MA231

Vector Analysis
WMS Revision Guide
Written by Khallil Benyattou

WMS
 ii MA231 Vector Analysis

Contents
1 Review of MA134 Geometry and Motion 1
1.1 Notation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1
1.2 The Tangential Line Integral . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1
1.3 Flux . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2

2 The Integral Theorems of Multivariable Calculus 3

2.1 Green’s Theorem . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3
2.2 Divergence Theorem . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3
2.3 Stokes’ Theorem . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4
2.4 Applications . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4

3 Introduction to Complex Analysis 5

3.1 Continuity & Convergence . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5
3.2 Complex Differentiation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5
3.3 Interpretations of Complex Valued Functions . . . . . . . . . . . . . . . . . . . . . . . . . . 6
3.4 Complex Line Integrals . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7
3.5 Linking Complex Analysis and Vector Analysis . . . . . . . . . . . . . . . . . . . . . . . . . 8
3.6 Power Series for Holomorphic Functions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9
3.7 Applications of Taylor’s and Cauchy’s Theorems . . . . . . . . . . . . . . . . . . . . . . . . 10

Introduction
Authors
Written by Khallil Benyattou to reflect the 2016 course taught by Dr. Mario Micallef.
General structure and chronology inspired by D.S. McCormick’s 2007 edition.

Feedback
Improvements can be sent via e-mail at revision.guides@warwickmaths.org
You can also enter suggestions into the society’s feedback form at http://tinyurl.com/WMSGuides

Website
You can find more revision guides and lots of other useful material on the WMS website:

http://www.warwickmaths.org

You can also find scanned lecture notes for 2013, by Alex Wendland, on Dropbox:

https://www.dropbox.com/sh/5m63moxv6csy8tn/0EaztHn5TH
History
First Edition: 2007
Most Recent: 2016

Author’s Note
I’d like to offer my sincere thanks to Christopher Nguyen for allowing me to view his scanned notes for
the module. Without his help with proof-reading, the quality of this guide wouldn’t be the same.
MA231 Vector Analysis 1

1 Review of MA134 Geometry and Motion

1.1 Notation
Notation’s always a tricky thing to get right. Sometimes it’s too suggestive and at others, too obscure.
This happens to be the case with multivariable calculus, most notably derivatives. Indeed, a conference
held in 1903 with the aim of agreeing upon standard notation for the partial derivative did the complete
opposite of what was intended. Long story short, there are now three more in existence. We aim to be
consistent with the notation used by the good doctor Mario Micallef himself.

• We will write x = (x1 , x2 , . . . , xn ) for a vector in Rn , and f : Rn → Rm for a function with components
f1 , . . . , fm : Rn → R.

• A function f : Rn → R is called a scalar field, and a function f : Rn → Rn is called a vector field.

• Given p, q ∈ Rn , a curve starting from p and ending at q is denoted by Cpq .

• For Ω ⊂ Rn , ∂Ω will denote the boundary of Ω, and Ω the closure of Ω.

Notably, we can observe that Ω = Ω ∪ ∂Ω.

Of course, notation elsewhere will vary but if you understand the general gist of the theorems and
results, interpreting them in other notation shouldn’t be too much of a problem.

1.2 The Tangential Line Integral

Definition 1.1. The tangential line integral of a vector field v along a curve Cpq that’s parameterised by
r : [a, b] → Rn , with r(a) = p and r(b) = q, is defined by
ˆ ˆ b
..=
dr
v · dr v(r(t)) · dt.
Cpq a dt

Note that this integral depends on the orientation of the path. However, it does not depend on the
parametrisation of the curve. In particular, we can parametrise the curve Cpq (of length l) by arc-length
s and re-write our tangential line integral as:
ˆ ˆ l
dr
v · dr = v(r(s)) · ds.
Cpq 0 ds

Definition 1.2. For f : Rn → R, the gradient of f is the vector field ∇f defined by

 
∂f ∂f ∂f
∇f ..= , ,..., .
∂x1 ∂x2 ∂xn

Definition 1.3. A gradient vector field is a vector field v : Rn → Rn such that v = ∇f for some f : Rn →
R. If this is the case, we call f a scalar potential 1 of v.

