of all the σalgebras of subsets of X which contain 1. We have:
1. G(1) is a σalgebra containing 1
2. if J is any σalgebra which contains 1 then
1 ⊂ G(1) ⊂ J
24
Thus G(1) is the smallest σalgebra of subsets of X which contains 1. G(1)
is called the σalgebra generated by 1.
Deﬁnition The σalgebra generated by the open sets of R is called the
algebra of Borel Sets of R.
Theorem 1.9.1. The σalgebra generated by the interval algebra 1 is the
algebra of Borel sets of R.
Proof. Let 1 be the collection of open sets of R
1. (a, b] =
¸
∞
n=1
(a, b +
1
n
) is the intersection of a countable family of open
sets. Therefore (a, b] ∈ G(1).
Therefore
1 ⊂ G(1)
and
G(1) ⊂ G(1)
2. let V be an open set in R. For each a ∈ V choose an interval I
a
∈ 1
with rational endpoints s.t. a ∈ I
a
⊂ V .
£
¢
¡
a
£
¢
Ia
V
V =
¸
a∈V
I
a
so V is the union of a countable family of elements of 1. Therefore
V ∈ G(1)
implies
1 ⊂ G(1) =⇒ G(1) ⊂ G(1)
So, combining these two results,
G(1) = G(1) = algebra of Borel sets
25
We have:
1 ⊂ Borel Sets in R ⊂ Lebesgue measurable Sets in R
The following theorem shows that any Lebesgue measurable set in R can
be obtained from a Borel set by removing a set of measure zero.
Theorem 1.9.2. Let E be a Lebesgue measurable subset of R. Then there
is a Borel set B containing E such that
B −E = B ∩ E
has measure zero, and hence
m(B) = m(E)
B
E
BE
Proof. 1. Suppose m(E) is ﬁnite. Let k be an integer > 0. Choose a
sequence
I
1
, I
2
, I
3
, . . .
in 1 such that
E ⊂
∞
¸
i=1
I
i
and s.t.
m(E) ≤
∞
¸
i=1
m(I
i
) ≤ m(E) +
1
k
Put B
k
=
¸
∞
i=1
I
i
then E ⊂ B
k
, B
k
is Borel, and
m(E) ≤ m(B
k
) ≤ m(E) +
1
k
Put B =
¸
∞
k=1
B
k
. Then E ⊂ B, B is Borel, and
m(E) ≤ m(B) ≤ m(E) +
1
k
for all k. So
m(E) = m(B) =⇒ m(B −E) = 0
since m(E) is ﬁnite.
26
2. Suppose m(E) = ∞.
−(k + 1) −k k k + 1
Put
E
k
= ¦x ∈ E : k ≤ [x[ < k + 1¦
Then
E =
∞
¸
k=0
E
k
is a countable disjoint union and m(E
k
) is ﬁnite.
For each integer k choose a Borel set B
k
s.t.
E
k
⊂ B
k
and
m(B
k
−E
k
) = 0
Put B =
¸
∞
k=1
B
k
. Then E ⊂ B, B is Borel, and
B −E ⊂
∞
¸
k=1
(B
k
−E
k
)
therefore
m(B −E) ≤
∞
¸
k=1
m(B
k
−E
k
) =
∞
¸
k=1
0 = 0
as required.
E
1
E
2
B
1
B
2
27
28
Chapter 2
Integration
2.1 Measure Space, Measurable sets
Deﬁnition We ﬁx a set X, a σalgebra M of subsets of X, and a measure
m on M. The triple (X, M, m) is then called a measure space, the elements
of M are called the measurable sets of the measure space.
Deﬁnition We call a function
f : X −→ [−∞, ∞]
measurable if
f
−1
(α, ∞] = ¦x ∈ X : f(x) > α¦
is a measurable ∀α ∈ R.
f
R
α
>α >α
Example The collection
¦E ⊂ R : f
−1
(E) is measurable¦
29
is a σalgebra containing the sets
¦(α, β] : α, β ∈ R¦
therefore contains the Borel sets. Therefore f
−1
is measurable for each Borel
set B ⊂ R.
2.2 Characteristic Function
Deﬁnition If E ⊂ X, we denote by χ
E
the function on X:
χ
E
(x) =
1 x ∈ E
0 x ∈ E
χ
E
is called the characteristic function of E.
E
R
1
E
Deﬁnition A real valued function
φ : X −→R
is called simple if it takes only a ﬁnite number of distinct values
a
1
, a
2
, . . . , a
n
(say). Each simple function φ can be written in a unique way as:
φ = a
1
χ
E
1
+ +a
n
χ
En
where a
1
, . . . , a
n
are distinct and
X = E
1
∪ ∪ E
n
is a disjoint union.
30
2.3 The Integral
Deﬁnition If φ is a nonnegative measurable simple function with φ =
a
1
χ
E
1
+ +a
n
χ
En
; X = E
1
∪ ∪E
n
disjoint union we deﬁne the integral
of φ w.r.t. the measure m to be
φdm = a
1
m(E
1
) + +a
n
m(E
n
) ∈ [0, ∞]
1
R
E E
a
1 2
2
a
Recall that 0.∞= 0, a.∞= ∞ if a > 0.
Deﬁnition if E is a measurable subset of X and φ is a nonnegative mea
surable simple function on X, we deﬁne the integral of φ over E w.r.t. the
measure m to be:
E
φdm =
φχ
E
dm
We note that:
1.
(φ +ψ) dm =
φdm+
ψ dm
2.
cφdm = c
φdm ∀c ≥ 0
To see 1. we put:
φ =
¸
a
i
χ
E
i
ψ =
¸
b
j
χ
F
j
Let ¦c
k
¦ be the set of distinct values of ¦a
i
+b
j
¦. Then
φ +ψ =
¸
c
k
χ
G
k
where G
k
=
¸
E
i
∩ F
j
, the union taken over ¦i, j : a
i
+b
j
= c
k
¦.
31
(φ +ψ) =
¸
c
k
m(G
k
)
=
¸
k
c
k
¸
{i,j:a
i
+b
j
=c
k
}
m(E
i
∩ F
j
)
=
¸
i,j
(a
i
+b
j
)m(E
i
∩ F
j
)
=
¸
a
i
m(E
i
∩ F
j
) +
¸
b
j
m(E
i
∩ F
j
)
=
¸
i
a
i
m(E
i
) +
¸
j
b
j
m(F
j
)
=
φdm+
ψ dm
which proves 1.
A very useful property of the integral is:
Theorem 2.3.1. Fix a simple nonnegative measurable function φ on X.
For each E ∈ M put
λ(E) =
E
φdm
Then λ is a measure on M.
Proof. Let
φ = a
1
χ
E
1
+ a
n
χ
En
for a
i
≥ 0. Then
λE =
E
φdm
=
φχ
E
dm
=
(a
1
χ
E
1
∩E
+ +a
n
χ
En∩E
) dm
= a
1
m(E
1
∩ E) + +a
n
m(E
n
∩ E)
= a
1
m
E
1
(E) + +a
n
m
En
(E)
therefore
λ = a
1
m
E
1
+ +a
n
m
En
is a linear combination of measures with nonnegative coeﬃcients. Therefore
λ is a measure.
32
Corollary 2.3.2. 1. If E =
¸
∞
n=1
E
n
is a countable disjoint union then
E
φdm =
∞
¸
n=1
En
φdm
2. If E =
¸
∞
n=1
E
n
is a countable increasing union then
E
φdm = lim
n→∞
En
φdm
3. If E =
¸
∞
n=1
E
n
is a countable decreasing intersection and
E
1
φdm <
∞ then
E
φdm = lim
n→∞
En
φdm
We can now deﬁne the integral of any nonnegative measurable function.
Deﬁnition Let f : X −→ [0, ∞] be a measurable function. Then we deﬁne
the integral of f w.r.t. the measure m to be:
f dm = sup
φ
φdm
where the sup is taken over all simple measurable functions φ such that:
0 ≤ φ ≤ f
f
φ
Deﬁnition If E is a measurable subset of X then we deﬁne the integral of
f over E w.r.t. m to be:
E
f dm =
fχ
E
dm
33
f
Ε
PSfrag replacements
fχ
E
We then have:
1. f ≤ g =⇒
f dm ≤
g dm
2. E ⊂ F =⇒
E
f dm ≤
F
f dm
2.4 Monotone Convergence Theorem
We can now prove our ﬁrst important theorem on integration.
Theorem 2.4.1. (Monotone Convergence Theorem, MCT) Let ¦f
n
¦ be a
monotone increasing sequence of nonnegative measurable functions. Then
limf
n
dm = lim
f
n
dm
Proof. We write f
n
↑ f to denote that ¦f
n
¦ is an increasing sequence of
functions with limf
n
= f. We need to prove that:
lim
f
n
dm =
f dm
1.
f
n
≤ f
n+1
≤ f
for all n. Therefore:
f
n
dm ≤
f
n+1
dm ≤
f dm
for all n. Therefore:
lim
f
n
dm ≤
f dm
34
2. Let φ be a simple measurable function s.t.:
0 ≤ φ ≤ f
f
f
φ
n
Let 0 < α < 1. For each integer n > 0 put
E
n
= ¦x ∈ X : f
n
(x) ≥ αφ(x)¦
so E
n
↑ X. Now:
n
f
E
n
f
αφ
En
αφdm ≤
En
f
n
dm ≤
f
n
dm
Let n →∞:
α
φdm ≤ lim
f
n
dm
is true ∀0 < α < 1. Therefore:
φdm ≤ lim
f
n
dm
so:
f dm ≤ lim
f
n
dm
hence
f dm = lim
f
n
dm
35
Deﬁnition In probability theory we have a measure space:
¸
X , M , P
sample events
space
with P(X) = 1. X is the sure event.
A measurable function:
f : X −→R
is called a random variable.
P¦x : f(x) ∈ B¦
is the probability that the random variable f takes value in the Borel set
B ⊂ R. If
f = a
1
χ
E
1
+ +a
n
χ
En
with a
1
, . . . , a
n
; E
1
, . . . , E
n
disjoint, and E
1
∪ ∪ E
n
= X, then the proba
bility that f takes value a
i
is
P¦x ∈ X : f(x) = a
i
¦ = P(E
i
)
and
f dP = a
1
P(E
1
) + +a
n
P(E
n
)
is the average value or the expectation of f.
We deﬁne the expectation of any random variable f to be:
E(f) =
f dP
2.5 Existence of Monotone Increasing Simple
Functions converging to f
In order to apply the MCT eﬀectively we need:
Theorem 2.5.1. Let f be a nonnegative measurable function f : X −→
[0, ∞]. Then there exists a monotone increasing sequence φ
n
of simple mea
surable functions converging to f
36
Proof. Put each integer n > 0:
φ
n
(x) =
k
2
n
k
2
n
≤ f(x) <
k+1
2
n
k = 0, 1, 2, . . . , n2
n
−1
n f(x) ≥ n
n
0
PSfrag replacements
k
2
n
k+1
2
n
Then
1. 0 ≤ φ
n
(x) ≤ φ
n+1
(x)
2. each φ
n
is simple and measurable
3. lim
n→∞
φ
n
(x) = f(x)
Theorem 2.5.2. Let f, g be nonnegative measurable functions mapping X
to [0, ∞] and c ≥ 0. Then
1.
cf dm = c
f dm
2.
(f +g) dm =
f dm+
g dm
Proof. Let φ
n
, ψ
n
be monotonic increasing sequences of nonnegative simple
functions with f = limφ
n
, g = limψ
n
. Then
1.
cf dm
MCT
= lim
cφ
n
dm = c lim
φ
n
dm
MCT
= = c
f dm
2.
(f + g) dm
MCT
= lim
(φ
n
+ ψ
n
) = lim
φ
n
dm + lim
ψ
n
dm
MCT
=
f dm+
g dm
This enables us to deal with series:
37
Theorem 2.5.3. Let f
n
be a sequence of nonnegative measurable functions
X −→[0, ∞]. Then:
(
∞
¸
n=1
f
n
) dm =
∞
¸
n=1
f
n
dm
Proof. put s
n
= f
1
+ +f
n
sum to n terms. s
n
is monotone increasing:
( lim
n→∞
s
n
) dm
MCT
= lim
n→∞
s
n
dm = lim
n→∞
n
¸
r=1
f
r
dm = lim
n→∞
n
¸
r=1
f
r
dm
Therefore:
(
∞
¸
r=1
f
r
) dm =
∞
¸
r=1
f
r
dm
Theorem 2.5.4. Let f : X −→[0, ∞] be a nonnegative measurable and put
λ(E) =
E
f dm
for each E ∈ M. Then λ is a measure in M.
Proof. Let E =
¸
∞
k=1
E
k
be a countable disjoint union. Then
λ(E) =
E
f dm
=
fχ
E
dm
=
(
¸
∞
k=1
fχ
E
k
) dm
MCT
=
¸
∞
k=1
fχ
E
k
dm
=
¸
∞
k=1
E
k
f dm
=
¸
∞
k=1
λ(E
k
)
therefore λ is countably additive, as required.
Corollary 2.5.5. 1. if E =
¸
∞
k=1
E
k
countable disjoint union then
E
f dm =
∞
¸
k=1
E
k
f dm
2. if E
k
↑ E then lim
k→∞
E
k
f dm =
E
f dm
3. If E
k
↓ E and
E
1
f dm < ∞, then lim
k→∞
E
k
f dm =
E
f dm
38
2.6 ’Almost Everywhere’
Deﬁnition Let f, g : X −→ [0, ∞]. Then we say that f = g almost every
where (a.e.) or f(x) = g(x) almost all x ∈ X (a.a.x) if ¦x ∈ X : f(x) = g(x)¦
has measure zero.
Theorem 2.6.1. Let f : X −→ [0, ∞] be nonnegative measurable. Then
f dm = 0 ⇔f = 0 a.e.
Proof. Put E = ¦x ∈ X : f(x) = 0¦
1. Put
E
n
= ¦x ∈ X : f(x) >
1
n
¦
for each integer n > 0, so
f >
1
n
χ
En
E
PSfrag replacements
1
n
E
n
Suppose that
f dm = 0. Then
0 =
f dm ≥
1
n
m(E
n
)
so
m(E
n
) = 0
for all n. But E
n
↑ E, so
m(E) = limm(E
n
) = lim0 = 0
therefore
f = 0 a.e.
2. Suppose f = 0 a.e., so m(E) = 0. Now,
0 ≤ f ≤ limnχ
E
39
E
PSfrag replacements
nχ
E
therefore
0 ≤
f dm
≤
limnχ
E
dm
MCT
= lim
nχ
E
dm
= limnm(E)
= lim0 = 0
so
f dm = 0
Corollary 2.6.2. Let f : X −→ [0, ∞] be nonnegative and measurable and
let E have measure zero. Then
E
f dm = 0
Proof.
fχ
E
= 0 a.e.
therefore
E
f dm =
fχ
E
dm = 0
Corollary 2.6.3. Let f, g : X −→ [0, ∞] be nonnegative and measurable
and f = g a.e. then
f dm =
g dm
40
Proof. Let f = g on E and m(E
) = 0 then
f dm =
E
f dm+
E
f dm =
E
g dm+
E
g dm =
g dm
Thus changing f on a set of measure zero makes no diﬀerence to
f dm.
Also
Theorem 2.6.4. If f
n
are nonnegative and f
n
↑ f a.e. then
lim
f
n
dm =
f dm
Proof. suppose f
n
≥ 0 and f
n
↑ f on E with m(E
) = 0. Then
f dm =
E
f dm+
E
f dm
MCT
= lim
E
f
n
dm+ 0
= lim
E
f
n
dm+
E
f
n
dm
= lim
f
n
dm
So far we have dealt with functions
X −→[0, ∞]
which are nonnegative, but have allowed the value ∞.
Now we look at functions
X −→R
which may be negative and we do not allow ∞ as a value.
Deﬁnition Let f : X −→R be measurable. Put
f
+
(x) =
f(x) f(x) ≥ 0
0 f(x) ≤ 0
f
−
(x) =
−f(x) f(x) ≤ 0
0 f(x) ≥ 0
41
f
PSfrag replacements
f
+
f
PSfrag replacements
f
−
Thus f = f
+
− f
−
and both f
+
, f
−
are nonnegative. We say that f is
integrable w.r.t. m if
f
+
dm < ∞ and
f
−
dm < ∞
and we write
f dm =
f
+
dm−
f
−
dm
and call it the integral of f (w.r.t. measure m).
If E is measurable we write
E
f dm =
fχ
E
dm =
E
f
+
dm−
E
f
−
dm
Theorem 2.6.5. Let f = f
1
−f
2
where f
1
, f
2
are nonnegative and measur
able and
f
1
dm < ∞ and
f
2
dm < ∞
Then f is integrable and
f dm =
f
1
dm−
f
2
dm
42
Proof. 1. f = f
1
−f
2
, therefore f
+
≤ f
1
; f
−
≤ f
2
. So
f
+
dm ≤
f
1
dm < ∞;
f
−
dm ≤
f
2
dm < ∞
therefore f is integrable.
2. f = f
1
−f
2
= f
+
−f
−
. So
f
1
+f
−
= f
+
+f
2
therefore
f
1
dm+
f
−
dm =
f
+
dm+
f
2
dm
so
f
1
dm−
f
2
dm =
f
+
dm−
f
−
dm =
f dm
as required.
Theorem 2.6.6. Let f, g be integrable and f = g a.e. Then
f dm =
g dm
Proof.
f
+
= g
+
a.e. =⇒
f
+
dm =
g
+
dm
f
−
= g
−
a.e. =⇒
f
−
dm =
g
−
dm
so
f dm =
f
+
dm−
f
−
dm =
g
+
dm−
g
−
dm =
g dm
So when integrating we can ignore sets of measure zero.
Theorem 2.6.7. Let f : X −→R be measurable. Then
1. f is integrable ⇔ [f[ is integrable
2. if f is integrable then
f dm
≤
[f[ dm
43
Proof. 1.
f integrable ⇔
f
+
dm < ∞ and
f
−
dm < ∞
⇔
(f
+
+f
−
) dm < ∞
⇔
[f[ dm < ∞
⇔ [f[ integrable
2.
f dm
=
f
+
dm−
f
−
dm
≤
f
+
dm+
f
−
dm
=
[f[ dm
Deﬁnition A complex valued function f : X −→C with
f(x) = f
1
(x) +if
2
(x)
(say) (f
1
, f
2
real) is called integrable if f
1
and f
2
are integrable and we deﬁne:
f dm =
f
1
dm+i
f
2
dm
Thus
'
f dm =
('f) dm
·
f dm =
(·f) dm
We have [f
1
[ ≤ [f[, [f
2
[ ≤ [f[, and therefore
[f[ integrable ⇔ [f
1
[ and [f
2
[ integrable ⇔ f
1
and f
2
integrable ⇔ f
integrable.
