This action might not be possible to undo. Are you sure you want to continue?
COMPLEX ANALYSIS
1
Douglas N. Arnold
2
References:
John B. Conway, Functions of One Complex Variable, SpringerVerlag, 1978.
Lars V. Ahlfors, Complex Analysis, McGrawHill, 1966.
Raghavan Narasimhan, Complex Analysis in One Variable, Birkh¨auser, 1985.
CONTENTS
I. The Complex Number System . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2
II. Elementary Properties and Examples of Analytic Fns. . . . . . . . . . . . . . . . 3
Diﬀerentiability and analyticity . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4
The Logarithm. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6
Conformality. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6
Cauchy–Riemann Equations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7
M¨obius transformations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9
III. Complex Integration and Applications to Analytic Fns. . . . . . . . . . . . . 11
Local results and consequences . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 12
Homotopy of paths and Cauchy’s Theorem. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 14
Winding numbers and Cauchy’s Integral Formula. . . . . . . . . . . . . . . . . . . . . . . . . 15
Zero counting; Open Mapping Theorem . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 17
Morera’s Theorem and Goursat’s Theorem . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 18
IV. Singularities of Analytic Functions. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 19
Laurent series. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 20
Residue integrals. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 23
V. Further results on analytic functions. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 26
The theorems of Weierstrass, Hurwitz, and Montel . . . . . . . . . . . . . . . . . . . . . . . 26
Schwarz’s Lemma . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 28
The Riemann Mapping Theorem. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 29
Complements on Conformal Mapping . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 31
VI. Harmonic Functions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 32
The Poisson kernel . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 33
Subharmonic functions and the solution of the Dirichlet Problem . . . . . . . . . 36
The Schwarz Reﬂection Principle . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 39
1
These lecture notes were prepared for the instructor’s personal use in teaching a halfsemester course
on complex analysis at the beginning graduate level at Penn State, in Spring 1997. They are certainly not
meant to replace a good text on the subject, such as those listed on this page.
2
Department of Mathematics, Penn State University, University Park, PA 16802.
Web: http://www.math.psu.edu/dna/.
1
2
I. The Complex Number System
R is a ﬁeld. For n > 1, R
n
is a vectorspace over R, so is an additive group, but doesn’t
have a multiplication on it. We can endow R
2
with a multiplication by
(a, b)(c, d) = (ac −bd, bc + ad).
Under this deﬁnition R
2
becomes a ﬁeld, denoted C. Note that (a/(a
2
+b
2
), −b/(a
2
+b
2
))
is the multiplicative inverse of (a, b). (Remark: it is not possible to endow R
n
with a ﬁeld
structure for n > 2.) We denote (0, 1) by i and identify x ∈ R with (x, 0), so R ⊂ C. Thus
(a, b) = a +bi, a, b ∈ R. Note that i
2
= −1. C is generated by adjoining i to R and closing
under addition and multiplication. It is remarkable that the addition of i lets us not only
solve the equation x
2
+ 1 = 0, but every polynomial equation.
For a and b real and z = a + bi we deﬁne Re z = a, Imz = b, ¯ z = a − bi, and
z = (a
2
+ b
2
)
1/2
. Then
Re z = (z + ¯ z)/2, Imz = (z − ¯ z)/(2i),
z
2
= z¯ z,
1
z
=
¯ z
z
2
,
z ±w = ¯ z ± ¯ w, zw = ¯ z ¯ w,
z/w = ¯ z/ ¯ w, z + w ≤ z +w.
The map θ → (cos θ, sin θ) deﬁnes a 2πperiodic map of the real line onto the unit
circle in R
2
. In complex notation this map is θ → cis θ := cos θ + i sin θ. Every nonzero
complex number can be written as r cis θ where r > 0 is uniquely determined and θ ∈ R
is uniquely determined modulo 2π. The number 0 is equal to r cis θ where r = 0 and θ
is arbitrary. The relation z = r cis θ determines the relations z → r which is simply the
function r = z and z → θ. The latter is denoted θ = arg θ. Note that for z = 0, arg θ is
determined modulo 2π (while arg 0 is arbitrary). We can normalize arg by insisting that
arg z ∈ (−π, π]. Note that if z
1
= r cis θ
1
and z
2
= r cis θ
2
then z
1
z
2
= r
1
r
2
cis(θ
1
+ θ
2
).
The latter formula just encapsulates the formula for the sine and cosine of a sum, and gives
arg z
1
z
2
= arg z
1
+ arg z
2
. In particular, ir cis θ = r cis(θ + π/2), so multiplication by i is
just the operation of rotation by π/2 in the complex plane. Multiplication by an arbitrary
complex number r cis θ is just rotation by arg θ followed by (or preceded by) dilation by a
factor r. Further, z
n
= r
n
cis(nθ). Every nonzero z ∈ C admits n distinct nth roots: the
nth roots of r cis θ are
n
√
r cis[(θ + 2πk)/n], k = 0, 1, . . . , n.
Lines and circles in the plane. Circles given by z − a = r where a ∈ C is the center
and r > 0 is the radius. If 0 = b ∈ C then the line through the origin in the direction
b is the set of all points of the form tb, t ∈ R, or all z with Im(z/b) = 0. If t ∈ R
and c > 0 then (t + ci)b = tb + cib represents a point in the half plane to the left of
b determined by the line tb, i.e., {z : Im(z/b) > 0} is the equation of that halfplane.
Similarly, {z : Im[(z − a)/b] > 0} is the translation of that halfplane by a, i.e., the half
plane determined by the line through a parallel to b and in the direction to the left of
b.
3
Stereographic projection determines a onetoone correspondence between the unit
sphere in R
3
minus the northpole, S, and the complex plane via the correspondence
z ↔
x
1
+ ix
2
1 −x
3
,
x
1
=
2 Re z
1 +z
2
, x
2
=
2 Imz
1 +z
2
, x
3
=
z
2
−1
z + 1
.
If we deﬁne C
∞
= C ∪ {∞}, then we have a onetoone correspondence between S and
C
∞
. This allows us to deﬁne a metric on C
∞
, which is given by
d(z
1
, z
2
) =
2z
1
−z
2

(1 +z
1

2
)(1 +z
2

2
)
, d(z, ∞) =
2
1 +z
2
.
II. Elementary Properties and Examples of Analytic Functions
For z = 1,
¸
N
n=0
z
n
= (1 − z
N+1
)/(1 − z). Therefore the geometric series
¸
∞
n=0
z
n
converges (to 1/(1−z)) if z < 1. It clearly diverges, in fact its terms become unbounded,
if z > 1.
Weierstrass MTest. Let M
0
, M
1
, . . . be positive numbers with
¸
M
n
< ∞ and suppose
that f
n
: X → C are functions on some set X satisfying sup
x∈X
f
n
(x) ≤ M
n
. Then
¸
∞
n=0
f
n
(x) is absolutely and uniformly convergent.
Theorem. Let a
0
, a
1
, · · · ∈ C be given and deﬁne the number R by
1
R
= limsup a
n

1/n
.
Then (1) for any a ∈ C the power series
¸
∞
n=0
a
n
(z − a)
n
converges absolutely for all
z −a < R and it converges absolutely and uniformly on the disk z −a ≤ r for all r < R.
(2) The sequence a
n
(z − a)
n
is unbounded for all z − a > R (and hence the series is
certainly divergent).
Thus we see that the set of points where a power series converges consists of a disk
z −a < R and possibly a subset of its boundary. R is called the radius of convergence of
its series. The case R = ∞ is allowed.
Proof of theorem. For any r < R we show absolute uniform convergence on D
r
= {z−a ≤
r}. Choose ˜ r ∈ (r, R). Then, 1/˜ r > limsup a
n

1/n
, so a
n

1/n
< 1/˜ r for all n suﬃciently
large. For such n, a
n
 < 1/˜ r
n
and so
sup
z∈D
r
a
n
(z −a)
n
 < (r/˜ r)
n
.
Since
¸
(r/˜ r)
n
< ∞ we get the absolute uniform convergence on D
r
.
If z − a = r > R, take ˜ r ∈ (R, r). Then there exist n arbitrarily large such that
a
n

1/n
≥ 1/˜ r. Then, a
n
(z −a)
n
 ≥ (r/˜ r)
n
, which can be arbitrarily large.
4
Theorem. If a
0
, a
1
, . . . ∈ C and lima
n
/a
n+1
 exists as a ﬁnite number or inﬁnity, then
this limit is the radius of convergence R of
¸
a
n
(z −a)
n
.
Proof. Without loss of generality we can suppose that a = 0. Suppose that z >
lima
n
/a
n+1
. Then for all n suﬃciently large a
n
 < a
n+1
z and a
n
z
n
 < a
n+1
z
n+1
.
Thus the series
¸
a
n
z
n
has terms of increasing magnitude, and so cannot be convergent.
Thus z ≥ R. This shows that lima
n
/a
n+1
 ≥ R.
Similarly, suppose that z < lima
n
/a
n+1
. Then for all n suﬃciently large a
n
 > a
n+1
z
and a
n
z
n
 > a
n+1
z
n+1
. Thus the series has terms of decreasing magnitude, and so, by
the previous theorem, z ≤ R. This shows that lima
n
/a
n+1
 ≤ R.
Remark. On the circle of convergence, many diﬀerent behaviors are possible.
¸
z
n
diverges
for all z = 1.
¸
z
n
/n diverges for z = 1, else converges, but not absolutely (this follows
from the fact that the partial sums of
¸
z
n
are bounded for z = 1 and 1/n ↓ 0).
¸
z
n
/n
2
converges absolutely on z ≤ 1. Sierpinski gave a (complicated) example of a function
which diverges at every point of the unit circle except z = 1.
As an application, we see that the series
∞
¸
n=0
z
n
n!
converges absolutely for all z ∈ C and that the convergence is uniform on all bounded sets.
The sum is, by deﬁnition, exp z.
Now suppose that
¸
∞
n=0
a
n
(z −a)
n
has radius of convergence R, and consider its formal
derivative
¸
∞
n=1
na
n
(z−a)
n−1
=
¸
∞
n=0
(n+1)a
n+1
(z−a)
n
. Now clearly
¸
n=0
a
n+1
(z−a)
n
has the same radius of convergence as
¸
∞
n=0
a
n
(z −a)
n
since
(z −a)
N
¸
n=0
a
n+1
(z −a)
n
=
N+1
¸
n=0
a
n
(z −a) −a
0
,
and so the partial sums on the left and right either both diverge for a given z or both
converge. This shows (in a roundabout way) that limsup a
n+1

1/n
= limsup a
n

1/n
=
1/R. Now lim(n+1)
1/n
= 1 as is easily seen by taking logs. Moreover, it is easy to see that if
limsup b
n
= b and limc
n
= c > 0, then limsup b
n
c
n
= bc. Thus limsup (n +1)a
n+1

1/n
=
1/R. This shows that the formal derivative of a power series has the same radius of
convergence as the original power series.
Diﬀerentiability and analyticity. Deﬁnition of diﬀerentiability at a point (assumes
function is deﬁned in a neighborhood of the point).
Most of the consequences of diﬀerentiability are quite diﬀerent in the real and complex
case, but the simplest algebraic rules are the same, with the same proofs. First of all,
diﬀerentiability at a point implies continuity there. If f and g are both diﬀerentiable at a
point a, then so are f ± g, f · g, and, if g(a) = 0, f/g, and the usual sum, product, and
quotient rules hold. If f is diﬀerentiable at a and g is diﬀerentiable at f(a), then g ◦ f
is diﬀerentiable at a and the chain rule holds. Suppose that f is continuous at a, g is
continous at f(a), and g(f(z)) = z for all z in a neighborhood of a. Then if g
(f(a)) exists
and is nonzero, then f
(a) exists and equals 1/g
(f(a)).
5
Deﬁnition. Let f be a complexvalued function deﬁned on an open set G in C. Then f
is said to be analytic on G if f
exists and is continuous at every point of G.
Remark. We shall prove later that if f is diﬀerentiable at every point of an open set in C
it is automatically analytic; in fact, it is automatically inﬁnitely diﬀerentiable. This is of
course vastly diﬀerent from the real case.
If Q is an arbitrary nonempty subset of C we say f is analytic on Q if it is deﬁned and
analytic on an open set containing Q.
We now show that a power series is diﬀerentiable at every point in its disk of convergence
and that its derivative is given by the formal derivative obtained by diﬀerentiating term
byterm. Since we know that that power series has the same radius of convergence, it
follows that a power series is analytic and inﬁnitely diﬀerentiable in its convergence disk.
For simplicity, and without loss of generality we consider a power series centered at zero:
f(z) =
¸
n
a
n
z
n
. Suppose that the radius of convergence is R and that z
0
 < R. We must
show that for any > 0, the inequality
f(z) −f(z
0
)
z −z
0
−g(z
0
)
≤
is satisﬁed for all z suﬃciently close to z
0
, where g(z) =
¸
∞
n=1
na
n
z
n−1
. Let s
N
(z) =
¸
N
n=0
a
n
z
n
, R
N
(z) =
¸
∞
n=N+1
a
n
z
n
. Then
f(z) −f(z
0
)
z −z
0
−g(z
0
)
≤
s
N
(z) −s
N
(z
0
)
z −z
0
−s
N
(z
0
)
+s
N
(z
0
) −g(z
0
) +
R
N
(z) −R
N
(z
0
)
z −z
0
=: T
1
+ T
2
+ T
3
.
Now s
N
(z
0
) is just a partial sum for g(z
0
), so for N suﬃciently large (and all z), T
2
≤ /3.
Also,
R
N
(z) −R
n
(z
0
)
z −z
0
=
∞
¸
n=N+1
a
n
z
n
−z
n
0
z −z
0
.
Now z
0
 < r < R for some r, and if we restrict to z < r, we have
a
n
z
n
−z
n
0
z −z
0
= a
n
z
n−1
+ z
n−2
z
0
+· · · + z
n−1
0
 ≤ a
n
nr
n−1
.