The Chain Rule. Given a function f : Rn → R and a path r : [a, b] → Rn , we can define a new map
h : [a, b] → R to describe how our scalar field acts along the path by h(t) ..= f (r(t)). Then
n
dh d X ∂f dxi dr
= (f (r(t))) = (r(t)) = ∇f (r(t)) · ,
dt dt i=1
∂xi dt dt
 
dr dx1 dxn
where r(t) = (x1 (t), . . . , xn (t)) and = ,..., .
dt dt dt

1 It’s worth mentioning that scalar potentials aren’t unique. This is because ∇f = ∇g whenever g(x) ..= f (x) + c for every

real number c. Although not every vector field is a gradient field, they happen to be quite useful as they allow us to easily
evaluate line integrals.
 2 MA231 Vector Analysis

FTC for Gradient Fields. Given a function f : Rn → R and a curve Cpq from p to q parameterised by
r : [a, b] → Rn , we have that ˆ
∇f · dr = f (q) − f (p).
Cpq
˛
n
Definition 1.4. A vector field v is called conservative if for all closed curves C ⊂ R , v · dr = 0.
C

Proposition 1.5. A vector field v : Rn → Rn is conservative iff ∀ p, q ∈ Rn , the tangential line integral
ˆ
v · dr
Cpq

is independent of the choice of path Cpq from p to q.

Theorem 1.6. A vector field v is a gradient field iff it’s conservative.

1.3 Flux
Definition 1.7. Given v = (a, b) ∈ R2 , define v⊥ ..= (b, −a). v⊥ is pronounced v perp and is the rotation
of v clockwise by 90 degrees.2

Definition 1.8. Given a regular curve C that’s parameterised by r(t) = (x(t), y(t)),
 
dx dy
• a tangent to C is given by r0 (t) = , and
dt dt
 
. 0⊥ dy dx
• a normal to C is given by N(t) .= r (t) = ,− .
dt dt

Definition 1.9. The flux of a planar vector field v(x, y) = (a(x, y), b(x, y)) across a curve Cpq parame-
terised by r : [a, b] → R2 is defined to be the integral
ˆ ˆ b
v · n ds = v(r(t)) · N(t) dt,
C a

where n is the unit normal to C and s is again the arc length parameter of C. The flux can be thought of
as a measure of how much the field runs across the oriented curve in question. This is as opposed to the
field running along the curve, where we discussed the tangential line integral.

Definition 1.10 (Coordinate Transformations in R3 ). Let f be a map from R3 to itself. If T : R3 → R3

defined by T (u, v, w) = (x(u, v, w), y(u, v, w), z(u, v, w)) is a re-parameterisation of Γ ⊂ R3 and maps Γ
into Ω ⊂ R3 , then3
˚ ˚  
 ∂(x, y, z) 
f dV = f   du dv dw
Ω Γ ∂(u, v, w) 

Definition 1.11. Given a vector field v : R3 → R3 and a surface S in R3 which is parameterised by

r : Ω → R3 , the flux of v through S is given by:
¨ ¨ ¨  
∂r ∂r ∂r ∂r
v · n dS = (v · n) ×
∂u ∂v
dA = ± v · × dA
S Ω Ω ∂u ∂v

where n is a unit normal to S whose direction ± will be indicated by the context.

2 This cannot be done in Rn for n > 3 as there is no notion of rotation by 90 degrees.

3 It’s understood that for the integral on the right, we substitute for (x, y, z) in terms of (u, v, w) in f .
MA231 Vector Analysis 3

2 The Integral Theorems of Multivariable Calculus

2.1 Green’s Theorem
Definition 2.1. If Ω ⊂ R2 is a region with outward unit normal n, then the positively oriented unit
tangent with respect to n is defined as t ..= −n⊥ .