We also have:
f, g integrable and c ∈ C =⇒ f +g and cf integrable, and
(f +g) dm =
f dm+
g dm
and
(cf) dm = c
f dm
44
Thus, the set L(X, R, m) of all integrable real valued functions on X is
a real vector space, and the set L(X, C, m) of all integrable complex valued
functions on X is a complex vector space. And on each space:
f −→
f dm
is a linear form.
If f : X −→C is integrable then
f dm =
f dm
e
iθ
(say) θ real
therefore
(real)
f dm
= e
−iθ
f dm
=
(real)
e
−iθ
f dm
=
'[e
−iθ
f] dm
≤
e
−iθ
f
dm
=
[f[ dm
therefore
f dm
≤
[f[ dm
2.7 Integral Notation
Deﬁnition When dealing with 1dimensional Lebesgue measure we write
[a,b]
f dm =
b
a
f(x) dx = −
a
b
f(x) dx
if a ≤ b. It follows that
b
a
f(x) dx +
c
b
f(x) dx =
c
a
f(x) dx
and
b
a
dx = b −a
45
∀a, b, c.
Notice that x is a dummy symbol and that
b
a
f(x) dx =
b
a
f(y) dy =
b
a
f(t) dt =
just as
n
¸
i=1
a
i
b
i
=
n
¸
j=1
a
j
b
j
=
2.8 Fundamental Theorem of Calculus
Theorem 2.8.1. Fundamental Theorem of Calculus Let f : R −→ R be
continuous and put
F(t) =
t
a
f(x) dx
then
F
(t) = f(t)
Proof. Let t ∈ R, let > 0. Then ∃δ > 0 s.t.
[f(t +h) −f(t)[ ≤ ∀[h[ ≤ δ
by continuity of f. Therefore
F(t+h)−F(t)
h
−f(t)
=
1
h
t+h
t
f(x) dx −
1
h
t+h
t
f(t) dx
=
1
h
t+h
t
[f(x) −f(t)] dx
≤
1
h
[h[ =
for all [h[ ≤ δ; h = 0, and hence the result.
Corollary 2.8.2. If G : R −→ R is C
1
(i.e. has a continuous derivative)
then
b
a
G
(x) dx = G(b) −G(a).
Proof. put F(t) =
t
a
G
(x) dx. Then
F
(t) = G
(t)
46
for all t. Therefore
F(t) = G(t) +c
for all t, where c is constant. Therefore
G(b) −G(a) = F(b) −F(a) =
b
a
G
(x) dx
as required.
Theorem 2.8.3. (Change of Variable)
Let
[t
3
, t
4
]
g
−→[t
1
, t
2
]
f
−→R
with f continuous and g ∈ C
1
; g(t
3
) = t
1
, g(t
4
) = t
2
. Then
t
2
t
1
f(x) dx =
t
4
t
3
f(g(y))g
(y) dy
Proof. Put
G(t) =
t
t
1
f(x) dx
then G
(t) = f(t), and therefore:
t
4
t
3
f(g(y))g
(y) dy =
t
4
t
3
G
(g(y)g
(y) dy
=
t
4
t
3
[
d
dy
G(g(y))] dy
= G(g(t
4
)) −G(g(t
3
))
= G(t
2
) −G(t
1
)
=
t
2
t
1
f(x) dx
as required.
To deal with sequences which are not monotone we need the concepts of
liminf and limsup.
Deﬁnition Let
¦a
n
¦ = a
1
, a
2
, a
3
, . . .
47
be a sequence in [−∞, ∞]. Put
b
n
= inf¦a
n
, a
n+1
, a
n+2
, . . .¦
c
n
= sup¦a
n
, a
n+1
, a
n+2
, . . .¦
Then
b
1
≤ b
2
≤ b
n
≤ b
n+1
≤ ≤ c
n+1
≤ c
n
≤ ≤ c
2
≤ c
1
Deﬁne:
liminf a
n
= limb
n
limsup a
n
= limc
n
Then
liminf a
n
≤ limsup a
n
and a
n
converges iﬀ liminf a
n
= limsup a
n
(= lima
n
).
2.9 Fatou’s Lemma
Theorem 2.9.1. (Fatou’s Lemma) Let f
n
be a sequence of nonnegative
measurable functions:
f
n
: X −→ [0, ∞]
Then
liminf f
n
dm ≤ liminf
f
n
dm
Proof. Put
g
r
= inf¦f
r
, f
r+1
, . . .¦
so that
≤ g
r
≤ g
r+1
≤
Now:
f
n
≥ g
r
∀n ≥ r
=⇒
f
n
≥
g
r
∀n ≥ r
=⇒ liminf
f
n
≥
g
r
∀r
=⇒ liminf
f
n
≥ lim
g
r
MCT
=
limg
r
=
liminf f
n
as required.
48
2.10 Dominated Convergence Theorem
Theorem 2.10.1. Lebesgue’s Dominated Convergence Theorem, DCT Let
f
n
be integrable and f
n
→f. Let
[f
n
[ ≤ g
for all n where g is integrable. Then f is integrable and
f = lim
f
n
Proof. [f[ = lim[f
n
[ ≤ g. Therefore [f[ is integrable, and therefore f is
integrable. Now
g ±f
n
≥ 0
therefore
liminf[g ±f
n
] ≤
FATOU
liminf
[g ±f
n
]
so
[g ±f] ≤
g + liminf ±
f
n
so
±
f ≤ liminf ±
f
n
Which gives:
:
f ≤ liminf
f
n
: −
f ≤ −limsup
f
n
therefore
f ≤ liminf
f
n
≤ limsup
f
n
≤
f
which is equivalent to:
lim
f
n
=
f
as required.
For series this leads to:
Theorem 2.10.2. Dominated convergence theorem for series Let f
n
be a
sequence of integrable functions such that
∞
¸
n=1
[f
n
[ dm < ∞
Then
¸
∞
n=1
f
n
is (equal a.e. to) an integrable function and
(
∞
¸
n=1
f
n
) dm =
∞
¸
n=1
f
n
dm
49
Proof. Put
s
n
= f
1
+f
2
+ +f
n
sum to n terms. Then
[s
n
[ ≤ [f
1
[ + +[f
n
[ ≤
∞
¸
r=1
[f
r
[ = g
(say). Then
g dm =
¸
∞
r=1
[f
r
[ dm
MCT
=
¸
∞
r=1
[f
r
[ dm < ∞
Therefore g is integrable (a.e. equal to an integrable fn). Therefore lims
n
is
integrable and
lims
n
dm
DCT
= lim
s
n
dm
therefore
∞
¸
r=1
f
r
dm =
∞
¸
r=1
f
r
dm
as required.
2.11 Diﬀerentiation under the integral sign
Another useful application.
Theorem 2.11.1. Diﬀerentiation under the integral sign Let f(x, t) be an
integrable function of x ∈ X for each a ≤ t ≤ b and diﬀerentiable w.r.t t.
Suppose
∂f
∂t
(x, t)
≤ g(x)
for all a ≤ t ≤ b, where g is integrable. Then
d
dt
f(x, t) dx =
∂f
∂t
(x, t) dx
Proof. Put
F(t) =
f(x, t) dx
50
Let a ≤ t ≤ b. Choose a sequence t
n
in [a, b] s.t. limt
n
= t and t
n
= t. Then,
by the Mean Value Theorem:
[f(x, t
n
) −f(x, t)[ =[t
n
−t[
∂f
∂t
(x, c(n, x, t))
≤ [t
n
−t[g(x)
Therefore
f(x, t
n
) −f(x, t)
t
n
−t
≤ g(x)
so
lim
F(tn)−F(t)
tn−t
= lim
f(x,tn)−f(x,t)
tn−t
dx
DCT
=
lim
f(x,tn)−f(x,t)
tn−t
dx
=
∂f
∂t
(x, t) dx
and therefore
dF
dt
=
∂f
∂t
(x, t) dx
as required.
51
52
Chapter 3
Multiple Integration
3.1 Product Measure
We have established the Lebesgue measure and Lebesgue integral on R. To
consider integration on
R
2
= R R
we use the concept of a product measure.
We proceed as follows:
Deﬁnition Let l be a measure on a σalgebra L of subsets of X. Let m be
a measure on a σalgebra ´ of subsets of Y .
Call
¦AB : A ∈ L, B ∈ ´¦
the set of rectangles in X Y
B
A
PSfrag replacements
A B
53
Lemma 3.1.1. Let AB =
¸
∞
i=1
A
i
B
i
be a rectangle written as a countable
disjoint union of rectangles. Then
l(A)m(B) =
∞
¸
i=1
l(A
i
)m(B
i
)
Proof. We have
χ
A×B
=
∞
¸
i=1
χ
A
i
×B
i
=⇒ χ
A
(x)χ
B
(y) =
∞
¸
i=1
χ
A
i
(x)χ
B
i
(y)
Fix x and integrate w.r.t. m term by term using the Monotone Convergence
Theorem:
χ
A
(x)m(B) =
∞
¸
i=1
χ
A
i
(x)m(B
i
)
Now integrate w.r.t. x using MCT to get:
l(A)m(B) =
∞
¸
i=1
l(A
i
)m(B
i
)
as required.
Deﬁnition Let / be the collection of all ﬁnite unions of rectangles in XY .
Each element of / is a ﬁnite disjoint union of rectangles. For each E ∈ /
such that
E =
∞
¸
i=1
A
i
B
i
is a countable disjoint union of rectangles we deﬁne:
π(E) =
∞
¸
i=1
l(A
i
)m(B
i
)
Theorem 3.1.2. π is welldeﬁned and is a measure on /.
Proof. 1. welldeﬁned
Suppose
E =
∞
¸
i=1
A
i
B
i
=
∞
¸
j=1
C
j
D
j
54
then
A
i
B
i
=
∞
¸
j=1
(A
i
∩ C
j
) (B
i
∩ D
j
)
Therefore, by Lemma 3.1.1
l(A
i
)m(B
i
) =
∞
¸
j=1
l(A
i
∩ C
j
)m(B
i
∩ D
j
)
and hence,
¸
∞
i=1
l(A
i
)m(B
i
) =
¸
∞
i=1
¸
∞
j=1
l(A
i
∩ C
j
)m(B
i
∩ D
j
)
=
¸
∞
j=1
l(C
j
)m(D
j
)
Dj
Bi
PSfrag replacements
A
i
C
j
D
i
B
j
2. countably additive Let E =
¸
∞
i=1
E
i
be a countable disjoint union with
E, E
i
∈ /.
E
i
=
n
i
¸
j=1
A
ij
B
ij
is a ﬁnite disjoint union (say). Then
E =
∞
¸
i=1
n
i
¸
j=1
A
ij
B
ij
Therefore
π(E) =
∞
¸
i=1
n
i
¸
j=1
l(A
ij
)m(B
ij
) =
∞
¸
i=1
π(E
i
)
Therefore π is countably additive, as required.
55
Deﬁnition The measure π on / extends to a measure (also denoted by π
and called the product of the measures l and m) on the σalgebra (denoted
L ´) of all subsets of X Y which are measurable w.r.t. π.
Similarly, if
(X
1
, M
1
, m
1
), . . . , (X
n
, M
n
, m
n
)
is a sequence of measure spaces then we have a product measure on a σ
algebra of subsets of
X
1
X
n
s.t.
π(E
1
E
n
) = m
1
(E
1
)m
2
(E
2
) . . . m
n
(E
n
)
In particular, starting with 1dimensional Lebesgue measure on R we get
ndimensional Lebesgue Measure on
R R = R
n
Example If we have two successive, independent events with independent
probability measures, P
1
, P
2
, then the probability of the ﬁrst event E and
the second event F is:
P(E F) = P[(E Y ) ∩ (X F)]
= P(E Y )P(X F) by independence
= P
1
(E)P
2
(F)
(i.e. product measure)
3.2 Monotone Class
Deﬁnition A nonempty collection M of subsets of X is called a monotone
class if
1. E
n
↑ E, E
n
∈ M =⇒ E ∈ M
2. E
n
↓ E, E
n
∈ M =⇒ E ∈ M
i.e. M is closed under countable increasing unions and under countable
decreasing intersections.
We have: M is a σalgebra =⇒ M is a monotone class. Therefore, there
are more monotone classes than σalgebras.
56
Deﬁnition If 1 is a nonempty collection of subsets of X and if M is the
intersection of all the monotone classes of subsets of X which contain 1, then
M is a monotone class, called the monotone class generated by 1.
M is the smallest monotone class containing 1 and
1 ⊂ monotone class ⊂ σalgebra
generated by 1 generated by 1
3.3 Ring of Subsets
Deﬁnition A nonempty collection R of subsets of X is called a ring of
subsets of X if:
E, F ∈ R =⇒
E ∪ F ∈ R
E ∩ F
∈ R
i.e. R is closed under ﬁnite unions and under relative complements.
Example The collection of all ﬁnite unions of rectangles contained in the
interior X of a ﬁxed circle in R
2
is a ring of subsets of X.
Note:
1. if R is a ring then R is closed under ﬁnite intersections since E ∩ F =
E ∩ (E ∩ F
)
.
2. if R is a ring then ∅ ∈ R since E ∩ E
= ∅
3. if R is a ring of subsets of X and if X ∈ R then R is an algebra of
subsets of X since E
= X ∩ E
, so R is closed under complements.
Theorem 3.3.1. (Monotone class lemma)
Let R be a ring of subsets of X and let M be the monotone class generated
by R.
Then M is a ring.
Proof. We have to show M is closed under ﬁnite unions and relative comple
ments. i.e. that:
E ∩ F
, E ∪ F, E
∩ F ∈ M
for all E, f ∈ M. So for each E ∈ M put
M
E
= ¦F ∈ M : E ∩ F
, E ∪ F, E
∩ F ∈ M¦
and we must show
M
E
= M
for all E ∈ M. Now
57
1. M
E
⊂ M by deﬁnition
2. M
E
is a monotone class, because:
F
n
↑ F, F
n
∈ M
E
=⇒ (E ∩ F
n
) ↓ (E ∩ F
) (E ∪ F
n
) ↑ (E ∪ F) (E
∩ F
n
) ↑ (E
∩ F)
with E ∩ F
n
E ∪ F
n
E
∩ F
n
all ∈ M
=⇒ E ∩ F
E ∪ F E
∩ F all ∈ M
(Mi monotone)
=⇒ F ∈ M
E
therefore M
E
is closed under countable increasing unions, and similarly
M
E
is closed under countable decreasing intersections.
3. E ∈ R
=⇒ R ⊂ M
E
since R is a ring. Therefore M
E
= M by (i), (ii) since
M is smallest monotone class containing R.
4. F ∈ M
E
∀E ∈ R, F ∈ M by (iii). Therefore E ∈ M
F
∀E ∈ R, F ∈
M.
Therefore
R ⊂ M
F
∀F ∈ M
and therefore
M
F
= M ∀F ∈ M
as required.
Corollary 3.3.2. Let M be the monotone class generated by a ring R, and
let X ∈ M.
Then M = ((R) the σalgebra generated by R.
Proof. M is a ring and X ∈ M
E ∈ M =⇒ E
= E
∩ X ∈ M
therefore M closed under compliments, and therefore M is an algebra.
58
Also, if E
n
is a sequence in M and E =
¸
∞
1
E
n
put
F
n
= E
1
∪ ∪ E
n
∈ M
then F
n
↑ E, and therefore E ∈ M.
Therefore, M is a σalgebra, hence the result.
Corollary 3.3.3. Let M be the monotone class generated by an algebra /.
Then M = ((/) the σalgebra generated by /.
Proof. X ∈ / =⇒ X ∈ M, and hence the result by previous corollary.
3.4 Integration using Product Measure
Let l be a measure on a σalgebra L of subsets of X.
Let m be a measure on a σalgebra ´ of subsets of Y .
Let / be the algebra of ﬁnite unions of rectangles AB; A ∈ L, B ∈ ´.
Let π be the product measure on the σalgebra ((/).
If
f : X −→ [0, ∞]
g : Y −→ [0, ∞]
F : X Y −→ [0, ∞]
are nonnegative measurable, write:
f dl =
X
f(x) dx
f dm =
Y
g(y) dy
F dπ =
X×Y
F(x, y) dxdy
If E ⊂ X Y write
E
x
= ¦y ∈ Y : (x, y) ∈ E¦
E
y
= ¦x ∈ X : (x, y) ∈ E¦
59
E
x
y
x
E
y
Theorem 3.4.1. if E ∈ G(A) and l, m are σﬁnite then:
π(E) =
X
m(E
x
) dx =
Y
l(E
y
) dy
Proof. To show
π(E) =
X
m(E
x
) dx (3.1)
1. Suppose l, m are ﬁnite measures: l(X) < ∞, m(Y ) < ∞. Let ^ be
the collection of all E ∈ ((/) s.t. Equation (3.1) holds. We will show
that N is a monotone class containing /:
/ ⊂ N ⊂ ((/)
and hence N = ((/) since by the corollary to the monotone class
lemma, ((/) is the monotone class generated by /.
(a) / ∈ N.
If E = AB is a rectangle then
m(E
x
) = χ
A
(x)m(B)
60
B
E
A
x
then
X
m(E
x
) dx = l(A)m(B) = π(E)
therefore E ∈ N. Now each element of / can be written as a
ﬁnite disjoint union of rectangles, therefore
/ ⊂ N
(b) Let E
n
↑ E, E
n
∈ N. So,
(E
n
)
x
↑ E
x
for each x ∈ X.
x
E
E
n
61
Then
X
m(E
x
) dx =
X
limm((E
n
)
x
) dx
MCT
= lim
X
m((E
n
)
x
) dx
= limπ(E
n
)
= π(E)
therefore E ∈ N, and N is closed under converging unions.
(c) Let E ∈ N. Then
X
m((E
)
x
) dx =
X
[m(Y ) −m(E
x
)] dx
= l(X)m(Y ) −π(E)x
= π(X Y ) −π(E)
= π(E
).
Therefore E
∈ N, and N is closed under complements. Therefore
N is a monotone class, and therefore N = ((/).
2. Suppose l, m are σﬁnite,
A
n
↑ X, B
n
↑ Y
(say) with
l(A
n
) < ∞, m(B
n
) < ∞
Then
Z
n
↑ (X Y )
where Z
n
= A
n
B
n
and π(Z
n
) < ∞.
Let E ∈ ((/). Then (E ∩ Z
n
) ↑ E, and
(E ∩ Z
n
)
x
↑ E
x
for all x ∈ X.
