MA 547 (Complex Analysis)

Summary of Lectures

Motivation for introducing complex numbers: The quadratic equation x2 + 1 = 0 has no

root p
in R. Using Cardano’s method, the roots of the cubic equation x3 = 15x + 4 are given by
√ p √
x = 3 2 + 11 −1 + 3 2 − 11 −1. Since x = 4 is clearly a root of the equation x3 = 15x + 4,
√
there was a need to give a meaning of −1 by enlarging the number system R.

The field of complex numbers: The field C of complex numbers is defined as the field having
the following properties:
(a) C contains (an isomorphic image of) the field R.
(b) C contains a root of the equation x2 + 1 = 0.
(c) If K is a subfield of C containing R and a root of the equation x2 + 1 = 0, then K = C.

Existence of C : The existence of the field C is provided by any of the following constructions
(models) of the field C.
(a) C = {(a, b) : a, b ∈ R}, where addition and multiplication are defined as follows:
(a, b) + (c, d) = (a + c, b + d) and (a, b).(c, d) = (ac − bd, ad + bc) for all (a, b), (c, d) ∈ C.
(b) C = {a+ib : a, b ∈ R}, where +, i are symbols and the addition and multiplication are defined
as follows:
(a + ib) + (c + id) = (a + c) + i(b + d) and (a + ib).(c + id) = (ac − bd) + i(ad + bc) for all
c + id ∈ C.
a + ib,  
a b
(c) C = : a, b ∈ R , with the usual addition and multiplication of matrices.
−b a
(d) C is the quotient ring R[x]hx2 + 1i .

Uniqueness of C : The uniqueness (up to isomorphism) of the field C can also be proved. In
fact, if C and C0 are fields satisfying the properties (a), (b) and (c) given in the definition of the
field of complex numbers, then there exists an isomorphism f : C → C0 which maps reals to reals
and also maps a root of the equation x2 + 1 = 0 to a root of the equation x2 + 1 = 0.

Real and imaginary parts: If i ∈ C is a root of the equation x2 + 1 = 0, then each z ∈ C can
be uniquely expressed as z = x + iy, where x, y ∈ R. We say that x = Re(z) and y = Im(z) are
respectively the real part and the imaginary part of z.
If z, w ∈ C, then Re(z + w) = Re(z) + Re(w) and Im(z + w) = Im(z) + Im(w), although Re(zw)
need not be equal to Re(z)Re(w) and Im(zw) need not be equal to Im(z)Im(w).

Order: A field F is said to be an ordered field if there exists a nonempty subset P of F satisfying
the following properties:
(a) P is closed under addition and multiplication, i.e. x + y ∈ P and xy ∈ P for all x, y ∈ P .
(b) (Law of trichotomy) For each x ∈ F , exactly one of x ∈ P or x = 0 or −x ∈ P holds.
In such case, we define a relation < on F as follows: For all x, y ∈ F , let x < y (equivalently,
y > x) if y − x ∈ P . We also write x ≤ y (equivalently, y ≥ x) if x < y or x = y. Note that ≤ is
a partial order on F .
Although partial orders (like, dictionary order) can be defined on the set C, the field C cannot be
made an ordered field.
C as a vector space: C is a vector space of dimension 1 over the field C. A basis for this vector
space is {1}.
C is a vector space of dimension 2 over the field R. A basis for this vector space is {1, i}.

Conjugate: For each z = x + iy ∈ C, the conjugate of z is defined as z = x − iy.

If z, w ∈ C, then
(a) Re(z) = z+z 2
and Im(z) = z−z
2i
.
(b) z = z.
(c) z + w = z + w and zw = z w.
More generally, if n ∈ N and z1 , . . . , zn ∈ C, then z1 + · · · + zn = z 1 + · · · + z n and
z1 . . . zn = z 1 . . . z n .
(d) ( wz ) = wz , provided that w 6= 0.
If p(z) = a0 + a1 z + · · · + an z n , where a0 , . . . , an ∈ R, and if z0 ∈ C is a root of the equation
p(z) = 0, then z 0 is also a root of the equation p(z) = 0.
The conjugation map z 7→ z from C to C is an automorphism of the field C whose restriction to R
is the identity map on R. Also, if f : C → C is an automorphism of the field C such that f (R) ⊂ R,
then f (z) = z for all z ∈ C or f (z) = z for all z ∈ C. However, there exist automorphisms of the
field C which are neither the identity map nor the conjugation map.

Linear map: A map T : C → C is C-linear iff there exists a ∈ C such that T (z) = az for all
z ∈ C.
A map T : C → C is R-linear iff there exist a, b ∈ C such that T (z) = az + bz for all z ∈ C.
Clearly, every C-linear map T : C → C is R-linear but an R-linear map T : C → C need not be
C-linear. For example, take T (z) = z for all z ∈ C.
However, an R-linear map T : C → C is C-linear iff T (i) = iT (1).
With respect to the ordered
 basis {1, i} of the vector space C (over R), for all a, b, c, d ∈ R, the
a b
2 × 2 matrix represents an R-linear map T : C → C and conversely every R-linear map
c d
 
a b
T : C → C has a matrix representation , where a, b, c, d ∈ R.
c d
An R-linear map T : C → C with above matrix representation is a C-linear map iff a = d and
b = −c.