Definition 2.2. Given a planar vector field v : R2 → R2 where v(x, y) ..= (a(x, y), b(x, y)), its curl4 is a
function given by
∂b ∂a
curl v = ∇ × v ..= − .
∂x ∂y

Green’s Theorem. Let Ω ⊂ R2 be a region, v(x, y) = (a(x, y), b(x, y)) be a planar vector field on Ω and
t be the positively oriented unit tangent to the boundary ∂Ω. Then,
¨ ˛
curl v dA = v · t ds.
Ω ∂Ω

We’ve been discussing regions a whole lot so it’s good to get a precise feel for what they are.

Definition 2.3. A subset Ω of Rn is called a region if there’s a map f : Rn → R satisfying the following
conditions:

(i) All partial derivatives of f are continuous.

(ii) The relation between Ω and f is Ω = {x ∈ Rn : f (x) < 0}.

(iii) ∇f (x) 6= 0 for all x ∈ ∂Ω ..= {x : f (x) = 0}.

If there is such an f , we call it a defining function of the region Ω. Note that f isn’t unique i.e. there
are many possible choices of defining function.

2.2 Divergence Theorem

Definition 2.4. Let v(x) = (v1 (x1 , . . . , xn ), . . . , vn (x1 , . . . , xn )) be a vector field on Rn . Then, the
divergence of v is defined to be:
n
X ∂vi
div v = ∇ · v ..=
i=1
∂x i

Divergence Theorem. Let Ω be a solid region in R3 and v : Ω → R3 be a vector field on Ω. Then

˚ ¨
∇ · v dV = v · n dA
Ω ∂Ω

where n is the unit outward normal to Ω.

The Divergence Theorem also applies to planar regions, Ω ⊂ R2 . This follows from Green’s Theorem
applied to v⊥ where n is an outward pointing unit normal from Ω.
¨ ˛
∇ · v dA = v · n ds
Ω ∂Ω

If the divergence of a vector field is locally constant in a region, we can say that the divergence is a
measure of the flux per unit volume of the vector field.

4 The reason for the notation ∇ × v is because the curl of a vector field v = (a, b, c) in R3 is given precisely by the

determinant below. This determinant isn’t a literal cross product of a “vector of differential operators” and the vector field,
but it’s reminiscent.  
 i j k 
.

curl v = ∂x ∂y ∂z 
. 
a b c
 4 MA231 Vector Analysis

An interesting case of 2.4 is when we take the divergence of a gradient field v = ∇f .

Definition 2.5. Let f : U → R be a scalar field for some open subset U of Rn . The Laplacian of f is
defined as
n
X ∂2f
∆f ..= ∇ · (∇f ) = .
i=1
∂x2i

Definition 2.6. A function f : U → R, where U ⊂ Rn is open, with ∆f = 0 is known as harmonic.

Lemma 2.7. Let f : R3 → R be harmonic. Then for all regions Ω ⊂ R3 , we have that
¨
∇f · n dA = 0.
∂Ω

2.3 Stokes’ Theorem

Definition 2.8. A curve or surface is regular if there exists a parameterisation whose derivative is well
defined and non-zero at all points. When we say regular, we mean that the object doesn’t have any corners
or cusps.

Definition 2.9. A surface S ⊂ R3 is regular if ∀ p ∈ S, ∃  > 0 and there’s a region Ω ⊂ R2 such that
B(p, , S) = {x ∈ S : kx − pk < } can be parameterised by a regular map r : Ω → R3 i.e. at all points in
Ω, we have that
∂r ∂r
n= × 6= 0.
∂u ∂v
So on a regular surface S, we can assign a unit normal vector given by n/knk on small pieces of the surface.