Since
¸
a
n
nr
n−1
is convergent, we have for N suﬃciently large and all z < r then
T
3
< /3. Now ﬁx a value of N which is suﬃciently large by both criteria. Then the
diﬀerentiability of the polynomial s
N
shows that T
1
≤ /3 for all z suﬃciently close to z
0
.
We thus know that if f(z) =
¸
a
n
z
n
, then, within the disk of convergence, f
(z) =
¸
na
n
z
n−1
, and by induction, f
(z) =
¸
n(n−1)a
n
z
n−2
, etc. Thus a
0
= f(0), a
1
= f
(0),
a
2
= f
(0)/2, a
3
= f
(0)/3!, etc. This shows that any convergent power series is the sum
of its Taylor series in the disk of convergence:
f(z) =
¸
f
n
(a)
n!
(z −a)
n
.
In particular, exp
= exp.
6
Lemma. A function with vanishing derivative on a region (connected open set) is constant.
Indeed, it is constant on disks, since we can restrict to segments and use the real result.
Then we can use connectedness to see that the set of points where it takes a given value
is open and closed.
Use this to show that exp(z) exp(a−z) ≡ exp a. Deﬁne cos z and sin z, get cos
2
+sin
2
≡
1, exp z = cos z + i sin z, cis θ = exp(iθ), exp z = exp(Re z) cis(Imz).
The Logarithm. If x = 0, the most general solution of exp z = w is z = log w+i arg w+
2πin, n ∈ Z. There is no solution to exp z = 0.
Deﬁnition. If G is an open set and f : G → C is a continuous function satisfying
exp(f(z)) = z, then f is called a branch of the logarithm on G.
Examples: C \ R
−
(principal branch), C \ R
+
, a spiral strip.
By the formula for the derivative of an inverse, a branch of the logarithm is analytic
with derivative 1/z.
A branch of the logarithm gives a branch of z
b
:= exp(b log z) (understood to be the
principal branch if not otherwise noted). Note that e
b
= exp b.
Note that diﬀerent branches of the logarithm assign values to log z that diﬀer by addition
of 2πin. Consequently diﬀerent branches of z
b
diﬀer by factors of exp(2πinb). If b is an
integer, all values agree. If b is a rational number with denominator d there are d values.
Conformality. Consider the angle between two line segments with a common vertex in
the complex plane. We wish to consider how this angle is transformed under a complex
map. The image will be two smooth curves meeting at a common point, so we have to
deﬁne the angle between two such curves.
Deﬁne a path in the complex plane as a continuous map γ : [a, b] →C. If γ
(t) ∈ C exists
at some point and is not zero, then it is a vector tangent to the curve at γ(t), and hence
its argument is the angle between the (tangent to the) curve and the horizontal. (Note:
if γ
(t) = 0, the curve may not have a tangent; e.g., γ(t) = t
2
− it
3
.) Let γ : [a, b] → C
and ˜ γ : [˜ a,
˜
b] → C be two smooth curves, such that γ(a) = ˜ γ(˜ a) = z, and γ
(a), ˜ γ
(˜ a) = 0.
Then arg ˜ γ
(˜ a)−arg γ
(a) measures the angle between the curves. Now consider the images
under f, e.g., of ρ(t) = f(γ(t)). Then ρ
(a) = f
(z)γ
(a), ˜ ρ
(˜ a) = f
(z)˜ γ
(˜ a). Therefore
arg ˜ ρ
(˜ a) −arg ρ
(a) = arg ˜ γ
(˜ a) +arg f
(z) −arg γ
(a) −arg f
(z), i.e., the angle is invariant
in both magnitude and sign. This says that an analytic mapping is conformal, whenever
its derivative is not zero. (Not true if the derivative is zero. E.g., z
2
doubles angles at the
origin.)
The conformality of an analytic map arises from the fact that in a neighborhood of a
point z
0
such a map behaves to ﬁrst order like multiplication by f
(z
0
). Since f
(z
0
) =
r exp iθ, near z
0
f simply behave like rotation by θ followed by dilation. By contrast, a
smooth map from R
2
→R
2
can behave like an arbitrary linear operator in a neighborhood
of a point.
At this point, show the computer graphics square1.mps and square2.mps.
7
Cauchy–Riemann Equations. Let f : Ω → C with Ω ⊂ C open. Abuse notation
slightly by writing f(x, y) as an alternative for f(x + iy). If f
(z) exists for some z =
x + iy ∈ Ω, then
f
(z) = lim
h∈R
h→0
f(z + h) −f(z)
h
= lim
h∈R
h→0
f(x + h, y) −f(x, y)
h
=
∂f
∂x
,
and
f
(z) = lim
h∈R
h→0
f(z + ih) −f(z)
ih
= lim
h∈R
h→0
−i
f(x, y + h) −f(x, y)
h
= −i
∂f
∂y
.
Thus complexdiﬀerentiability of f at z implies not only that the partial derivatives of f
exist there, but also that they satisfy the Cauchy–Riemann equation
∂f
∂x
= −i
∂f
∂y
.
If f = u + iv, then this equation is equivalent to the system
∂u
∂x
=
∂v
∂y
,
∂v
∂x
= −
∂u
∂y
.
Another convenient notation is to introduce
∂f
∂z
:=
1
2
∂f
∂x
−i
∂f
∂y
,
∂f
∂¯ z
:=
1
2
∂f
∂x
+ i
∂f
∂y
.
(These are motivated by the equations x = (z + ¯ z)/2, y = (z − ¯ z)/(2i), which, if z and ¯ z
were independent variables, would give ∂x/∂z = 1/2, ∂y/∂z = −i/2, etc.)
In terms of these, the Cauchy–Riemann equations are exactly equivalent to
∂f
∂¯ z
= 0,
which is also equivalent to
∂f
∂z
=
∂f
∂x
.
Thus, if f is analytic on Ω, then the partial derivaties of f exist and are continuous on
Ω, and the CauchyRiemann equations are satisﬁed there. The converse is true as well:
Theorem. If the partial derivatives of f exist and are continuous on Ω, and the Cauchy
Riemann equations are satisﬁed there, then f is analytic on Ω.
8
Proof. Let z = x + iy ∈ Ω. We must show that f
(z) exists. Let r be small enough that
the disk of radius r around z belongs to Ω and choose h = s + it with 0 < s + it < r.
Then, by the mean value theorem,
f(z + h) −f(z)
h
=
f(x + s, y + t) −f(x, y)
s + it
=
f(x + s, y + t) −f(x, y + t)
s
s
s + it
+
f(x, y + t) −f(x, y)
t
t
s + it
=
∂f
∂x
(x + s
∗
, y + t)
s
s + it
+
∂f
∂y
(x, y + t
∗
)
t
s + it
,
where s
∗
 < s and t
∗
 < t. Note that
∂f
∂x
(x + s
∗
, y + t) −
∂f
∂x
(x, y) → 0
as h → 0, and similarly for the second partial derivative. Moreover s/(s + it) stays
bounded (by 1). Thus
f(z + h) −f(z)
h
=
∂f
∂x
(x, y)
s
s + it
+
∂f
∂y
(x, y)
t
s + it
+ R(h),
where lim
h→0
R(h) = 0. Now if the Cauchy–Riemann equations hold, then the the ﬁrst
two terms on the right hand side sum to ∂f/∂x(x, y), which is independent of h, so the
limit exists and is equal to ∂f/∂x(x, y).
Remark. The theorem can be weakened to say that if f is continuous on Ω and the partial
derivatives exist and satisfy the Cauchy–Riemann equations there (without assuming that
the partial derivatives are continuous), then the complex derivative of f exists on Ω (which
is equivalent to f being analytic on Ω. This is the Looman–Menchoﬀ Theorem. See
Narasimhan, Complex Analysis in One Variable, for a proof. We do need at least continuity,
since otherwise we could take f to be the characteristic function of the coordinate axes.
Note that
∂
∂z
∂
∂¯ z
=
∂
∂¯ z
∂
∂z
=
1
4
∆.
This shows that any analytic function is harmonic (equivalently, its real and imaginary
parts are harmonic). It also shows that the conjugate of an analytic function, while not
analytic, is harmonic.
If u(x, y) = Re f(x + iy), then u is harmonic. Conversely, if u is harmonic on a region
Ω, does there exist a harmonic conjugate, i.e., a function v on Ω such that f = u + iv is
analytic? Clearly, if there exists v it is determined up to addition of a real constant (or
f is determined up to addition of an imaginary constant). The book gives an elementary
construction of the harmonic conjugate in a disk and in the whole plane. We sketch the
idea of a more general proof. We require that the domain satisfy the condition that if γ
is any piecewise simple smooth closed curve in Ω, then γ is the boundary of a subset G of
9
Ω. In other words, Ω has no holes (is simplyconnected). Our proof follows directly from
the following result, which holds on simplyconnected domains: let f : Ω → R
2
be a C
1
vectorﬁeld. Then there exists v : Ω →R such that f = ∇v if and only if ∂f
1
/∂y = ∂f
2
/∂x.
The “only if” part is obvious. To prove the “if” part, ﬁx a point (x
0
, y
0
) in Ω and for any
(x, y) ∈ Ω, let γ
(x,y)
be a piecewise smooth path in Ω from (x
0
, y
0
) to (x, y), and let τ be
the unit tangent to the path pointing from (x
0
, y
0
) towards (x, y). Deﬁne
v(x, y) =
γ
(x,y)
f · τ ds.
It is essential that this quantity doesn’t depend on the choice of path, or, equivalently,
that
γ
f · τ ds = 0
for all piecewise smooth simple closed paths γ. But
γ
f · τ ds =
γ
(f
2
, −f
1
) · nds =
∇· (f
2
, −f
1
) dxdy = 0,
by the divergence theorem. It is easy to check that ∇v = f.
For a nonsimply connected region, there may exist no harmonic conugate. E.g., if
u = log z on C \ {0}, then u + i arg z is analytic on C \ {z ≤ 0}, but cannot be extended
to an analytic function on C \ {0}.
M¨obius transformations.
Deﬁnition. Let a, b, c, d be complex numbers with ad = bc. Then the mapping
S(z) =
az + b
cz + d
, z ∈ C, z = −d/c,
is a M¨obius transformation.
Remarks. 1) Note that if ad = bc the same expression would yield a constant. 2) The
coeﬃcents aren’t unique, since we can multiply them all by any nonzero complex constant.
To each MT we associate the nonsingular matrix
a b
c d
of its coeﬃcients, which is
determined up to a nonzero multiple. 3) The linear (but nonconstant) polynomials are
MTs, namely the ones for which c = 0. 4) If z = −d/c then S(z) ∈ C and S
(z) =
(ad − bc)/(cz + d)
2
= 0. 5) We can deﬁne S(−d/c) = ∞ and S(∞) = a/c, so S can be
viewed of as a map from C
∞
into itself (which is, as we will now see, 11 and onto).
If S and T are MTs, then so is S ◦ T, its coeﬃcient matrix being the product of the
coeﬃcient matrices of S and T. Any MTs admits an inverse, namely the transform with
the inverse coeﬃcient matrix. With the understanding that S(∞) = a/c and S(−d/c) = ∞
(or, if c = 0, then S(∞) = ∞, the MT maps C
∞
11 onto itself. This is evident from the
10
existence of its inverse S
−1
(z) = (dz −b)/(−cz +a). Moreover, the composition of M¨obius
transformations is again one, so they form a group under composition. If S is a nonlinear
MT we can write it as
az + b
z + d
= a +
b −ad
z + d
.
This shows that any MT can be written as a composition of translations, rotations and
dilations (multiplication by a complex number), and inversion (reciprocals).
Since the geometric action of translation, dilation, and rotation are quite evident, let’s
consider inversion. It takes the unit disk D to C
∞
\
¯
D. We now show that inversion maps
lines and circles to other lines and circles.
Consider the equation z −p = kz −q where k > 0, p, q ∈ C. This is
(x −a)
2
+ (y −b)
2
= k
2
[(x −c)
2
+ (y −d)
2
]
which is clearly the equation of a circle if k = 1 and the equation of a line if k = 1. In either
case we have a circle in the Riemann sphere, which we call C
1
. Substituting 1/z for z and
doing some simple manipulations gives the equation z −1/p = (kq/p)z −1/q, which
is another such circle, C
2
. Thus we have 1/z ∈ C
1
⇐⇒ z ∈ C
2
, showing that inversion
maps circles in the sphere onto other circles in the sphere. Since the same property is
evident for translation, dilation, and rotation, we conclude that this property holds for all
MT.
Some further analysis, which will be omitted, establishes two further properties. 1) A
MT is orientiation preserving in the sense that, if we traverse a circle in the order of three
distinct points on it, z
1
, z
2
, z
3
, the region to the left of the circle will map to the region
to the left of the image circle, with respect to the image orientation. 2) A MT preserves
the property of two points being symmetric with respect to a circle, i.e., lying on the same
ray from the center, and such that the geometric mean of their distances from the center
equals the radius.
Let z
2
, z
3
, z
4
be distinct points in C. Then it is easy to see that there is a unique MT
taking these points to 1, 0, and ∞, respectively, namely
Sz =
z −z
3
z −z
4
z
2
−z
4
z
2
−z
3
.
We can also handle, as a special case, the possibility that one of the z
i
is ∞. We infer
from this and the invertibility of the MTs that given two ordered triples of distinct points
in C
∞
, there is a unique MT taking the ﬁrst onto the second.
If we have any two disks or halfplanes, or one of each, we can map one to the other by
a MT, and can arrange that the image of any three points on the boundary of one go to
three points on the boundary of the other.
The book makes heavy use of the the notation (z
1
, z
2
, z
3
, z
4
) (cross ratio) for the image
of z
1
under the transformation taking z
2
, z
3
, z
4
to 1, 0, and ∞, respectively, so
(z
1
, z
2
, z
3
, z
4
) =
z
1
−z
3
z
1
−z
4
z
2
−z
4
z
2
−z
3
.