Complex plane: By identifying z = x + iy ∈ C with the point (x, y) in the xy-plane, we get a
one-one correspondence between C and the points in the xy-plane.
The complex numbers can also be represented as vectors in the xy-plane, where 0 ∈ C is identified
−→
with the zero vector and z = x + iy ∈ C \ {0} is identified with the vector OP , where O = (0, 0)
and P = (x, y).
When the complex numbers are represented as points (or, vectors) in the xy-plane, the xy-plane
is called the complex plane or the z-plane or the Gaussian plane and the diagram is referred to as
the Argand diagram. Also, in this case, the x-axis is called the real axis and the y-axis is called
the imaginary axis.

Modulus:
p For each z = x + iy ∈ C, the modulus or the absolute value of z is defined as
|z| = x + y 2 .
2

If z, w ∈ C, then
(a) |z| ≥ 0 and |z| = 0 ⇔ z = 0.
(b) |z| = |z|.
(c) |Re(z)| ≤ |z| and |Im(z)| ≤ |z|.
(d) |zw| = |z||w|.
  = |z| , provided that w 6= 0.
z
(e) w |w|
(f) (Triangle inequality) |z + w| ≤ |z| + |w| with equality iff either w = 0 or else z = tw for some
t ∈ R with t ≥ 0.
More generally, if n ∈ N and z1 , . . . , zn ∈ C, then |z1 + · · · + zn | ≤ |z1 | + · · · + |zn |.
 
(g)  |z| − |w|  ≤ |z − w| with equality iff either w = 0 or else z = tw for some t ∈ R with t ≥ 0.
(h) (Parallelogram law) |z + w|2 + |z − w|2 = 2(|z|2 + |w|2 ).

Equation of a circle: If z0 = x0 + iy0 ∈ C and r > 0, then the equation |z − z0 | = r represents

the circle in the complex plane with centre (x0 , y0 ) and radius r.

Equation of a straight line: The general equation of a straight line in the complex plane is
az + az + c = 0, where a ∈ C \ {0} and c ∈ R.
−→
Polar representation: Let z = x + iy ∈ C \ {0} be represented by the vector OP in the complex
p −→
plane. If r = x2 + y 2 and θ is an angle which OP makes with the positive real axis, then
x = r cos θ, y = r sin θ and so z = r(cos θ + i sin θ). We say that θ is an argument or amplitude of
z and we define arg(z) = {θ ∈ R : z = |z|(cos θ + i sin θ)}.
The unique θ ∈ arg(z) satisfying θ ∈ (−π, π] is called the principal argument of z and it is denoted
by Arg(z). Thus arg(z) = {Arg(z) + 2nπ : n ∈ Z}.
If z, w ∈ C \ {0}, then arg(zw) = arg(z) + arg(w) and arg wz = arg(z) − arg(w).

However, if z, w ∈ C \ {0}, then Arg(zw) need not be equal to Arg(z) + Arg(w) (for example, take
z = −1, w = i) and Arg wz need not be equal to Arg(z) − Arg(w) (for example, take z = −1 and

w = −i).

Square roots: Let w = a + ib ∈ C. In order to find all z = x + iy ∈ C such that z 2 = w, we

solve the simultaneous equations x2 − y 2 = a and 2xy = b.
√
If b = 0 and a ≥ 0, then z = ± a.
√
z = ±i −a.
If b = 0 and a < 0, thenp
1
√ b
p √ 
If b 6= 0, then z = ± 2
√ 2 2 2
a + a + b + i |b| −a + a + b . 2

Thus for each w ∈ C \ {0}, there exist exactly two distinct z ∈ C such that z 2 = w.
p√ p√ 
Example: The square roots of 1 + i are ± √12 2+1+i 2−1 .

de Moivre’s formula: (cos θ + i sin θ)n = cos nθ + i sin nθ for all θ ∈ R, n ∈ Z.

√ √
Using de Moivre’s formula, we get (−1 + i 3)100 = 2100 cos 200π 200π

99
3
+ i sin 3
= 2 (−1 + i 3).

nth roots: Let w = a + ib ∈ C \ {0}. The method of finding square roots of w, as described
above, is not helpful in finding nth roots of w for n > 2. However, using de Moivre’s formula, we
1
find that there exist exactly n distinct nth roots of w and these are |w| n cos 2kπ+α 2kπ+α


n
+ i sin n
,
where α = Arg(w) and k ∈ {0, 1, . . . , n − 1}. All these nth roots of w are the vertices of a regular
1
n-gon which is inscribed in the circle |z| = |w| n in the complex plane.
1
Example: The seventh roots of 1 + i are 2 14 cos (8n+1)π + i sin (8n+1)π


28 28
for n = 0, 1, . . . , 6.

C as a metric space: If du (z, w) = |z − w| for all z, w ∈ C, then du is a metric on C, which is

called the usual metric on C. Thus (C, du ) is a metric space.
If E(6= ∅) ⊂ C, then E is also a metric space with respect to the (restriction to E × E of the)
metric du .
In the following, we give definitions and results in the metric space (C, du ). However, they can
also be formulated in the metric space (E, du ).