Definition 2.10. A regular surface S ⊂ R3 is orientable if a choice of unit normal can be continuously
assigned on all of S.

Stokes’ Theorem5 . Let S be a regular surface in R3 , oriented by a continuous choice of unit normal n,
and let v be a vector field on R3 . Let t be a unit tangent to ∂S which is positively oriented with respect
to n in the sense that n × t points into S. Then,
¨ ˆ
curl v · n dA = v · t ds.
S ∂S

2.4 Applications
Lemma 2.11. Let Ω be a solid region in R3 and f : Ω → R3 be continuously differentiable on Ω. Then6
∇ × (∇f ) = 0 in Ω.
Proof. Let S be an arbitrary surface. Applying Stokes’ Theorem to ∇f on S gives
¨ ˆ
FTC
∇ × (∇f ) · n dA = ∇f · t ds = f (q) − f (p) = 0
S ∂S

Lemma 2.12. Let Ω be a solid region in R3 and v : Ω → R3 be a vector field. Then ∇ · (∇ × v) = 0.

Proof. ˚ ¨
∇ · (∇ × v) dV = ∇ × v · n dA by the Divergence Theorem
Ω ∂Ω
ˆ
= v · t ds via Stokes’ Theorem
∂(∂S)

= 0 since ∂ (∂S) = ∅.

As Ω is arbitrary, we’re forced to conclude that ∇ · (∇ × v) = 0.

5 This is Green’s Theorem for orientable surfaces with boundary in R3 .

6 Take note that the calculus identity ∇ × (∇ · ) = 0 for some function is closely related to the topological fact that
∂ (∂ · ) = ∅ for some surface.
MA231 Vector Analysis 5

Definition 2.13. A region Ω ⊂ R3 is said to have trivial 1st homology if, inside of Ω, every closed curve
C is the boundary of some surface S.

Proposition 2.14. If Ω ⊂ R3 has trivial 1st homology and v : Ω → R3 satisfies ∇ × v = 0, then v is

conservative and admits a scalar potential.

Definition 2.15. A region Ω ⊂ R3 is said to have trivial 2nd homology if, inside of Ω, every closed surface
S is the boundary of some region E.

Proposition 2.16. If Ω ⊂ R3 has trivial 2nd homology and v : Ω → R3 satisfies ∇ · v = 0, then v

admits a vector potential i.e. ∃ w : Ω → R3 such that v = ∇ × w.

3 Introduction to Complex Analysis

Our next aim is to understand the calculus of functions f : C → C. Writing f (x + iy) = u(x, y) + iv(x, y)
where u, v : R2 → R, we see that there is a one-to-one correspondence between functions f : C → C and
vector fields of the form (u, v) : R2 → R2 . We’ll capitalise upon this link to apply the Divergence Theorem
and Stokes’ Theorem to derive very useful results.

3.1 Continuity & Convergence

The definitions of convergence and continuity for complex functions are analogous to their real counterparts.
However, as one would expect, R ⊂ C so considering intervals no longer suffices. Instead, we consider balls:

Definition 3.1. For z ∈ C and ε > 0, the open ball of radius ε around z is B(z, ε) = {w ∈ C : |w − z| < ε}.
Definition 3.2. A set E ⊂ C is open if ∀z ∈ E, ∃ r > 0 such that B(z, r) ⊂ E.
 From now on, E will always denote an open subset of C.

Definition 3.3. Let (zn )n∈N be a sequence in C where zn = xn + iyn . Then (zn ) converges to some z ∈ C
if |zn − z| → 0 as n → ∞. More precisely, this means that ∀ ε > 0, ∃ N ∈ N such that whenever n > N ,
zn ∈ B(z, ε). As usual, z is called the limit of the sequence and we write that zn → z.

The usual rules for limits of sequences follow from this definition. If (zn ) → z and (wn ) → w then
(zn + wn ) → z + w, (zn wn ) → zw and zn /wn → z/w so long as w 6= 0.

