Chapter 1

Holomorphic functions

In this chapter, we will study the notion of complex differentiability for complex-valued functions
of the complex variable. Functions that are complex-differentiable will be called holomorphic.
Although the definition of complex differentiability resembles incredibly that of real differentia-
bility,it turns out that the former is much stronger than the latter. For example, holomorphic
functions are infinitely many times differentiable (the existence of one complex derivative implies
that of infinitely many ones) and they admit a power series expansion.
We begin by recalling the usual definition of continuity.
Let Ω ⊂ ƒ and f : Ω → ƒ. We say that f is continuous at z0 ∈ ƒ if

∀ ϵ > 0 ∃ δ > 0 : | f (z) − f (z0 )| < ϵ ∀z ∈ Ω such that |z − z0 | < δ.

Equivalently, f is continuous at z0 ∈ Ω if for every sequence {zn }n ⊂ Ω converging to z, we have

that f (zn ) → f (z0 ) as n → ∞.
Using the decomposition z = x + iy into its real and imaginary part, the function f of the complex
variable can be also seen as a function of the two variables (x, y). Since the notion of convergence
in ƒ and ’2 are the same, f is continuous with respect to the variable z if and only if it is continuous
with respect to the variables (x, y). We also observe that, if f is continuous, then the real-valued
function z 7→ | f (z)| is continuous.
We say that the function f has a maximum in Ω if there exists z0 ∈ Ω such that

| f (z)| ≤ | f (z0 )|, ∀z ∈ Ω.

We say that f has a minimum in Ω if the same property as above, but with the reversed inequality,
holds true. Analogous to the case of functions of the real variable, we have the following results.

Theorem 1.1. A continuous function on a compact set Ω is bounded and attains its maximum and
minimum in Ω.

Proof. Left to the reader. Hint: first prove that f (Ω) is compact. □

Theorem 1.2. A continuous function on a compact set is uniformly continuous.

1
2 CHAPTER 1. HOLOMORPHIC FUNCTIONS

1.1 Holomorphic functions

Let Ω ⊂ ƒ be an open set and f : Ω → ƒ. We say that f is complex differentiable at z0 ∈ Ω if

the following limit exists
f (z0 + h) − f (z0 )
lim
h→0 h
The limit is denoted by f ′ (z0 ) or ddzf (z0 ) and it is a complex number. We point out that, in the above
limit, h ∈ ƒ and it approaches 0 from all directions.
We say that f is holomorphic on Ω if and only if it is holomorphic at every point of Ω and
f ′ : Ω → ƒ. In other words, f is holomorphic if it is “C 1 in the complex sense”. We will see
later in this notes that the assumption on the continuity of f ′ is actually redundant and that it will
follow from the existence of the complex derivative itself.
If F ⊂ ƒ is a closed set, we say that f is holomorphic on F if it is holomorphic on an open set
containing F. If f is holomorphic on all ƒ, we say that f is entire.
As for functions of the real variable, we observe that complex differentiability implies continuity.
In fact, if f is holomorphic at z0 then

f (z0 + h) − f (z0 ) − ah = hψ(h)

where a = f ′ (z0 ) and ψ is a function such that ψ(h) → 0 as h → 0.

Using the above definition, it is easy to prove that elementary operations with holomorphic func-
tions preserve holomorphicity.

Proposition 1.1. Let f, g : Ω → ƒ be holomorphic. Then:

a. f + g is holomorphic and ( f + g)′ = f ′ + g′ ;

b. f g is holomorphic and ( f g)′ = f ′ g + f g′ ;

c. if g(z0 ) , 0, f /g is holomorphic at z0 and

f ′ g − f g′
( f /g)′ (z0 ) = (z0 )
g2

Proof. Left as exercise. □

Proposition 1.2. Let Ω, Ω̃ be open sets and g : Ω → Ω̃, f : Ω̃ → ƒ be holomorphic functions.

Then f ◦ g : Ω → ƒ is holomorphic and

( f ◦ g)′ (z) = f ′ (g(z))g′ (z), ∀z ∈ Ω.

Example 1. The function f (z) = z is holomorphic on any open set in ƒ (hence entire) and f ′ (z) =
1. Furthermore, any polynomial function

p(z) = a0 + a1 z + · · · + an zn

is entire and
p′ (z) = a1 + · · · + nan zn−1 .
1.2. THE CAUCHY-RIEMANN EQUATIONS 3

Example 2. The function f (z) = 1/z is homolorphic on any open set of ƒ not containing the origin
and f ′ (z) = −1/z2 .