From the deﬁnition, (Tz
1
, Tz
2
, Tz
3
, Tz
4
) = (z
1
, z
2
, z
3
, z
4
) for any MT T.
Example: z → (1 −z)/(1 + z) takes the right half plane onto the unit disk.
11
III. Complex Integration and Applications to Analytic Functions
If γ : [a, b] → C is piecewise smooth, and f is a continuous complexvalued function
deﬁned on the image of γ, deﬁne
γ
f(z) dz =
b
a
f(γ(t))γ
(t) dt.
This is simply the line integral of f along γ. It can be deﬁned analogously to the Riemann
integral as the limit of sums of the form
¸
(f ◦ γ)(τ
k
)[γ(t
k
) −γ(t
k−1
)], so is the Riemann–
Stieltjes integral of f ◦ γ with respect to γ. Using this deﬁnition it can be extended to
rectiﬁable paths, i.e., ones for which γ is only of bounded variation.
If φ is a piecewise smooth increasing function from [c, d] onto [a, b] and we let ρ = γ◦φ (so
ρ is the same path as γ but with a diﬀerent parametrization), then
ρ
f(z) dz =
γ
f(z) dz.
From the fundamental theorem of calculus we see that if F is any analytic function on
a neighborhood of γ, then
γ
F
(z) dz = F(β) −F(α) where α = f(a), β = f(b). If γ is a
closed path, this is zero.
As a consequence, if γ = re
iθ
, then
γ
z
n
dz = 0 for all n ∈ Z, n = −1.
By direct calculation we have
γ
z
−1
dz = 2πi.
More generally, let z
0
 < r and consider
γ
1
z−z
0
dz. Let z
±
be the point on the circle
z − z
0
 = r with arg(z
±
− z
0
) = ±(π − ). Clearly, as → 0, z
±
→ z
1
, the point on the
circle with the same imaginary part as z
0
and negative real part. Now let γ
be the curve
running from z
−
to z
+
on the circle. In a neighborhood of γ
the function Log(z −z
0
) is
a primitive of 1/(z −z
0
). So
γ
1
z −z
0
dz = Log(z
+
−z
0
) −Log(z
−
−z
0
) = log r + i(π −) −log r + i(π −).
Taking the limit as ↓ 0 gives
γ
1
z −z
0
dz = 2πi.
Local results and consequences.
Theorem (Cauchy’s integral formula for a disk). Let f be analytic on a closed disk
and let γ be a path around the boundary of the disk. Then
f(z) =
1
2πi
γ
f(w)
w −z
dw
12
for all z in the open disk.
Proof. Consider the function
g(t) =
γ
f(z + t(w −z))
w −z
dw.
Then
g
(t) =
γ
f
(z + t(w −z)) dw,
and, since F(w) = f(z +t(w−z))/t is a primitive for f
(z +t(w−z)), g
(t) ≡ 0. Therefore,
γ
f(w)
w −z
dw = g(1) = g(0) = f(z)
γ
1
w −z
dw = 2πif(z).
As an easy corollary, we have:
Cauchy’s Theorem for a disk. Let f be analytic on a closed disk and let γ be a path
around the boundary of the disk. Then
γ
f(w) dw = 0.
Proof. Fix z in the open disk and apply Cauchy’s formula to the function w → f(w)(w −
z).
We can now show that an analytic (i.e., continuously complex diﬀerentiable) function
has a power series expansion around each point in its domain. (And we already know the
converse is true.)
Theorem. Let f be analytic in B(a; R). Then f(z) =
¸
a
n
(z −a)
n
for z −a < R, for
some a
n
∈ C.
Proof. Fix z and choose r < R with z − a < r. Let γ be the path around w − a = r.
Then (z −a)/(w −a) < 1, so the Weierstrass Mtest implies that the series
¸
1
(w −a)
n+1
(z −a)
n
=
1
w −z
,
converges absolutely and uniformly in w for w −a = r. Since f is bounded on the circle,
the same sort of convergence holds for
¸
f(w)
(w −a)
n+1
(z −a)
n
=
f(w)
w −z
.
Therefore we can interchange summation and integraton to get
f(z) =
1
2πi
γ
f(w)
w −z
dw =
¸
a
n
(z −a)
n
,
where
a
n
=
γ
f(w)
(w −a)
n+1
dw.
13
Corollary (of proof ). Under the same hypotheses
f
(n)
(a) =
n!
2πi
γ
f(w)
(w −a)
n+1
dw.
Corollary (Cauchy’s estimate). If f is analytic in B(a; R), then
f
(n)
(a) ≤
n! max
B(a;R)
f
R
n
.
Proof. Evident if R is replaced by r < R. Then pass to limit.
Corollary (Maximum Modulus Principle). An analytic function on a region which
achieves a maximum modulus in the region is constant.
Proof. By the Cauchy estimate for n = 0, if the function achieves its maximum modulus
at a point, it achieves it everywhere on a disk containing the point. Thus, the set of
points were the maximum modulus is achieved is open. It is also clearly closed. Hence the
function has constant modulus. It follows from the Cauchy–Riemann equations that the
function itself is constant.
Corollary (Liouville’s Theorem). A bounded entire function is constant.
Proof. Cauchy’s estimate for n = 1 implies f
≡ 0.
Corollary (Fundamental Theorem of Algebra). Every nonconstant polynomial has
a root in C.
Theorem. If f is analytic on a region then there does not exist a point a such that
f
(n)
(a) = 0 for all n, unless f is identically zero on the region.
Proof. The set of such a is closed, and using the Taylor expansion, is open.
Corollary. If f is analytic on a region Ω and not identically zero, then for each a ∈ Ω
there exists a unique integer p ∈ N and a function g, analytic in a neighborhood of a and
nonzero at a, such that
f(z) = (z −a)
p
g(z).
Proof. f(z) =
¸
c
n
(z − a)
n
with c
n
= f
(n)
(a)/n!. Let p be the least integer with c
p
= 0
(which exists by the preceeding theorem), and g(z) =
¸
∞
n=p
c
n
(z −a)
n−p
.
The number p is the multiplicity of a as a root of f.
Corollary. If f is analytic on a region and not identically zero, then the roots of f are
isolated.
Corollary. An analytic function on a region is completely determined by its value on any
set of points containing a limit point.
14
Homotopy of paths and Cauchy’s Theorem.
Deﬁnition. Let γ
0
, γ
1
: [0, 1] → G be two closed paths in a region G ⊂ C. We say that
the paths are homotopic in G if there exists a continuous function Γ : [0, 1] × [0, 1] → G
such that for each s Γ( · , s) is a closed path with Γ( · , 0) = γ
0
and Γ( · , 1) = γ
1
.
This is an equivalence relation, written γ
0
∼ γ
1
in G.
Theorem. Suppose that γ
0
and γ
1
are two homotopic piecewise smooth closed curves in
G and that f is analytic on G. Then
γ
0
f(z) dz =
γ
1
f(z) dz.
If we assume that the homotopy Γ is C
2
, the result is easy: let
g(s) =
1
0
f(Γ(t, s))
∂Γ
∂t
(t, s) dt.
Then
g
(s) =
1
0
∂
∂s
[f(Γ(t, s))
∂Γ
∂t
(t, s)] dt =
1
0
∂
∂t
[f(Γ(t, s))
∂Γ
∂s
(t, s)] dt = 0,
since both integrands equal
f
(Γ(s, t))
∂Γ
∂t
∂Γ
∂s
+ f(Γ(s, t))
∂
2
Γ
∂t∂s
.
The proof without assuming smoothness is based on Cauchy’s theorem for a disk and a
polygonal approximation argument.
Sketch of proof. Since Γ is continuous on the compact set [0, 1] × [0, 1], it is uniformly
continuous. Moreover, there exists > 0 so that G contains a closed disk of radius about
each point in the range of Γ. Therefore, we can ﬁnd an integer n, so that Γ maps each of
the squares [j/n, (j + 1)/n] ×[k/n, (k + 1)/n] into the disk of radius about Γ(j/n, k/n).
Let ρ
k
be the closed polygonal path determined by the points Γ(j/n, k/n), j = 0, 1, . . . , n.
It is easy to conclude from Cauchy’s theorem for a disk that
ρ
k
f(z) dz =
ρ
k+1
f(z) dz,
and also that
γ
0
f(z) dz =
ρ
0
f(z) dz,
γ
1
f(z) dz =
ρ
n
f(z) dz.
15
Deﬁnition. A closed path in G is homotopic to 0 in G if it is homotopic to a constant
path.
Corollary. If γ is a piecewise smooth path which is homotopic to 0 in G, and f is analytic
in G, then
γ
f(z) dz = 0.
Deﬁnition. A region G is simplyconnected if every closed path in G is homotopic to 0.
As an immediate consequence of the theorem above, we have Cauchy’s theorem (previ
ously proved only for G a disk).
Cauchy’s Theorem. If G is is a simplyconnected region in the complex plane then
γ
f(z) dz = 0
for every function f analytic in G and every piecewise smooth closed curve γ in G.
Winding numbers and Cauchy’s Integral Formula. Intuitive meaning of winding
number: if a pole is planted in the plane at a point a and a closed curve is drawn in
the plane not meeting a, if the curve were made of welllubricated rubber and were to be
contracted as small as possible, if it contracts to a point other than a its winding number
is 0. If it contracts around the pole planted at a, its winding number is the number of
times it circles the pole (signed by orientation).
Deﬁnition. Let γ be a piecewise smooth closed curve in C and a ∈ C\γ. Then the index
of γ with respect to a, or the winding number of γ about a is deﬁned to be
n(γ; a) =
1
2πi
γ
1
z −a
dz.
Demonstrate how the winding number can be calculated by drawing a negatively di
rected ray from a and integrating over pieces of the curve between successive intersections
of the curve with the ray.
Proposition. The index n(γ; a) is an integer for all a ∈ C \ γ. It is constant on each
connected component of C \ γ.
Proof. Assuming γ is parametrized by [0, 1], deﬁne
g(t) =
t
0
γ
(s)
γ(s) −a
ds.
16
An easy direct calculation shows that
d
dt
exp[−g(t)][γ(t) −a] = 0,
so, indeed, exp[−g(t)][γ(t) −a] = γ(0) −a. Taking t = 1,
exp[−2πin(γ; a)][γ(1) −a] = γ(0) −a,
or, exp[−2πin(γ; a)] = 1, which implies that n(γ; a) is indeed integral.
The function is obviously continuous, and being integer, it is constant on connected
components. Clearly also, it tends to 0 as a tends to ∞, so it is identically 0 on the
unbounded component.
The next result is an immediate corollary of the invariance of path integrals of analytic
functions under homotopy.
Proposition. If γ
0
and γ
1
are paths which are homotopic in C \ {a} for some point a,
then n(γ
0
; a) = n(γ
1
; a).
Cauchy’s Integral Formula. Let γ a piecewise smooth curve in a region G which is null
homotopic there, and let f be an analytic function on G. Then
n(γ; a)f(a) =
1
2πi
γ
f(z)
z −a
dw, a ∈ G\ γ.
Proof. For a ∈ G\ γ ﬁxed, let
g(z) =
f(z) −f(a)
z −a
, z ∈ G\ {a},
f
(a), z = a.
It is easy to show that g is continuous (even analytic!) on G (show the continuity as a
homework exercise). Applying Cauchy’s theorem to g gives the desired result immediately.
Remark. This presentation is quite a bit diﬀerent from the book’s. The book proves
Cauchy’s formula under the assumption that γ is homologous to 0, i.e., that n(γ; a) = 0
for all a ∈ C \ G. It is easy to see that homotopic to 0 implies homologous to 0, but the
reverse is not true, so the book’s result is a bit stronger. It is true that all piecewise smooth
closed curves in region are homologous to zero iﬀ G is simplyconnected (see Theorem 2.2
of Chapter 8).
The book also goes on to consider chains, which are unions of curves and proves that
f(a)
k
¸
j=1
n(γ
j
; a) =
1
2πi
k
¸
j=1
γ
j
f(w)
w −z
dz,
if the chain γ
1
+ · · · + γ
n
is homologous to zero, i.e.,
¸
k
j=1
n(γ
j
; z) = 0 for all z ∈ C \ G.
This sort of result is useful, e.g., in the case of an annulus. However we can deal with this
solution in other ways (by adding a line connecting the circles).
17
Zero counting; Open Mapping Theorem.
Theorem. Let G be a region, γ a curve homotopic to 0 in G and f an analytic function
on G with zeros a
1
, a
2
, . . . , a
m
repeated according to multiplicity. Then
1
2πi
γ
f
(z)
f(z)
dz =
m
¸
k=1
n(γ; a
k
).
Proof. We can write f(z) = (z − a
1
)(z − a
2
) . . . (z − a
m
)g(z) where g(z) is analytic and
nonzero on G. Then
f
(z)
f(z)
=
1
z −a
1
+
1
z −a
2
+· · ·
1
z −a
m
+
g
(z)
g(z)
,
and the theorem follows.
If the a
i
are the solutions of the equation f(z) = α, then
1
2πi
γ
f
(z)
f(z) −α
dz =
m
¸
k=1
n(γ; a
k
).
Note that we may as well write the integral as
1
2πi
f◦γ
1
w −α
dw.
Theorem. If f is analytic at a and f(a) = α with ﬁnite multiplicity m ≥ 1, then there
exists , δ > 0 such that for all 0 < ζ −α < δ, the equation f(z) = ζ has precisely m roots
in z −a < and all are simple.
Proof. That the number of roots is m comes from the continuity of the integral as long
as ζ remains in the same connected component of C \ f ◦ γ as α, and the fact that it is
integer valued. We can insure that the roots are simple by taking small enough to avoid
a root of f
.
Open Mapping Theorem. A nonconstant analytic function is open.
Proof. This is an immediate corollary.
18
Morera’s Theorem and Goursat’s Theorem.