Definition 3.4. For E ⊂ C, f : E → C is continuous at z ∈ E if zn → z =⇒ f (zn ) → f (z).

Equivalently, f is continuous at z if ∀ε > 0, ∃ δz > 0 such that w ∈ B(z, δz ) =⇒ f (w) ∈ B(f (z), ε).
Again, the usual rules for continuous functions follow. If f, g : E → C are continuous at z, then so
are f + g, f g and f /g. For the last case, we of course require that g(z) 6= 0. We’re now ready to define
differentiability in the complex plane.

3.2 Complex Differentiation

Definition 3.5. For E ⊂ C, f : E → C is complex differentiable 7 at z ∈ E if the following limit exists
(and is finite)
lim f (z + h) − f (z) .
h→0
in C h

If the limit does indeed exist, we call the limiting value the derivative of f at z, written f 0 (z).
Note that this condition is more restrictive than in the real case. This is because there are more ways
for a sequence (hn ) to tend to 0 in C. The following examples serve to illustrate this point more concretely.

7 This f (z + hn ) − f (z)
is equivalent to saying that the limit lim exists for every null sequence (hn )n∈N in C.
n→∞ hn
 6 MA231 Vector Analysis

Example 3.6. Let f : C → C defined by:

1. f (z) = z is differentiable at every point of C, and f 0 (z) = 1.

2. f (z) = z is not differentiable anywhere in the complex plane. Consider the difference quotient

f (z + hn ) − f (z) z + hn − z hn
δ ..= = = .
hn hn hn
In the case that hn = 1/n, δ = 1. If hn = i/n, then δ = −i/i = −1. Since these two differ, the limit
doesn’t exist and f is accordingly not differentiable at any point z ∈ C.

I’m beginning to sound like a parrot but the usual rules for differentiation i.e. the product, quotient
and chain rules remain valid for complex differentiation.

3.3 Interpretations of Complex Valued Functions

There are three ways to view complex valued functions:

(i) In the naı̈ve algebraic way that we have been.

(ii) As functions F : R2 → R2 where we make the following correspondences:
• z = x + iy ∈ C ←→ (x, y) ∈ R2

• f (z) = u(z) + iv(z) ∈ C ←→ F (x, y) = (u(x, y), v(x, y)) ∈ R2

(iii) f : E → C may be viewed as a vector field f : E → R2 . If f = u + iv, then f is the vector field8
f (x, y) = (u(x, y), −v(x, y)).

Using our correspondence, we’ll look at complex and real differentiability. Let the complex number
h = h1 + ih2 correspond to the real vector h = (h1 , h2 ).

Definition 3.7. A function F : E → R2 is Fréchet differentiable at (x0 , y0 ) ∈ E if there is a bounded

linear map A(x0 ,y0 ) : R2 → R2 such that

kF (x0 + h1 , y0 + h2 ) − F (x0 , y0 ) − A(x0 ,y0 ) hk

lim = 0.
khk→0 khk

The linear map A(x0 ,y0 ) is then called the Fréchet derivative of F at (x0 , y0 ). Using the standard basis of
R2 , our linear map A(x0 ,y0 ) has a matrix representation. In fact, it’s the familiar Jacobian matrix!
 
∂u ∂u
 ∂x (x0 , y0 ) ∂y (x0 , y0 )
A(x0 ,y0 ) =   = DF(x0 , y0 )
 
 ∂v ∂v 
(x0 , y0 ) (x0 , y0 )
∂x ∂y

Given f = u + iv, we can say that f is continuous if and only if u and v are both continuous.
However, the same is not true of complex differentiation. As a counterexample, consider f (z) = z so that
f (x + iy) = x − iy. Letting u(x, y) = x and v(x, y) = −y, we see that both u and v are differentiable but
f is most certainly not complex differentiable according to 3.6. Naturally, as mathematicians we want to
see what’s up; we seek conditions on u and v that guarantee the differentiability of f !