Example 3. The function f (z) = z̄ is not holomorphic. In fact,

f (z0 + h) − f (z0 ) h̄
= ,
h h
in particular the above quantity equals 1 if h ∈ ’ and −1 if h ∈ i’ (i.e. if h is purely imaginary).
Hence the limit when h → 0 does not exist.

A function f : Ω → Ω̃ between two open sets Ω, Ω̃ ⊂ ƒ is said to be a holomorphic isomorphism

if it is homolorphic, bijective and if its inverse f −1 : Ω̃ → Ω is holomorphic. If Ω = Ω̃, f is said
to be a automorphism.

1.2 The Cauchy-Riemann equations

Example 3 is the first clear example of the big difference between real and complex differentiation.
The function f (z) = z̄ corresponds, in fact, to the map F(x, y) = (x, −y) which is infinitely many
times differentiable as a function from ’2 to ’2 . The existence of the (real) derivative of F does
not imply the existence of the (complex) derivative of f .
Let us look at the matter more closely. Given a function f : Ω → ƒ, we write it in terms of its real
and imaginary part
f (z) = u(z) + iv(z), u, v : Ω → ’.
We can associate to f the mapping F(x, y) = (u(x, y), v(x, y)), defined on an open subset of ’2 with
values in ’2 . We want to understand the necessary and sufficient conditions on u, v (as functions
on ’2 ) for the function f to be holomorphic.
Let us first assume f to be holomorphic at some point z = x + iy. Then

f (z + w) − f (z) = f ′ (z)w + w ψ(w)

where ψ(w) → 0 as w → 0. We write f ′ (z) = a + ib and w = h + ik so that

f ′ (z)w = (ah − bk) + i(bh + ak).

In terms of F, the above relations become

F(x + h, y + k) − F(x, y) = (ah − bk, bh + ak) + ψ1 (h, k)h + ψ2 (h, k)k

! !
a −b h
= + ψ1 (h, k)h + ψ2 (h, k)k
b a k

where ψ1 (h, k), ψ2 (h, k) are functions that tend to 0 when (h, k) approaches the origin. The above
relation implies that F is differentiable, with
!
a −b
dF(x,y) =
b a
4 CHAPTER 1. HOLOMORPHIC FUNCTIONS

On the other hand, F(x, y) = (u(x, y), v(x, y)) and its differential is the Jacobian matrix

∂u/∂x ∂u/∂y
!
JF (x, y) = dF(x,y) =
∂v/∂x ∂v/∂y

The equality between the two matrices shows that

∂u ∂v ∂u ∂v
= and =−
∂x ∂y ∂y ∂x

These are the so-called Cauchy-Riemann equations. In particular, one has that

∂u ∂u
f ′ (z) = −i
∂x ∂y

Conversely, let u(x, y), v(x, y) be two continuously differentiable functions in the sense of real
variables which satisfy the Cauchy-Riemann equations. Define the function

f (z) = f (x + iy) = u(x, y) + iv(x, y).

Then it is immediately verified by reversing the above steps that f is holomorphic.

Remark 1. The Cauchy-Riemann equations written above can be recovered in several different
ways. For instance, suppose f is holomorphic and f (x + iy) = u(x, y) + iv(x, y). We can take
h = h1 ∈ ’ in the definition of complex-derivative and get
u(x + h1 , y) + iv(x + h1 , y) − u(x, y) − iv(x, y) ∂u ∂v
f ′ (z) = lim = (x, y) + i (x, y).
h1 →0 h1 ∂x ∂x
Analogously, if we take h = ih2 ∈ i’
u(x, y + h2 ) + iv(x, y + h2 ) − u(x, y) − iv(x, y) 1 ∂u ∂v
f ′ (z) = lim = (x, y) + (x, y).
h2 →0 ih2 i ∂y ∂y
This implies
∂u ∂v 1 ∂u ∂v
(x, y) + i (x, y) = (x, y) + (x, y)
∂x ∂x i ∂y ∂y
and hence the Cauchy-Riemann equations, after separating the real and imaginary parts of both
sides and using that 1/i = −i.
Another possible approach is the following. If we regard f as a functions from an open subset of
’2 to ƒ, the complex-differentiability at z = x + iy is equivalent to the existence of two complex
numbers α, β (the partial derivatives of f with respect to x and y respectively) such that

f (z + h) − f (z) = f (x + h1 , y + h2 ) − f (x, y) = αh1 + βh2 + hψ(h), h = h1 + ih2

where ψ(h) → 0 as h → 0. Since 2h1 = h + h̄ and 2ih2 = h − h̄, the above equality can be rewritten
as
α − iβ α + iβ
f (z + h) − f (z) = h+ h̄ + hψ(h).
2 2i
In other words,
f (z + h) − f (z) α − iβ α + iβ h̄
= + + ψ(h)
h 2 2i h
1.2. THE CAUCHY-RIEMANN EQUATIONS 5

and since the limit of h̄/h does not exist when h approaches 0, we must have that