Morera’s Theorem. If G is a region and f : G → C a continuous function such that
T
f dz = 0 for each triangular path T in G, then f is analytic.
Proof. Since it is enough to show f is analytic on disks, we can assume G is a disk. Let a
be the center of the disk, and deﬁne F(z) =
[a,z]
f. By hypothesis
F(z) −F(z
0
)
z −z
0
=
1
z −z
0
[z,z
0
]
f(w) dw =
1
0
f(z
0
+ t(z −z
0
))dt.
Continuity then implies that F
(z
0
) = f(z
0
), so f has a primitive, so is analytic.
Before proceeding to Goursat’s Theorem, we consider another result that uses an argu
ment similar to Morera’s Theorem.
Theorem. Let G be a region and f : G → C a continuous function. Then f admits a
primitive on G if and only if
γ
f = 0 for all piecewise smooth closed curves γ in G.
Proof. Only if is clear. For the if direction, ﬁx a ∈ G and deﬁne F(z) =
γ(a,z)
f where
γ(a, z) is any piecewise smooth path from a to z. Reasoning as in Morera’s theorem
F
= f.
In particular, if G is a simplyconnected region then every analytic function admits a
primitive. If 0 / ∈ G, then 1/z admits a primitive. Adjusting the primitive by a constant
we can assure that it agrees with a logarithm of z
0
at some point z
0
. It is then easy to
check by diﬀerentiation that exp(F(z))/z ≡ 1. Thus
Proposition. If G is a simply connected domain, 0 ∈ G, then a branch of the logarithm
may be deﬁned there.
(As an interesting example, take G to be the complement of the spiral r = θ.)
We can at long last show that a function which is complex diﬀerentiable in a region
(but without assuming it is continuously diﬀerentiable) is analytic.
Goursat’s Theorem. Let G be a region and f : G → C complex diﬀerentiable at each
point of G. Then f is analytic.
Proof. We will show that the integral of f vanishes over every triangular path. Let ∆
0
be
a closed triangle with boundary T
0
. Using repeated quadrisection we obtain triangles
∆
0
⊂ ∆
1
⊂ . . .
with boundaries T
n
such that

T
0
f ≤ 4
n

T
n
f.
19
Let d = diamT
0
, p = perimT
0
. Then diamT
n
= d/2
n
, perimT
n
= p/2
n
. By compactness
¸
∆
n
= ∅. Let z
0
be in the intersection (it’s unique).
From the diﬀerentiability of f at z
0
we have: given > 0
f(z) −f(z
0
) −f
(z
0
)(z −z
0
) ≤ z −z
0

for all z in some disk around z
0
, therefore for all z ∈ ∆
n
, some n. Since
T
n
f(z) dz =
T
n
z dz = 0,

T
n
f = 
T
n
[f(z) −f(z
0
) −f
(z
0
)(z −z
0
)]dz
≤ diamT
n
perimT
n
= dp/4
n
.
Therefore 
T
0
f ≤ dp, and, since was arbitrary, it must vanish.
IV. Singularities of Analytic Functions
Deﬁnition. If f is analytic in B(a; R) \ {a}, for some a ∈ C, R > 0, but not in B(a; R).
We say f has an isolated singularity at a. If there exists g analytic on some disk centered
at a which agrees with f except at a, we say f has a removable singularity.
Theorem. Suppose f has an isolated singularity at a. It is removable if and only if
lim
z→a
(z −a)f(z) = 0.
Proof. “Only if” is clear. For “if” let g(z) = (z −a)f(z), z = a, g(a) = 0. Using Morera’s
theorem (separating the cases a inside the triangle, on the boundary, or in the interior), it
is easy to see that g is analytic. Therefore g(z) = (z −a)h(z) for h analytic, and h is the
desired extension of f.
Corollary. If lim
z→a
f(z) ∈ C, or even f(z) is bounded in a punctured neighborhood of a,
or even f(z) ≤ Cz−a
−(1−)
for some > 0, some C, then f has a removable singularity.
Now either lim
z→a
f(z) exists as a ﬁnite real number, equals +∞, or does not exist as
a ﬁnite number or inﬁnity. The ﬁrst case is that of a removable singularity. The second is
called a pole, the third, an essential singularity.
Note that if f has a pole at a, then 1/f(z) has a removable singularity there and extends
to a with value 0. Using this we see that f has a pole at a if and only if f(z) = (z−a)
−m
g(z)
for some integer m > 0 and g analytic in a neighborhood of a. Thus
f(z) =
∞
¸
n=−m
a
n
(z −a)
n
in a punctured neighborhood of a.
The function e
1/z
has an essential singularity since
lim
z↓0
e
1/z
= +∞, lim
z↑0
e
1/z
= 0.
20
Theorem (CasoratiWeierstrass). If f has an essential singularity at a, then the image
under f of any punctured disk around a is dense in C.
Proof. WLOG, assume a = 0. If the conclusion is false, there is a punctured disk around
0 in which f(z) stays a ﬁxed positive distance away from some number c ∈ C. Consider
the function g(z) = [f(z) −c]/z. It tends to ∞ as z → 0, so it has a pole at 0. Therefore
z
m
[f(z) − c] → 0 as z → 0 for suﬃciently large m. Therefore, z
m
f(z) → 0, which is not
possible if f has an essential singularity.
Laurent series.
Deﬁnition. Suppose f is analytic on {z : z > R} for some R. Then we say that f has
an isolated singularity at ∞. It is removable if f(1/z) has a removable singularity at 0.
We easily see that f has a removable singularity at ∞ ⇐⇒ ∃ lim
z→∞
f(z) ∈ C ⇐⇒
f is bounded in a neighborhood of inﬁnity.
Note that if f has a removable singularity at ∞ then we can expand f in a power series
in 1/z, convergent for z > R uniformly and absolutely for z ≥ ρ > R.
We say that a series of the form
¸
∞
n=−∞
a
n
(z − a)
n
converges if both the series
¸
∞
n=1
a
−n
(z − a)
−n
and
¸
∞
n=0
a
n
(z − a)
n
converge (and their sum is the value of the
doubly inﬁnite summation).
Lemma. Let γ be a piecewise smooth curve and g a continuous complexvalued function
on γ. Deﬁne
f(z) =
γ
g(w)
z −w
dw, z / ∈ γ.
The f is analytic on C \ γ.
Proof. Fix z / ∈ γ. Then
g(w)
ζ−w
−
g(w)
z−w
ζ−z
= −
g(w)
(ζ −w)(z −w)
→ −
g(w)
(z −w)
2
as ζ → z with the convergence uniform in w ∈ γ. Thus f is diﬀerentiable at each z ∈
C \ γ.
Let f be deﬁned and analytic in an annulus R
1
< z < R
2
. Deﬁne
f
2
(z) =
1
2πi
w=r
f(w)
w −z
dw,
where R
2
> r > z (it doesn’t matter which such r we take, by Cauchy’s theorem). Then
f
2
is analytic in z < R
2
and so has a convergent power series there:
f
2
(z) =
∞
¸
n=0
a
n
z
n
.
21
Also, let
f
1
(z) = −
1
2πi
w=r
f(w)
w −z
dw
where now r is taken with R
1
< r < z. Then this is analytic in z > R
1
and tends to 0
at ∞. Thus
f
1
(z) =
∞
¸
n=1
a
−n
z
−n
.
Now, it is easy to use Cauchy’s integral formula to see that f = f
1
+ f
2
in the annulus.
This shows that
f(z) =
∞
¸
n=−∞
a
n
z
n
,
with the convergence absolute and uniform on compact subsets of the annulus. Note that
such an expansion is unique, since we can easily recover the a
n
:
a
m
=
1
2πi
¸
n∈Z
a
n
z=r
z
n
z
m+1
dz =
1
2πi
z=r
f(z)
z
m+1
dz
where r is any number in (R
1
, R
2
) (result is independent of choice of r since two diﬀerent
circles are homotopic in the annulus).
Summarizing:
Theorem (Laurent Expansion). If f is an analytic function on the annulus G =
ann(0; R
1
, R
2
) for some 0 ≤ R
1
< R
2
≤ ∞, then
f(z) =
∞
¸
n=−∞
a
n
z
n
for all z in the annulus. The convergence is uniform and absolute on compact subsets of
the annulus and the coeﬃcients a
n
are uniquely determined by the formula above.
If f has an isolated singularity at a, it has a Laurent expansion in a punctured disk at
a:
f(z) =
∞
¸
n=−∞
a
n
(z −a)
n
, 0 < z −a < R.
We deﬁne
ord
a
(f) = inf{n ∈ Z : a
n
= 0}.
Thus f has a removable singularity ⇐⇒ ord
a
(f) ≥ 0, and it extend to a as 0 ⇐⇒
ord
a
(f) > 0, in which case this is the order of the 0. ord
a
(f) = +∞ ⇐⇒ f is identically
0 near a. ord
a
(f) = −∞ ⇐⇒ f has an essential singularity at a. ord
a
(f) < 0 but ﬁnite
⇐⇒ f has a pole, in which case −ord
a
(f) is the order of the pole.
22
We call a
−1
the residue of f at a, denoted Res(f; a). Note that, if f has an isolated
singularity at a, then for r suﬃciently small,
Res(f; a) =
1
2πi
z−a=r
f(z) dz.
Give pictoral proof that if G is simplyconnected domain, f analytic on G except for iso
lated singularies, and γ is a simple closed curve that encloses singularities at a
1
, a
2
, . . . , a
m
(necessarily ﬁnite), then
1
2πi
γ
f(z) dz =
m
¸
j=1
Res(f; a
j
).
a simple closed curve in G .
More precisely:
Theorem (Residue Theorem). Let G be an open set, E a discrete subset of G, and
γ a nullhomotopic piecewise smooth closed curve in G which doesn’t intersect E. Then
{a ∈ En(γ; a) = 0} is ﬁnite and
1
2πi
γ
f(z) dz =
¸
a∈E
n(γ; a) Res(f; a).
for all functions f which are analytic in G\ E.
Proof. Let F : [0, 1] ×[0, 1] → G be a homotopy from γ to a constant. The image K of F is
compact, so has a ﬁnite intersection with E. If a ∈ E does not belong to this intersection,
the homotopy takes place in the the complement of a, showing that the n(γ; a) = 0.
Let a
1
, . . . , a
m
be the points in the intersection, and let g
i
be the singular part of f at
a
i
. The f −
¸
g
i
is analytic on K (after removing the removable singularities at the a
i
),
so
γ
f −
¸
g
i
= 0. The result follows easily.
We will proceed to two easy but useful consequences of the residue theorem, the argu
ment principle and Rouch´e’s theorem. First, however, we consider the application of the
residue theorem to the computation of deﬁnite integrals.
Residue integrals.
First, how to practically compute residues. For essential singularities it can be very
diﬃcult, but for poles it is usually easy. If f has a simple pole at a, then f(z) = g(z)/(z−a)
with g analytic at a and Res(f; a) = g(a) (and thus Cauchy’s therem may be viewed as
the special case of the residue theorem where f has a simple pole). If f has a double pole
at a, then
f(z) = g(z)/(z −a)
2
=
g(a)
(z −a)
2
+
g
(a)
z −a
+· · · ,
23
so Res(f; a) = g
(a). Similarly, if f has a pole of order m with f(z) = g(z)/(z − a)
m
, we
expand g as a Taylor series around a and can read oﬀ Res(f; a). (It is
1
(m−1)!
g
(m−1)
(a),
but it is easier to remember the procedure than the formula.)
Example 1: evaluate
∞
−∞
1
(x
2
+ a
2
)(x
2
+ b
2
)
dx, 0 < a < b.
Use a semicircle centered at the origin in upper half plane and let radius tend to inﬁnity to
get the answer of π/[ab(a + b)]. Same technique works for all rational functions of degree
≤ −2.
Example 2: evaluate
2π
0
sin
4
(x) dx.
Write as
γ
1
iz
z+1/z
2i
4
dz, with γ the unit circle, to get the answer 3π/4. This approach
applies to rational functions in sin and cos.
Example 3: evaluate
∞
0
1
t
1/2
+ t
3/2
dt.
Use a branch of the z
−1/2
with the positive real axis as the branch cut, and integrate
around this cut and over a large circle centered at the origin (avoid origin with a small
circle). The answer is π. This approach applies to integrals of the form
∞
0
t
α
r(t)dt where
α ∈ (0, 1) and r(t) is a rational function of degree ≤ −2 with at worst a simple pole at 0.
Example 4: evaluate
∞
0
sin x
x
dx
We will evaluate
lim
X→+∞
δ↓0
−δ
−X
+
X
δ
e
iz
z
dz.
The imaginary part of the quantity inside the limit is 2
X
δ
(sin x)/xdx. Note: it is essential
that when taking the limit we excise a symmetric interval about zero. (The ordinary
integral doesn’t exist: e
ix
/x has a nonintegrable singularity near the origin, even though
its imaginary part, (sin x)/x doesn’t.) On the other hand, it is not necessary that we cut
oﬀ the integrand at symmetric points −X and X (we could have used −X
1
and X
2
and
let them go independently to zero).
We use a somewhat fancy path: starting at −X we follow the xaxis to −δ, then a
semicircle in the upperhalf plane to δ, then along the xaxis to X, then along a vertical
segment to X + iX, then back along a horizontal segment to −X + iX, and then down
a vertical segment back to −X. The integrand e
z
/z has no poles inside this curve, so
its contour integral is 0. It is easy to see that the integral is bounded by 1/X on each
of the vertical segments and by 2e
−X
on the horizontal segment. On the other hand,
24
e
iz
/z = 1/z +analytic and the integral of 1/z over the semicircle (integrated clockwise) is
easily seen to be −πi. We conclude that the limit above equals πi, whence
∞
0
sin x
x
dx =
π
2
.
Example 5: evaluate
∞
¸
n=−∞
1
(n −a)
2
, a ∈ R \ Z.