Open and closed balls: Let z0 ∈ C and r > 0. Then Br (z0 ) = {z ∈ C : |z − z0 | < r} is the
open ball (or, open disk) in C with centre z0 and radius r. Also, Br [z0 ] = {z ∈ C : |z − z0 | ≤ r}
is the closed ball (or, closed disk) in C with centre z0 and radius r.
We denote the open unit disk B1 (0) = {z ∈ C : |z| < 1} by D.

Open and closed sets: A subset G of C is called an open set in C if for each z ∈ G, there exists
r > 0 such that Br (z) ⊂ G.
A subset F of C is called a closed set in C if C \ F is an open set in C.
Let E(6= ∅) ⊂ C. Then S ⊂ E is open (respectively, closed) in E iff S = G ∩ E for some open
(respectively, closed) set G in C.

Examples:
(a) Every open ball in C is an open set in C and every closed ball in C is a closed set in C.
(b) {z ∈ C : 1 < Re(z) < 2} is an open set in C and {z ∈ C : Im(z) ≥ 0} is a closed set in C.
(c) {z ∈ C : 1 < |z| ≤ 2} is neither open nor closed in C.
(d) E = {z ∈ C : |z| ≤ 1} is both open and closed in E ∪ {z ∈ C : |z − 3| < 1}.

Equivalent metrics on C : Let d1 (z, w) = |Re(z) − Re(w)| + |Im(z) − Im(w)| and

d∞ (z, w) = max{|Re(z) − Re(w)|, |Im(z) − Im(w)|} for all z, w ∈ C. Then d1 and d∞ are metrics
on C and the three metrics du , d1 , d∞ on C are equivalent, i.e. each of the three metric spaces
(C, du ), (C, d1 ) and (C, d∞ ) has the same class of open sets.

Bounded set and bounded function: A subset E of C is said to be bounded if there exists
r > 0 such that E ⊂ Br [0], i.e. |z| ≤ r for all z ∈ E. A set in C which is not bounded is called
unbounded.
For example, {z ∈ C : (Re(z))2 + |Im(z)| < 5} is a bounded set in C and
{z ∈ C : |Re(z)| + Im(z) < 1} is an unbounded set in C
If Ω is a nonempty subset of C, then a function f : Ω → C is said to be bounded if f (Ω) is a
bounded subset of C, i.e. if there exists r > 0 such that |f (z)| ≤ r for all z ∈ Ω.

Convergence of sequence: A sequence in C is a function f : N → C, which we usually denote

by (zn ) by writing zn = f (n) for each n ∈ N.
A sequence (zn ) in C is called convergent if there exists z ∈ C such that for every ε > 0, there
exists n0 ∈ N satisfying |zn − z| < ε for all n ≥ n0 .
Such a z (if it exists) is unique and is called the limit of (zn ), which is denoted by zn → z or
lim zn = z.
n→∞
We note that every convergent sequence in C is bounded.
 
(1+i)n
Example: The sequence n!
is convergent with limit 0.



Proposition: A sequence (zn ) in C is convergent iff both the sequences Re(zn ) and Im(zn )
in R are convergent.
Also, in such case lim zn = lim Re(zn ) + i lim Im(zn ).
n→∞ n→∞ n→∞

Cauchy sequence: A sequence (zn ) in C is called a Cauchy sequence if for every ε > 0, there
exists n0 ∈ N such that |zm − zn | < ε for all m, n ≥ n0 .
Proposition: The metric space (C, du ) is complete, i.e. every Cauchy sequence in C is convergent.

Bolzano-Weierstrass theorem: Every bounded sequence in C has a convergent subsequence.

For example, the sequence (cos n + i sin n) has a convergent subsequence although the sequence
itself is not convergent.

n
P
Series in C : A series in C is a pair (zn ), (sn ) , where (zn ) is a sequence in C and sn = zk
k=1
∞
P
for all n ∈ N. A series in C is usually denoted by zn .
n=1
∞
P
The series zn is said to be convergent if the sequence (sn ) is convergent. In such case the sum
n=1
∞
P ∞
P
of the series zn is defined as zn = lim sn .
n=1 n=1 n→∞
∞
P ∞
P
The series zn is said to be absolutely convergent if the series |zn | is convergent.
n=1 n=1
Every absolutely convergent series is convergent but the converse is not true.

Examples:
1
(a) The geometric series 1 + z + z 2 + · · · converges (absolutely) with sum 1−z for z ∈ D and
diverges (i.e. does not converge) for z ∈ C \ D.
∞ ∞
1
{z ∈ C : z n 6= −1}, then the series
T P
(b) If z ∈ 1+z n
is convergent iff |z| > 1.
n=1 n=1

Limit point: Let E ⊂ C. We say that z ∈ C is a limit point of E in C if for every r > 0,
Br (z) ∩ (E \ {z}) 6= ∅.
Using Bolzano-Weierstrass theorem for sequences, it follows that every infinite bounded subset of
C has a limit point in C.