The thought process is to begin by supposing that f is differentiable and consider the following difference
quotients, incrementing f in the directions of the coordinate axes. This means that we’d be looking at
h = ε and h = εi as ε → 0 in R for the x and y axes respectively.

8 The significance of the negative sign for the second entry of the vector field is explained in section 3.5.
MA231 Vector Analysis 7

The complex differentiability of f tells us that the following two limits are equal (and in fact all limits
as h → 0 in C are too).
f (z + ε) − f (z) ∂u ∂v
f 0 (z) = lim = +i
ε→0 ε ∂x ∂x

f (z + εi) − f (z) ∂v ∂u
f 0 (z) = lim = −i
ε→0 εi ∂y ∂y

Equating real and imaginary parts, we arrive at the Cauchy-Riemann equations: ux = vy and vx = −uy .
This set of equations is a necessary condition for the complex differentiability of f . If we add continuity
of the partial derivatives, we get that it is also a sufficient condition:

Theorem 3.8. A function f : E → C is complex differentiable at z0 = x0 + y0 ∈ E if and only if

F : E → R2 , defined by F (x + iy) = (u(x, y), v(x, y)), is Fréchet differentiable as a map from R2 to R2 and
DF(x0 , y0 ) is of the form
 
a −b
b a

where a, b ∈ R. In particular, f satisfies the Cauchy–Riemann equations and

∂u ∂u
f 0 (z0 ) = a + ib = (x0 , y0 ) + i (x0 , y0 ).
∂x ∂y

Definition 3.9. A map f : E → C is holomorphic on E if it is complex differentiable at all points of E.

Holomorphicity requires that the set over which f is C-differentiable is open.

3.4 Complex Line Integrals

Having studied the differentiation of complex functions, we now rear our heads toward integration.

Definition 3.10. Consider an oriented curve C in the complex plane, parameterised by z : [a, b] → C
defined by z(t) = x(t) + iy(t). Given a continuous function f : C → C, we define the following integrals:

ˆ ˆ b
dz dz dx dy
(i) f dz ..= f (z(t)) dt where = +i
C a dt dt dt dt
ˆ ˆ b
dz dz dx dy
(ii) f dz ..= f (z(t)) dt where = −i
C a dt dt dt dt

ˆ ˆ b     s 2  2
 dz   dz  dx dy
(iii) f |dz| ..= f (z(t))   dt where   = +
C a dt dt dt dt

Example 3.11. Let f (z) = 1/z where z ∈ C \ {0} and C be the unit circle centred at the origin traversed
counterclockwise. We can parameterise C with z : [0, 2π] → C defined by z(t) = eit . By invoking the
parameterisation at (∗), we have that
‰ ‰ 2π
1 (∗) 1
dz = ieit dt = 2πi.
|z|=1 z 0 eit

Upon re-arranging, we arrive at an important identity; a complicated way of getting 1.

‰
1 1
dz = 1
2πi |z|=1 z
 8 MA231 Vector Analysis

3.5 Linking Complex Analysis and Vector Analysis

Recall that to f = u + iv, we associate the vector field f = (u, −v). Now we’ll explain the meaning behind
the negative sign on v in f . For a curve C, parameterised by t 7→ z(t) = x(t) + iy(t), we can expand the
following integral and separate it into its real and imaginary parts.
ˆ ˆ b   dx 
 dy
f dz = u(x(t), y(t)) + iv(x(t), y(t)) dt +i
C a dt dt
ˆ b  ˆ b 
dx dy dy dx
= u(x(t), y(t)) − v(x(t), y(t)) dt + i u(x(t), y(t)) + v(x(t), y(t)) dt
a dt dt a dt dt
ˆ b   ˆ b  
dx dy dy dx
= (u, −v) · , dt + i (u, −v) · ,− dt
a dt dt a dt dt
ˆ ˆ  ⊥
dr dr
= f· dt + i f· dt
C dt C dt
ˆ ˆ
= f · dr + i f · n ds
C C

= (Tangential line integral of f along C) + i (Flux of f across C)

It happens to be the case that the Cauchy-Riemann equations are linked to the notions of divergence
and curl from the first half of the course. This is a consequence of basic re-arrangement!