Use a rectangular path with opposite corners ±(n+1/2+ni), n a positive integer and eval
uate
γ
(z −a)
−2
cot πz dz, a / ∈ Z. The poles of the integrand are at a and the integers and
the residue is −π/ sin
2
(πa) and 1/[π(n−a)
2
]. Since sin x+iy = sin xcosh y +i cos xsinh y
and cos x + iy = cos xcosh y −i sin xsinh y, we get
 cot x + iy
2
=
cos
2
x + sinh
2
y
sin
2
x + sinh
2
y
,
so the  cot  equals 1 on the vertical segments and, if n is large, is bounded by 2 on the
horizontal segments. It follows easily that the integral goes to 0 with n, so
∞
¸
n=−∞
1
(n −a)
2
=
π
2
sin
2
πa
.
To state the argument principle we deﬁne:
Deﬁnition. A complex function on an open set G is called meromorphic if it is analytic
on G except for a set of poles.
Note that the pole set of a meromorphic function is discrete (but may be inﬁnite). A
meromorphic function is continuous when viewed as taking values in C
∞
.
Suppose f is meromorphic at a (i.e., on a neighborhood of a). Then Res(f/f
; a) =
ord
a
(f). Indeed, f(z) = (z − a)
m
g(z) with m = ord
a
(f) and g(a) = 0, and f
/f =
m/(z −a) + g
/g. The residue theorem then immediately gives
Theorem (Argument Principle). Let G be an open set, f a meromorphic function
on G, and γ a nullhomotopic piecewise smooth closed curve in G which doesn’t intersect
either the zero set Z or the pole set P of f. Then
1
2πi
γ
f
f
=
¸
a
n(γ; a) ord
a
(f)
where the sum is over Z ∪ P and contains only ﬁnitely many terms.
Note that the left hand side is nothing other than n(f ◦ γ; 0), which is behind the name
(how much does the argument change as we traverse this curve). We saw this theorem
previously in the case where f was analytic, in which case the integral counts the zeros of
f. For meromorphic f we count the poles as negative zeros.
Our ﬁnal application of the residue theorem is Rouch´e’s theorem.
25
Theorem (Rouch´e). Let G be an open set, f and g meromorphic functions on G, and γ a
nullhomotopic piecewise smooth closed curve in G which doesn’t intersect Z
f
∪P
f
∪Z
g
∪P
g
.
If
f(z) + g(z) < f(z) +g(z), z ∈ γ,
then
¸
a∈Z
f
∪P
f
n(γ; a) ord
a
(f) =
¸
a∈Z
g
∪P
g
n(γ; a) ord
a
(g).
Proof. The hypothesis means that on γ the quotient f(z)/g(z) is not a nonnegative real
number. Let l be a branch of the logarithm on C \ [0, ∞). Then l(f/g) is a primitive of
(f/g)
/(f/g) = f
/f − g
/g which is analytic in a neighborhood of γ. Hence,
γ
f
/f =
γ
g
/g and the theorem follows from the argument principle.
Corollary (Classical statement of Rouch´e’s theorem). Suppose f and φ are analytic
on a simplyconnected region G and that γ is a simple closed piecewise smooth curve in G.
If φ < f on γ, then f + φ has the same number of zeros as f in the region bounded by
γ.
Proof. (f + φ) + (−f) = φ < f ≤ f + φ + −f.
Example: If λ > 1, then the equation z exp(λ − z) = 1 has exactly one solution in the
unit disk (take f = z exp(λ −z), φ = −1).
V. Further results on analytic functions
The theorems of Weierstrass, Hurwitz, and Montel. Consider the space of contin
uous functions on a region in the complex plane, endowed with the topology of uniform
convergence on compact subsets. (In fact this is a metric space topology: we can de
ﬁne a metric such that a sequence of continuous functions converge to another continuous
function in this metric, if and only if the sequence converges uniformly on compact sub
sets.) Weierstrass’s theorems states that the space of analytic functions on the region is
a closed subspace and diﬀerentiation acts continuously on it. Montel’s theorem identiﬁes
the precompact subsets of the space of analytic functions.
Theorem (Weierstrass). Let G ⊂ C be open, f
n
: G → C, n = 1, 2, . . . , analytic,
f : G → C. If f
n
→ f uniformly on compact subsets of G, then f is analytic. Moreover
f
n
converges to f
uniformly on compact subsets of G.
Proof. If T is a triangular path,
T
f = lim
n→∞
T
f
n
= 0,
so f is analytic by Morera’s theorem.
26
Let γ be the boundary of a disk around z contained in G.
limf
n
(z) = lim
1
2πi
γ
f
n
(w)
(w −z)
2
dz =
1
2πi
γ
f(w)
(w −z)
2
dz = f
(z).
Combining Weierstrass’s Theorem and the Argument Principle we get:
Theorem (Hurwitz). Let G be a domain in C, f
n
, f analytic functions on G with
f
n
→ f uniformly on compact subsets. If each of the f
n
is nowherevanishing, then f is
either nowherevanishing or identically zero.
Proof. If f is not identically zero, but f(a) = 0, we can choose r > 0 so that
¯
B(a, r) ⊂ G,
and a is the only zero of f on this disk. Letting γ be the boundary of the disk, the argument
principle then tells us that
γ
f
n
/f
n
= 0, but
γ
f
/f = 2πi ord
a
(f) = 0. But, in view of
Weierstrass’s Theorem the hypothesis tells us that f
n
/f
n
converges to f
/f uniformly on
γ, so we have a contradiction.
Corollary. Let G be a domain in C, f
n
, f analytic functions on G with f
n
→ f uniformly
on compact subsets. If each of the f
n
is injective, then f is either injective or identically
constant.
Proof. If f is not identically constant, but there exists distinct a, b with f(a) = f(b),
then we can choose r > 0 so that B(a, r) ∪ B(b, r) ⊂ G, and B(a, r) ∩ B(b, r) = ∅. Now
f
n
(z) − f(a) converges to f(z) − f(a) uniformly on compact subsets of B(a, r) and the
latter function has a zero but is not identically zero there, so for all suﬃciently large n,
f
n
(z) = f(a) for some z ∈ B(a, r), and, by the same argument, for some z ∈ B(b, r). This
contradicts the injectivity of the f
n
.
Theorem (Montel). Let G ⊂ C be open, F a family of analytic functions on G. Suppose
that the functions f
1
, f
2
, . . . are uniformly bounded on each compact subset of G. Then
there exists a subsequence which converges uniformly on compact subsets.
Proof. Fix a ∈ G and let d(a) = dist(a, ∂G)/2. First we show that we can ﬁnd a subse
quence that converges uniformly on B(a, d(a)/2). Now
f
m
(z) =
∞
¸
n=0
c
n
(f
m
)(z −a)
n
, z ∈
¯
B(a, d(a)).
where c
n
(f) = f
(n)
(a)/n!. By the Cauchy estimates
c
n
(f
m
) ≤ M/d(a)
n
where M is the uniform bound for the f
m
on
¯
B(a, d(a)).
Since c
1
(f
m
) is uniformly bounded in C, we can ﬁnd a subsequence f
m
1
(j)
for which
c
1
(f
m
1
(j)
) converges as j → ∞. Then we can ﬁnd a further subsequence so that c
2
(f
m
2
(j)
)
converges, as well, etc. For the diagonal sequence g
j
= f
m
j
(j)
c
n
(g
j
) converges for all n.
27
Claim: g
j
converges uniformly on B(a, d(a)/2). To prove this, we must show that for
any > 0
g
j
−g
k
 ≤ on B(a, d(a)/2).
Now, for any N ∈ N, z ∈ B(a, d(a)/2),
g
j
(z) −g
k
(z) ≤
N
¸
n=1
c
n
(g
j
) −c
n
(g
k
)z −a
n
+
∞
¸
n=N+1
2M
d(a)
n
z −a
n
≤
N
¸
n=1
c
n
(g
j
) −c
n
(g
k
)[d(a)/2]
n
+
2M
2
N
.
Given > 0 we can choose N large enough that
2M
2
N
and then choose j and k large enough
that c
n
(g
j
) −c
n
(g
k
) ≤ /[2(N + 1)[d(a)/2]
n
] for 0 ≤ n ≤ N. This proves the claim.
Fix a ∈ G and let d(a) = dist(a, ∂G)/2. First we show that we can ﬁnd a subsequence
that converges uniformly on B(a, d(a)/2). Now
f
m
(z) =
∞
¸
n=0
f
(n)
m
(a)
n!
(z −a)
n
, z ∈
¯
B(a, d(a)).
By the Cauchy estimates
f
(n)
m
(a) ≤ Mn!/d(a)
n
where M is the uniform bound for the f
m
on
¯
B(a, d(a)). In particular, sup
m
f
(n)
m
(a) < ∞
for each n.
Thus we can ﬁnd a subsequence f
m
0
(j)
for which f
m
0
(j)
(a) converges as j → ∞. Then
we can ﬁnd a further subsequence so that f
m
2
(j)
(a) converges as well, etc. For the diagonal
sequence g
j
= f
m
j
(j)
g
(n)
j
(a) converges for all n.
Claim: g
j
converges uniformly on B(a, d(a)/2). To prove this, we must show that for
any > 0
g
j
−g
k
 ≤ on B(a, d(a)/2).
Now, for any N ∈ N, z ∈ B(a, d(a)/2),
g
j
(z) −g
k
(z) ≤
N
¸
n=0
g
(n)
j
(a) −g
(n)
k
(a)z −a
n
/n! +
∞
¸
n=N+1
2M
d(a)
n
z −a
n
≤
N
¸
n=1
g
(n)
j
(a) −g
(n)
k
(a)z −a
n
/n! +
2M
2
N
.
Given > 0 we can choose N large enough that 2M/2
N
≤ /2 and can then choose j and
k large enough that g
(n)
j
(a) − g
(n)
k
(a) ≤ n!/{2(N + 1)[d(a)/2]
n
} for 0 ≤ n ≤ N. This
proves the claim and shows that we can ﬁnd a subsequence which converges uniformly on
B(a, d(a)/2).
28
Now we may choose a countable sequence of points a
i
∈ G such that G =
¸
∞
i=1
B(a
i
, d(a
i
)/2) (e.g., all points in G with rational real and imaginary parts). Given the
sequence of functions f
m
uniformly bounded on compact sets, we may ﬁnd a subsequence
f
n
1
(j)
which converges uniformly on B(a
1
, d(a
1
)/2). Then we may take a subsequence of
that sequence which converges uniformly on B(a
2
, d(a
2
)/2), etc. The diagonal subsequence
h
j
converges uniformly on each B(a
i
, d(a
i
)/2).
We complete the proof by showing that h
j
converges uniformly on all compact subsets.
Indeed this is immediate, because any compact subset of G is contained in a ﬁnite union
of the B(a
i
, d(a
i
)/2).
Schwarz’s Lemma. Schwarz’s lemma is an easy, but very useful, consequence of the
maximum modulus principle. The most diﬃcult part is to remember that the German
analyst Herman Schwarz (actually Karl Herman Amandus Schwarz, 18431921), inventor
of the “Schwarz inequality,” “Schwarz lemma,” and “Schwarz alternating method,” spells
his name without a “t”, while the French analyst, Laurent Schwartz, 1915–, inventor of
the theory of distributions, uses a “t”.
Suppose that an analytic map f maps a disk of radius about a into the disk of
radius δ about f(a). We can use this to estimate the derivative f
(a) and the stretching
f(b) −f(a)/b −a. To keep the statement simple, we translate in the domain and range
space so that a = f(a) = 0 and dilate in the domain and range so that = δ = 1. We then
get:
Theorem (Schwarz’s Lemma). Suppose f maps the open unit disk into itself leaving
the origin ﬁxed. Then f(z) ≤ z for all z in the disk, and if equality holds for any nonzero
point, then f(z) = cz for some c of modulus 1. Also f
(0) ≤ 1 and if equality holds, then
f(z) = cz for some c of modulus 1.
Proof. The function g(z) = f(z)/z has a removable singularity at the origin, and extends
analytically to the disk with g(0) = f
(0). It satisﬁes g(z) ≤ 1/r on the circle of radius
r < 1, so by the maximum modulus theorem it satisﬁes this condition on the disk of radius
r. Letting r tend to one gives the result. Equality implies g is constant.
From a homework exercise (§III .3, no. 10), we know that the most general M¨obius
transformations of the open unit disk D into itself is given by
z → c
z −a
1 − ¯ az
with a in the disk and c = 1. We can use Schwarz’s lemma to show that this is most
general onetoone analytic map of the disk onto itself.
First consider the special case f(0) = 0. Then Schwarz shows that f
(0) ≤ 1 with
equality if and only if f(z) = cz some c = 1. Applying Schwarz to f
−1
gives 1/f
(0) =
(f
−1
)
(0) ≤ 1. Thus equality does indeed hold.
Returning to the general case, say f(a) = 0. Then we can apply the special case above
to f ◦ φ
−a
to get f(φ
−a
(z)) = cz or f(z) = cφ
a
(z).
29
The Riemann Mapping Theorem. We begin by deriving some easy consequences of
simpleconnectivity. Recall that every analytic function on a simplyconnected region
admits a primitive. It follows that if f is a nowherevanishing function on a simply
connected domain, then there exists an analytic function F on that domain such that
f = exp(F). Indeed, let g be primitive of f
/f. It follows immediately that f exp(−F) is
constant, and so adjusting g by an additive constant, f = exp(F). Another consequence
is the squareroot property: if f is a nowherevanishing analytic function on the domain,
then there exists an analytic function g with [g(z)]
2
= f(z). Indeed, we can just take
g = exp(F/2). For future use we remark that if there exists an analytic isomorphism of a
region G
1
onto a region G
2
, then G
1
has the squareroot property if and only if G
2
does.
We now turn to a lemma of Koebe.
Lemma. Let D = B(0, 1) and let G D be a region containing the origin and having
the square root property. Then there exists an injective analytic map r : G → D such that
r(0) = 0, r(z) > z, z ∈ G\ {0}.