Closure: Let E ⊂ C. The closure of E in C is defined as

T
E = {F : F is a closed set in C, E ⊂ F }.
Note that E is the smallest closed set in C containing E and also E = E ∪ E 0 , where E 0 is the set
of all limit points of E in C.
It can be shown that E = {z ∈ C : Br (z) ∩ E 6= ∅ for all r > 0}
= {z ∈ C : there exists a sequence (zn ) in E such that zn → z}.
If E = C, then E is said to be dense in C. For example, {x + iy : x, y ∈ Q} is dense in C.

Sequential criterion of closed set: A set E ⊂ C is closed in C iff for every z ∈ C and for every
sequence (zn ) in E converging to z, z ∈ E.
Using this criterion, it can be easily shown that {z ∈ C : Re(z) ≥ 0} is a closed set in C and
{z ∈ C : Im(z) < 0} is not closed in C.

Interior: Let E ⊂ C. The interior of E in C is defined as

E 0 = {G : G is an open set in C, G ⊂ E}.
S

Note that E 0 is the largest open set in C contained in E. It can be seen that
E 0 = {z ∈ E : Br (z) ⊂ E for some r > 0}.

Boundary: The boundary of a set E ⊂ C is defined as ∂E = E ∩ C \ E = E \ E 0 .

Thus ∂E = {z ∈ C : for every r > 0, Br (z) ∩ E 6= ∅ and Br (z) ∩ (C \ E) 6= ∅}.

Examples: Let z0 ∈ C and r > 0. Then Br (z0 ) = Br [z0 ], Br [z0 ]0 = Br (z0 ) and
∂Br (z0 ) = ∂Br [z0 ] = {z ∈ C : |z − z0 | = r}.
In particular, D = {z ∈ C : |z| ≤ 1} and ∂D = ∂D = {z ∈ C : |z| = 1}.
Function on C : Let f : Ω ⊂ C → C. Considering Ω0 = {(x, y) ∈ R2 : x+iy ∈ Ω} ⊂ R2 , we define
u : Ω0 → R and v : Ω0 → R respectively by u(x, y) = Re(f (x + iy)) and v(x, y) = Im(f (x + iy))
for all (x, y) ∈ Ω0 . We say that u and v are the real part and the imaginary part of f . Note that
f (z) = u(x, y) + iv(x, y) for all z = x + iy ∈ Ω.

Continuity: A function f : Ω ⊂ C → C is said to be continuous at z0 ∈ Ω if for every ε > 0,

there exists δ > 0 such that |f (z) − f (z0 )| < ε for all z ∈ Ω satisfying |z − z0 | < δ.
We say that f : Ω → C is continuous if f is continuous at each point of Ω.

Examples: Each of the the functions z 7→ Re(z), z 7→ Im(z), z 7→ z and z 7→ |z| from C to C is
continuous.

Proposition: Let u and v be respectively the real part and the imaginary part of f : Ω ⊂ C → C.
Then f is continuous at z0 = x0 + iy0 ∈ Ω iff both u and v are continuous at (x0 , y0 ).

Sequential criterion of continuity: A function f : Ω ⊂ C → C is continuous at z0 ∈ Ω iff for

every sequence (zn ) in Ω converging to z0 , f (zn ) → f (z0 ).
Re(z)
(
Example: The function f : C → C, defined by f (z) = if z 6= 0,
z
1 if z = 0,
is not continuous at 0.

Combinations of continuous functions: Let f : Ω ⊂ C → C and g : Ω → C be continuous at

z0 ∈ Ω. Then the functions f + g : Ω → C and f g : Ω → C are continuous at z0 . Moreover, if
g(z0 ) 6= 0, then there exists δ > 0 such that the function fg : Ω ∩ Bδ (z0 ) → C is continuous at z0 .
It follows that every polynomial function on C is continuous and every rational function
p
q
: {z ∈ C : q(z) 6= 0} → C is continuous.
Again, let f : Ω1 ⊂ C → C and g : Ω2 ⊂ C → C such that f (Ω1 ) ⊂ Ω2 . If f is continuous at
z0 ∈ Ω1 and g is continuous at f (z0 ), then the composite function g ◦ f : Ω1 → C is continuous at
z0 .

Examples:
(a) If f : Ω ⊂ C → C is continuous at z0 ∈ Ω, then the functions z 7→ f (z) and z 7→ |f (z)| from
Ω to C are continuous
 z+1 at z0 .
z 2 +4
if z ∈ C \ {2i, −2i},
(b) Let f (z) =
1 if z ∈ {2i, −2i}.
Then f : C → C is continuous on C \ {2i, −2i} and discontinuous at 2i and −2i.

Proposition: For a function f : Ω ⊂ C → C, the following statements are equivalent.

(a) f is continuous.
(b) For every open set G in C, f −1 (G) is open in Ω.
(c) For every closed set F in C, f −1 (F ) is closed in Ω.

Examples: {z ∈ C : 1 < |z + 2| < 3} is an open set in C and {z ∈ C : Re(z 2 + 3iz) = 6} is a

closed set in C.