62
x
E
Z
n
So:
X
m(E
x
) dx =
X
limm((E ∩ Z
n
)
x
) dx
MCT
= lim
X
m((E ∩ Z
n
)
x
) dx
(since Z
n
has ﬁnite measure)
= limπ(E ∩ Z
n
)
= π(E)
as required.
3.5 Tonelli’s Theorem
Theorem 3.5.1. (Repeated integral of a nonnegative funtcion)
Let F : X Y −→[0, ∞] be nonnegative measurable. Then
X
¸
Y
F(x, y) dy
dx =
X×Y
F(x, y) dxdy =
Y
¸
X
F(x, y) dx
dy
(notation as before.)
63
Proof. To show
X
¸
Y
F(x, y) dy
dx =
X×Y
F(x, y) dxdy (3.2)
1. Equation 3.2 holds for F = χ
E
, since
X
¸
Y
χ
E
(x, y) dy
dx =
X
m(E
x
) dx = π(E) =
X×Y
χ
E
(x, y) dxdy
therefore Equation 3.2 also holds for any simple function.
2. (General Case)
∃ monotone increasing sequence of nonnegative measurable functions
with F = limF
n
X
Y
F(x, y) dy
dx =
X
Y
limF
n
(x, y) dy
dx
MCT
=
X
lim
Y
F
n
(x, y) dy
dx
MCT
= lim
X
Y
F
n
(x, y) dy
dx
(1.)
= lim
X×Y
F
n
(x, y) dxdy
MCT
=
X×Y
limF
n
(x, y) dxdy
=
X×Y
F(x, y) dxdy
3.6 Fubini’s Theorem
Theorem 3.6.1. (Repeated Integral of an integrable function)
Let F be an integrable function on X Y . Then:
X
¸
Y
F(x, y) dy
dx =
X×Y
F(x, y) dxdy =
Y
¸
X
F(x, y) dx
dy
Where
Y
F(x, y) dy is equal to an integrable function of x a.e., and
X
F(x, y) dx
is equal to an integrable function of y a.e.
64
Proof.
X×Y
F(x, y) dxdy =
X×Y
F
+
(x, y) dxdy −
X×Y
F
−
(x, y) dxdy
Tonelli
=
X
Y
F
+
(x, y) dy
dx −
X
Y
F
−
(x, y) dy
dx
Therefore
Y
F
+
(x, y) dy, and
Y
F
−
(x, y) dy are each ﬁnite a.a.x., and each
has a ﬁnite integral w.r.t. x. (*)
So:
F(x, y) = F
+
(x, y) −F
−
(x, y)
is integrable w.r.t y a.e., and:
Y
F(x, y) dy =
Y
F
+
(x, y) dy −
Y
F
−
(x, y) dy
for a.a.x, and is an integrable function of x (by (*)), with:
X
Y
F(x, y) dy
dx =
X
Y
F
+
(x, y) dy
dx −
X
Y
F
−
(x, y) dy
dx
=
X×Y
F(x, y) dxdy
Theorem 3.6.2. Let f be an integrable function R −→R. Then
1.
f(x +c) dx =
f(x) dx
2.
f(cx) dx =
1
c
f(x) dx (c = 0)
Proof. These are true for f = χ
E
, because:
1.
χ
E
(x +c) dx =
χ
E−c
(x) dx = m(E −c) = m(E) =
χ
E
(x) dx
2.
χ
E
(cx) dx =
χ1
c
E
(x) dx = m(
1
c
E) =
1
[c[
m(E) =
1
[c[
χ
E
(x) dx
Therefore, these are true for f simple, and therefore true for f nonnegative
measurable (by MCT), since ∃f
n
simple, s.t. f
n
↑ f. Therefore, these are
true for f = f
+
−f
−
integrable.
65
We now see how to deal with integration on change of variable on R
n
.
Recall
f(cx) dx =
1
c
f(x) dx.
Theorem 3.6.3. (Linear change of variable)
Let R
n
A
−→R
n
be an invertable matrix, then
f(Ax) dx =
1
[ det A[
f(x) dx
Proof. we can reduce A to the unit matrix I by a sequence of elementary
row operations.
1. To replace row i by row i + c row j, multiply A by:
N =
¸
¸
¸
¸
¸
¸
¸
¸
¸
¸
1
.
.
.
.
.
.
.
.
.
1 c
.
.
.
1
.
.
.
with c in the i,j position.
2. To interchange row i and row j, multiply A by:
P =
¸
¸
¸
¸
¸
¸
¸
¸
¸
¸
¸
¸
¸
¸
¸
¸
¸
¸
¸
1
.
.
.
.
.
.
1
.
.
.
.
.
.
0 1
.
.
. 1
.
.
.
.
.
.
.
.
. 1
.
.
.
1 0
1
.
.
.
1
66
3. To replace row i by c row j, multiply A by:
D =
¸
¸
¸
¸
¸
¸
¸
¸
¸
¸
1
.
.
.
1
c
1
.
.
.
1
Therefore, there exists matrices B
1
, . . . , B
k
, each of type N, P or D s.t.
B
1
B
2
B
k
A = I
and therefore,
A = B
−1
k
B
−1
2
B
−1
1
is a product of matrices of type N, P or D.
Now, if the theorem holds for matrices A, B then it also holds for AB
since:
f(ABx) dx =
1
[ det B[
f(Ax) dx =
1
[ det B[[ det A[
f(x) dx =
1
[ det AB[
f(x) dx
therefore, it is suﬃcient to prove it for matrices of type N, P, D.
1. Let
N =
¸
¸
¸
¸
¸
¸
¸
¸
1 c 0 0
0 1 0 0
0 0 1 0
.
.
.
.
.
.
1
(say). det N = 1. Then
f(Nx) dx =
f(x
1
+cx
2
, x
2
, . . . , x
n
) dx
1
dx
2
dx
n
=
f(x
1
, x
2
, . . . , x
n
) dx
1
dx
2
dx
n
(by Fubini, and translation invariance)
=
1
 det N
f(x) dx
67
2. Let
P =
¸
¸
¸
¸
¸
¸
¸
¸
0 1 0 0
1 0 0 0
0 0 1 0
.
.
.
.
.
.
1
(say). det P = −1. Then
f(Px) dx =
f(x
2
, x
1
, . . . , x
n
) dx
1
dx
2
dx
n
=
f(x
1
, x
2
, . . . , x
n
) dx
1
dx
2
dx
n
(by Fubini)
=
1
 det P
f(x) dx
3. Let
D =
¸
¸
¸
¸
¸
¸
¸
¸
c 0 0 0
0 1 0 0
0 0 1 0
.
.
.
.
.
.
1
(say). det N = c. Then
f(Dx) dx =
f(cx
1
, x
2
, . . . , x
n
) dx
1
dx
2
dx
n
=
1
c
f(x
1
, x
2
, . . . , x
n
) dx
1
dx
2
dx
n
(by Fubini)
=
1
 det D
f(x) dx
Corollary 3.6.4. if E ⊂ R
n
is measurable and R
n
A
−→R
n
is a linear homo
morphism then
m(A(E)) = [ det A[m(E)
Proof.
m(E) =
χ
E
(x) dx
=
χ
A(E)
(AX) dx
=
1
 det A
χ
A(E)
(x) dx
=
1
 det A
m(A(E))
68
as required.
69
70
Chapter 4
Diﬀerentiation
4.1 Diﬀerentiation
If R
f
−→R is a real valued function od a real variable then the derivative of
f at a is deﬁned to be:
f
(a) = lim
h−→0
f(a +h) −f(a)
h
(4.1)
We want to deﬁne the derivative f
(a) when f is a vectorvalued function of
a vector variable:
f : M −→ N
where m, N are real (or complex) vector spaces.
We cannot use Equation (4.1) directly since we don’t know how to divide
f(a +h) −f(a), which is a vector in N, by h, which is a vector in M
So, we rewrite Equation (4.1) as:
f(a +h) = f(a)
....
(constant)
+ f
(a)h
. .. .
(linear in h)
+ φ(h)
....
(remainder)
where
lim
h−→0
φ(h)
h
=
f(a +h) −f(a)
h
−f
(a)
This suggests that we take M, N to be normed spaces and deﬁne f
(a) to
be a linear operator such that
f(a +h) = f(a) +f
(a)h +φ(h)
where
lim
h−→0
φ(h)
h
= 0
71
(h)
a
h
a+h
f
f(a)
f’(a) h
a+h
f(a+h)
φ
This f
(a)h is the linear approx to the change in f when variable changes by
h from a to a +h.
4.2 Normed Space
Deﬁnition let M be a real or complex vector space. Then M is called a
normed space if a function . exists,:
M −→R
x −→ x
is given on M (called the norm on M), such that:
1. x ≥ 0
2. x = 0 ⇔ x = 0
3. αx = [α[ x ∀ scalar α
4. x +y ≤ x +y (triangle inequality)
Example 1. R
n
with (α
1
, . . . , α
n
) =
α
2
1
+ +α
2
n
is called the Eu
clidean Norm on R
n
.
2. C
n
with (α
1
, . . . , α
n
) =
[α
1
[
2
+ +[α
n
[
2
is called the Hilbert
Norm on C
n
.
3. C
n
with (α
1
, . . . , α
n
) = max¦[α
1
[, . . . , [α
n
[¦ is called the sup norm
on C
n
.
4. if (X, M, m) is a measure space the the set of integrable functions
L
(X, R, m) with
f =
[f[ dm
is called the L
norm (functions are to be regarded as equal if they are
equal a.e.)
72
4.3 Metric Space
Deﬁnition a set X is called a metric space if a function
D : X X −→R
is given (called a metric on X) such that
1. d(x, y) ≥ 0
2. d(x, y) = 0 ⇔ x = y
3. d(x, z) ≤ d(x, y) +d(y, z). This is known as the triangle inequality
Deﬁnition If a ∈ X and r > 0 the we write
B
X
(a, r) = ¦x ∈ X : d(a, x) < r¦
and call it the ball in X centre a, radius r.
Example if M is a normed space then M is also a metric space with
d(x, y) = x −y
4.4 Topological space
Deﬁnition a set X is called a topological space id a collection 1 of subsets
of X is given (called the topology on X) such that:
1. ∅ and X belong to 1
2. if ¦V
i
¦
i∈I
is any family of elements of 1 then
¸
i∈I
v
i
belongs to 1. i.e.
1 is closed under unions
3. if U and V belong to 1 then U ∩V belongs to 1. i.e. 1 is closed under
ﬁnite intersections.
We call the elements of 1 the open sets of the topological space X, or open
in X.
Example Let X be a metric space. Then X is a topological space where we
deﬁne a set V to be open in V if V ⊂ X and each a ∈ V ∃ r > 0 s.t.
B
X
(a, r) ⊂ V
73
a
r
V
Theorem 4.4.1. if X is a metric space then each c ∈ X and s > 0 the ball
B
X
(c, s) is open in X.
Proof. Let a ∈ B
X
(c, s). Put r = s −d(a, c) > 0.
Then
x ∈ B
X
(a, r)
=⇒ d(x, a) < r = s −d(a, c)
=⇒ d(x, a) +d(a, c) < s
=⇒ d(x, c) < s
=⇒ x ∈ B
X
(c, s)
74
a
r
c
x
s
PSfrag replacements
B
X
(a, r)
B
X
(c, s)
and, therefore BX(a, r) ⊂ B
X
(c, s), as required.
4.5 Continuous map of topological spaces
Deﬁnition a map f : X −→ Y of topological spaces is called continuous if:
V open in Y =⇒ f
−1
V open in X
Theorem 4.5.1. let X, Y be metric spaces, then:
f : X −→ Y
is continuous if and only if,
for each a ∈ X, > 0
∃ δ > 0 s.t.
d(x, a) < δ =⇒ d(f(x), f(a)) <
(4.2)
i.e.
fB
X
(a, δ) ⊂ B
Y
(f(a), )
Proof. 1. Let f be continuous. Let a ∈ X, > 0. Then B
Y
(f(a), ) is
open in Y . Therefore, f
−1
B
Y
(f(a), ) is open in X.
75
a
δ
ε
f(a)
PSfrag replacements
f
−1
B
Y
(f(a), )
B
Y
(f(a), )
∃δ > 0 s.t.
B
X
(a, δ) ⊂ f
−1
B
Y
(f(a), )
and
fB
X
(a, δ) ⊂ B
Y
(f(a), )
therefore Equation (4.2) holds.
2. Let Equation (4.2) hold. Let V be open in Y , a ∈ f
−1
V , and so
f(a) ∈ V .
Now, ∃ > 0 s.t.
B
Y
(f(a), ) ⊂ V
f(a)
V
a
PSfrag replacements
f
−1
V
and ∃δ > 0 s.t.
fB
X
(a, δ) ⊂ B
Y
(f(a), ) ⊂ V
therefore,
B
X
(a, δ) ⊂ f
−1
V
so, f
−1
V is open in X, and f is continuous.
76
4.6 Homeomorphisms
Deﬁnition a map f : X −→Y of topolgical spaces is called homeomorphism
if:
1. f is bijective
2. f and f
−1
are continuous
Deﬁnition X is homeomorphic to Y (topologically equivalent) if ∃ a home
omorphism X −→ Y . i.e. ∃ a beijective map under which the open sets of
X correspond to the open sets of Y .
Deﬁnition a property P of a topological space is called a topological prop
erty if X has property P and X homeomorphic to Y =⇒ Y has property
P.
Example ’compactness’ is a topolgical property
Note: we have a category with topological spaces as objects, and con
tinuous maps as morphisms.
4.7 Operator Norm
Deﬁnition Let M, N be ﬁnite dimensional normed spaces. Then we can
make the vector space
L(M, N)
of all linear operators T : M −→ N, into a normed space by deﬁning:
T = sup
x∈M
x=0
Tx
x
T is called the operator norm of T.
We have:
1. If x = 0 and y =
x
x
then
y =
1
x
x = 1
77
x
PSfrag replacements
x
x
i.e. y has unit norm. Therefore
Tx
x
=
T
x
x
= Ty
and
T = sup
y∈M
y=1
Ty
T
y
y
PSfrag replacements
T
2. if α is a scalar then
αT = sup
y=1
αTy = [α[ sup
y=1
Ty = [α[T
3. if S, T : M −→N then
S+T = sup
y=1
(S+T)y ≤ sup
y=1
(Sy+Ty) ≤ sup Sy+supTy = S+T
78
4.
T = 0 ⇐⇒ sup
x=0
Tx
x
= 0
⇐⇒ Tx = 0 ∀x
⇐⇒ Tx = 0 ∀x
⇐⇒ T = 0
(by 2,3,4 the operator norm is a norm).
Theorem 4.7.1. 1. if T : M −→N and x ∈ M then
Tx ≤ T x
2. If L
T
−→M
S
−→N then
ST ≤ S T
Proof. 1.
Tx
x
≤ T
∀x = 0, be deﬁnition. therefore,
Tx ≤ T x
for all x
2.
ST = sup
y=1
STy
≤ sup S Ty (by 1.)
≤ sup
y=1
S T y (by 1.)
= S T
Note:
1. if M is ﬁnite dimensional then every choice of norm on M deﬁnes the
same topology on M
2. a sequence x
r
of points in a topological space X is said to converge to
a ∈ M if, for each open set V containing a ∃N s.t. x
r
∈ V ∀ r ≥ N.
For a normed space M this is the same as: for each > 0 ∃N s.t.
x
r
−a < ∀ r ≥ N.
79
4.8 Diﬀerentiation
Deﬁnition let
M ⊃ V
f
−→N
where M, N are normed spaces and V is open in M. Let a ∈ V . THen f is
diﬀerentiable at a if ∃ a linear operator
M
f
(a)
−→ N
such that
f(a +h) = f(a0) +f
(a)h +φ(h)
where
φ(h)
h
−→ 0 as h −→0
Theorem 4.8.1. if f is diﬀerentiable at a the the operator f
(a) is uniquely
determined by:
f
(a)h = lim
t−→0
f(a+th)−f(a)
t
= directional derivative of f at a along h
=
d
dt
f(a +th)
t=0
a
a+th
a+h
f
(a) is called the derivative of f at a
Proof.
f(a +th) = f(a) +f
(a)th +φ(th)
therefore,
f(a +th) −f(a)
t
−f
(a)h
=
φ(th)
t
=
φ(th)
th
h
tends to 0 as t −→ 0.
80
f’(a) h
a
a+th
a+h h
f
f(a)
f(a+th)
f(a+h)
Example 1. f : R −→R, f(x) = x
3
. Then
f(a + h) = (a +h)
3
= a
3
....
f(a)
+3a
2
h
....
f
(a)h
+3ah
2
+h
3
. .. .
φ(h)
so f
(a) = 3a
2
.
2. f : R
n×n
−→R
n×n
, f(X) = X
3
.
f(A+H) = (A +H)
3
= A
3
....
f(A)
+A
2
H +AHA+HA
2
. .. .
f
(A)H
+AH
2
+HAH +H
2
A +H
3
. .. .
φ(H)
taking (say) Euclidean norm on R
n
and operator norm on R
n×n
=
L(R
n
, R
n
).
φ(H)
H
=
AH
2
+HAH+H
2
A+H
3
H
≤
3A H
2
H
= 3A H −→ 0
as H −→0
therefore, f is diﬀerentiable at A, and
f
(A) : R
n×n
−→R
n×n
is given by
f
(A)H = A
2
H +AHA+HA
2
81
Theorem 4.8.2. Let R
n
⊃ V
f
−→R
m
be a diﬀerentiable function with
f = (f
1
, . . . , f
m
) = (f
i
)
Then
f
=
∂f
i
∂x
j
Proof. For each a ∈ V we have
f
(a) : R
n
−→R
m
is a linear operator. Therefore, f
(a) is an mn matrix whose j
th
column is
f
(a)e
j
= lim
t−→0
f(a+te
j
)−f(a)
t
=
∂f
∂x
j
(a)
=
∂f
1
∂x
j
(a), . . . ,
∂f
m
∂x
j
(a)
therefore,
f
(a) =
∂f
i
∂x
j
(a)
for all a ∈ V . Therefore,
f
=
∂f
i
∂x
j
which is the Jacobian matrix.
Theorem 4.8.3. Let M ⊃ V
f
−→N
1
N
k
where
f(x) =
f
1
(x), . . . , f
k
(x)
Then f
1
, . . . , f
k
diﬀerentiable at a ∈ V =⇒ f is diﬀerentiable at a, and
f
(a)h =
f
1
(a)h, . . . , f
k
(a)h
.
Proof. (k = 2)
M ⊃ V
f
−→N
1
N
2
f(x) =
f
1
(x), f
2
(x)
Take any norms on M, N
1
, N
2
and deﬁne a norm on N
!