Proof. Let a ∈ D \ G and let b be one of the square roots of −a. Deﬁne
q(z) = φ
−a
([φ
−b
(z)]
2
)
(where φ
α
(z) := (z −α)/(1 − ¯ αz)). Clearly q(0) = 0, and q is not just multiplication by a
constant (in fact, it is not 11 on D), so Schwarz’s lemma implies
q(z) < z, 0 = z ∈ D.
Next, note that φ
a
is an analytic map of G into D which never vanishes, so admits an
analytic square root g(z), which we may determine uniquely by insisting that g(0) = b.
Deﬁning
r = φ
b
◦ g,
we have r(0) = 0, and q(r(z)) ≡ z. Thus if 0 = z ∈ G, r(z) = 0 and z = q(r(z)) <
r(z).
Two regions in the complex plane are called conformally equivalent if they are analyti
cally isomorphic, that is, if there exists a onetoone analytic mapping of the ﬁrst domain
onto the second (whose inverse is then automatically analytic).
Riemann Mapping Theorem. Let G C be simplyconnected. Then G is conformally
equivalent to the open unit disk.
Proof. The only consequence of simpleconnectivity that we shall use in the proof is that
G has the squareroot property.
The structure of the proof is as follows: ﬁrst we use the hypotheses to exhibit an injective
analytic map of h
0
of G onto a domain G
0
which satisﬁes the hypotheses of Koebe’s lemma.
Then we use an extremal problem to deﬁne an injective analytic map f of G
0
onto a domain
30
G
1
contained in the disk. If G
1
ﬁlls the entire disk we are done; otherwise, G
1
satisﬁes the
hypotheses of Koebe’s lemma and we use the lemma to contradict the extremality of f.
Let a ∈ C \ G and let g be an analytic square root of z − a. It is easy to see that
g is injective and that if w ∈ g(G) then −w / ∈ g(G). Pick w
0
∈ g(G) and then r > 0
such that B(w
0
, r) ⊂ g(G) (Open Mapping Theorem). Therefore B(−w
0
, r) ∩ g(G) = ∅.
Set h(z) = [g(z) + w
0
]
−1
so that h : G →
¯
B(0, 1/r) is analytic and injective. Then set
h
0
(z) = r[h(z) − h(z
0
)]/3 where z
0
∈ G. This map is an analytic isomorphism of G onto
a connected open set G
0
D containing 0.
Let F be the set of all injective analytic functions f : G
0
→ D with f(0) = 0. Fix
0 = w
0
∈ G
0
, and let α = sup
f∈F
f(w
0
). Note α ∈ (0, 1]. We claim that the supremum
is achieved. Indeed, let f
n
∈ F, f
n
(w
0
) → α. Since the f
n
are uniformly bounded
(by 1), Montel’s theorem assures us that there is a subsequence f
n
k
which converges
uniformly on compact subsets to some analytic function f on G
0
. Obviously f(0) = 0
and f(w
0
) = α > 0, so f is not constant, and since the f
n
are injective, so is f. Clearly
f ≤ 1 on G
0
, and by the maximum modulus principle, f < 1. Thus f ∈ F and achieves
the supremum.
We complete the theorem by showing that f is an isomorphism of G
0
onto D (so f ◦ h
0
is the desired analytic isomorphism of G onto D). We already know that f is analytic and
injective, so we need only show that f maps onto D. If, to the contrary, G
1
:= f(G
0
) D
then (since G
1
inherits the squareroot property from G
0
which inherits it from G) Koebe’s
lemma gives an injective analytic function r : G
1
→ D with r(0) = 0, r(z) > z for z = 0.
Then r ◦ f ∈ F, but r(f(w
0
)) > f(w
0
), a contradiction.
We now note that the squareroot property is equivalent to simpleconnectivity. Indeed
we have seen that simpleconnectivity implies the squareroot property. On the other hand,
if G satisﬁes the square root property, then either G = C, which is simplyconnected, or
G is isomorphic to the disk (by the proof of the Riemann mapping theorem above), and
so is simplyconnected.
If G C is a simplyconnected region and z
0
∈ G is arbitrary we may take any analytic
isomorphism of G onto the disk and follow it with a suitably chosen M¨obius transformation
(z → c(a −z)/(1 −¯ az), a < 1, c = 1, to obtain an analytic isomorphism f of G onto D
satisfying
f(z
0
) = 0, f
(z
0
) > 0.
If f
1
and f
2
were two such analytic isomorphisms, then g = f
2
◦ f
−1
1
is an analytic isomor
phism of the disk onto itself with g(0) = 0, g
(0) > 0, and it easily follows that g is the
identity and f
1
= f
2
.
Complements on Conformal Mapping. There are many interesting questions in con
formal mapping which we will not have time to investigate, but I will quickly state some
key results.
Extension to boundary. If the boundary of a simplyconnected region is a simple closed
curve, then it can be proved that a conformal map of the region onto the open disk extends
to a topological homeomorphism of the closure of the region on to the closed disk.
31
Constructive conformal mapping. Conformal is used to obtain explicit solutions to
problems involving analytic and harmonic functions. For example, Joukowski used the
conformal mapping f ◦ g where
f(z) = 2
1 + z
1 −z
, g(z) =
z −1
z + 1
2
,
applied to a circle passing through −1 and containing 1 in its interior. The resulting region,
called a Joukowski airfoil, looks like a crosssection of a wing, and by adjusting the circle,
one can adjust the exact shape of the airfoil. Using the explicit conformal map to the
disk and explicit formulas for the solution of Laplace’s equation on (the exterior of) a disk
(similar to those that will be derived in the next section, one can calculate things like the
drag and lift of the airfoil, and so this was a useful approach to airfoil design.
An important class of regions are the polygons. The Schwarz–Christoﬀel formula is an
explicit formula for the conformal map of a given polygon onto the disk. The formula
involves a few complex parameters which are determined by the vertices of the polygon.
Given the vertices, one cannot usually compute the parameters analytically, but they are
not diﬃcult to compute numerically (with a computer).
Multiply connected regions. For lack of time we will not study conformal mapping of
multiply connected regions, but we just state, without proof, the result for doubly con
nected regions (regions whose complement in the Riemann sphere consists of two connected
components, i.e., roughly regions in the plane with one hole). It can be shown (with ba
sically the tools at our disposal) that any such region is conformally equivalent to the
annulus 1 < z < R for some R > 1. The value of R is uniquely determined. In particular,
two such annuli with diﬀerent values of R
1
are not conformally equivalent.
VI. Harmonic Functions
A realvalued C
2
function on an open subset of C is called harmonic if its Laplacian
vanishes. We have seen previously that the real part of an analytic function is harmonic
and that if the domain is simplyconnected every realvalued harmonic function u admits a
harmonic conjugate, i.e., a realvalued function v, for which u+iv is analytic. In particular,
we see that every harmonic function is C
∞
.
Suppose u is harmonic on G, a ∈ G. If r > 0 is such that
¯
B(a, r), then we can choose
a disk D with
¯
B(a, r) ⊂ D ⊂ G, and since D is simplyconnected u = Re f on D for some
analytic function f on D. Cauchy’s integral formula then gives
f(a) =
1
2πi
z−a=r
f(z)
z −a
dz =
1
2π
2π
0
f(a + re
iθ
)dθ.
Taking real parts we get:
Mean Value Theorem. If u : G →R is harmonic and
¯
B(a, r) ⊂ G, then
(MVP) u(a) =
1
2π
2π
0
u(a + re
iθ
)dθ.
32
A harmonic function satisﬁes the Maximum Principle: it does not assume its maximum
on a region unless it is constant there. Indeed the conclusion is true for any continuous
function with the meanvalue property (MVP), or even just satisfying the inequality u(a) ≤
(2π)
−1
2π
0
u(a + re
iθ
)dθ (since this implies that the set of points were the maximum is
achieved is open, and it is obviously closed). The reverse inequality similarly shows that a
harmonic function satisﬁes the Minimum Principle.
Given a bounded domain G and a function f : ∂G →R, the Dirichlet problem consists
of ﬁnding a function u :
¯
G → R such that u is harmonic on G, continuous on
¯
G, and
coincident with f on ∂G. An important consequence of the maximum and minimum
principles is uniqueness for the Dirichlet problem. If u
1
and u
2
are both harmonic on
G, continuous on
¯
G, and agree on ∂G, then u
1
− u
2
is harmonic and vanishes on the
boundary, but it takes its maximum and minimum on the boundary, so it is identically
zero, i.e., u
1
= u
2
.
When applying the maximum principle to functions which are not known to extend
continuously to the boundary this easytoprove consequence is useful: If u satisﬁes
the maximum principle on a bounded open set and there is a number K such that
limsup
z→z
0
u(z) ≤ K, then u ≤ K on G. [Proof: Let M = sup
z∈G
u(z) and take
z
n
∈ G with u(z
n
) → M. Pass to a subsequence with z
n
→ z
0
∈
¯
G. If z
0
∈ G, then
u takes its maximum there, so is constant and obviously M ≤ K. Otherwise z
0
∈ ∂G, so
K ≥ limsup
z→z
0
u(z) ≥ lim
n
u(z
n
) = M.]
The Poisson kernel. For 0 ≤ r < 1, θ ∈ R let
P
r
(θ) = Re
1 + re
iθ
1 −re
iθ
=
1 −r
2
1 −2r cos θ + r
2
,
deﬁne the Poisson kernel. Note that
P
r
(θ) = Re(1 + 2
∞
¸
n=1
r
n
e
inθ
) =
∞
¸
n=−∞
r
n
e
inθ
.
Lemma. For 0 ≤ r < 1, P
r
(θ) is a smooth, positive, even, 2πperiodic function of θ with
mean value 1. Moreover if θ > 0, then
lim
r↑1
P
r
(θ) = 0, θ ∈ R \ 2πZ,
and if δ > 0, then the convergence is uniform over θ ∈ R such that θ − 2πn ≥ δ for all
n ∈ Z.
Proof. Elementary (use the power series expansion for the integral and the formula
P
r
(θ) =
1 −r
2
(1 −r)
2
+ 2r(1 −cos θ)
for the limits).
Now let D be the open unit disk and ∂D its boundary. The next theorem shows how
to solve the Dirichlet problem for D using the Poisson kernel:
33
Theorem. Given f : ∂D →R continuous, deﬁne
u(z) = f(z), z ∈ ∂D,
u(re
iθ
) =
1
2π
2π
0
P
r
(θ −t)f(e
it
) dt, 0 ≤ r < 1, θ ∈ R.
The u is continuous on
¯
D and harmonic on D.
Proof.
u(re
iθ
) = Re
1
2π
2π
0
f(e
it
)
¸
1 + re
i(θ−t)
1 −re
i(θ−t)
dt
= Re
1
2π
2π
0
f(e
it
)
¸
e
it
+ re
iθ
e
it
−re
iθ
dt
or
u(z) = Re
1
2π
2π
0
f(e
it
)
¸
e
it
+ z
e
it
−z
, z ∈ D.
The integral deﬁnes an analytic function of z (since the integrand is continuous in t and z
and analytic in z), so u, being the real part of an analytic function, is harmonic.
It remains to show that lim
z∈D
z→e
it
0
u(z) = f(e
it
0
) for t
0
∈ R. Now
u(re
iθ
) −f(e
it
0
) =
1
2π
t
0
+π
t
0
−π
P
r
(θ −t)[f(e
it
) −f(e
it
0
)]dt,
so it suﬃces to show that the last integral can be made arbitrarily small by taking θ
suﬃciently close to t
0
and r < 1 suﬃciently close to 1. Now given > 0 we can choose
δ > 0 so that f(e
it
) −f(e
it
0
) ≤ if t −t
0
 ≤ δ. Split the interval of integration as
{t : t −t
0
 ≤ π} = I
1
∪ I
2
:= {t : t −t
0
 ≤ δ} ∪ {t : δ ≤ t −t
0
 ≤ π}.
On I
1
, the integrand is bounded by P
r
(θ −t), so the integral is bounded by
2π
I
1
P
r
(θ −t)dt ≤
2π
2π
0
P
r
(θ −t)dt ≤ .
Now suppose that θ is taken suﬃciently near t
0
, namely θ −t
0
 ≤ δ/2. Then for t ∈ I
2
,
θ −t ≥ t −t
0
 −θ −t
0
 ≥ δ/2, θ −t ≤ t −t
0
 +θ −t
0
 ≤ π + δ/2
Thus P
r
(θ −t) → 0 uniformly for t ∈ I
2
as r ↑ 1. Since f is bounded, the entire integrand
does so as well. Thus if θ −t
0
 ≤ δ/2 we may take r suﬃciently close to 1 that the integral
over I
2
is bounded by .
It is now easy to deduce that the mean value property characterizes harmonic functions.
34
Theorem. If u : G →R is a continuous function satisfying the mean value property, then
it is harmonic.
Proof. It is enough to show that u is harmonic on each disk with closure in G. From the
Poisson kernel construction we can construct a harmonic function with the same boundary
values as u on this disk. The diﬀerence between this function and u satisﬁes the mean
value property and is zero on the boundary of the disk, so it vanishes on the disk.
Corollary. If a sequence of harmonic functions converges uniformly on compact subsets,
then the limit function is harmonic.
From the expressions
P
r
(θ) =
1 −r
2
(1 + r)
2
−2r(1 + cos θ)
=
1 −r
2
(1 −r)
2
+ 2r(1 −cos θ)
we have
1 −r
1 + r
≤ P
r
(θ) ≤
1 + r
1 −r
.
Now suppose u is harmonic on B(0, R) and nonnegative, and re
iθ
is an arbitrary point
of B(0, R). Choose > 0 small enough that r < R −, and set U(z) = u((R −)z). U is
harmonic on a neighborhood of
¯
B(0, 1), so
u(re
iθ
) = U([r/(R −)]e
iθ
) =
1
2π
2π
0
P
r/(R−)
(θ −t)U(e
it
) dt.