Limit: Let z0 ∈ C be a limit point of Ω ⊂ C and let f : Ω → C. We say that ` ∈ C is a limit

of f at z0 if for every ε > 0, there exists δ > 0 such that |f (z) − `| < ε for all z ∈ Ω satisfying
0 < |z − z0 | < δ.
Such an `, when it exists, is unique and we write lim f (z) = `.
z→z0
Remark: A function f : Ω ⊂ C → C is continuous at z0 ∈ Ω iff either z0 is an isolated point of
Ω (i.e. there exists δ > 0 such that Bδ (z0 ) ∩ Ω = {z0 }) or else lim f (z) = f (z0 ).
z→z0

Proposition: Let z0 ∈ C be a limit point of Ω ⊂ C and let u and v be respectively the real part
and the imaginary part of f : Ω → C. Then lim f (z) exists (in C) iff both lim u(x, y) and
z→z0 (x,y)→(x0 ,y0 )
lim v(x, y) exist (in R).
(x,y)→(x0 ,y0 )
Further, in such case lim f (z) = lim u(x, y) + i lim v(x, y).
z→z0 (x,y)→(x0 ,y0 ) (x,y)→(x0 ,y0 )

Sequential criterion of limit: Let z0 ∈ C be a limit point of Ω ⊂ C and let f : Ω → C. Then

lim f (z) = ` ∈ C iff for every sequence (zn ) in Ω \ {z0 } converging to z0 , f (zn ) → `.
z→z0

zRe(z) z
Examples: lim = 0 but lim does not exist (in C).
z→0 |z| z→0 z

Compactness: A subset E of C is called compact if for every class {Gα : α ∈ Λ} of open sets in
S
C satisfying E ⊂ Gα , there exist α1 , . . . , αn ∈ Λ of such that E ⊂ Gα1 ∪ · · · ∪ Gαn .
α∈Λ
It can be shown that E ⊂ C is compact iff every sequence in E has a subsequence converging to
a point of E.

Example: C and D are not compact.

Heine-Borel theorem: A subset E of C is compact if E is closed and bounded in C.

Also, every compact set in C is closed and bounded in C.

Example: D and ∂D are compact.

Continuity and compactness: Let Ω be a nonempty compact set in C. If f : Ω → C is

continuous, then f (Ω) is compact (and hence f (Ω) is closed and bounded in C) and there exist
z1 , z2 ∈ Ω such that |f (z1 )| = sup{|f (z)| : z ∈ Ω} and |f (z2 )| = inf{|f (z)| : z ∈ Ω}.

Example: There is no continuous function from D onto C or from ∂D onto D.

Uniform continuity: A function f : Ω ⊂ C → C is said to be uniformly continuous if for every

ε > 0, there exists δ > 0 such that |f (z) − f (w)| < ε for all z, w ∈ Ω satisfying |z − w| < δ.
If f : Ω → C is uniformly continuous, then f is continuous. However, the converse is not true in
general, as an example below shows.

Sequential criterion of uniform continuity: A function f : Ω ⊂ C → C is uniformly contin-

uous iff for every sequences (zn ) and (wn ) in Ω with |zn − wn | → 0, |f (zn ) − f (wn )| → 0.
1 1
Example: If f (z) = z−2i and g(z) = 1−z for all z ∈ D, then f : D → C is uniformly continuous
and g : D → C is continuous but not uniformly continuous.

Proposition: If Ω is a nonempty compact subset of C and f : Ω → C is continuous, then

f : Ω → C is uniformly continuous.

Connectedness: A subset E of C is said to be disconnected if there exist nonempty disjoint

open sets G and H in E such that E = G ∪ H.
E ⊂ C is called connected if it is not disconnected.
Note that a nonempty subset E of C is connected iff ∅ and E are the only subsets of E which are
both open and closed in E.
If E ⊂ C, then a maximal connected subset of E is called a component of E.
Proposition: Let Ω be a nonempty connected set in C. If f : Ω → C is continuous, then f (Ω) is
connected.

Example: Since D is connected (as seen below) and Z is disconnected, there is no continuous
function from D onto Z.

Path: Let E be a nonempty subset of C. A path in E is a continuous map γ : [a, b] → E for

some a, b ∈ R with a < b.
If γ : [a, b] → E is a path in E, then the trace (or, range) of γ is the set {γ} = {γ(t) : t ∈ [a, b]},
which is a compact and connected subset of E.
If z, w ∈ E and γ : [a, b] → E is a path such that γ(a) = z and γ(b) = w, then we say that γ joins
z to w, or γ is a path in E from z to w, or γ is a path in E with initial point z and final point w.
If γ1 : [a, b] → E and γ2 : [a, b] → E are paths in E such that γ1 (b) = γ2 (a), then the sum (or,
join/union) of  γ1 and γ2 is the path γ1 + γ2 : [a, b] → E, defined by
γ1 (2t − a) if t ∈ a, a+b
2 
,
(γ1 + γ2 )(t) =
γ2 (2t − b) if t ∈ a+b
2
, b .
If z, w ∈ C and γ(t) = (1 − t)z + tw for all t ∈ [0, 1], then γ is called the line segment in C
joining z to w. In this case we denote {γ} by [z, w]. If γ1 , . . . , γn : [0, 1] → C are finitely many
line segments in C such that γj (1) = γj+1 (0) for j = 1, . . . , n − 1, then γ1 + · · · + γn is called a
polygonal path (or, a polygon) in C.