N
2
by
(y
1
, y
2
) = y
1
 +y
2

82
Then
f(a +h) = (f
1
(a +h), f
2
(a +h))
=
f
1
(a) +f
1
(a)h +φ
1
(h), f
2
(a) +f
2
(a)h +φ
2
(h)
= (f
1
(a), f
2
(a)) +
f
1
(a)h, f
2
(a)h
. .. .
linear in h
+
φ
1
(h), φ
2
(h)
. .. .
remainder
Now:
 (φ
1
(h), φ
2
(h)) 
h
=
φ
1
(h)
h
+
φ
2
(h)
h
tends to 0 as h −→0. Therefore, f is diﬀerentiable at am and
f
(a)h =
f
1
(a)h, f
2
(a)h
as required.
4.9 Notation
1. Given a function of n variables
R
n
⊃ V
f
−→R
V open, we shall denote by
x
1
, x
2
, . . . , x
n
the usual coordinate functions on R
n
and shall often denote the partial
derivative of f at a w.r.t. j
th
variable by:
∂f
∂x
j
(a) = lim
t−→0
f(a
1
,a
2
,...,a
j
+t,...,an)−f(a
1
,a
2
,...,a
j
,...,an)
t
=
d
dt
f(a +te
j
)
t=0
Notice that the symbol x
j
does not appear in the deﬁnition: it is a
’dummy symbol’ indicating deriv w.r.t. j
th
’slot’, sometimes written as
f,
j
(a).
2. given a function
R
2
⊃ V
f
−→R
83
V open, we often denote by x, y the usual coordinate functions instead
of x
1
, x
2
and write
∂f
∂x
for
∂f
∂x
1
∂f
∂y
for
∂f
∂x
2
etc.
3. given
M ⊃ V
f
−→N
V open, if f is diﬀerentiable at a, ∀a ∈ V , we say that f is diﬀerentiable
on V and call the function on V :
f
: a −→ f
(a)
the derivative of f. We write:
f
(2)
= (f
)
f
(3)
= ((f
)
)
and call f C
r
if f
(r)
exists and is continuous, f C
∞
if f
(r)
exists ∀r.
4.10 C
r
Functions
Theorem 4.10.1. Let
R
n
⊃ V
f
−→R
V open. Then f is C
1
⇐⇒
∂f
∂x
i
exists and is continuous for i = 1, . . . , n.
Proof. (case n = 2)
R
2
⊃ V
f
−→R, f(x, y)
1. if f is C
1
, then
∂f
∂x
,
∂f
∂y
exist and
f
=
∂f
∂x
,
∂f
∂y
therefore, f
is continuous, and therefore,
∂f
∂x
and
∂f
∂y
are continuous
84
2. Let
∂f
∂x
,
∂f
∂y
exists and be continuous on V . Then
f(a +h) = f(a) +
∂f
∂x
(a)h
1
+
∂f
∂y
(a)h
2
+φ(h)
(say) where h = (h
1
, h
2
). Now
φ(h) = f(a +h) −f(a +h
2
e
2
) −
∂f
∂x
(a)h
1
+f(a +h
2
e
2
) −f(a) −
∂f
∂y
(a)h
2
=
∂f
∂x
(m
1
)h
1
−
∂f
∂x
(a)h
1
+
∂f
∂y
(m
2
)h
2
−
∂f
∂x
(a)h
2
(say) by mean value theorem.
PSfrag replacements
m
1
m
2
a
a +h
a +h
2
e
2
Therefore,
φ(h)
h
≤
∂f
∂x
(m
1
) −
∂f
∂x
(a)
[h
1
[
h
+
∂f
∂y
(m
2
) −
∂f
∂y
(a)
[h
2
[
h
−→0
as h −→ 0 since
∂f
∂x
,
∂f
∂y
are continuous at a.
Therefore, f is diﬀerentiable at a and f
(a) =
∂f
∂x
(a),
∂f
∂y
(a)
. Also,
f
=
∂f
∂x
,
∂f
∂y
is continuous.
85
Corollary 4.10.2. f is C
r
⇐⇒
∂
r
f
∂x
i
1···∂x
ir
exists and is continuous for each
i
1
, . . . , i
r
.
Theorem 4.10.3. if
R
n
⊃ V
f
−→R V open
is C
2
then
∂
2
f
∂x
i
∂x
j
is a symmetric matrix.
Proof. Let
R
n
⊃ V
f
−→R
be C
2
. Need to show:
∂
2
f
∂x∂y
=
∂
2
f
∂y∂x
V
a+h
b
b+h
a x
+
 +

Let (a, b) ∈ V . Let h = 0, k = 0 be such that the closed rectangle
(a, b), (a +h, b), (a +h, b +k), (a, b +k)
is contained in V . Put
g(x) = f(x, b +k) −f(x, b)
86
Then
f(a +h, b +k) −f(a +h, b)
−f(a, b +k) +f(a, b) = g(a +h) −g(a)
= hg
(c) some a ≤ c ≤ a +h by MVT
= h
∂f
∂x
(c, b +k) −
∂f
∂x
(c, b)
= hk
∂
2
f
∂y∂x
(c, d) some b ≤ d ≤ b +k by MVT
and, similarly:
f(a +h, b +k) −f(a +h, b)
−f(a, b +k) +f(a, b) = kh
∂
2
f
∂x∂y
(c
, di
)
for some a ≤ c
≤ a +h and b ≤ d
≤ b +k. Therefore
∂
2
f
∂y∂x
(c, d) =
∂
2
f
∂x∂y
(c
, d
)
Now, let (h, k) −→ (0, 0). Then (c, d) −→ (a, b), and (c
, d
) −→ (a, b).
Therefore,
∂
2
f
∂y∂x
(a, b) =
∂
2
f
∂x∂y
(a, b)
by continuity of
∂
2
f
∂y∂x
and
∂
2
f
∂x∂y
4.11 Chain Rule
Theorem 4.11.1. (Chain Rule for functions on ﬁnite dimensional real or
complex vector spaces)
Let L, M, N be ﬁnite dimensional real or complex vector spaces. Let U be
open in L, V open in M and let
U
g
−→ V
f
−→N
Let g be diﬀerentiable at a ∈ U, f be diﬀerentiable at g(a). Then the compo
sition f g is diﬀerentiable at a and
(f g)
(a) = f
(g(a)) g
(a) operator product
87
Proof. we have
f[g(a
h
)] = f [g(a) +Th +φ(h)] where T = g
(a)
and
φ(h)
h
−→0
as h −→0
= f[g(a) +y] where y = Th +φ(h)
so y ≤ T h +φ(h) −→0
as h −→0
= f(g(a)) +Sy +yψ(y) where S = f
(g(a))
and ψ(h) −→0
as y −→0
= f(g(a) +S(Th +φ(h) +yψ(y)
= f(g(a)) +STh +Sφ(h) +yψ(y)
Now
Sφ(h) +yψ(y)
h
≤ S
φ(h)
h
+
T +
φ(h)
h
ψ(y)
tends to 0 as h −→0. Therefore, f g is diﬀerentiable at a, and
(f g)
(a) = ST = f
(g(a)) g
(a)
Example
d
dt
f(g
1
(t), . . . , g
n
(t)) = (f g)
(t) = f
(g(t))
. .. .
1×n
g
(t)
....
n×1
where
R ⊃ U
g
−→V
f
−→R
V open in R
n
, and f, g diﬀerentiable. Then
f
(g(t))g
(t) =
∂f
∂x
1
(g(t)), . . .
∂f
∂x
n
(g(t))
¸
¸
d
dt
g
1
(t)
.
.
.
d
dt
g
n
(t)
=
∂f
∂x
1
(g(t))
d
dt
g
1
(t) + +
∂f
∂x
n
(g(t))
d
dt
g
n
(t)
is the usual chain rule.
88
Chapter 5
Calculus of Complex Numbers
5.1 Complex Diﬀerentiation
Deﬁnition let
C ⊃ V
f
−→C
be a complex valued function of a complex variable deﬁned on open V . C is
a 1dimensional complex normed space.
Let a ∈ V . Then f is diﬀerentiable at a as a function of a complex
variable if ∃ a linear map of complex spaces:
C
f
(a)
−→C
s.t.
f(a +h) = f(a) +f
(a)h +φ(h)
where
φ(h)
h
−→0 as h −→ 0. f
(a) is a 1 1 complex matrix, i.e. a complex
number, and
f(a +h) −f(a)
h
−f
(a)
=
[φ(h)[
[h[
which tends to 0 as h −→0. Therefore
f
(a) = lim
h−→0
f(a +h) −f(a)
h
We call f
(a) the derivative of f at a. f
is called holomorphic on V if
f
(a) exists ∀ a ∈ V . We write f =
df
dy
.
Example
f : C −→C, f(z) = z
n
89
then,
lim
h−→0
f(z+h)−f(z)
h
= lim
h−→0
(z+h)
n
−z
n
h
= lim
h−→0
[nz
n−1
+
n(n−1)
2
z
n−2
h + higher powers h]
= nz
n−1
therefore, f is holomorphic in C and f
(z) = nz
n−1
.
Example
f(z) = z
then,
f(z +h) −f(z)
h
=
h
h
=
1 h real
−1 h pure imaginary
which does not converge as h −→0.
The usual rules for diﬀerentiation apply:
1.
d
dy
(f +g) =
df
dy
+
dg
dy
2.
d
dy
fg = f
dg
dy
+g
df
dy
3.
d
dy
f
g
=
g
df
dy
−f
dg
dy
g
2
if g = 0
4.
d
dz
f(g(z)) = f
(g(z)) g
(z) chain rule
By deﬁnition C = R
2
z = x +iy = (x, y)
and the operation of mult by i =
√
−1 C
i
−→C is the operator R
2
i
−→R
2
with matrix
0 −1
1 0
, because
ie
1
= i = (0, 1) = 0e
1
+e
2
ie
2
= ii = −1 = −e
1
+ 0e
2
= (−1, 0)
Let f is (real) diﬀerentiable on V then each a ∈ v the operator
C
f
(a)
−→C i.e R
2
f
(a)
−→R
2
90
preserves addition and commutes with mult by real scalars, for each a ∈ V ,
and f is holomorphic in V iﬀ f
(a) also commutes with multiplication by
complex scalars. ⇐⇒ f
(a) also commutes with multiplication by i.
⇐⇒
∂u
∂x
∂u
∂y
∂v
∂x
∂v
∂y
0 −1
1 0
=
0 −1
1 0
∂u
∂x
∂u
∂y
∂v
∂x
∂v
∂y
⇐⇒
∂u
∂y
−
∂u
∂x
∂v
∂y
−
∂v
∂x
=
−
∂v
∂x
−
∂v
∂y
∂u
∂x
∂u
∂y
⇐⇒
∂u
∂x
=
∂v
∂y
∂v
∂x
= −
∂u
∂y
CauchyRiemann Equations
so, therefore: f = u + iv is holomorphic in V ⇐⇒ f is realdiﬀerentiable
on V and satisﬁes the CauchyRiemann Equations.
Example f(z) = z
2
= (x +iy)
2
= x
2
−y
2
+ 2ixy. then
u = x
2
−y
2 ∂u
∂x
= 2x =
∂v
∂y
v = 2xy
∂v
∂x
= 2y = −
∂u
∂x
Note, if f = u +iv is holomorphic in V then
∂f
∂y
(a) = lim
h−→0
f(a+h)−f(a)
h
= lim
t−→0
f(a+t)−f(a)
t
= lim
t−→0
f(a+it)−f(a)
it
=
∂
∂x
[u +iv] =
1
i
∂
∂y
[u +iv]
therefore,
∂f
∂z
=
∂u
∂x
+i
∂v
∂x
=
∂v
∂y
−i
∂u
∂y
this also gives the CauchyRiemann Equations.
5.2 Path Integrals
Deﬁnition C
1
map
[t
1
, t
2
]
α
−→V ⊂ R
2
V open
t −→ α(t)
91
is called a (parametrical) path in R fom α(t
1
) to α(t
2
) and if f, g are complex
valued continuous on V we write
α
(f dx +g dy) =
t
2
t
1
[f(α(t))
d
dt
x(α(t)) +g(α(t))
d
dt
y(α(t))] dt
and call it the integral of the diﬀerential form f dx+g dy over the path α. If
[s
1
, s
2
]
σ
−→[t
1
, t
2
]
is C
1
with σ(s
1
) = t
1
, σ(s
2
) = t
2
and β(s) = α(σ(s)) then β is called a
reparameterisation of α.
We have
β
(f dx +g dy) =
s
2
s
1
[f(β(s))
d
ds
x(β(s)) +g(β(s))
d
ds
y(β(s))] ds
=
s
2
s
1
[f(α(σ(s)))
d
ds
x(α(σ(s))) +g(α(σ(s)))
d
ds
y(α(σ(s)))] ds
=
s
2
s
1
[f(α(σ(s)))(x α)
(σ(s)) + g(α(σ(s)))(y α)
(σ(s))]σ
(s) ds
=
t
2
t
1
[f(α(t))
d
dt
x(α(t)) +g(α(t))
d
dt
y(α(t))] dt
=
α
(f dx +g dy)
therefore, the integral over α is independent of parmaterisation.
If we take
[t
1
, t
2
]
σ
−→ [t
1
, t
2
]
with σ(s) = t
1
+ t
2
− s, and β(s) = α(σ(s)) we have σ(t
2
) = t
1
, σ(t
1
) = t
2
then β is same path but traversed in the opposite direction and
β
(f dx +g dy) =
t
1
t
2
[f(α(t))
d
dt
x(α(t)) +g(α(t))
d
dt
y(α(t))] ds
= −
α
(f dx +g dy)
Deﬁnition If f is C
1
on V we write
df =
∂f
∂x
dx +
∂f
∂y
dy
A diﬀerential form of this type is called exact, and is called the diﬀerential
of f.
92
If α is a path from a to b, α(t
1
) = a, α(t
2
) = b, then
α
df =
t
2
t
1
[f(α(t))
d
dt
x(α(t)) +g(α(t))
d
dt
y(α(t))] dt
=
t
2
t
1
[
d
dt
f(α(t))] dt
= f(α(t
2
)) −f(α(t
1
))
= f(b) −f(a) so we have path independence
= change of value of f along α
If α is a closed path, (i.e. a = b) then
α
df = 0
If f(z) = u(x, y) + iv(x, y), for z = x + iy with u, v real, put dz = dx + idy
and
f(z) dz = [u +iv][dx +idy]
= (u dx −v dy) +i(v dx +u dy)
then
α
f(z) dz =
α
(u dx −v dy) +i
α
(v dx +u dy)
=
t
2
t
1
¸
[u(α(t)) +iv(α(t))]
d
dt
[x(α(t)) +iy(α(t))]
¸
dt
=
t
2
t
1
f(α(t)) α
(t) dt
If f is holomorphic then
df = du +i dv =
∂u
∂x
dx +
∂u
∂y
dy +i
∂v
∂x
dx +i
∂v
∂y
dy
=
∂u
∂x
dx −
∂v
∂x
dx +i
∂v
∂x
dx +i
∂u
∂x
dy
=
∂u
∂x
+i
∂v
∂x
( dx +i dy)
= f
(z) dz
therefore, if f is holomorphic on V open then
1. for any path α in V from a to b:
α
f
(z) dz = f(b) −f(a)
path independence
93
2. for any closed path α in V :
α
f
(z) dz = 0
Example 1. if α path from a to b
α
z
n
dz =
α
[
d
dz
z
n+1
n + 1
] dz =
b
n+1
−a
n+1
n + 1
n = 1
in particular,
2. if α is a closed path,
α
z
n
dz = 0 n = 1
But:
3. for the closed path α(t) = e
it
, 0 ≤ t ≤ 2π
α
dz
z
=
2π
0
1
e
it
ie
it
dt =
2π
0
i dt = 2πi = 0
therefore,
1
z
cannot be the derivative of a holomorphic function on C.
Deﬁnition if [t
1
, t
2
]
α
−→C is a path in C then we write
L(α) =
t
2
t
1
[α
(t)[ dt
and call it the length of α.
If β = α σ is a reparameterisation of α with α
(s) > 0
L(β) =
s
2
s
1
[β
(s)[ ds
=
s
2
s
1
[α
(σ(s))σ
(s)[ ds
=
t
2
t
1
[α
(t)[ dt
= L(α)
94
Theorem 5.2.1. (Estimating a complex integral)
Let f be bounded on α:
[f(α(t))[ ≤ M
(say), t
1
≤ t ≤ t
2
. then
α
f(z) dz
≤ ML(α)
Proof.
α
f(z) dz
=
t
2
t
1
f(α(t))α
(t) dt
≤
t
2
t
1
[f(α(t))[ [α
(t)[ dt
≤ M
t
2
t
1
[α
(t)[ dt
= ML(α)
5.3 Cauchy’s Theorem for a triangle
If f(z) = u(x, y) +i v(x, y) is holomorphic then
α
f(z) dz =
α
(u dx −v dy) +i
(v dx +u dy)
and
∂
∂y
u =
∂
∂x
(−v),
∂
∂y
v =
∂
∂x
u (by CauchyRiemann).
Therefore, the neccessary conditions for pathindependence are satisﬁed,
but not always suﬃcient, e.g.f(z) =
1
z
. However, we have
Theorem 5.3.1. (Cauchy’s Theorem for a triangle)
Let C ⊃ V
f
−→ C be holomorphic on open V , and let T be a triangle
(interior plus boundary δT) ⊂ V . Then
δT
f(z) dz = 0
Proof. write
i(T) =
δt
f(z) dz
95
and join the midpoints of the sides to get 4 triangles
S
1
, S
2
, S
3
, S
4
then
i(T) =
4
¸
j=1
i(S
j
)
and therefore
[i(T)[ ≤
4
¸
j=1
[i(S
j
)[ ≤ 4[i(T
1
)[
(say) where T
1
is one of S
1
, . . . , S
4
.
1
T
T
Repeat the process to get a sequence of triangles
T
1
, T
2
, . . . , T
n
, . . .
with
[i(T)[ ≤ 4
n
[i(T
n
)[
Let
¸
∞
j=1
T
j
= ¦c¦. Then
i(T
n
) =
δTn
f(z) dz
=
δTn
[f(c) +f
(c)(z −c) +[z −c[φ(z −c)] dz
=
δTn
[z −c[ φ(z −c) dz where [φ(z −c)[ −→0 as [z −c[ −→0
96
n
c
z
T
Let > 0. Choose δ > 0 s.t. [φ(z −c)[ < ∀ [z −c[ < δ. Let L =length of T.
Choose n s.t. length of T
n
=
l
”
n
< δ.