Using the nonnegativity of u, the mean value property, and the upper bound for P
r
, we
get
u(re
iθ
) ≤
1 + r/(R −)
1 −r/(R −)
u(0).
Letting ↓ 0 this gives
u(re
iθ
) ≤
R + r
R −r
u(0).
We obtain a lower bound in a similar fashion. Of course we can translate the disk so it is
centered at an arbitrary point a. Thus we have proven:
Theorem (Harnack’s Inequalities). If u is nonnegative and harmonic on B(a, R),
then
R −r
R + r
u(a) ≤ u(a + re
iθ
) ≤
R + r
R −r
u(a)
for 0 ≤ r < R.
For example, if u is harmonic and nonnegative on the disk of radius R about a then
u(z) ∈ [u(a)/3, 3u(a)] on the disk of radius R/2 about a.
35
Theorem (Harnack). Let 0 ≤ u
1
≤ u
2
≤ . . . be a sequence of harmonic functions on a
region G. Then either lim
n
u
n
(z) = +∞ uniformly on compact subsets of G or there is an
harmonic function u with lim
n
u
n
(z) = u(z) uniformly on compact subsets of G.
Proof. For any point a ∈ G let D be a disk centered at a contained in G. Then there
exists a constant C > 1 with
u
n
(z) ≤ Cu
n
(a), u
n
(a) ≤ Cu
n
(z), z ∈ D.
Consequently, the numbers u
n
(a) remain bounded as n → ∞ if and only if u
n
(z) remains
bounded for all z ∈ D. This shows that the set of points where u
n
stays bounded is both
open and closed, so it is either the empty set or all of G. In the ﬁrst case, u
n
(z) → +∞
for all z ∈ G, and the above estimate shows that the convergence is uniform on a covering
set of discs, hence on all compact subsets.
In the second case, lim
n
u
n
(a) ∈ C for all a, and since
u
n
(z) −u
m
(z) ≤ C[u
n
(a) −u
m
(a)], n ≥ m, z ∈ D,
the sequence is uniformly Cauchy on D so converges uniformly on D. Again, uniform
convergence on compact sets follows.
Subharmonic functions and the solution of the Dirichlet problem.
Deﬁnition. Let u be a continuous realvalued function on a region G. If
u(a) ≤
1
2π
2π
0
u(a + re
iθ
) dθ
whenever
¯
B(a, r) ⊂ G, we say that u is subharmonic.
As mentioned earlier, a subharmonic function satisﬁes the maximum principle. Other
simple properties are:
• harmonic functions are subharmonic
• the sum of two subharmonic functions is subharmonic
• the pointwise maximum of two subharmonic functions is subharmonic
The next lemma states that subharmonicity is a local property:
Lemma. Suppose that u is subharmonic in a neighborhood of each point of a region G.
Then u is subharmonic on G.
Proof. If
¯
D =
¯
B(a, r) ⊂ G, let ˜ u be the continuous function on
¯
D which agrees with u on
∂D and is harmonic on D. From the Poisson kernel representation, we have
˜ u(a) =
1
2π
2π
0
u(a + re
iθ
) dθ.
Now u − ˜ u is subharmonic in a neighborhood of each point of D, and it follows by the
usual connectedness argument that the points where it achieves its maximum on D is
either empty or all of D. Hence its maximum on
¯
D is achieved on the boundary, so is 0.
This shows that u(a) ≤ ˜ u(a) as desired.
One more property we shall use:
36
Lemma. Suppose that u is subharmonic in a region G and
¯
D =
¯
B(a, r) ⊂ G. Deﬁne ˜ u
on
¯
D as the continous function on
¯
D which agrees with u on ∂D and is harmonic on D,
and set ˜ u = u on G\
¯
D. Then ˜ u is subharmonic on G.
Proof. Clearly it is continuous and subharmonic in D and G \
¯
D. So it suﬃces to show
that for each point a of ∂D, ˜ u is bounded by its mean on any circle around a. Now u − ˜ u
is subharmonic on D and 0 on ∂D, so u ≤ ˜ u on D, and therefore on G. Thus
˜ u(a) = u(a) ≤
1
2π
2π
0
u(a + re
iθ
) dθ ≤
1
2π
2π
0
˜ u(a + re
iθ
) dθ.
Note that
Let G be a bounded region. If u is a harmonic function on G continuous up to the
boundary, and v is any subharmonic function that is less than u on the boundary, then
the maximum principle implies that v ≤ u in G. Hence it is reasonable to try to solve
the Dirichlet problem by seeking the largest subharmonic function that doesn’t exceed the
given boundary values. This is the approach of O. Perron.
Let f be a bounded (at this point not necessarily continuous) realvalued function on
∂G. Deﬁne the Perron family P(f, G) as the set of subharmonic functions v on G for which
() limsup
z→a
v(z) ≤ f(a) for all a ∈ ∂G.
Note that P(f, G) contains all constants not exceeding min f, so it is not empty.
From the maximum principle we obtain:
Lemma. If v ∈ P(f, G) then v ≤ sup f on G.
Now deﬁne the Perron function
u(z) =
sup
v∈P(f,G)
v(z), z ∈ G,
f(z), z ∈ ∂G.
We will show that (1) the Perron function is harmonic on G, and (2) under mild restrictions
on G, it extends continuously to
¯
G with value f on ∂G. Thus we will have exhibited a
solution to the Dirichlet problem.
Proof that the Perron function is harmonic. Let D be any disk with closure in G and
let a be any point in the disk. Choose functions v
n
∈ P(f, G) with v
n
(a) → u(a). Let
V
n
= max(v
1
, . . . , v
n
), and then deﬁne
˜
V
n
to be V
n
oﬀ D and to be harmonic on D and
continuous on
¯
D. We know that V
n
and
˜
V
n
are subharmonic, and clearly the V
n
and thus
the
˜
V
n
are nondecreasing. All belong to P(f, G) so are bounded by u, but v
n
≤ V
n
≤
˜
V
n
,
so
˜
V
n
(a) ↑ u(a). By Harnack’s Theorem,
˜
V
n
converges to an harmonic function V on D.
37
We know that V (a) = u(a). We will now show that if b is any other point of D then
V (b) = u(b). This will show that u is harmonic in D as desired.
Choose functions φ
n
∈ P(f, G) with φ
n
(b) → u(b), and set w
n
= max(φ
n
, v
n
) ∈ P(f, G).
Construct W
n
and
˜
W
n
from w
n
in analogy with V
n
and
˜
V
n
. Then w
n
≥ v
n
, W
n
≥ V
n
,
˜
W
n
≥
˜
V
n
, and W
n
converges to a harmonic function W ≥ V with W(b) = u(b). Also
W(a) = u(a) = V (a), so V −W is a nonpositive harmonic function on D which achieves
its maximum at a. Therefore V ≡ W on D, so indeed V (b) = u(b).
We now turn to the continuity of the Perron solution at the boundary.
Deﬁnition. Let G be a bounded region and z
0
∈ ∂G. A continuous function φ on
¯
G
which is harmonic on G is called a barrier function for G at z
0
if it is positive on
¯
G\ {z
0
}
and zero at z
0
.
For example, suppose that there is a line through z
0
such that
¯
G\ {z
0
} lies inside one
of the open halfplanes determined by the line. We may write the halfplane as Im[(z −
z
0
)/b] > 0 for some b ∈ C, and so Im[(z −z
0
)/b] is a barrier function for G at z
0
.
Next consider the function
1 −1/z where we use the principal branch of the square
root, deﬁned on C\(−∞, 0]. This function is analytic on C\[0, 1] and extends continuously
to 1 with value 0, and its real part is everywhere positive on C \ [0, 1]. Thus if z
0
= 1
belongs to ∂G, and is the only point of intersection of the interval [0, 1] with
¯
G, we
can use Re
1 −1/z as a barrier function for G at z
0
. Similarly, if z
0
is any boundary
point for which there exists a line segment which intersects
¯
G only at z
0
, then there is a
barrier function for G at z
0
. Thus barrier functions exist at all boundary points of smooth
domains, polygonal domains, even on a disk minus a segment, and on many other domains.
A standard example of a domain without a barrier at some boundary point is a punctured
disk.
Theorem. Let G be a bounded domain, f a bounded function on ∂G, and u the corre
sponding Perron function. If z
0
is a point in ∂G possessing a barrier function and f is
continuous at z
0
, then lim
z→z
0
u(z) = f(z
0
).
Proof. First we show that for any > 0, there is a harmonic function w on G, continuous
on
¯
G, which satisﬁes w > f on ∂G and w(z
0
) = f(z
0
) + . Indeed, let D be a disk about
z
0
such that f(z) −f(z
0
) < for z ∈ ∂G∩ D. Let M denote the maximum of f on ∂G
and m denote to the minimum of the barrier function φ on G\ D, so m > 0. Set
w(z) =
φ(z)
m
[M −f(z
0
)] + f(z
0
) + =
¸
φ(z)
m
−1
[M −f(z
0
)] + M + .
Then for z ∈ ∂G∩ D,
w(z) ≥ f(z
0
) + > f(z),
while for z ∈ ∂G\ D,
w(z) > M ≥ f(z).
38
So w > f on the boundary as claimed, and obviously w(z
0
) = f(z
0
) + .
It follows that if v ∈ P(f, G), then v ≤ w on G. Since this is true of all such v, we
have u ≤ w on G, and so limsup
z→z
0
u(z) ≤ w(z
0
) = f(z
0
) + . Since was arbitrary,
limsup
z→z
0
u(z) ≤ f(z
0
).
Next let
v(z) = f(z
0
) − −
φ(z)
m
[M + f(z
0
)] =
¸
1 −
φ(z)
m
[M + f(z
0
)] −M −.
Now we get v < f on ∂G and v(z
0
) = f(z
0
) − . Since v ∈ P(f, G), u ≥ v on G, and so
liminf
z→z
0
u(z) ≥ f(z
0
) −. Letting tend to zero we have liminf
z→z
0
u(z) ≥ f(z
0
).
Corollary. Let G be a bounded region in C which possesses a barrier at each point of its
boundary, and let f be a continuous function on ∂G. Then the Perron function for f on
G solves the Dirichlet problem.
The Schwarz Reﬂection Principle. To state the reﬂection principles, we introduce
the following terminology: a region G is symmetric with respect to the real axis if z ∈ G
implies ¯ z ∈ G.
Reﬂection Principle for Harmonic Functions. Let G be a region which is symmetric
with respect to the real axis and deﬁne G
+
, G
−
, and G
0
as the intersection of G with the
upper halfplane, lower halfplane, and real axis, respectively. If u is a continuous real
valued function on G
+
∪ G
0
, which is harmonic on G
+
and zero on G
0
, then u admits a
unique extension to a harmonic function on all of G. The extension is given by u(z) =
−u(¯ z) for z ∈ G
−
.
Proof. If such an extension exists it is certainly unique, so it suﬃces to show that the stated
extension deﬁnes a harmonic function in G. It certainly deﬁnes a continuous extension
which is harmonic in G
+
∪G
−
, so it suﬃces to show that it is harmonic in a neighborhood
of z
0
∈ G
0
. By translating and extending we can assume that z
0
= 0 and that G contains
the closed unit disk,
¯
D. Deﬁne U on
¯
D as the solution of the Dirichlet problem with
boundary data u
∂D
(where u has been extended to G
−
by the formula above). From the
Poisson kernel representation we have U = 0 on the real interval (−1, 1). Thus U = u on
the entire boundary of the upper halfdisk, and the entire boundary of the lower halfdisk.
Since both U and u are harmonic in the halfdisks, they coincide, and thus u, like U is
harmonic in the whole disk.
Reﬂection Principle for Analytic Functions. Let G, G
+
, G
−
, and G
0
be as above.
If f is a continuous complexvalued function on G
+
∪ G
0
, which is analytic on G
+
and
real on G
0
, then f admits a unique extension to a harmonic function on all of G. The
extension is given by f(z) = f(¯ z) for z ∈ G
−
.
Proof. We could base a proof on the previous result applied to Imf and harmonic conju
gates, but it is also easy to verify this directly using Morera’s theorem.
39
Of course the line of symmetry could be any line, not just the real axis. Thus if G is
symmetric with respect to any line, deﬁned and analytic on one side of the line, and real
on the line, it can be extended to be analytic on the whole domain. The proof can be
found by translating and rotating the domain to the standard case.
We can also translate and rotate the image, so it is not necessary that f be real on the
symmetry line, it is suﬃcient that it map it into some other line. An important application
is to analytic continuation. For example, if f maps a rectangle into itself analytically and
takes the edges into the edges, we can apply the reﬂection principle repeatedly to extend
f to an entire function.
The reﬂection principles can be applied for symmetries with respect to circles as well
as for lines.
Theorem. Let f be analytic on the unit disk, continuous on its closure, and real on the
boundary. Then f admits a unique extension to the entire plane. The extension is given
by f(z) = f(1/¯ z) for z > 1.
Proof. Let g(z) = f((1 + iz)/(1 − iz)), so g maps the upper half plane into C and is real
on the real axis. Applying the reﬂection principle to g gives an extension of f with
f
1 + iz
1 −iz
= f
1 + i¯ z
1 −i¯ z
for z in the lower half plane, or, equivalently,
f(w) = f(1/w)
for w in the exterior of the disk.
As an application of the reﬂection principle consider an analytic function on the unit
disk which extends continuously to the closure taking the unit circle into itself. We may
extend this function to C
∞
by f(z) = 1/f(1/¯ z) for z > 1. (We obtain this by applying
the standard reﬂection principle to g = φ
−1
◦ f ◦ φ where φ(z) = (1 + iz)/(1 − iz).) The
extended function is analytic everywhere except at the points symmetric to the zeros of
f, where it has a pole. Inﬁnity is either a removable singularity or a pole according to
whether f is 0 at 0 or not. Thus the extended function is meromorphic on C
∞
with a
ﬁnite number of poles. It follows that f is rational.