α + iβ α − iβ
=0 and f ′ (z) = .
2i 2
The above relations read as
∂f ∂f 1∂ f ∂f 
+i = 0, and f ′ (z) = −i
∂x ∂y 2 ∂x ∂y

and the first one yields the Cauchy-Riemann equations.

The above remark suggests the introduction of the following operators

∂ 1 ∂ ∂ ∂ 1 ∂ ∂
∂= := −i , ∂¯ = := +i .
∂z 2 ∂x ∂y ∂z̄ 2 ∂x ∂y

The reason for the above notation is for the chain rule to hold. In fact, if we write x and y as
functions of z, z̄
z + z̄ z − z̄
x= , y=
2 2i
then formally we have
∂f ∂ f ∂x ∂ f ∂y 1 ∂ f 1 ∂f
= + = +
∂z ∂x ∂z ∂y ∂z 2 ∂x 2i ∂y
∂f ∂ f ∂x ∂ f ∂y 1 ∂ f 1 ∂f
= + = −
∂z̄ ∂x ∂z̄ ∂y ∂z̄ 2 ∂x 2i ∂y

We then have the following

Proposition 1.3. f is holomorphic at z0 if and only if

∂f
(z0 ) = 0
∂z̄
in which case
∂f
f ′ (z0 ) = (z0 ).
∂z
Moreover, if f is holomorphic and F(x, y) is its associated map, F is differentiable in the sense of
real variables and
det(JF (x0 , y0 )) = | f ′ (z0 )|2 , z0 = x0 + iy0 .

Proof. The last part of the statement follows from the fact that
 ∂u ∂v ∂u ∂v 
det(JF (x0 , y0 )) = − (x0 , y0 )
∂x ∂y ∂y ∂x

and from the Cauchy-Riemann equations, which yield

 ∂u 2  ∂u 2
det(JF (x0 , y0 )) = + = | f ′ (z0 )|2 .
∂x ∂y
□
6 CHAPTER 1. HOLOMORPHIC FUNCTIONS

1.2.1 Harmonic functions

We say that a function u(x, y) is harmonic if it is real valued , twice continuously differentiable
and
∂2 u ∂2 u
+ = 0.
∂x2 ∂y2
One defines the Laplacian operator as

∂2 ∂2
∆ := +
∂x2 ∂y2
so u is harmonic if and only if it is real valued and ∆u = 0.
Suppose now that f : Ω → ƒ is twice continuously differentiable in the usual sense with respect
to the real variables (x, y). One can prove (see exercise 5 from worksheet 2) that
∂ ∂ ∂ ∂
4 f =4 f = ∆ f.
∂z ∂z̄ ∂z̄ ∂z
If we also assume f to be holomorphic, then from Proposition 1.3 we derive that f , as well as its
real and imaginary part, are harmonic. The converse is also true, we state it below without proof.
Theorem 1.3. Let Ω be a “simply connected” open set and let u be harmonic on Ω. Then there
exists a holomorphic function f on Ω such that u = Re( f ). The difference between two such
functions is a pure imaginary constant.

The last statement of the theorem tells us that if f and g are two holomorphic functions such that
Re( f ) = Re(g) = u, then f − g is constant and such constant is purely imaginary. The students can
prove themselves such statement, see also exercise 2 from Worksheet 2.
In the above statement, there is also the notion of simply connected open set. See the

1.3 Angles under holomorphic maps

One geometric property of holomorphic functions is that they preserve angles between curves.
Let Ω ⊂ ƒ be an open set and γ be a curve in Ω, i.e.

γ : [a, b] → Ω.

We write
γ(t) = x(t) + iy(t)
and assume that γ is differentiable so that

γ′ (t) = x′ (t) + iy′ (t).