Path connectedness: A subset E of C is said to be path connected if for each z, w ∈ E, there

exists a path in E from z to w.
If Ω is a nonempty path connected set in C and if f : Ω → C is continuous, then f (Ω) is path
connected.

Proposition: Every path connected set in C is connected.

However, a connected set in C need not be path connected. For example,
x + i sin x1 : 0 < x ≤ 1 ∪ {ix : −1 ≤ x ≤ 1} is connected but not path connected.


Examples of connected and disconnected sets:

(a) Every convex set in C is (path) connected. In particular, C and {z ∈ C : Re(z) > 1} are
(path) connected. Also, every open ball and every closed ball in C are (path) connected.
Consequently D and D are (path) connected.
(b) C \ {0} and ∂D are (path) connected.
(c) D ∪ {z ∈ C : |z + 2| ≤ 1} is (path) connected.
(d) D ∪ {z ∈ C : Im(z) = 1} is (path) connected.
(e) D ∪ {z ∈ C : |z − 3| ≤ 1} is disconnected.

Proposition: Let E be an open and connected set in C. Then E is path connected.

In fact, if z, w ∈ E, then there exists a polygonal path in E joining z to w. Moreover, each of the
line segments forming the polygonal path can be taken to be parallel to either the real axis or the
imaginary axis.

Domain: A (nonempty) open and connected set in C is called a domain in C.

Differentiability: A function f : Ω ⊂ C → C is said to be differentiable at z0 ∈ Ω0 if

lim f (z)−f
z−z0
(z0 )
(or, equivalently, lim f (z0 +h)−f
h
(z0 )
) exists (in C).
z→z0 h→0
If f is differentiable at z0 , then the derivative of f at z0 is defined as
f 0 (z0 ) = lim f (z)−f
z−z0
(z0 )
= lim f (z0 +h)−f
h
(z0 )

.
z→z0 h→0
If Ω is open in C, then f : Ω → C is called differentiable if f is differentiable at each point of Ω.

Example:
(a) If f (z) = z 2 for all z ∈ C, then f : C → C is differentiable and f 0 (z) = 2z for all z ∈ C.
(b) If f (z) = z for all z ∈ C, then f : C → C is not differentiable at any point of C.
(c) If f (z) = |z|2 for all z ∈ C, then f : C → C is differentiable only at 0 and f 0 (0) = 0.

Proposition: If f : Ω ⊂ C → C is differentiable at z0 ∈ Ω0 , then f is continuous at z0 .

The converse of this proposition is not true, in general. For example, if f (z) = |z| for all z ∈ C,
then f : C → C is continuous but not differentiable at 0.

Combinations of differentiable functions: Let f : Ω ⊂ C → C and g : Ω → C be differentiable

at z0 ∈ Ω0 . Then the functions f + g : Ω → C and f g : Ω → C are differentiable at z0 . Moreover,
if g(z0 ) 6= 0, then there exists δ > 0 such that Bδ (z0 ) ⊂ Ω and the function fg : Bδ (z0 ) → C
differentiable at z0 .
It follows that every polynomial function on C is differentiable and every rational function
p
q
: {z ∈ C : q(z) 6= 0} → C is differentiable.
 z+3
z 2 −3z+2
if z ∈ C \ {1, 2},
Example: Let f (z) =
1 if z ∈ {1, 2}.
Then f : C → C is differentiable on C \ {1, 2} and not differentiable at 1 and 2.

Chain rule: Let f : Ω1 ⊂ C → C and g : Ω2 ⊂ C → C be such that f (Ω1 ) ⊂ Ω2 . If f is

differentiable at z0 ∈ Ω01 and g is differentiable at f (z0 ) ∈ Ω02 , then g ◦ f : Ω1 → C is differentiable
at z0 and (g ◦ f )0 (z0 ) = g 0 (f (z0 ))f 0 (z0 ).

Remark: The following criterion of differentiability is useful in proving the chain rule.
A function f : Ω ⊂ C → C is differentiable at z0 ∈ Ω0 iff there exist a function ϕ : Ω → C such
that ϕ is continuous at z0 and f (z) − f (z0 ) = (z − z0 )ϕ(z) for all z ∈ Ω.
Also, in such case f 0 (z0 ) = ϕ(z0 ).