Then
[i(T
n
)[ ≤
l
2
n
l
2
n
=
l
2
4
n
therefore [i(T)[ ≤ l
2
, and therefore i(T) = 0.
Deﬁnition A set V ⊂ C is called starshaped if ∃a ∈ V s.t.
[a, z] ⊂ V ∀ z ∈ V
a
z
V
Theorem 5.3.2. Let C ⊃ V
f
−→ C be a holomorphic function on an open
starshaped set. Then f = F
for some complexdiﬀerentable function F on
V and hence:
α
f(z) dz = 0
for each closed curve α in V .
Proof. Choose a ∈ V s.t. [a, z] ⊂ V ∀ z ∈ V . Put F(w) =
w
a
f(z) dz. Then
97
w+h
a
w
F(w +h) =
w+h
a
f(z) dz
=
w
a
f(z) dz +
w+h
w
f(z) dz by Cauchy
= F(w) +f(w)h +
w+h
w
[f(z) −f(w)] dy
. .. .
φ(h)
Let > 0. Choose δ > 0 w.t. [f(z) −f(w)[ < ∀[z −w[ < δ. Therefore
[φ(h)[ ≤ [h[
for all [h[ < δ, and therefore
lim
h−→0
[φ(h)[
[h[
= 0
Therefore, F is diﬀerentiable at w and F
(w) = f(w) as required.
5.4 Winding Number
We have seen that
circle about o
dz
z
= 2πi
More generally:
98
Theorem 5.4.1. Let a ∈ C ad α : [t
1
, t
2
] −→C−¦a¦ be a closed path. The
1
2πi
α
dz
z −a
is an integer, called the winding number of α about a
Proof. put
β(t) =
t
t
1
α
(s)
α(s) −a
ds
so β
(t) =
α
(t)
α(t)−a
. Then
d
dt
[α(t) −a]e
−β(t)
= [α
(t) −β
(t)¦α(t) −a¦] e
−β(t)
= 0
therefore, [α(t)−a]e
−β(t)
is a constant function of t and is equal to [α(t
1
)−a],
since β(t
1
) = 0. Therefore
e
β(t)
=
α(t) −a
α(t
1
) −a
therefore,
e
β(t
2
)
=
α(t
2
) −a
α(t
1
) −a
= 1
and therefore, β(t
2
) = 2nπi, with n an integer.
a
PSfrag replacements
e
β(t)
β(t)
2nπi
99
Therefore,
α
dz
z −a
=
t
2
t
1
α
(s)
α(s) −a
ds = β(t
2
) = 2nπi
as required.
Theorem 5.4.2. Let C be a circle and a ∈ C. Then
1. if a is inside C, then C has winding number 1 about a:
C
dz
z −a
= 2πi
a
C
2. if a is outside C then winding number about a is 0.
C
dz
z −a
= 0
a
C
Proof. 1. Let a be inside C, let C
1
be a circle, centre a inside C. Let
α, β, γ be the closed paths shown, each is contained in an open star
shaped set on which
1
z−a
is holomorphic.
100
1
a
α
β
γ
C
C
α
dz
z−a
+
β
dz
z−a
+
γ
dz
z−a
=
C
dz
z−a
−
C
1
dz
z−a
0 + 0 = 0 =
C
−
C
1
therefore,
C
dz
z −a
=
C
1
dz
z −a
=
2π
0
ire
iθ
re
iθ
dθ = 2πi
(put z = a +re
iθ
).
2. Let a be outside C, then C is contained in an open starshaped set on
which
1
z−a
is holomorphic. Therefore,
C
dz
z −a
= 0
5.5 Cauchy’s Integral Formula
Theorem 5.5.1. (Cauchy’s Integral Formula)
Let f be holomorphic on open V in C. Let w ∈ V . Then, for any circle
C around w, such that C and it’s interior is contained in V , we have:
f(w) =
1
2πi
C
f(z)
z −w
dz
(Thus the values of f on any circle uniquely determine the values of f inside
the circle)
101
w
C
V
Proof. Let > 0. Choose a circle C
1
centre w, radius r (say) inside C such
that [f(z) −f(w)[ ≤ ∀z ∈ C
1
by continuity of f.
w
α
β
γ
C
C
1
Let α, β, γ be as indicated. Then, by integrating
f(z)
z−w
:
C
−
C
1
=
α
+
β
+
γ
= 0 + 0 + 0 = 0
since each of α, β, γ are contained in a starshaped set on which
f(z)
z−w
is a
holomorphic function of z. Therefore,
C
f(z)
z−w
dz =
C
1
f(z)
z−w
dz
=
C
1
f(w)
z−w
dz +
C
1
f(z)−f(w)
z−w
dz
= 2πi f(w) + 0
because:
C
1
f(z) −f(w)
z −w
dz
≤
r
2πr = 2π
for all > 0. Hence result.
102
Let C ⊃ V
f
−→ C be holomorphic on V open. Let w ∈ V . Pick a circle
C around w such that C and it’s interior ⊃ V . We have:
f(w) =
1
2πi
C
f(z)
z −w
dz
diﬀerentiating w.r.t. w under the integral sign:
f
(w) =
1
2πi
C
f(z)
(z −w)
2
dz
w
r z
C
PSfrag replacements
w
1
Justiﬁed since ∃ r > 0 s.t.
f(z)
(z −w
1
)
2
≤
[f(z)[
r
2
for all w
1
on an open set containing w
1
, which is integrable w.r.t. z.
Repeating n times gives:
f
(n)
(w) =
n!
2πi
C
f(z)
(z −w)
n+1
dz (5.1)
Thus, f holomorphic on open V =⇒ f has derivatives of all orders and
f
(n)
(w) is given by Equation(5.1) for any suitable circle C (or any suitable
closed curve) around w.
103
5.6 Termbyterm diﬀerentiation, analytic func
tions, Taylor series
Theorem 5.6.1. Let ¦f
n
(z)¦ be a sequence of functions which is uniformly
convergent to the function f(z) on a path α. Then
lim
n−→∞
α
f
n
(z) dz =
α
f(z) dz
Proof. Let > 0. Choose N s.t.
[f
n
(z) −f(z)[ ≤ ∀n ≥ N ∀ z = α(t) t
1
≤ t ≤ t
2
(deﬁnition of uniform convergence). Then
α
f
n
(z) dz −
α
f(z) dz
≤
α
[f
n
(z) −f(z) dz[
≤ l(α) ∀ n ≥ N
hence the result.
Theorem 5.6.2. (term by term diﬀerentiation)
Let
¸
∞
i=1
f
n
(z) be a series of holomorphic functions on an open set V ,
which converges uniformly on a circle C which, together with it’s interior, is
contained in V .
Then the series converges inside C to a holomorphic function F (say)
and
f
(k)
=
∞
¸
n=1
f
(k)
n
inside C
Proof. Let w be inside C, then λ = inf
z∈C
[z −w[ > 0.
=⇒
1
[z −w[
k+1
≤
1
λ
k+1
∀ z ∈ C
Therefore,
¸
∞
n=1
fn(z)
(z−w)
k+1
converges uniformly to
f(z)
(z−w)
k+1
z ∈ C. Therefore,
∞
¸
n=1
C
f
n
(z)
(z −w)
k+1
dz =
C
f(z)
(z −w)
k+1
dz
and therefore,
∞
¸
n=1
f
(k)
n
(w) = f
(n)
(w)
as required.
104
Deﬁnition Let C ⊃ V
f
−→ C with V opn. Then f is called analytic on V
if, for each a ∈ V ∃ r > 0 and constants c
0
, c
1
, c
2
, . . . ∈ C, such that:
f(z) = c
0
+c
1
(z −a) = c
2
(z −a)
2
+ ∀ z s.t. [z −a[ < r
The R.H.S. is called the Taylor series of f about a.
Lemma 5.6.3. If 0 < ρ < r then the series converges uniformly on
¦z : [z −a[ ≤ ρ¦
r
R
a
PSfrag replacements
ρ
Proof. Let 0 < ρ < R < r.
¸
c
n
(z − a)
n
converges for z − a = R, and
therefore,
¸
c
n
R
n
converges. Therefore, ∃K s.t. [c
n
R
n
[ ≤ K ∀n. Therefore
[c
n
(z −a)
n
[ ≤ [c
n
[ [z −a[
n
≤
K
R
n
ρ
n
= K
ρ
R
n
for all [z −a[ ≤ ρ.
Therefore, we can diﬀerentiate term by term n times:
f
(n)
(z) = n!c
n
+ (n + 1)!c
n+1
[z −a[ + higher powers of z −a
Therefore, f
(n)
(a) = n!c
n
, and therefore, c
n
=
1
n!
f
(n)
(a).
The Taylor coeﬃcents are uniquely determined, f is C
∞
, and
f(z) =
∞
¸
n=1
1
n!
f
(n)
(z −a)
n
on [z −a[ < r
Thus, f analytic on V =⇒ f holomorphic on V .
Conversely,
105
Theorem 5.6.4. Let C ⊃ V
f
−→ C be holomorphic on open V . Let C be
any circle, centre a s.t. C and it’s interior are contained in V . Then f has
a Taylor series about a convergent inside C. Hence f is analytic on V .
Proof. Let w be inside C. Then
f(w) =
1
2πi
C
f(z)
z−w
dz
=
1
2πi
C
f(z)
(z−a)−(w−a)
dz
=
1
2πi
C
f(z)
(z−a)[1−
w−a
z−a
]
dz
=
1
2πi
C
f(z)
z−a
¸
∞
n=1
w−a
z−a
n
dz
=
¸
∞
n=1
(w −a)
n
1
2πi
C
f(z)
(z −a)
n=1
dz
. .. .
cn
=
¸
∞
n=0
1
n!
f
(n)
(c)(w −a)
n
as required.
106
Chapter 6
Further Calculus
6.1 Mean Value Theorem for Vectorvalued
functions
Theorem 6.1.1. (Mean value theorem for vector valued functions)
Let M, N be ﬁnite dimensional normed spaces and let
M ⊃ V
f
−→ N
be a C
1
function, where V is open in M. Let x, y ∈ V and [x, y] ⊂ V . Then
f(x −f(y) ≤ k x −y
where k = sup
z∈[x,y]
f
(z)
Proof.
f(y) −f(x) =
1
0
d
dt
f[ty + (1 −t)x] dt
=
1
0
f
[ty + (1 −t)x]
. .. .
operator
(y −x)
. .. .
vector
dt
Therefore,
f(y) −f(x) ≤
1
0
f
[ty + (1 −t)x] y −x dt
≤
1
0
k y −x dt
= k y −x
as required.
107
6.2 Contracting Map
Theorem 6.2.1. Let M be a ﬁnite dimensional real vector space with norm.
Let
M ⊃ V
f
−→M
be a C
1
function, V open in M, f(0) = 0, f
(0) = 1.
Let 0 < < 1, and let B be a closed ball centre O s.t.
1 −f
(x) ≤ ∀x ∈ B
Then
1.
f(x) −f(y) ≥ (1 −) x −y ∀ x, y ∈ B (6.1)
Thus f
B
is injective.
2. (1 −)B ⊂ f(B) ⊂ (1 +)B
r
B
f
PSfrag replacements
(1 −)r
(1 +)r
f(B)
Proof. Let r =radius of B
1.
f
(x) = 1 + (f
(x) −1) ≤ 1 + ∀x ∈ B
therefore,
f(x) = f(x) −f(0)
MV T
≤ (1 +)x ≤ (1 +)r
for all x ∈ B. Therefore, f(B) ⊂ (1 +)B
108
2.
(1 −f)(x) −(1 −f)(y)
MV T
≤ x −y ∀ x, y ∈ B
therefore, x −y[ −f(x) −f(y) ≤ x −y, and so
f(x) −f(y) ≥ (1 −)x −y
and hence Eqn(6.1).
3. To show (1−)B ⊂ f(B). Let a ∈ (1−)B, deﬁne g(x) = x−f(x) +a.
Then
g
(x) = 1 −f
(x) ≤ ∀x ∈ B
therefore,
g(x) −g(y)
MV T
≤ x −y
therefore, g is a contracting map (shortens distances).
Also,
g(x) = g(x) −g(0) +a since a = g(0)
≤ g(x) −g(0) +a
≤ x +a
≤ r + (1 −)r ∀ x ∈ B
= r
therefore, g(x) ∈ B ∀ x ∈ B. So g maps B into B and is contracting.
Therefore, by the contraction mapping theorem ∃x ∈ B s.t. g(x) = x.
i.e. x −f(x) +a = x
i.e. f(x) = a. Therefore, a ∈ f(B), and (1 −)B ⊂ f(B) as required.
6.3 Inverse Function Theorem
Deﬁnition Let M ⊃ V
f
−→ W ⊂ N where M, N are ﬁnite dimensional
normed spaces. Then f is called a C
r
diﬀeomorphism if
1. V open in M, W open in N
2. V
f
−→W is bijective
109
3. f and f
−1
are C
r
V is C
r
diﬀeomorphic to W if ∃ a C
r
diﬀeomorphism V −→W.
Example
R
f
−→ (0, ∞)
f(x) = e
x
is C
∞
, and f
−1
(x) = lnx is C
∞
. Therefore, f is a C
∞
diﬀeomor
phism.
PSfrag replacements
x = ln y
y = e
x
Theorem 6.3.1. (Inverse Function Theorem)
Let M ⊃ V
f
−→ N be C
r
where M, N are ﬁnite dimensional normed
spaces and V is open in M. Let a ∈ V be a point at which
f
(a) : M −→N
is invertible. Then ∃ open neighbourhood W of a such that
f
W
: W −→f(W)
is a C
r
diﬀeomorphism.
110
M
N
f(W)
a
W
Proof. Let T be the inverse of f
(a) and let F be deﬁned by
F(x) = Tf(x +a) −Tf(a)
We have:
F(0) = Tf(a) −Tf(a) = 0
0
U
+a
a
f
F
f(a)
0
T f(a)
 T f(a)
T
F(U)
U+a
f(U+a)
We prove that F maps an open neighbourhood U of 0 onto an open
neighbourhood F(U) of 0. It the follows that f maps open U +a onto open
111
f(U +a) by a C
r
diﬀeomorphism. Now
F
(x) = T f
(x +a)
=⇒ F
(0) = T f
(a) = 1
M
Choose a closed ball B centre 0 of positive radius s.s.
F
(x) −1
M
 ≤
1
2
∀ x ∈ B
and also s.t. det F
(x) = 0. Then by the previous theorem (with =
1
2
) we
have:
F
B
is injective
and
F(x) −F(y) ≥
1
2
x −y ∀ x, y, ∈ B
and
1
2
B ⊂ F(B)
B
U
F(B)
F
PSfrag replacements
1
2
B
F
−1
therefore, F
−1
:
1
2
B −→ B is welldeﬁned and continuous. Let B
0
be the
interior of B, an open set. Put U = F
−1
(
1
2
B
0
) ∩ B
0
, an open set. F(U) is
open since F
−1
is continuous. So F
U
: U −→ F(U) is a homeomorphism of
open U onto open F(U).
Let G be it’s inverse. To show G is C
r
.
112
F
G
U
y
y+l
x
x+h
F(U)
Let x, x +h ∈ F(U), G(x) = y, G(x +h) = y +l (say), and h ≥
1
2
l.
Let F
(y) = S. Then F(y + h) = F(y) +Sl +φ(l), where
φ(l)
l
−→ 0 as
l −→ 0.
=⇒ x +h = x +Sl +φ(l),
=⇒ l = S
−1
h −S
−1
φ(l)
=⇒ G(x +h) = y +l = G(x) +S
−1
h −S
−1
φ(l)
Now,
S
−1
φ(l)
h
≤ S
−1

φ(l)
l
l
h
−→ 0
as h −→0 since l ≤ 2h =⇒
l
h
≤ 2.
So, G is diﬀerentiable at x ∀x ∈ F(U) and
G
(x) = S
−1
= [F
(y)]
−1
= [F
(G(x))]
−1
It follows that if G is C
s
for some 0 ≤ s < r then G
is C
s
since G
is a
composition of C
s
functions
F
, G, []
−1
and therefore G is C
s+1
. So
G C
s
=⇒ G C
s+1
∀0 ≤ s < r
therefore G is C
r
as required.
113
Example Let f, g be C
r
on an open set containing (a, b)
W
x a
y
b
W’
g(a,b)
t
f(a,b)
z
Let
∂(f, g)
∂(x, y)
=
∂f
∂x
∂f
∂y
∂g
∂x
∂g
∂y
= 0
at (a, b). Then (f, g) maps an open neighbourhood W of (a, b) onto an open
neighbourhood W
of (f(a, b), g(a, b) by a C
r
diﬀeomorphism. Therefore, for
each z, t ∈ W
∃ unique (x, y) ∈ W such that
z = f(x, y)
t = g(x, y)
and x = h(z, t), y = k(z, t) where h and k are C
r
.
114
Chapter 7
Coordinate systems and
Manifolds
7.1 Coordinate systems
Deﬁnition Let X be a topological space. Let V be open in X. A sequence
y = (y
1
, . . . , y
n
)
of realvalued functions on V is called an ndimensional coordinate system
on X with domain V if
X ⊃ V
y
−→y(V ) ⊂ R
n
x :−→y(x) = (y
1
(x), . . . , y
n
(x))
is a homeomorphism of V onto an open set y(V ) in R
n
.
x V
X
y(V)
y(x)
PSfrag replacements
R
n
y
n
(x)
y
1
(x)
115
Example On the set V = ¦y = 0 or x > 0¦ in R
2
the functions r, θ given
by
x = r cos θ
y = r sin θ
−π < θ < π
map V homomorphically onto an open set in R
2
.
r
(x,y)
v
PSfrag replacements
−π
θ
π
(r, θ)
θ
θ
Therefore, (r, θ) is a 2dimensional coordinate system on R
2
with domain
V .
Deﬁnition If y = (y
1
, . . . , y
n
) is a coordinate system with domain V and if
f is a realvalued function on V then
f = F(y
1
, . . . , y
n
)
for a unique function F on y(V ) s.t. F = f y
−1
We call f a C
r
function of y
1
, . . . , y
n
if F is C
r
and we write
∂f
∂y
i
=
∂F
∂x
i
(y
1
, . . . , y
n
)
and call it the partial derivative w.r.t Y
i
in the coord system y
1
, . . . , y
n
.