For our purposes, we will always assume that γ′ (t) , 0 for all t ∈ [a, b]. From a geometric
standpoint, γ′ (t) can be interpreted as the tangent vector to the curve γ at the point γ(t) ∈ Ω. If γ
and η are two curves passing through the same point z0 ∈ Ω, say

z0 = γ(t0 ) = η(t1 ),
1.3. ANGLES UNDER HOLOMORPHIC MAPS 7

the tangent vectors γ′ (t0 ) and η′ (t1 ) of the two curves at z0 determine an angle θ, which is defined
to be the angle between the two curves at z0 .
If f : Ω → ƒ is a holomorphic function, the composition f ◦ γ defines a new curve in Ω, whose
tangent vector at every point f (γ(t)) is given by

( f ◦ γ)′ (t) = f ′ (γ(t))γ′ (t).

Analogously for the composition f ◦ η. The curves f ◦ γ and f ◦ η pass through the point f (z0 )
and their tangent vector at that point is given, respectively, by

f ′ (z0 )γ′ (t0 ) and f ′ (z0 )η′ (t1 ).

We have the following

Theorem 1.4. If f ′ (z0 ) , 0, the angle between the curves γ, η at z0 is the same as the angle
between the curves f ◦ γ and f ◦ η at f (z0 ).

Proof. Geometrically speaking, the tangent vectors under f are changed by multiplication with
f ′ (z0 ), which can be represented in polar coordinates as a dilation and a rotation, so preserves the
angles.
A proof in terms of sine and cosine of the angle between the curves goes as follows.
Let z = a + ib and w = h + ik be complex numbers. Then

zw̄ = ah + bk + i(bh − ak).

We define the scalar product between z and w as follows

⟨z, w⟩ = Re(zw̄)

and observe that it coincides with the scalar product in ’2 between vectors (a, b) and (h, k).
Let θ(z, w) be the angle between vectors z, w. Then

⟨z, w⟩
cos θ(z, w) =
|z||w|
and
 π  ⟨z, −iw⟩
sin θ(z, w) = cos θ(z, w) − = .
2 |z||w|
Let f ′ (z0 ) = α. Then
⟨αz, αw⟩ = Re(αzᾱw̄) = |α|2 Re(z, w)
from which immediately follows that

cos θ(αz, αw) = cos(z, w), sin θ(αz, αw) = sin(z, w).

Since sine and cosine uniquely determine the angle, the result follows. □

A map which preserves the angle is called conformal. The above theorem then states that holo-
morphic maps with non-zero derivative are conformal.
8 CHAPTER 1. HOLOMORPHIC FUNCTIONS

1.4 Power Series

So far, we have only given rational functions as examples of holomorphic functions. Another,
very important, class of holomorphic functions is given by power series. At this point, we warmly
encourage the student to review what learned about complex power series in the course MAA207
Series of Functions, Differential Equations.
A power series is an expansion of the form
∞
X
an zn
n=0

where an ∈ ƒ. To test for the absolute convergence of this series, one should investigate the
convergence of the real series
X∞
|an ||z|n .
n=0

We recall that, if the above series converges for some z0 ∈ ƒ, then it converges in the whole disc
|z| ≤ |z0 |. Furthermore, we have the following

Theorem 1.5. Given a power series n an zn , there exists 0 ≤ R ≤ ∞ such that

P

(i) if |z| < R the series converges absolutely;

(ii) if |z| > R the series diverges.

Moreover, if we use the convention that 1/0 = ∞ and 1/∞ = 0, then R is given by the Hadamard’s
formula
1/R = lim sup |an |1/n
n→∞

The number R is called radius of convergence of the power series and the disc |z| < R is called
disc of convergence. In general, nothing can be said about the convergence on the boundary of
the disc |z| = R, on which one may have convergence or divergence depending on the situation.

Proposition 1.4. Let {an }n be a sequence of complex numbers such that

|an+1 |
lim = R.
n→∞ |an |
Then
lim |an |1/n = R,
n→∞

n an z .
in other words R is the radius of convergence of the power series
P n

Proof. See Worksheet 3, exercise 1. □

Remark 2. The above definition and result can be also given for power series expansions at points
z0 different than the origin X
an (z − z0 )n .
n
1.4. POWER SERIES 9

Example 4. The prime example of power series is the complex exponential

∞ n
z
X z
e :=
n=0
n!

The student will prove that the above series converges absolutely for all z ∈ ƒ, i.e. R = ∞. The
converge is uniform for all |z| ≤ r.
When z is real, the above definition coincides with the usual exponential function.
Example 5. Similarly, we define the complex trigonometric functions as follows
∞
X z2n+1
sin z := (−1)n
n=0
(2n + 1)!
∞
X z2n
cos z := (−1)n
n=0
(2n)!