Complex differentiability and real differentiability: Let u and v be respectively the real
part and the imaginary part of f : Ω ⊂ C → C and let z0 = x0 + iy0 ∈ Ω0 .
(a) If f is differentiable at z0 , then all the partial derivatives ux , uy , vx and vy exist at (x0 , y0 )
and ux (x0 , y0 ) = vy (x0 , y0 ) and uy (x0 , y0 ) = −vx (x0 , y0 ).
Also, in this case f 0 (z0 ) = ux (x0 , y0 ) + ivx (x0 , y0 ).
(b) If all the partial derivatives ux , uy , vx and vy exist at (x0 , y0 ) and ux (x0 , y0 ) = vy (x0 , y0 ) and
uy (x0 , y0 ) = −vx (x0 , y0 ), then fneed not be differentiable at z0 . For example, for the function
z2
if z 6= 0,
f : C → C, defined by f (z) = z
0 if z = 0,
ux (0, 0) = 1 = vy (0, 0) and uy (0, 0) = 0 = −vx (0, 0) but f is not differentiable at 0.
(c) f is differentiable at z0 iff u and v are differentiable at (x0 , y0 ) and ux (x0 , y0 ) = vy (x0 , y0 ) and
uy (x0 , y0 ) = −vx (x0 , y0 ).
In particular, therefore, if there exists δ > 0 such that ux , uy , vx , vy exist on Bδ ((x0 , y0 )),
are continuous at (x0 , y0 ) and ux (x0 , y0 ) = vy (x0 , y0 ) and uy (x0 , y0 ) = −vx (x0 , y0 ), then f is
differentiable at z0 .
(d) Let Ω0 = {(x, y) ∈ R2 : x + iy ∈ Ω} and let f˜(x, y) = (u(x, y), v(x, y)) for all (x, y) ∈ Ω0 . Then
f is differentiable (as a function of a complex variable) at z0 iff f˜ : Ω0 → R2 is differentiable
(as a function of two real variables) at (x0 , y0 ) and the R-linear map f˜0 (x0 , y0 ) : R2 → R2 is a
C-linear map (i.e. a linear map over the field R2 ).
 Put in an equivalent way, f is differentiable (as a function of a complex variable) at z0 iff f˜
is differentiable (as a function of two real variables) at (x0 , y0 ) and ux (x0 , y0 ) = vy (x0 , y0 ) and
uy (x0 , y0 ) = −vx (x0 , y0 ).
The equations ux = vy and uy = −vx are known as the Cauchy-Riemann equations.

Example:
(a) If f (z) = y + ix for all z = x + iy ∈ C, then f : C → C is not differentiable at any point of C.
(b) If f (z) = ex (cos y + i sin y) for all z = x + iy ∈ C, then f : C → C is differentiable and
f 0 (z) = f (z) for all z ∈ C.
(c) If f (z) = zRe(z) for all z ∈ C, then f : C → C is differentiable only at 0 and f 0 (0) = 0.

Looman-Menchoff theorem: Let u and v be the real part and the imaginary part of a continu-
ous function f : Ω → C, where Ω is an open set in C. If ux (x, y) = vy (x, y) and uy (x, y) = −vx (x, y)
for all (x, y) ∈ R2 with x + iy ∈ Ω, then f is differentiable (on Ω).
Montel stated (and Tolstov proved) that the continuity of f in the above result can be replaced
by boundedness of f .

Polar forms of Cauchy-Riemann equations: Let Ω be an open set in C and let f : Ω → C be

such that f (z) = u(x, y)+iv(x, y) = u(r, θ)+iv(r, θ) for all z = x+iy = r(cos θ +i sin θ) ∈ Ω\{0}.
The Cauchy-Riemann equations ux (x, y) = vy (x, y), uy (x, y) = −vx (x, y), expressed in polar forms,
become ur (r, θ) = 1r vθ (r, θ), 1r uθ (r, θ) = −vr (r, θ) for all z = x + iy = r(cos θ + i sin θ) ∈ Ω \ {0}.

Holomorphic function: If Ω is an open set in C, then a function f : Ω → C is said to be

holomorphic (or, analytic) if f is differentiable (on Ω).
If Ω is an open set in C, then the set of all holomorphic functions from Ω to C is denoted by H(Ω).
A holomorphic function f : C → C is also called an entire function.
We say that a function f : Ω ⊂ C → C is holomorphic at z0 ∈ Ω if there exists δ > 0 such that
Bδ (z0 ) ⊂ Ω and f is differentiable on Bδ (z0 ).

Example: If f (z) = e−x (cos y − i sin y) and g(z) = x3 + iy 3 for all z = x + iy ∈ C, then f : C → C
is holomorphic (hence entire) and g : C → C is not holomorphic at any point of C.

Proposition: Let Ω be a domain in C. If f : Ω → C is holomorphic and f 0 (z) = 0 for all z ∈ Ω,

then f is a constant function.

Remark: The connectedness of Ω in  the above proposition is, in general, necessary. For example,
1 if |z| < 1,
f : C \ ∂D → C, defined by f (z) =
2 if |z| > 1,
0
is not a constant function although f (z) = 0 for all z ∈ C \ ∂D.
In the above proposition, if Ω is not assumed to be connected, then we can conclude that f is
constant on each component of Ω.

Proposition: Let f : Ω → C be holomorphic, where Ω is a domain in C. If either of the functions

Re(f ) or Im(f ) or |f | on Ω is a constant, then f is a constant function.

Failure of mean value theorem: A direct generalization of the mean value theorem (similarly,
Rolle’s theorem) in R to C is not valid, in general.
For example, if f (z) = z 3 for all z ∈ C, then f : C → C is holomorphic, but there is no c ∈ [1, i]
(the line segment from 1 to i) such that f (i) − f (1) = (i − 1)f 0 (c).