∂u ∂v ∂ ∂
= ⇐⇒ (u) + (−v) = 0 ⇐⇒ ∇ · f = 0
∂x ∂y ∂x ∂y
∂u ∂v ∂ ∂
=− ⇐⇒ (−v) + (−u) = 0 ⇐⇒ curl f = 0
∂y ∂x ∂x ∂y

Definition 3.12. A vector field f that satisfies curl f = 0 is called irrotational. If f satisfies ∇ · f = 0,
then it is called incompressible.

Proposition 3.13. If f = u+iv satisfies the Cauchy-Riemann equations, then u and v are both harmonic.

Proof.
       
∂ ∂u ∂ ∂u CR∂ ∂v ∂ ∂v
∆u = + = + − =0
∂x ∂x ∂y ∂y ∂x ∂y ∂y ∂x

Cauchy’s Theorem. If f : E → C is holomorphic and C ⊂ E is a closed curve that is the boundary of

a region Ω ⊂ E, then ˆ
f dz = 0.
C

Proof. As in the beginning of this section, we can write

ˆ ˆ ˆ
f dz = f · dr + i f · n ds
C ∂Ω ∂Ω
¨ ¨
= curl f dA + i ∇ · f dA
Ω Ω
| {z } | {z }
Green’s Thm. Div. Thm.

= 0 by the Cauchy-Riemann equations.

It’s important that our region Ω be simply connected i.e. that it doesn’t contain any holes. Otherwise,
Cauchy’s Theorem doesn’t apply. Below is an example on C \ {0}.
MA231 Vector Analysis 9

Example 3.14. Recall the integral identity from example 3.11. This doesn’t contradict Cauchy’s Theorem
because {z : |z| = 1} isn’t the boundary of any region Ω in E ..= C \ {0}. If {z : |z| = 1} were the boundary
of some region, the region would need to contain the origin.

Estimation Lemma. If we have a continuous function f : E → C and a curve C ⊂ E, then

ˆ  ˆ
 
 f dz  6
  |f | |dz| .
C C

The estimation lemma will come in handy in the final exam. Make sure to apply it very carefully, only
when comparing the magnitudes (absolute values) of complex numbers. Make certain that you’re well
acquainted with the standard and reverse triangle inequalities. They’ll be of paramount importance when
trying to bound integrands from above when trying to evaluate integrals.

Evaluation Lemma. If f : E → C is continuous at z0 ∈ E, then

˛
1 f (z)
f (z0 ) = lim dz.
r→0 2πi |z−z |=r (z − z0 )
0

Cauchy’s Theorem concerns itself with a region Ω in which f is holomorphic at every point. Does this
remain the case if f isn’t holomorphic at a single point in Ω? How about several points? To answer these
questions, we must add a few more conditions and this leads us nicely into the discussion of Cauchy’s
integral formulae.

Cauchy’s Integral Formula. Suppose that f : E → C is holomorphic and z0 ∈ E. Choose a ρ > 0 such
that B(z0 , ρ) ⊂ E. Let C ⊂ E be a closed curve such that there exists a region Ω ⊂ E \ {z0 } with ∂Ω =
C ∪ (−∂B(z0 , ρ)). Then for every z0 ∈ B(z0 , ρ), with the contour integral being taken counterclockwise,
we have that ‰
1 f (z)
f (z0 ) = dz.
2πi C (z − z0 )

Cauchy’s integral formula expresses the fact that a holomorphic function on a closed disk is completely
determined by its values on the disk’s boundary.