The above series converge absolutely in ƒ and are uniform on all bounded discs |z| ≤ r. They give
extensions of the sine and cosine functions to complex numbers.
A simple calculation exhibits a connection between the trigonometric functions and the complex
exponential
eiz − e−iz eiz + e−iz
sin z = , cos z = .
2i 2
The above relations are called the Euler formulas.
Example 6. Define
∞
X zn
f (z) := (−1)n−1
n=1
n
This series converges absolutely when |z| < 1 and coincides with log(1 + x) when z = x is real. We
can therefore define log z for |z − 1| < 1 as

log z = f (z − 1).

We say that a function f is analytic at z0 if there exists a power series

∞
X
an (z − z0 )n
n=0

and some r > 0 such that the series converges absolutely for |z − z0 | < r and for such z
∞
X
f (z) = an (z − z0 )n .
n=0

The above power series expressing f in a neighborhood of z0 is uniquely determined. We also say
that f has a power series expansion at z0 , meaning that the values of f at points z nearby z0 are
given by an absolutely convergent power series as above.
If Ω is an open set, we say that f is analytic on Ω if it is analytic at every point in Ω. If S is any
arbitrary set, we say that f is analytic on S if f is analytic on an open set containing S .
Simple operations with analytic functions preserve analyticity.
10 CHAPTER 1. HOLOMORPHIC FUNCTIONS

Proposition 1.5. (i) If f, g are analytic on Ω, so are f + g and f g. The quotient f /g is analytic
on the open subset of z ∈ Ω on which g(z) , 0;

(ii) If g : Ω → Ω̃ and f : Ω̃ → ƒ are analytic, then f ◦ g is analytic.

Proof. Left as an exercise. □

Power series are of course analytic in their disc of convergence.

Theorem 1.6. Let f (z) = n an zn be a power series whose radius of convergence is R. Then f is
P
analytic on the open disc D(0, R).

Proof. We have to show that f has a power series expansion at any arbitrary z0 ∈ D(0, R). Let
then s > 0 be such that |z0 | + s < R. We write

z = z0 + (z − z0 )

and replace it in the power series to get

∞ ∞ ∞ n !
X X
n X X n n−k
an z =
n
an z0 + (z − z0 ) = an z0 (z − z0 )k .
n=0 n=0 n=0 k=0
k

The series in the above right hand side is uniformly convergent on |z − z0 | < s (why?). For such z,
the order of the summation can hence be interchanged to get
∞
∞ X ! 
X n n−k 
f (z) =  an z0  (z − z0 )k .

k=0 n=k
k

Remark 3. Making a translation, the theorem shows that if f has a power series expansion on a
disc D(z0 , r), that is, X
f (z) = an (z − z0 )n
for |z − z0 | < r, then f is analytic on this disc.

(i) Let f (z) = n an zn be a non-constant power series with radius of conver-

P
Theorem 1.7.
gence R > 0. Suppose f (0) = 0. Then there exists r > 0 such that f (z) , 0 for all |z| < r.

(ii) Let f (z) = n an zn and g(z) = n bn zn be two power series with positive radii of conver-
P P
gence. Suppose f (z) = g(z) on an infinite set having 0 as an adherence point. Then an = bn
for all n. In other words, the two power series coincide.

Proof. (i) Since f (z) is non-constant there exists n ∈ N such that an , 0. We set N := min{n ∈
N : an , 0} and write
f (z) = aN zN (1 + g(z))
where g(z) → 0 as z → 0. Therefore, there exists r > 0 such that |g(z)| < 1/2 for all |z| < r
and consequently

| f (z)| > |an ||z|n |1 + g(z)| > (1/2)|an ||z|n > 0, ∀ z ∈ D(0, r).
1.4. POWER SERIES 11

(ii) Let h(z) := f (z) − g(z) = n (an − bn )zn . By the hypothesis h(z) = 0 on an infinite set of
P
points having 0 as an adherence point. From point (i), we deduce that h is constant on the
smallest disc of convergence, and more precisely h(z) = 0.
□

Corollary 1.1. Let f (z) = n an zn be a power series with positive radius of convergence. If
P
f (z) = 0 on an infinite set of points having 0 as an adherence point, then an = 0 for all n.