Harmonic function: If Ω is an open set in R2 , then ϕ : Ω → R is called a harmonic function if

all the second order partial derivatives ϕxx , ϕyy , ϕxy , ϕyx of ϕ are continuous on Ω and
ϕxx (x, y) + ϕyy (x, y) = 0 ( Laplace’s equation) for all (x, y) ∈ Ω.
For example, if ϕ(x, y) = x2 − y 2 for all (x, y) ∈ R2 , then ϕ : R2 → R is a harmonic function.

Proposition: Let Ω be an open set in C. If f : Ω → C is holomorphic, then Re(f ) : Ω0 → R and

Im(f ) : Ω0 → R are harmonic functions, where Ω0 = {(x, y) ∈ R2 : x + iy ∈ Ω}.

Example: There is no holomorphic function f : C → C such that Re(f (x + iy)) = x2 + y 2 for all
x + iy ∈ C.

Harmonic conjugate: Let Ω be a domain in R2 and let Ω0 = (x + iy : (x, y) ∈ Ω}. If u : Ω → R

and v : Ω → R are harmonic functions such that u + iv : Ω0 → C is holomorphic, then v is said to
be a harmonic conjugate of u.
If v1 and v2 are harmonic conjugates of a harmonic function u : Ω → R, where Ω is a domain in
R2 , then v1 − v2 : Ω → R is a constant function.

Existence of harmonic conjugate:

(a) If Ω is a domain in R2 and u : Ω → R is harmonic, then u need not have a harmonic conjugate
v : Ω → R. For example, take u(x, y) = 21 log(x2 + y 2 ) for all (x, y) ∈ R2 \ {(0, 0)}.
x 2 2
However, if u(x, y) = x2 +y 2 for all (x, y) ∈ R \ {(0, 0)}, then u : R \ {(0, 0)} → R has a
y
harmonic conjugate v : R2 \ {(0, 0)} → R, where v(x, y) = − x2 +y 2
2 for all (x, y) ∈ R \ {(0, 0)}.

(b) If Ω is either R2 or a disk in R2 or a rectangle in R2 with sides parallel to the axes and if
u : Ω → R is harmonic, then u has a harmonic conjugate v : Ω → R.
If u(x, y) = 2x(1 − y) for all (x, y) ∈ R2 , then all the harmonic conjugates v : R2 → R of the
harmonic function u : R2 → R are given by v(x, y) = x2 − y 2 + 2y + c for all (x, y) ∈ R2 , where
c ∈ R. The corresponding holomorphic functions f : C → C are given by f (z) = iz 2 + 2z + ic
for all z ∈ C, where c ∈ R.
Similar method can be used to find the real part of a holomorphic function when its imaginary
part is given.
∞
an (z − z0 )n , where an ∈ C for all
P
Power series: A power series in C is a series of the form
n=0
n ∈ N ∪ {0}, z0 ∈ C and z is a complex variable. Such a series is said to be centred at z0 .
∞
an (z − z0 )n converges for z = z0 .
P
Every power series
n=0
∞
an z n (i.e.
P
By replacing z − z0 by z, it is sufficient to consider only power series of the form
n=0
series centred at 0).

Example:
∞
z n converges absolutely for all z ∈ D and diverges for all z ∈ C \ D.
P
(a) The power series
n=0
∞
zn
P
(b) The power series n!
converges absolutely for all z ∈ C.
n=0
∞
n!z n converges only for z = 0.
P
(c) The power series
n=0
∞
zn
P
(d) The power series n2
converges absolutely for all z ∈ D and diverges for all z ∈ C \ D.
n=1
∞
zn
P
(e) The power series n
converges absolutely for all z ∈ D, converges (but not absolutely) for
n=1
all z ∈ ∂D \ {1} and diverges for all z ∈ (C \ D) ∪ {1}.
 ∞
an z n converges for z = z1 ∈ C\{0}, then it converges absolutely
P
Proposition: If a power series
n=0
∞
an z n diverges for z = z2 ∈ C, then it diverges
P
for all z ∈ C with |z| < |z1 |. If the power series
n=0
for all z ∈ C with |z| > |z2 |.
∞
an z n , there exists a unique R with
P
Radius of convergence: For every power series
n=0
∞
n
P
0 ≤ R ≤ ∞ such that the series an z converges absolutely if |z| < R and diverges if |z| > R.
n=0
∞
an z n .
P
We define R to be the radius of convergence of the power series
n=0
∞ ∞ ∞ ∞ ∞
zn zn zn
zn, n!z n ,
P P P P P
Example: The radii of convergence of the power series n!
, n2
and n
n=0 n=0 n=0 n=1 n=1
are 1, ∞, 0, 1 and 1 respectively.
∞
an z n , then the series can converge
P
Remark: If R is the radius of convergence of a power series
n=0
for all or some or none of z satisfying |z| = R, as the examples given above show.
∞
an z n is given
P
Cauchy-Hadamard formula: The radius of convergence R of a power series
n=0
1 1
by R
= lim sup |an | n .
n→∞