116
Note:
∂f
∂y
i
(a) =
∂F
∂x
i
(y
1
(a), . . . , y
n
(a))
=
d
dt
F(y(a) +te
i
)
t=0
=
d
dt
f(α
i
(t))
t=0
= rate of change of f along curve α
i
in V
a
PSfrag replacements
α
i
(t)
y(a)
y(a) +te
i
where α
i
is the curve given by:
y(α
i
(t)) = y(a) +te
i
= (y
1
(a), . . . , y
i
(a) +t, . . . , y
n
(a))
therefore, along curve α
i
all coords y
1
, . . . , y
n
are constant except i
th
coord
y
i
and α
i
(t) is parameterised by change t in i
th
coord y
i
.
α
i
is called the i
th
coordinate curve at a.
7.2 C
r
manifold
Deﬁnition Let y
1
, . . . , y
n
with domain V , and z
1
, . . . , z
n
with domain W
be two coordinate systems on X. Then these two systems are called C
r

compatible if each z
i
is a C
r
function of y
1
, . . . , y
n
, and each y
i
is a C
r
function of z
1
, . . . , z
n
on V ∩ W.
We call X an ndimensional C
r
manifold if a collection of ndimensional
coordinate systems is given whose domains cover X and which are C
r

compatible.
117
Example the 2sphere S
2
given by:
x
1
+y
1
+z
2
= 1 in R
3
The function x, y with domain ¦z > 0¦ is a 2dimensional coordinate system
on S
2
.
y
z
x
Similarly the functions x, z with domain ¦y > 0¦ is a 2dimensional co
ordinate system
On the overlap ¦y > 0, z > 0¦ we have:
x = x x = x
y =
1 −x
2
−y
2
z =
1 −x
2
−y
2
these systems are C
∞
compatible. In this way, we make S
2
into a C
∞

manifold.
7.3 Tangent vectors and diﬀerentials
Deﬁnition Let a ∈ X, X a manifold. Let y
1
, . . . , y
n
be coords on X at a
(i.e. domain an open neighbourhood of a). Then, the linear operators
∂
∂y
1
a
, ,
∂
∂y
n
a
(7.1)
118
act on the diﬀerentiable functions f at a (i.e. functions f which are real
valued diﬀerentiable functions of y
1
, . . . , y
n
on an open neighbourhood of a)
and are deﬁned by:
∂
∂y
j
a
f =
∂f
∂y
j
(a)
The operators (7.1) are linearly independent since
¸
α
j
∂
∂y
j
a
y
i
= α
j
∂y
i
∂y
j
(a) = α
j
δ
i
j
= α
i
(7.2)
=⇒ α
j
∂
∂y
j
a
= 0 =⇒ α
i
= 0 ∀i (7.3)
The real vector space with basis (7.1) is denoted T
a
X and is called the tangent
space of X at a. If v ∈ T
a
X then (7.2) shows that v has components vy
i
w.r.t. basis (7.1)
Deﬁnition If α(t) is a curve in X then the velocity vector ˙ α(t) ∈ T
α(t)
X is
the tangent vector at α(t) given by taking rate of change along α(t) w.r.t. t:
PSfrag replacements
α(t)
˙ α(t)
i.e.
˙ α(t)f =
d
dt
f(α(t))
=
d
dt
F(y
1
(α(t)), . . . , y
n
(α(t)) if f = F(y
1
, . . . , y
n
)
=
∂F
∂x
j
(y
1
(α(t)), . . . , y
n
(α(t)))
d
dt
y
j
(α(t))
=
∂f
∂y
j
(α(t))
d
dt
y
j
(α(t))
=
¸
d
dt
y
j
(α(t))
∂
∂y
j
α(t)
f
119
therefore,
˙
α(t) is the tangent vector with components
d
dt
y
j
(α(t)) = ˙ y
j
(t)
i.e.
˙ α(t) = ˙ y
j
(t)
∂
∂y
j
α(t)
The tangent space T
a
X does not depend on the choice of (compatible) coor
dinates at a since if z
1
, . . . , z
n
is another coordinate system at a, then
∂
∂z
j
a
= velocity vector of j
th
coordinate curve of z
1
, . . . , z
n
at a
=
∂y
i
∂z
j
∂
∂y
i
a
∈ T
a
X
therefore,
∂
∂z
1
a
, ,
∂
∂z
n
a
is also a basis for T
a
X, and
∂y
i
∂z
j
(a) is the transition matrix from z
1
, . . . , z
n
to
y
1
, . . . , y
n
.
Note also that the curve α(t) with coordinates:
y(α(t)) = (y
1
(a) +α
1
t, . . . , y
n
(a) +α
n
t)
has velocity vector
˙ α(0) = α
1
∂
∂y
1
a
+ α
n
∂
∂y
n
a
Therefore, every tangent vector (at a) is the velocity vector of some curve,
and vice versa.
Deﬁnition Let a ∈ X and f be a diﬀerentiable function at a. Then for each
v ∈ T
a
X with v = ˙ α(t) (say), deﬁne
'df
a
, v` = vf = ˙ α(t)f =
d
dt
f(α(t)) = rate of change of f along v
Thus df
a
is a linear form on T
a
X, called the diﬀerential of f at a df
a
measures
the rate of change of f at a. If y
1
, . . . , y
n
are coordinates on X at a then
dy
i
a
,
∂
∂y
j
a
=
∂
∂y
j
a
y
i
=
∂y
i
∂y
j
(a) = δ
i
j
120
therefore, dy
1
a
, . . . , dy
n
a
is the basis of T
a
X
∗
which is dual to the basis
∂
∂y
1
a
, ,
∂
∂y
n
a
of T
a
X.
Thus the linear form dy
i
a
gives the i
th
component of tangent vectors at a:
'dy
i
a
, v` = i
th
component of v = vy
i
Also, df
a
has components:
df
a
,
∂
∂y
j
a
=
∂
∂
j
a
f =
∂f
∂y
j
(a)
therefore,
df
a
=
∂f
∂y
j
(a) dy
j
a
(which is called the chain rule for diﬀerentials.)
7.4 Tensor Fields
From now on assume manifolds, functions are C
∞
.
Deﬁnition Let W be an open set in a manifold X. A tensor ﬁeld S on X
with domain W is a function on W:
x −→ S
x
which assigns to each x ∈ W a tensor of ﬁxed type over the tangent space
T
a
X at x. e.g.
S
x
: T
x
X (T
x
X)
∗
T
x
X −→R
We can add tensor ﬁelds, contract them, form tensor products and wedge
products by carrying out these operations at each point x ∈ W. e.g.
(R +S)
x
= R
x
+S
x
(R ⊗S)
x
= R
x
⊗S
x
Deﬁnition A tensor ﬁeld with no indices is called a scalar ﬁeld (i.e. a real
valued function); a tensor ﬁeld with one upper index is called a vector ﬁeld;
a tensor ﬁeld with r skewsymmetric lower indices is called a diﬀerential r
form; a tensor ﬁeld with two lower indices is called a metric tensor if it is
symmetric and nonsingular at each point.
121
If y
1
, . . . , y
n
is a coordinate system with domain W and S is a tensor ﬁeld
on W with (say) indices of type downupdown, then
S = α
i
j
k
dy
i
⊗
∂
∂y
k
⊗dy
k
where the scalar ﬁelds α
i
j
k
are the components of S w.r.t. the coordinates
y
i
.
If f is a scalar ﬁeld, then df is a diﬀerential 1form. If v is a vector ﬁeld,
then vf is the scalar ﬁeld deﬁned by:
(vf)
x
= v
x
f = 'df
x
, v
x
` = 'df, v`
x
i.e. vf = 'df, v` = rate of change of f along v.
If ([) is a metric tensor on X then for any two vector ﬁelds u, v with
common domain W we deﬁne the scalar ﬁeld (u[v) by
(u[v)
x
= (u
x
[v
x
)
x
if v is a vector ﬁeld then we can lower it’s index to get a diﬀerential 1form
ω such that
'ω, u` = (v[u)
Conversely, raising the index of a 1form gives a vector ﬁeld. Raising the
index of df gives a vector ﬁeld gradf called the gradient of f:
(gradf[u) = 'df, u` = uf = rate of change of f along u
If ([) is positive deﬁnite then for u = 1 we have:
[(gradf[u)[ ≤ gradf
Thus the maximum rate of change of f is gradf and is attained in the
direction of gradf.
A metric tensor ([) deﬁnes a ﬁeld ds
2
of quadratic forms called the
associated line element by:
ds
2
(v) = (v[v) = v
2
if metric is positive deﬁnite
If y
i
are coordinates with domain W then, on W:
each vector ﬁeld u = α
i ∂
∂y
i
with components α
i
each diﬀerential rform ω = ω
i
1
...ir
dy
i
1
∧ ∧ dy
ir
each diﬀerential 1form ω = ω
i
dy
i
'ω, u` = α
i
ω
i
122
if f is a scalar ﬁeld df =
∂f
∂y
i
dy
i
'df, u` = uf = α
i
∂f
∂y
i
if
∂
∂y
i
∂
∂y
j
= g
ij
then
([) = g
ij
dy
i
⊗dy
j
and
ds
2
= g
ij
dy
i
dy
j
gradf has components
g
ij
∂f
∂y
j
therefore,
gradf = g
ij
∂f
∂y
j
∂
∂y
i
so
gradf = (gradf[gradf) = 'df, gradf` = g
ij
∂f
∂y
i
∂f
∂y
j
Example On R
3
with the usual coordinate functions x, y, z the usual metric
tensor is
([) = dx ⊗dx +dy ⊗dy +dz ⊗dz
with components
g
ij
=
¸
¸
¸
¸
¸
1 0 0
0 1 0
0 0 1
= g
ij
∂
∂x
,
∂
∂y
,
∂
∂z
are orthonormal vector ﬁelds
df =
∂f
∂x
dx +
∂f
∂y
dy +
∂f
∂y
dy
and
gradf =
∂f
∂x
∂
∂x
+
∂f
∂y
∂
∂y
+
∂f
∂z
∂
∂z
The line element is
ds
2
= (dx)
2
+ (dy)
2
+ (dz)
2
123
If r, θ, φ are spherical polar cordinates on R
3
then
x = r sin θ cos φ
y = r sin θ sin φ
z = r cos θ
therefore,
ds
2
= (dx)
2
+ (dy)
2
+ (dz)
2
= (sin θ cos φdr +r cos θ cos φdθ −r sin θ sin φdφ)
2
+(sin θ sin φdr +r cos θ sin φdθ +r sin θ cos φdφ)
2
+(cos θdr −r sin θdθ)
2
= (dr)
2
+r
2
(dθ)
2
+r
2
sin
2
θ(dφ)
2
therefore,
g
ij
=
¸
1 0 0
0 r
2
0
0 0 r
2
sin
2
θ
and
g
ij
=
¸
1 0 0
0
1
r
2
0
0 0
1
r
2
sin
2
θ
so
df =
∂f
∂r
dr +
∂f
∂θ
dθ +
∂f
∂φ
dφ
and df has components:
∂f
∂r
,
∂f
∂θ
,
∂f
∂φ
therefore, gradf has components
∂f
∂r
,
1
r
2
∂f
∂θ
,
1
r
2
sin
2
θ
∂f
∂φ
therefore,
gradf =
∂f
∂r
∂
∂r
+
1
r
2
∂f
∂θ
∂
∂θ
+
1
r
2
sin
2
θ
∂f
∂φ
∂
∂φ
124
7.5 Pullback, Pushforward
Deﬁnition Let X and Y be manifolds and
X
φ
−→Y
a continuous map. Then for each scalar ﬁeld f on Y we have a scalr ﬁeld
φ
∗
f = f φ
on X (we assume that φ
∗
f is C
∞
for each C
∞
f).
φ
∗
f is called the pullback of f to X under φ.
X
Y
a
f
PSfrag replacements
φ
φ
∗
f
φ(a)
(φ
∗
f)(x) = f(φ(x))
For each a ∈ X we have a linear operator
T
a
X
φ∗
−→T
a
Y
If v = ˙ α(t) ∈ T
a
X then we deﬁne φ
∗
v by:
[φ
∗
v]f =
d
dt
f(φ(α(t))) = ˙ α(t)[f φ] = v[f φ] = v[φ
∗
f]
for each scalar ﬁeld f on Y at φ(a).
PSfrag replacements
a
α(t)
˙ α(t)
φ
φ(α(t))
φ
∗
˙ α(t)
φ(a)
125
Thus the velocity vector of α(t) pushes forward to the velocity vector of
φ(α(t))
Theorem 7.5.1. Let X
φ
−→ Y and let y
1
, . . . , y
n
be coordinates on X at a
and let
φ
1
(x), . . . , φ
n
(x)
be the coordinates of φ(x) w.r.t. a coordinate system z
1
, . . . , z
n
on Y at φ(a).
Then the pushforward
T
a
X
φ∗
−→T
φ(a)
Y
has matrix
∂φ
i
∂y
j
(x)
Proof.
i
th
component of φ
∗
∂
∂y
j
a
=
φ
∗
∂
∂y
j
a
z
i
=
∂
∂y
j
a
[φ
∗
z
i
]
=
∂φ
i
∂y
j
(a)
since φ
i
(x) = z
i
(φ(x)), so φ
i
= φ
∗
z
i
.
Thus the matrix of φ
∗
is the same as the matrix of the derivative in the
case R
n
−→ R
n
. The pushforward φ
∗
is often called the derivative of φ at
a.
The chain rule for vector spaces
(ψ φ)
(x) = ψ
(φ(x)) φ
(x)
corresponds, for manifolds, to:
Theorem 7.5.2. (Chain rule for maps of manifolds; functorial property of
the pushforward)
Let X
φ
−→Y
ψ
−→Z be maps of manifolds. Then
(ψ φ)
∗
= ψ
∗
φ
∗
Proof. Let ˙ α(t) be a tangent vector on X, then:
(ψ φ)
∗
˙ α(t) = velocity vector of ψ(φ(α(t)))
= ψ
∗
[velocity vector of φ(α(t))]
= ψ
∗
[φ
∗
˙ α(t)]
as required.
126
Deﬁnition Let X
φ
−→ Y be a map of manifolds and ω a tensor ﬁeld on Y
with r lower indices. Then we deﬁne the pullback of ω to X under φ to be
the tensor ﬁeld φ
∗
ω on X having r lower indices and deﬁned by:
(φ
∗
ω)
x
[v
1
, . . . , v
r
] = ω
φ(x)
[φ
∗
v
1
, . . . , φ
∗
v
r
]
all v
1
, . . . , v
r
∈ T
x
X, ∀x ∈ X. i.e.
(φ
∗
ω)
x
= φ
∗
[ω
φ(x)
]
We note the φ
∗
on tensor ﬁelds preserves all the algebraic operations such
as additions, tensor products, wedge products.
We have:
Theorem 7.5.3. if X
φ
−→Y and f is a scalar ﬁeld on Y , then
φ
∗
df = dφ
∗
f
i.e., the pullback commutes with diﬀerentials
Proof. Let v ∈ T
x
X. Then
'(φ
∗
df)
x
, v` = '(df)
φ(x)
, φ
∗
v`
= [φ
∗
v]f
= v[φ
∗
f]
= '(dφ
∗
f)
x
, v`
therefore, φ
∗
df)
x
= (dφ
∗
f)
x
∀x ∈ X, and therefore φ
∗
df = dφ
∗
f.
7.6 Implicit funciton theorem
Theorem 7.6.1 (Implicit Function Theorem). (on the solution spaces
of l equations in n variables)
Let f = (f
1
, . . . , f
l
) be a sequence of C
r
realvalued functions on an open
set V in R
n
. Let c = (c
1
, . . . , c
l
) ∈ R
l
and let
X = ¦x ∈ V : f(x) = c, rankf
(x) = l¦
Then for each a ∈ X we can select n − l of the usual coordinate functions
x
1
, . . . , x
n
:
x
l+1
, . . . , x
n
127
(say) so that on an open neighbourhood of a in X they form a coordinate
system on X. Any two such coordinate systems are C
r
compatible. Thus X
is an (n −l)dimensional C
r
manifold.
Proof.
f
=
¸
¸
∂f
1
∂x
1
∂f
1
∂x
l
∂f
1
∂x
n
.
.
.
.
.
.
.
.
.
∂f
l
∂x
1
∂f
l
∂x
l
∂f
l
∂x
n
l×n
Let a ∈ X. THen f
(a) has rank l, so the matrix f
(a) has l linearly inde
pendent columns: the ﬁrst l columns (say). Put
F = (f
1
, . . . , f
l
, x
l
+ 1, . . . , x
n
)
then
F
=
¸
¸
¸
¸
¸
¸
¸
¸
¸
¸
¸
∂f
1
∂x
1
∂f
1
∂x
l
0
.
.
.
.
.
.
∂f
l
∂x
1
∂f
l
∂x
l
0
0 0 1 0
.
.
.
.
.
.
.
.
.
0 0 0 1
l×n
therefore,
det F
(a) =
∂f
1
∂x
1
(a)
∂f
1
∂x
l
(a)
.
.
.
.
.
.
∂f
l
∂x
1
(a)
∂f
l
∂x
l
(a)
= 0
PSfrag replacements
R
l
R
n−l
R
l
R
l
R
n−l
R
n−l
a
W
F(W)
G
F
X
U
c
128
By the inverse function theorem, F maps an open neighbourhood W of a onto
an open neighbourhood F(W) by a C
r
diﬀeomorphism with inverse G (say).
Note that F, and hence also G leave the coordinates x
l+1
, . . . , x
n
unchanged.
F maps W ∩X homeomorphically onto ¦c¦U where U is open in R
n−l
.
Thus (x
l+1
, . . . , x
n
) maps W ∩X homeomorphically onto U and is there
fore an (n − l)dimensional coordinate system on X with domain W ∩ X.
Also, if
G = (G
1
, . . . , G
l
, x
l+1
, . . . , x
n
)
on F(W) then
x
1
= G
1
(c
1
, . . . , c
l
, x
l+1
, . . . , x
n
)
.
.
.
.
.
.
x
l
= G
l
(c
1
, . . . , c
l
, x
l+1
, . . . , x
n
)
on W ∩ X. Therefore x
1
, . . . , x
l
and C
r
functions of x
l+1
, . . . , x
n
on W ∩ X.
Hence any two such coordinate systems on X are C
r
compatible.
Note: the proof shows that if (say) the ﬁrst l columns of the matrix
∂f
i
∂x
j
(a) are linearly independent, then the l equations in n unknowns
f
1
(x
1
, . . . , x
n
) = c
1
.
.
.
.
.
.
f
l
(x
1
, . . . , x
n
) = c
l
determine x
1
, . . . , x
n
as C
r
functions of x
l+1
, . . . , x
n
on an open neighbour
hood of a in the solution space.
Example 1.
R
n
⊃ V
f
−→R
a C
r
function.
f
=
∂f
∂x
i
1×n
=
∂f
∂x
1
, . . . ,
∂f
∂x
n
= ∇f
are the components of the gradient vector ﬁeld.