There is a tight connection between holomorphic and analytic functions. We shall, in fact, prove
that these two classes of functions coincide, meaning that holomorphic functions are analytic and
conversely. That analytic functions are holomorphic is a consequence of the following theorem.
Theorem 1.8. The power series f (z) = n an zn defines a holomorphic function on its disc of
P
convergence. The derivative of f is also a power series, obtained by differentiating the series of f
term by term
∞
X
f (z) =
′
nan zn−1 .
n=1
Moreover, f′ has the same radius of convergence as f .

Proof. The assertion about the radius of convergence follows simply by the fact that lim supn n1/n =
1. We denote this radius by R. It then remains to prove that the function
∞
X
g(z) := nan zn−1
n=1

is the complex derivative of f on the disc of convergence.

For any fixed N ∈ N, we introduce the partial sum and the tail of the power series as follows
N
X ∞
X
S N (z) := an zn , and T N (z) := an zn ,
n=0 n=N+1

and observe that S N is a holomorphic function. For |z0 | < r < R and h sufficiently close to the
origin, we write
f (z0 + h) − f (z0 ) S N (z0 + h) − S N (z0 )
− g(z0 ) = − S ′N (z0 ) + S ′N (z0 ) − g(z0 )
h h
T N (z0 + h) − T N (z0 )
+ .
h
On the one hand, since g(z) converges uniformly on any disc strictly included in the disc of con-
vergence, we have that S ′N (z0 ) → g(z0 ) and the rate of convergence is independent of z0 . In other
words, for a fixed ϵ > 0 there exists N1 ∈ N such that for all N ≥ N1

|S ′N (z0 ) − g(z0 )| < ϵ.

On the other hand, we recall that we have an − bn = (a − b)(an−1 + an−2 b + · · · + abn−2 + bn−1 ) for
any two numbers a, b, therefore

T N (z0 + h) − T N (z0 ) X∞
(z0 + h)n − zn0 X∞
≤ |an | ≤ |an |nrn−1 ,
h n=N+1
h n=N+1
12 CHAPTER 1. HOLOMORPHIC FUNCTIONS

where we have used that |z0 | < r and |z0 + h| < r if |h| is chosen sufficiently small. The expression
on the right hand side is the tail of a convergent series since g converges absolutely for |z| < R.
Consequently, there exists N2 ∈ N such that for all N ≥ N2
T N (z0 + h) − T N (z0 )
< ϵ.
h
Finally, if we fix N ≥ max (N1 , N2 ), we deduce from the definition of derivative the existence of
δ > 0 such that
S N (z0 + h) − S N (z0 )
− S ′N (z0 ) < ϵ
h
whenever |h| < δ.
Summing all up, we have proved that for any fixed |z0 | < r < R and ϵ > 0 there exists δ > 0 such
that for |h| < δ we have
f (z0 + h) − f (z0 )
− g(z0 ) < 3ϵ
h
concluding that g(z) is the derivative of f (z) on any disc D(0, r) with r < R, hence on disc of
convergence. □

The above theorem not only proves that power series are holomorphic on their disc of convergence,
but also that they are infinitely many times differentiable. Successive applications of the theorem
yield in fact the following

Corollary 1.2. A power series is infinitely complex differentiable in its disc of convergence, and
the higher derivatives are also power series obtained by termwise differentiation.

That holomorphic functions are analytic is a much deeper result that we will study in a further
chapter. It shows, in particular, the great difference between real and complex differentiation as
it implies that a holomorphic function (i.e. a function that has the first complex-derivative) is
actually infinitely many times differentiable. The existence of one derivative is enough to ensure
the existence of all the others.

Appendix to Chapter 1

Let γ and η be two curves in an open set Ω. After a reparametrization, we assume that they are
defined on the same interval [a, b]. We say that γ is homotopic to η if there exists a continuous
function
Ψ : [a, b] × [c, d] → Ω
defined on a rectangle [a, b] × [c, d] such that

Ψ(t, c) = γ(t), Ψ(t, d) = η(t), ∀t ∈ [a, b].

For each fixed s ∈ [c, d], we see the function ψ s (t) = Ψ(t, s) as a continuous curve, and we may
view the family of all ψ s as a continuous deformation of γ to η.
We say that the homotopy Ψ leaves the endpoint fixed if

Ψ(a, s) = γ(a), Ψ(b, s) = γ(b), ∀s ∈ [c, d].

1.4. POWER SERIES 13

In general, one always assumes that the homotopy under consideration leaves the endpoint fixed.
Let Ω be an open set. We say that Ω is simply connected if it is connected and any closed path is
homotopic to a point. Informally speaking, Ω is simply connected when there are no “holes” and
any closed curve can be “continuously shrank” to a point.