Taylor’s Theorem. Let f : E → C be holomorphic and suppose that B(z0 , R) ⊂ E. Then there exists a
sequence of complex numbers an such that for all z ∈ B(z0 , R), the power series below converges and is
equal to f (z):
∞
X
f (z) = an (z − z0 )n .
n=0

For each n ∈ N0 , the an ’s are defined by the nth Cauchy Integral Formula (abbreviated as CIn ) given by
ˆ
1 f (z)
an ..= dz.
2πi |z−z0 |=R (z − z0 )n+1
f (n) (z0 )
In particular, f is infinitely often C-differentiable on B(z0 , R) and an = .
n!

3.6 Power Series for Holomorphic Functions

Example 3.15. As an extension of the power series you’ve studied in your first year analysis courses, the
following functions C → C also have their own power series representations:
∞ ∞ ∞
1 X X zn X z 2n+1
= z n where |z| < 1 z
e = sinh(z) =
1 − z n=0 n=0
n! n=0
(2n + 1)!

∞ ∞ ∞
X (−1)n z 2n+1 X (−1)n z 2n X z 2n
sin(z) = cos(z) = cosh(z) =
n=0
(2n + 1)! n=0
(2n)! n=0
(2n)!
 10 MA231 Vector Analysis

Example 3.16. The objective is to find the power series expansion of 1+e2z about z0 = πi/2. Accordingly,
we want to find a sequence of terms an that puts our original function 1 + e2z into the form
∞  n
X πi
an z − .
n=0
2

In order to do this, we need to re-arrange our term involving the exponential to align it about πi/2.
∞
n ∞
n
2z 2(z− πi ) +πi 2(z− πi )
X 2n z − πi2
X 2n z − πi2
1+e =1+e 2 =1−e 2 =1− =−
n=0
n! n=1
n!

3.7 Applications of Taylor’s and Cauchy’s Theorems

Liouville’s Theorem. Suppose that f : C → C is holomorphic and bounded. Then f is constant.
Proof. For any z ∈ C and for any R > 0, the 1st Cauchy Integral Formula gives
ˆ
0 1 f (w)
f (z) = dw.
2πi |w−z|=R (w − z)2

Our second condition is that f is bounded i.e. there exists an M > 0 satisfying |f (z)| 6 M for all z ∈ C.
By the Estimation Lemma, we have that
ˆ
0 M M
|f (z)| 6 |dz| = .
2πR2 |w−z|=R R

Letting R → ∞ tells us that f 0 (z) = 0 for all z ∈ C so f is constant.

Fundamental Theorem of Algebra. Every non-constant polynomial p(z) = a0 +a1 z +· · ·+an z n ∈ C[z]
has at least one root.

Proof. Suppose that p(z) is non-constant and has no roots. A function f : C → C is said to be entire if it
is complex differentiable at every point of C. Then 1/p is an entire function and we need only prove that
it is bounded and thus constant. If an 6= 0 for n > 1, z 6= 0, we can write

p(z)  a0 a1 an−1 
= + + · · · + + an .
zn zn z n−1 z
 
 p(z)  |an |
Note that the first term converges to 0 as z → ∞ and for sufficiently large |z|,  n  >
 and
z 2
n
 
 = |z|
 1  1 2 1 z→∞
 p(z)  |p(z)| |z|n 6 |an | |z|n −→ 0


Finally, 1/p is bounded and, by Liouville’s theorem, constant i.e. 1/p = c ⇐⇒ p = 1/c which contradicts
our original assumption that p is non-constant. Hence, p has at least one root.

Closing Remarks
Vector Analysis isn’t a very proof oriented course. Instead, it would benefit you to have a working knowl-
edge of the theorems and be able to make different types of calculations quickly. The exam itself is quite
time-pressured so although it’s easier said than done, try to keep a calm mind9 and not rush. Haste makes
waste.

Good luck,
Khallil

9 TM04 for all you Pokémon fans.