If ∇f is nonzero at each point of the space X of solutions of the
equations:
f(x
1
, . . . , x
n
) = c
(one equation in n unknowns) then X is an (n − 1)dimensional C
r

manifold.
129
If (say)
∂f
∂x
1
(a) = 0 at a ∈ X then
x
2
, x
3
, . . . , x
n
are coordinates on an open neighbourhood W of a in X and x
1
is a C
r
function of x
2
, . . . , x
n
on W.
2. R
n
⊃ V
f
1
,f
2
−→ R two C
r
functions.
∂f
i
∂x
j
=
∂f
1
∂x
1
∂f
1
∂x
n
∂f
2
∂x
1
∂f
2
∂x
n
If the two rows ∇f
1
, ∇f
2
are linearly independent at each point of the
space X of solutions of the two equations:
f
1
(x
1
, . . . , x
n
) = c
1
f
2
(x
1
, . . . , x
n
) = c
2
(two equations in n unknowns) then X is an (n − 2)dimensional C
r
manifold. If (say)
∂(f
1
, f
2
)
∂(x
1
, x
2
)
= 0 at a ∈ X
then x
3
, x
4
, . . . , x
n
are coordinates on an open neighbourhood W of a
in X and x
1
and x
2
are C
r
functions of x
3
, . . . , x
n
on W.
7.7 Constraints
If f
1
, . . . , f
l
are C
r
functions on an open set V in R
n
and c = (c
1
, . . . , c
n
) we
consider the equations
f
1
= c
1
, . . . , f
l
= c
l
as a system of constraints which are satisﬁed by points on the constraint
manifold:
X =
x ∈ V : f
1
(x) = c
1
, . . . , f
l
(x) = c
l
, rank
∂f
i
∂x
j
(x) = l
Let X
i
−→ R
n
be the inclusion map i(x) = x∀x ∈ X. Then for each scalar
ﬁeld f on V :
(i
∗
f)(x) = f(i(x)) = f(x) ∀x ∈ X
130
the pullback i
∗
f is the restriction of f to X. We call i
∗
f the constrained
function F. If ω is a diﬀerential rform on V the i
∗
ω is called the constrained
rform ω. Similarly if ([) is a metric tensor on V and ds
2
the line element
the i
∗
([) is the constrained matrix and i
∗
ds
2
is the constrained line element.
For each a ∈ X the pushforward
T
a
X
i∗
−→T
a
R
n
is injective and enables us to identify T
a
X with a vector subspace of T
a
R
n
.
The constrained metric and constrained line element are just the restrictions
to T
a
X of the matrix and line element, for each a ∈ X.
Example Let X be the constraint surface in R
3
:
f(x, y, z) = c f C
r
We have:
df =
∂f
∂x
dx +
∂f
∂y
dy +
∂f
∂z
dz unconstrained
Pulling back to X, where f is constant, we get:
0 =
∂f
∂x
dx +
∂f
∂y
dy +
∂f
∂z
dz constrained
If (say)
∂f
∂z
= 0 at a ∈ X then, by the implicit function theorem, the con
strained functions x, y are coordinates on X in a neighbourhood W of a and
constrained z is a C
r
function of x, y on W.
z = F(x, y)
say. Now
dz = −
∂f
∂x
∂f
∂z
dx −
∂f
∂y
∂f
∂z
dy
Therefore,
∂z
∂x
y
=
∂F
∂x
= −
∂f
∂x
∂f
∂z
∂z
∂y
x
=
∂F
∂y
= −
∂f
∂y
∂f
∂z
The usual line element on R
3
is
ds
2
= (dx)
2
+ (dy)
2
+ (dz)
2
unconstrained
131
pulling back to X gives:
ds
2
= (dx)
2
+ (dy)
2
+
−
∂f
∂x
∂f
∂z
dx −
∂f
∂y
∂f
∂z
dy
2
=
¸
1 +
∂f
∂x
∂f
∂z
2
(dx)
2
+ 2
∂f
∂x
∂f
∂y
(
∂f
∂y
)
2
dxdy +
¸
1 +
∂f
∂y
∂f
∂z
2
(dy)
2
The coeﬃcients give the components of the constrained metric on X with
respect to the coordinates x and y
Example using spherical polar coordinates on R
3
the usual line element is
ds
2
= (dr)
2
+r
2
(dθ)
2
+r
2
sin
2
θ(dφ)
2
Pulling back to the sphere S
2
by the constraint:
r = const
we have:
ds
2
= r
2
(dθ)
2
+r
2
sin
2
(dφ)
2
Thus the constrained line element has components:
r
2
0
0 r
2
sin
2
θ
7.8 Lagrange Multipliers
A scalar ﬁeld f on a manifold X has a critical point a ∈ X if df
a
= 0.
i.e.
∂f
∂y
i
(a) = 0 for a coordinate system y
i
at a. (e.g. a local maximum or
minimum or saddle point)
Problem given a scalar ﬁeld F on V open in R
n
, to ﬁnd the critical point
of constrained F, where the constraints are:
f
1
= c
1
, . . . , f
l
= c
l
and f
i
are C
r
functions on V
Method Take f
1
, . . . , f
l
, x
l+1
, . . . , x
n
(say) as coordinates on R
3
as in the
proof of the implicit function theorem, so
dF =
l
¸
i=1
∂f
∂f
i
df
i
+
n
¸
i=l+1
∂f
∂x
i
dx
i
unconstrained
132
Thus
dF = 0 +
n
¸
i=l+1
∂F
∂x
i
dx
i
constrained
This is zero at a critical point a, so
∂F
∂x
i
(a) = 0, i = l + 1, . . . , n, and
hence
dF
a
=
l
¸
i=1
∂F
∂f
i
(a)df
i
at a critical point a of constrained F. If we put
λ
i
=
∂F
∂f
i
(a) = rate of change of F at a with respect to the i
th
constraint
we have:
dF = λ
1
df
1
+ +λ
l
df
l
at a where the scalars λ
1
, . . . , λ
l
are called Lagrange multipliers.
Since dF is a linear combination of df
1
, . . . , df
l
at a we can take the
wedge product to get the equations:
dF ∧ df
1
∧ ∧ df
l
= 0
f
1
= c
1
, . . . , f
l
= c
l
which must hold at any critical point of constrained F.
7.9 Tangent space and normal space
If f is constant on the constraint manifold X then
df = 0 constrained
therefore,
'df, ˙ α(t)` = 0
for all curves in X. Hence the system of l linear equations
df
1
= 0, . . . , df
l
= 0
give the tangent space to X at each point. Also if ([) is a metric tensor
with domain V then
(gradf[ ˙ α(t)) = 'df, ˙ α(t)` = 0
133
for all curves in X. Therefore gradf is orthogonal to the tangent space to X
at each point. Hence
gradf
1
, . . . , gradf
l
is a basis for the normed space to X at each point.
At a critical point of constrained F we have:
dF = λ
1
df
1
+ +λ
l
df
l
and raising the index gives:
gradF = λ
1
gradf
1
+ +λ
l
gradf
l
thus gradF is normal to the constraint manifold X at each critical point of
constrained F.
7.10 ?? Missing Page
1. =⇒ ω is closed.
However ω is not exact because it’s integral around circle α(t) =
(cos t, sin t) 0 ≤ t ≤ 2π is
α
ω =
2π
0
[cos t cos t −(sin t)(−sin t)]
cos
2
t + sin
2
t
dt =
2π
0
dt = 2π = 0
Note: on R
2
−¦negative xaxis¦ we have:
x = r cos θ −π < θ < π
y = r sin θ
therefore,
ω =
r cos θ[sin θ dr +r cos θ dθ] −r sin θ[cos θ dr −r sin θ dθ]
r
2
cos
2
θ + r
2
sin
2
θ
= dθ
so.
α
ω is pathindependent provided α does not cross the ve xaxis.
α
ω = θ(b) −θ(a), and ω is called the angleform
2. Let ω be a diﬀerential nform with domain V open in R
n
ω = f(x
1
, . . . , x
n
)dx
1
∧ ∧ dx
n
(say) x
i
usual coord
then we deﬁne:
V
ω =
V
f(x
1
, . . . , x
n
)dx
1
dx
2
. . . dx
n
Lebesgue integral x
i
dummy
134
7.11 Integral of Pullback
Theorem 7.11.1. Let V
φ
−→φ(V ) be a C
1
diﬀeomorphism of open V in R
n
onto open φ(V ) in R
n
, with det φ
> 0. Let ω be an nform on φ(V ). Then
V
φ
∗
ω =
φ(V )
ω
[so writing 'ω, V ` to denote integral of ω over V we have: 'φ
∗
ω, V ` = 'ω, φV `
adjoint]
Proof.
PSfrag replacements
φ
∗
ω
V
ω
φV
φ
Let
ω = f(x
1
, . . . , x
n
)dx
1
∧ ∧ dx
n
φ = (φ
1
, . . . , φ
n
)
i.e. φ
i
(x) = x
i
(φ(x)) i.e. φ
i
= x
i
φ = φ
∗
x
i
.
Then
φ
∗
ω = f(φ
1
, . . . , φ
n
)dφ
1
∧ ∧ dφ
n
= f(φ(x))
∂φ
1
∂x
i
1
dx
i
1
∧ ∧
∂φ
n
∂x
in
dx
in
= f(φ(x)) det
∂φ
i
∂x
j
(x)
dx
1
∧ ∧ dx
n
therefore,
V
φ
∗
ω =
V
f(φ(x)) det φ
(x)dx
1
dx
2
. . . dx
n
=
φ(V )
f(x) dx
by the general change of variable theorem for multiple integrals (still to be
proved).
135
Corollary 7.11.2. Let y
1
, . . . , y
n
be positively oriented coordinates with do
main V open in R
n
. Then
V
f(y
1
, . . . , y
n
)dy
1
∧ ∧ dy
n
. .. .
nondummy y
i
with wedge
=
y(V )
f(y
1
, . . . , y
n
)dy
1
dy
2
. . . dy
n
. .. .
Lebesgue integral over y(V ). y
1
, . . . , yn dummy. No wedge
Proof.
PSfrag replacements
y
∗
x
i
V
y(V )
y = (y
1
, . . . , y
n
)
x
i
y
∗
x
i
= x
i
y = y
i
therefore,
LHS =
V
y
∗
[f(x
1
, . . . , x
n
) dx
1
∧ cdots ∧ dx
n
]
=
y(V )
f(x
1
, . . . , x
n
) dx
1
∧ ∧ dx
n
=
y(V )
f(x
1
, . . . , x
n
) dx
1
dx
2
. . . dx
n
x
1
, . . . , x
n
dummy
= RHS
7.12 integral of diﬀerential forms
To deﬁne
X
ω where ω is an nform and X is an ndimensional manifold
(e.g. ω is a 2form, X a surface) we need some topological notions:
1. A topological space X is called Hausdorﬀ if, for each a, b ∈ X, a = b, ∃
open disjoint V, W s.t. a ∈ V, b ∈ W.
136
PSfrag replacements
a
b
V
W
X
2. A set A ⊂ X is called closed in X if it’s complement in X
A
= ¦x ∈ X : x ∈ Z¦
is open in X.
PSfrag replacements
A
X
A
3. If A ⊂ X then the intersection of all the closed subsets of X which
contain A is denoted
¯
A and is called the closure of A in X;
¯
A is the
smallest closed subset of X which contains A.
4. Let S be a tensor ﬁeld on a manifold X. THe closure in X of the set
¦x ∈ X : S
x
= 0¦
is called the support of S, denoted supp S.
PSfrag replacements
S = 0
X
supp S
Theorem 7.12.1. Let X be a Hausdorﬀ manifold and let A ⊂ X be compact.
Then ∃ scalar ﬁelds
F
1
, . . . , F
k
on X s.t.
137
1. F
i
≥ 0
2. F
1
+ +F
k
= 1 on A ( partition of unity)
3. each supp F
i
is contained in the domain V
i
of a coordinate system on
X.
Proof. (sketch) Let
g(t) =
e
−
1
t
t > 0
0 t ≤ 0
PSfrag replacements
1
g(t)
let
b(t) =
g[1 −t
2
]
g(1)
PSfrag replacements
1
1
1
−1
b(t)
b is C
∞
, supp b = [−1, 1], 0 ≤ b ≤ 1. (b is for bump).
Let a ∈ A. Pick a coordinate system y on X at a with domain V . Pick
a ball centre y(a) radius 2r > 0 contained in y(V ). Put
h(x) =
b
y(x)−y(a)
r
if x ∈ V
0 if x ∈ V
¸
138
then h is C
∞
, supp h ⊂ V , 0 ≤ h ≤ 1, h(a) = 1; ’bump’ at a.
Let W = ¦x ∈ X : h(x) > 0¦. then W is an open neighbourhood of a
and h > 0 on W.
Thus, for each a ∈ A we have an open neighbourhood W
a
of a and a
scalar ﬁeld f
a
s.t. h
a
> 0 on W
a
, h
a
is C
∞
, supp h
a
⊂ a coord domain V
a
,
0 ≤ h
a
≤ 1.
PSfrag replacements
a
W
a
V
a
y(a)
r
2r
R
n
h
a
is a bump at a, for each a ∈ A. Since A is compact we can select a
ﬁnite number of points a
1
, . . . , a
k
such that W
a
1
, . . . , W
a
k
cover A. THen put
F
i
=
ha
i
ha
1
+···+ha
k
ifh
a
i
= 0
0 ifh
a
i
= 0
¸
to get the required scalar ﬁelds F
1
, . . . , F
k
.
139
PSfrag replacements
h
a
1
h
a
2
X
F
1
F
2
F
1
+F
2
= 1
7.13 orientation
Deﬁnition Let y
1
, . . . , y
n
with domain V , z
1
, . . . , z
n
with domain W be two
coordinate systems on a manifold X. Then they have the same orientation
if
∂(y
1
, . . . , y
n
)
∂(z
1
, . . . , z
n
)
= det
∂y
i
∂z
j
> 0
on V ∩ W.
Since
∂
∂z
j
=
∂y
i
∂z
j
∂
∂y
i
this means that
∂
∂z
1
a
, ,
∂
∂z
n
a
has the same orientation in T
a
X as
∂
∂y
1
a
, ,
∂
∂y
n
a
each a ∈ V ∩ W.
We call X oriented if a family of mutually compatible coordinate systems
is given on X, whose domains cover X and any two of which have the same
orientation.
Example x = r cos θ, y = r sin θ. Then
∂(x,y)
∂(r,θ)
= r > 0. Therefore x, y and
r, θ have the same orientation.
140
Deﬁnition Let ω be a diﬀerential nform with compact support on an ori
ented ndimensional Hausdorﬀ manifold X. We deﬁne
X
ω
the integral of ω over X as follows:
1. Suppose supp ω ⊂ V where V is the domain of positively oriented
coordinates y
1
, . . . , y
n
and ω = f(y
1
, . . . , y
n
) dy
1
∧ ∧dy
n
on V . Then
deﬁne:
x
ω =
y(V )
f(y
1
, . . . , y
n
) dy
1
. . . dy
n
dummy y
i
PSfrag replacements
V
W
X
z(W)
ω
2
ω
1
φ
y
z
y(V )
This is independent of the choice of coordinates since if supp ω ⊂ W
where W is the domain of positively oriented coordinates z
1
, . . . , z
n
then
ω = f(y
1
, . . . , y
n
) dy
1
∧ ∧ dy
n
= g(z
1
, . . . , z
n
) dz
1
∧ ∧ dz
n
(say) on V ∩ W
= y
∗
[f(x
1
, . . . , x
n
) dx
1
∧ ∧ dx
n
= z
∗
[g(x
1
, . . . , x
n
) dx
1
∧ ∧ dx
n
= y
∗
ω
1
= z
∗
ω
2
141
(say). So ω
1
= φ
∗
ω2 where φ = z y
−1
: y(V ∩ W) −→ z(V ∩ W).
Therefore,
y(V )
f(x
1
, . . . , x
n
) dx
1
. . . dx
n
=
y(V )
ω
1
=
y(V ∩W)
ω
1
=
z(V ∩W)
ω
2
=
z(W)
ω
2
=
z(W)
g(x
1
, . . . , x
n
) dx
1
. . . dx
n
as required.
2. Choose a partition of unity
F
1
+ +F
k
= 1
on supp ω. Put ω
i
= F
i
ω. Then
ω
1
+ +ω
k
= ω
and supp ω
i
⊂ supp F
i
⊂ V
i
where V
i
is the domain of a coordinates
system. Deﬁne
X
ω =
X
ω
1
+ +
X
ω
k
using (1.). This is independent of the choice of partition of unity since
if
G
1
+ +G
l
= 1 on supp ω
then
¸
l
j=1
X
G
j
ω =
¸
l
j=1
X
¸
k
i=1
F
i
G
j
ω
=
¸
l
j=1
¸
k
i=1
X
F
i
G
j
ω
similarly
=
¸
k
i=1
X
F
i
ω
Deﬁnition If A ⊂ X is a Borel set we deﬁne
A
ω =
X
χ
A
ω
Example to ﬁnd the area of the surface X:
x
2
+y
2
+z = 2, z > 0
We have:
2xdx + 2y dy +dz = 0
142
constrained to X. Therefore,
ds
2
= (dx)
2
+ (dy)
2
+ (2xdx + 2y dy)
2
constrained to X
= (1 + 4x
2
)(dx)
2
+ 8xy dxdy + (1 + 4y
2
)(dy)
2
g
ij
=
1 + 4x
2
4xy
4xy 1 + 4y
2
therefore,
√
g =
det g
ij
=
1 + 4x
2
+ 4y
2
. So, area element is
1 + 4x
2
+ 4y
2
dx∧
dy. Therefore:
area =
1 + 4x
2
+ 4y
2
dx ∧ dy
=
√
1 + 4r
2
r dr ∧ dθ
=
2π
0
√
2
0
r
√
1 + 4r
2
dr
dθ
= 2π
2
3
1
8
(1 + 4r
2
)
3
2
√
2
0
=
π
6
[27 −1]
=
13
3
π
143
144
Chapter 8
Complex Analysis
8.1 Laurent Expansion
Theorem 8.1.1. Let f be holomorphic on V − ¦a¦ where V is open and
a ∈ V . Let C be a circle, centre a, radius r s.t. C and it’s interior is
contained in V . Then ∃ ¦c
n
¦ n = 0, ±1, ±2, . . . ∈ C s.t.
f(z) =
∞
¸
n=−∞
c
n
(z −a)
n
in 0 < [z −a[ < r
The RHS is called the Laurent series of f about a. The coeﬃcient c
n
are
uniquely determined by:
c
n
=
1
2πi
C
f(z)
(z −a)
n+1
dz
Proof. Let w be inside C. Choose circles C
1
, C
2
as shown with centres a, w:
PSfrag replacements
C
a
w
C
1
C
2
145
Then
f(w) =
1
2πi
C
2
f(z)
z−w
dz
=
1
2πi
C
f(z)
z−a
d −
1
2πi
C
1
f(z)
z−w
dz
=
1
2πi
C
f(z)
(z−a)−(w−a)
dz +
1
2πi
C
1
f(z)
(w−a)−(z−a)
dz
=
1
2πi
C
f(z)
(z−a)[1−
w−a
z−a
]
dz +
1
2πi
C
1
f(z)
(w−a)[1−
z−a
w−a
]
dz
=
1
2πi
C
f(z)
(z−a)
¸
∞
n=0
w−a
z−a
n
dz +
1
2πi
C
1
f(z)
(w−a)
¸
∞
n=0
z−a
w−a
n
dz
=
¸
∞
n=0
(w −a)
n
1
2πi
C
f(z)
(z −a)
n+1
dz
. .. .
+ve powers (w −a)
+
¸
∞
n=0
(w −a)
−n−1
1
2πi
C
!
f(z)
(z −a)
−n
dz
. .. .
ve powers (w −a)
as required.
Deﬁnition If f is holomorphic in V −¦a¦ anf
f(z) =
∞
¸
n=−∞
c
n
(z −a)
n
= +
c
−2
(z −a)
2
+
c
−1
z −a
. .. .
p(z)
+c
0
+c
1
(z −a) +
is the Laurent series of f at a. p(z) =
¸
−1
n=−∞
c
n
(z−a)
n
is called the principal
part of f at a. p(z) is holomorphic on C −¦a¦.
f(z) −p(z) is holomorphic in V (deﬁning it’s value at 0 to be c
0
).
For any closed curve α in C −¦a¦:
α
p(z) dz =
α
+
c
−2
(z−a)
2
+
c
−1
z−a
dz
= + 0 +c
1
α
dz
z−a
= 2πi Res(f, a) W(α, a)
where W(α, a) =
1
2πi
α
dz
z−a
is the winding number of α about a. and
Res(f, a) = c
−1
is the residue of f at a
146
8.2 Residue Theorem
Theorem 8.2.1 (Residue Theorem). Let f be holomorphic on V −¦a
1
, . . . , a
n
¦
where a
1
, . . . , a
n
are distinct points in a starshaped (or contractible) open set
V . Then for any closed curve α in V −¦a
1
, . . . , a
n
¦ we have:
α
f(z) dz = 2πi
n
¸
j=1
Res(f, a
j
) W(α, a
j
)
Proof. Let p
1
, . . . , p
n
be the principal parts of f at a
1
, . . . , a
n
respectively.
Then f−(p
1
+ +p
n
) is holomorphic on V . Therefore,
α
[f −(p
1
+ +p
n
)] dz =
0. So,
f(z) dz =
n
¸
j=1
α
p
j
(z) dz = 2πi
n
¸
j=1
Res(f, a
j
)W(α, a
j
)
as required.
Deﬁnition If f is holomorphic on a neighbourhood of a, excluding possibly
a itself, and if the Laurent expansion has at most a ﬁnite number of negative
powers:
f(z) =
¸
∞
r=n
c
r
(z −a)
r
= c
n
(z −a)
n
+c
n+1
(z −a)
n+1
+ c
n
= 0
= (z −a)
n
[c
n
+c
n+1
(z −a) + ]
= = (z −a)
n
f
1
(z) f
1
holomorphic on a nbd of a and f
1
(a) = 0
then we say that f has order n at a.
e.g. order 2:
f(z) = c
2
(z −a)
2
+c
3
(z −a)
3
+ c
2
= 0
order 3:
f(z) =
c
−2
(z −a)
3
+
c
−2
(z −a)
2
+
c
1
(z −a)
+c
0
+c
1
(z −a) + c
−3
= 0
If n > 0 we call a a zero of f.
If n < 0 we call a a pole of f.
n = 1: simple zero; n = 2: double zero.
147
n = −1: simple pole; n = −2: double pole.
If f has order n and g has order m:
f(z) = (z −a)
n
f
1
(z) f
1
(a) = 0
g(z) = (z −a)
n
g
1
(z) g
1
(a) = 0
then:
f(z) g(z) = (z −a)
n+m
f
1
(z) g
1
(z) fg has order n +m
f(z)
g(z)
= (z −a)
n−m
f
1
(z)
g
1
(z)
f
g
has order n −m
The residue theorem is very useful for evaluating integrals:
Example to evaluate
∞
0
x sinx
x
2
+a
2
, a > 0 we put
f(z) =
z e
iz
z
2
+a
2
[e
iz
is easier to handle than sin z]
=
z e
iz
(z−ia)(z+ia)
Simple poles at ia, −ia.
PSfrag replacements
ia
−ia
−R R
α
Choose a closed contour that goes along xaxis −R to R (say) then loops
around a pole, upper semicircle α (say).
R
−R
xe
ix
x
2
+a
2
dx +
α
z e
iz
z
2
+a
2
dz = 2πi Res(f, ia)
148
1. if f(z) =
c
z−ia
+ c
0
+ c
1
(z − ia) + on neighbourhood of ia then
(z − ia)f(z) = c + c
0
(z − ia) + c
1
(z − ia)
2
+ on neighbourhood of
ia. Then lim
z−→ia
(z −ia)f(z) = c = Res(f, ia). Therefore
Res(f, ia) = c = lim
z−→ia
f(z)
z−ia
= lim
z−→ia
z e
iz
z+ia
=
ia
2ia
e
−a
=
1
2
e
−a
2. α(t) = Re
it
0 ≤ t ≤ π = R(cos t +i sin t). Therefore
α
z e
iz
z
2
+a
2
dz
=
π
0
Re
it
e
iR(cos t+i sin t)
iRe
it
R
2
e
2it
+a
2
dt
≤
R
2
R
2
−a
2
π
0
e
−Rsin t
dt
which −→0 as R −→∞ since
lim
R→∞
π
0
e
−Rsint
dt
DCT
=
π
0
lim
R→∞
e
−Rsin t
dt = 0
Therefore
∞
−∞
x(cos x +i sin x)
x
2
+a
2
dx + 0 = 2πi
e
−a
2
so
∞
−∞
xsin x
x
2
+a
2
dx = πe
−a
Example to calculate
∞
−∞
sinx
x
dx put f(z) =
e
iz
z
holomorphic except for
simple pole at z = 0.
PSfrag replacements
−R R
α
−r r
β
149
Choose a closed contour along xaxis from −R to −r, loops around 0 by
β(t) = re
it
then r to R then back along α(t) = Re
it
.
−r
−R
e
ix
x
dx +
β
e
iz
z
dz +
R
r
e
ix
x
dx +
α
e
iz
z
dz = 0
1. Res(f, 0) = lim
z→0
(z −0)f(z) = lim
z→0
e
iz
= 1
2.
α
e
iz
z
dz
=
π
0
e
iR(cos t+i sin t)
iRe
it
Re
it
dt
≤
π
0
e
−Rsin t
dt
which −→0 as R −→∞.
3.
e
iz
z
=
1
z
+g(z), g holomorphic in neighbourhood of 0. Therefore
β
e
iz
z
dz =
β
dz
z
+
β
g(z) dz
Let sup [g(z)[ = M (say on a closed ball centre 0 redius δ > 0 (say).
β
g(z) dz
≤ Mπr −→0 as r −→ 0
if r ≤ δ
β
dz
z
= −
π
0
ire
it
re
it
dt = −
π
0
i dt = −iπ ∀r
therefore
lim
r−→0
β
e
iz
z
dz = −iπ
therefore
0
−∞
e
ix
x
dx −iπ +
∞
0
e
ix
x
dx + 0 = 0
so
∞
−∞
cos x +i sin x
x
dx = iπ
and
∞
−∞
sin x
x
dx = π
150
8.3 Uniqueness of analytic continuation
Theorem 8.3.1 (Uniqueness of analytic continuation). Let f, g be holo
morphic on a connected open set V . Let a ∈ V and let ¦z
k
¦ be a sequence in
V (= a) converging to a s.t.
f(z
k
) = g(z
k
) ∀k
Then f = g on V .
Proof. Put F = f −g, so F(z
k
) = 0. To show F = 0 on V .
1. F is holomorphic at a. Therefore F(z) = b
0
+b
1
(z −a)+b
2
(z −a)
2
+
on [z −a[ < R (say). F(a) = limF(z
k
) = 0. Therefore b
0
= 0.
Suppose we know that b
0
, b
1
, . . . , b
m−1
are all zero.
F(z) = (z −a)
m
[b
m
+b
m+1
(z −a) +b
m+2
(z −a)
2
+ ]
= (z −a)
m
F
m
(z)
(say). F(z
k
) = 0, so F
m
(z
k
) = 0 and F
m
(a) = 0. Therefore b
m
= 0.
b
r
= 0∀r and F = 0 on an open ball centre a.
2. Put V = W∪W
where W = ¦z ∈ V : F = 0 on an open ball centre z¦
and W
= V − W. Then W is open, and a ∈ W by 1. Therefore W
is nonempty. Suppose c ∈ W
and c is a noninterior point of W
.
Then each integer r > 0∃ w
r
∈ W s.t. [w
r
− c[ <
1
r
, F(w
r
) = 0 and
limw
r
= c. Therefore F = 0 on an open ball centre c by 1., and c ∈ W.
Therefore W
is empty.
151
152
Chapter 9
General Change of Variable in
a multiple integral
9.1 Preliminary result
Theorem 9.1.1. Let R
n
⊃ V
φ
−→ R
n
be C
1
, V open, a ∈ V , det φ
(a) = 0.
Then
lim
m(φB)
m(B)
= [ det φ
(a)[
where the limit is taken over cubes B containing a with radius B −→0.
Proof. Let   be the sup norm on R
n
.
(α
1
, . . . , α
n
) = max([α
1
[, . . . , [α
n
[)
so a ball, radius r is a cube, with side 2r.
Let 0 < < 1. Put T(x) = [φ
(x)]
−1
. Fix a closed cube J containing
a and put k = sup
x∈J
T(x). Choose δ > 0 s.t. φ
(x) − φ
(y) ≤
k
all
x −y ≤ 2δ; x, y ∈ J.
If B is a cube ⊂ J containing a of radius ≤ δ. and centre c (say).
Consider:
T(c)φ has derivative T(c)φ
(x) which equals 1 at x = c and
T(c)φ
(x) −1 = T(c)φ
(x) −T(c)φ
(c)
≤ T(c φ
(x) −φ
(c)
≤ k
k
=
153
therefore (1 − )B
1
⊂ T(c)φB ⊂ (1 + )B
1
, where B
1
is a translate of B to
new centre T(c)φ(x). Therefore
(1 −)
n
m(B) ≤ [ det T(c)[m(φB) ≤ (1 +)
n
m(B)
=⇒ (1 −)
n
≤ [ det T(c)[
m(φB)
m(B)
≤ (1 +)
n
=⇒ lim[ det T(c)[
m(φB)
m(B)
= 1
=⇒ [ det T(a)[ lim
m(φB)
m(B)
= 1
Therefore:
lim
m(φB)
m(B)
=
1
[ det T(a)[
= [ det φ
(a)[
as required
Theorem 9.1.2. Let f be a continuous real valued function on an open
neighbourhood of a in R
n
. THen
lim
B
f(x) dx
m(B)
= f(a)
where the limit is taken over cubes B containing a with radius B −→ 0.
Proof. Let > 0. Choose δ > 0 s.t. [f(x) − f(a)[ < ∀ x − a < δ. Then
each cube B containing a of radius ≤
δ
2
we have:
B
f(x) dx
m(B)
−f(a)
=
B
[f(x) −f(a)] dx
m(B)
≤
m(B)
m(B)
=
hence result.
Recall that the σalgebra generated by the topology of R
n
is called the
collection of Borel Sets in R
n
.
Theorem 9.1.3. Let A be a Borel set in R
r
, B be a Borel set in R
s
. Then
AB is a Borel set in R
r+s
.
Proof. For ﬁxed V open in R
r
, the sets:
¦V W : W open in R
s
¦ (9.1)
154
are all open in R
r+s
. Therefore the σalgebra generated by Eqn(9.1):
¦V B : B Borel in R
s
¦
consists of Borel sets in R
r+s
. Hence for ﬁxed B Borel in R
s
, the set:
¦V B : V open in R
r
(9.2)
are all Borel sets in R
r+s
. Therefore the σalgebra generated by Eqn(9.2):
¦AB : A Borel in R
r
¦
consists of Borel sets in R
r+s
. Therefore A B is Borel for each A Borel in
R
r
, B Borel in R
s
.
Theorem 9.1.4. Let E be a lebesgue measurable subset of R
n
. THen there
is a Borel set B containing E such that
B −E = B ∩ E
has measure zero, and hence m(B) = m(E).
Proof. We already know this is true for n = 1. So we use induction on n.
Assume true for n −1. Let E ⊂ R
n
, E measurable.
1. Let m(E) < ∞. Let k be an integer > 0.
E ⊂ R
n−1
R
choose a sequence of rectangles
A
1
B −1, A
2
B
2
, . . .
covering E with A
i
⊂ R
n−1
measurable and B
i
⊂ R measurable, and
such that
m(E) ≤
∞
¸
i=1
m(A
i
)m(B
i
) ≤ m(E) +
1
k
By the induction hypothesis choose Borel sets C
i
, D
i
s.t.
A
i
⊂ C
i
, B
i
⊂ D
i
, m(A
i
) = m(C
i
), m(B
i
) = m(D
i
)
Put B
k
=
¸
∞
i=1
C
i
D
i
. Then E ⊂ B
k
, B
k
is Borel and m(E) ≤
m(B
k
) ≤
¸
∞
i=1
m(C
i
)m(d
i
) ≤ m(E) +
1
k
.
Put B =
¸
∞
k=1
B
k
. Then E ⊂ B, B is Borel and m(E) = m(B).
Therefore m(B ∩ E
) = m(B) −m(E) = 0 as required.
155
2. Let m(E) = ∞. Put E
k
= ¦x ∈ E : k ≤ [x[ < k + 1¦. E =
¸
∞
k=0
E
k
is
a countable disjoint union, and m(E
k
) < ∞. For each integer k choose
by 1. Borel B
k
s.t. E
k
⊂ B
k
and m(B
k
∩ E
k
) = 0
Put B =
¸
B
k
. Then E ⊂ B and m(B ∩ E
) = m(B) − m(E) as
required. ???????????
9.2 General change of variable in a multiple
integral
Theorem 9.2.1 (General change of variable in a multiple integral).
Let R
n
⊃ V
φ
−→ W ⊂ R
n
be a C
1
diﬀeomorphism of open V onto open W.
Let f be integrable on E. Then
W
f(x) dx =
V
f(φ(x))[ det φ
(x)[ dx (Lebesgue Integrals) (9.3)
Proof. 1. we have f = f
+
−f
−
with f
+
, f
−
≥ 0. Therefore, it is suﬃcient
to prove Eqn(9.3) for f ≥ 0.
2. if f ≥ 0 then ∃ a monotone increasing sequence of nonnegative simple
functions ¦f
n
¦ such that f = limf
n
so, using the monotone convergence
theorem, it is suﬃcient to prove Eqn(9.3) for f simple.
3. if f is simple then f =
¸
k
i=1
a
i
χ
E
i
with ¦E
i
¦ Lebesgue measurable,
so it is suﬃcient to prove Eqn(9.3) with f = χ
E
with E Lebesgue
measurable.
4. if E Lebesgue measurable ∃ Borel B s.t. E ⊂ B and Z = B − E has
measure zero. Therefore, χ
E
= χ
B
− χ
Z
, and it is suﬃcient to prove
Eqn(9.3) for f = χ
B
, B Borel, and for f = χ
Z
, Z measure zero.
5. If f = χ
E
with E Borel then E = φF with F Borel and Eqn(9.3)
reduces to
W
χ
φF
(x) dx =
V
χ
F
(x)[ det φ
(x)[ dx
i.e.
m(φF) =
F
[ det φ
(x)[ dx (9.4)
If Eqn(9.4) holds for each rectangle F ⊂ V then Eqn(9.4) holds for F ∈
ring R of ﬁnite disjoint unions of rectangles ⊂ V .
156
Therefore Eqn(9.4) holds for F ∈ monotone class generated by R.
Therefore Eqn(9.4) holds for F ∈ σalgebra generated by R (monotone
class lemma). So Eqn(9.4) holds for any Borel set F ⊂ V
6. to show Eqn(9.4) holds for any rectangle F ⊂ V : for each rectangle
F ⊂ V put
λ(F) = m(φF) −
F
[ det φ
(x)[ dx
Then, λ is additive:
λ(
¸
B
j
) =
¸
λ(B
j
)
for a disjoint union.
Must prove λ(F) = 0 for each rectangle F. Suppose B is a rectangle
for which λ(B) = 0.
Suppose B is a cube. [λ(B)[ > 0. Therefore, ∃ > 0 s.t. [λ(B)[ ≥
m(B).
Divide B into disjoint subcubes, each of
1
2
the edge of B. For one such,
B
1
say,
[almbda(B
1
)[ ≥ m(B
1
)
Sivide again
[λ(B
2
)[ ≥ m(B
2
)
continuing we get a decreasing sequence of cubes ¦B
k
¦ converging to a
say, with
lim
h−→∞
[λ(B
k
)[
m(B
k
)
≥ > 0
But
lim
h→∞
λ(B
k
)
m(B
k
)
= lim
h→∞
m(φB
k
)−
B
k
 det φ
(x) dx
m(B
k
)
= [ det φ
(a)[ −[ det φ
(a)[
= 0
a contradiction. Therefore λ(B) = 0 for all cubes B. Therefore,
m(φB) =
B
[ det φ
(x)[ dx for all cubes B as required.
7. Now suppose Z has measure zero, and choose a Borel set B s.t.
Z ⊂ B ⊂ V
and s.t. B has measure zero.
157
Then
m(φZ) ≤ m(φB) =
B
[ det φ
(x)[ dx = 0
since B has measure zero. Therefore φZ has measure zero.
Therefore under a C
1
diﬀeomorphism we have:
Z of measure zero =⇒ φZ of measure zero
therefore f = χ
Z
with Z of measure zero =⇒ f φ = χ
φ
−1
Z
with
φ
−1
Z of measure zero.
Therefore, Eqn(9.3) holds for f = χ
Z
since then both sides are zero.
This completes the proof.
158
This action might not be possible to undo. Are you sure you want to continue?