2
I. The Complex Number System R is a ﬁeld. For n > 1, Rn is a vectorspace over R, so is an additive group, but doesn’t have a multiplication on it. We can endow R2 with a multiplication by (a, b)(c, d) = (ac − bd, bc + ad). Under this deﬁnition R2 becomes a ﬁeld, denoted C. Note that (a/(a2 + b2 ), −b/(a2 + b2 )) is the multiplicative inverse of (a, b). (Remark: it is not possible to endow Rn with a ﬁeld structure for n > 2.) We denote (0, 1) by i and identify x ∈ R with (x, 0), so R ⊂ C. Thus (a, b) = a + bi, a, b ∈ R. Note that i2 = −1. C is generated by adjoining i to R and closing under addition and multiplication. It is remarkable that the addition of i lets us not only solve the equation x2 + 1 = 0, but every polynomial equation. For a and b real and z = a + bi we deﬁne Re z = a, Im z = b, z = a − bi, and ¯ 2 2 1/2 z = (a + b ) . Then Re z = (z + z )/2, ¯ Im z = (z − z )/(2i), ¯ z ¯ 1 = 2, z2 = z z , ¯ z z z ± w = z ± w, zw = z w, ¯ ¯ ¯¯ z + w ≤ z + w.
z/w = z /w, ¯ ¯
The map θ → (cos θ, sin θ) deﬁnes a 2πperiodic map of the real line onto the unit circle in R2 . In complex notation this map is θ → cis θ := cos θ + i sin θ. Every nonzero complex number can be written as r cis θ where r > 0 is uniquely determined and θ ∈ R is uniquely determined modulo 2π. The number 0 is equal to r cis θ where r = 0 and θ is arbitrary. The relation z = r cis θ determines the relations z → r which is simply the function r = z and z → θ. The latter is denoted θ = arg θ. Note that for z = 0, arg θ is determined modulo 2π (while arg 0 is arbitrary). We can normalize arg by insisting that arg z ∈ (−π, π]. Note that if z1 = r cis θ1 and z2 = r cis θ2 then z1 z2 = r1 r2 cis(θ1 + θ2 ). The latter formula just encapsulates the formula for the sine and cosine of a sum, and gives arg z1 z2 = arg z1 + arg z2 . In particular, ir cis θ = r cis(θ + π/2), so multiplication by i is just the operation of rotation by π/2 in the complex plane. Multiplication by an arbitrary complex number r cis θ is just rotation by arg θ followed by (or preceded by) dilation by a factor r. Further, z n =√n cis(nθ). Every nonzero z ∈ C admits n distinct nth roots: the r nth roots of r cis θ are n r cis[(θ + 2πk)/n], k = 0, 1, . . . , n. Lines and circles in the plane. Circles given by z − a = r where a ∈ C is the center and r > 0 is the radius. If 0 = b ∈ C then the line through the origin in the direction b is the set of all points of the form tb, t ∈ R, or all z with Im(z/b) = 0. If t ∈ R and c > 0 then (t + ci)b = tb + cib represents a point in the half plane to the left of b determined by the line tb, i.e., {z : Im(z/b) > 0} is the equation of that halfplane. Similarly, {z : Im[(z − a)/b] > 0} is the translation of that halfplane by a, i.e., the halfplane determined by the line through a parallel to b and in the direction to the left of b.
3
Stereographic projection determines a onetoone correspondence between the unit sphere in R3 minus the northpole, S, and the complex plane via the correspondence z↔ x1 = 2 Re z , 1 + z2 x1 + ix2 , 1 − x3 2 Im z x2 = , 1 + z2
x3 =
z2 − 1 . z + 1
If we deﬁne C∞ = C ∪ {∞}, then we have a onetoone correspondence between S and C∞ . This allows us to deﬁne a metric on C∞ , which is given by d(z1 , z2 ) = 2z1 − z2  (1 + z1 2 )(1 + z2 2 ) , d(z, ∞) = 2 1 + z2 .
II. Elementary Properties and Examples of Analytic Functions
n N +1 For z = 1, )/(1 − z). Therefore the geometric series n=0 z n n=0 z = (1 − z converges (to 1/(1 − z)) if z < 1. It clearly diverges, in fact its terms become unbounded, if z > 1. N ∞
Weierstrass MTest. Let M0 , M1 , . . . be positive numbers with Mn < ∞ and suppose that fn : X → C are functions on some set X satisfying supx∈X fn (x) ≤ Mn . Then ∞ n=0 fn (x) is absolutely and uniformly convergent. Theorem. Let a0 , a1 , · · · ∈ C be given and deﬁne the number R by 1 = lim sup an 1/n . R
n Then (1) for any a ∈ C the power series n=0 an (z − a) converges absolutely for all z − a < R and it converges absolutely and uniformly on the disk z − a ≤ r for all r < R. (2) The sequence an (z − a)n is unbounded for all z − a > R (and hence the series is certainly divergent). ∞
Thus we see that the set of points where a power series converges consists of a disk z − a < R and possibly a subset of its boundary. R is called the radius of convergence of its series. The case R = ∞ is allowed. Proof of theorem. For any r < R we show absolute uniform convergence on Dr = {z −a ≤ r}. Choose r ∈ (r, R). Then, 1/˜ > lim sup an 1/n , so an 1/n < 1/˜ for all n suﬃciently ˜ r r n large. For such n, an  < 1/˜ and so r sup an (z − a)n  < (r/˜)n . r
z∈Dr
Since
(r/˜)n < ∞ we get the absolute uniform convergence on Dr . r
If z − a = r > R, take r ∈ (R, r). Then there exist n arbitrarily large such that ˜ 1/n ≥ 1/˜. Then, an (z − a)n  ≥ (r/˜)n , which can be arbitrarily large. r r an 
then f (a) exists and equals 1/g (f (a)). and consider its formal ∞ ∞ derivative n=1 nan (z−a)n−1 = n=0 (n+1)an+1 (z−a)n . by the previous theorem. and g(f (z)) = z for all z in a neighborhood of a. On the circle of convergence. This shows that lim an /an+1  ≤ R. and so the partial sums on the left and right either both diverge for a given z or both converge. This shows that the formal derivative of a power series has the same radius of convergence as the original power series. but the simplest algebraic rules are the same. Sierpinski gave a (complicated) example of a function which diverges at every point of the unit circle except z = 1. ∈ C and lim an /an+1  exists as a ﬁnite number or inﬁnity. Thus the series has terms of decreasing magnitude. then lim sup bn cn = bc. Deﬁnition of diﬀerentiability at a point (assumes function is deﬁned in a neighborhood of the point). First of all. and so cannot be convergent. Thus the series an z n has terms of increasing magnitude. and quotient rules hold.4 Theorem. Suppose that z > lim an /an+1 . f /g. we see that the series zn n! n=0 converges absolutely for all z ∈ C and that the convergence is uniform on all bounded sets. Remark. . Now suppose that n=0 an (z − a)n has radius of convergence R. then g ◦ f is diﬀerentiable at a and the chain rule holds. and so. f · g. z n /n2 converges absolutely on z ≤ 1. Thus lim sup (n + 1)an+1 1/n = 1/R. z ≤ R. by deﬁnition. Diﬀerentiability and analyticity. a1 . exp z. Suppose that f is continuous at a. . diﬀerentiability at a point implies continuity there. If f is diﬀerentiable at a and g is diﬀerentiable at f (a). Then for all n suﬃciently large an  > an+1 z and an z n  > an+1 z n+1 . suppose that z < lim an /an+1 . product. Then if g (f (a)) exists and is nonzero. Then for all n suﬃciently large an  < an+1 z and an z n  < an+1 z n+1 . if g(a) = 0. Moreover. Most of the consequences of diﬀerentiability are quite diﬀerent in the real and complex case. Now clearly n=0 an+1 (z−a)n ∞ has the same radius of convergence as n=0 an (z − a)n since N N +1 ∞ ∞ (z − a) n=0 an+1 (z − a) = n=0 n an (z − a) − a0 . Proof. If a0 . If f and g are both diﬀerentiable at a point a. Now lim(n+1)1/n = 1 as is easily seen by taking logs. z /n diverges for z = 1. This shows (in a roundabout way) that lim sup an+1 1/n = lim sup an 1/n = 1/R. As an application. and the usual sum. it is easy to see that if lim sup bn = b and lim cn = c > 0. but not absolutely (this follows from the fact that the partial sums of z n are bounded for z = 1 and 1/n ↓ 0). This shows that lim an /an+1  ≥ R. Thus z ≥ R. and. . then this limit is the radius of convergence R of an (z − a)n . else converges. Similarly. many diﬀerent behaviors are possible. g is continous at f (a). with the same proofs. . Without loss of generality we can suppose that a = 0. then so are f ± g. z n diverges n for all z = 1. The sum is.
a1 = f (0). This shows that any convergent power series is the sum of its Taylor series in the disk of convergence: f (z) = In particular. a2 = f (0)/2. Let sN (z) = f (z) − f (z0 ) sN (z) − sN (z0 ) − g(z0 ) ≤ − sN (z0 ) z − z0 z − z0 RN (z) − RN (z0 ) =: T1 + T2 + T3 . we have for N suﬃciently large and all z < r then T3 < /3. For simplicity. then. we have an n z n − z0 n−1 = an z n−1 + z n−2 z0 + · · · + z0  ≤ an nrn−1 . a3 = f (0)/3!. the inequality f (z) − f (z0 ) − g(z0 ) ≤ z − z0 is satisﬁed for all z suﬃciently close to z0 . Thus a0 = f (0). Then f is said to be analytic on G if f exists and is continuous at every point of G. Then the diﬀerentiability of the polynomial sN shows that T1 ≤ /3 for all z suﬃciently close to z0 . f n (a) (z − a)n . exp = exp. z − z0 z − z0 n=N +1 Now z0  < r < R for some r. If Q is an arbitrary nonempty subset of C we say f is analytic on Q if it is deﬁned and analytic on an open set containing Q.5 Deﬁnition. ∞ n RN (z) − Rn (z0 ) z n − z0 = an . etc. it follows that a power series is analytic and inﬁnitely diﬀerentiable in its convergence disk. f (z) = n(n−1)an z n−2 . Also. We shall prove later that if f is diﬀerentiable at every point of an open set in C it is automatically analytic. T2 ≤ /3. z − z0 Since an nrn−1 is convergent. This is of course vastly diﬀerent from the real case. it is automatically inﬁnitely diﬀerentiable. in fact. Since we know that that power series has the same radius of convergence. and without loss of generality we consider a power series centered at zero: f (z) = n an z n . n! . We now show that a power series is diﬀerentiable at every point in its disk of convergence and that its derivative is given by the formal derivative obtained by diﬀerentiating termbyterm. and if we restrict to z < r. Suppose that the radius of convergence is R and that z0  < R. Remark. f (z) = n−1 nan z . and by induction. We thus know that if f (z) = an z n . Let f be a complexvalued function deﬁned on an open set G in C. etc. + sN (z0 ) − g(z0 ) + z − z0 Now sN (z0 ) is just a partial sum for g(z0 ). RN (z) = n=N +1 an z . within the disk of convergence. Then ∞ n=1 nan z n−1 . where g(z) = N ∞ n n n=0 an z . We must show that for any > 0. Now ﬁx a value of N which is suﬃciently large by both criteria. so for N suﬃciently large (and all z).
If b is an integer. Then we can use connectedness to see that the set of points where it takes a given value is open and closed. Deﬁne cos z and sin z. This says that an analytic mapping is conformal. There is no solution to exp z = 0. At this point. ˜ → C be two smooth curves. Deﬁnition.mps. such that γ(a) = γ (˜) = z. If x = 0.) The conformality of an analytic map arises from the fact that in a neighborhood of a point z0 such a map behaves to ﬁrst order like multiplication by f (z0 ). and hence its argument is the angle between the (tangent to the) curve and the horizontal. exp z = exp(Re z) cis(Im z). Use this to show that exp(z) exp(a − z) ≡ exp a. ˜ a b] ˜ a ˜ a Then arg γ (˜)−arg γ (a) measures the angle between the curves. γ (˜) = 0. (Note: if γ (t) = 0.mps and square2. b] → C and γ : [˜.. By contrast. since we can restrict to segments and use the real result. Consequently diﬀerent branches of z b diﬀer by factors of exp(2πinb).g. all values agree. of ρ(t) = f (γ(t)).) Let γ : [a. e. then it is a vector tangent to the curve at γ(t). Since f (z0 ) = r exp iθ. i. get cos2 + sin2 ≡ 1. it is constant on disks. the curve may not have a tangent. A branch of the logarithm gives a branch of z b := exp(b log z) (understood to be the principal branch if not otherwise noted). Examples: C \ R− (principal branch). Now consider the images ˜ a under f . Then ρ (a) = f (z)γ (a). Indeed. whenever its derivative is not zero. n ∈ Z. b] → C. . The image will be two smooth curves meeting at a common point. The Logarithm.g. the most general solution of exp z = w is z = log w+i arg w + 2πin.g. show the computer graphics square1. Deﬁne a path in the complex plane as a continuous map γ : [a. If G is an open set and f : G → C is a continuous function satisfying exp(f (z)) = z. C \ R+ .6 Lemma. A function with vanishing derivative on a region (connected open set) is constant. and γ (a). γ(t) = t2 − it3 . We wish to consider how this angle is transformed under a complex map. E. a branch of the logarithm is analytic with derivative 1/z. e. Therefore ˜ a γ a arg ρ (˜) − arg ρ (a) = arg γ (˜) + arg f (z) − arg γ (a) − arg f (z). the angle is invariant ˜ a ˜ a in both magnitude and sign. ρ (˜) = f (z)˜ (˜). If b is a rational number with denominator d there are d values. a spiral strip.. a smooth map from R2 → R2 can behave like an arbitrary linear operator in a neighborhood of a point. so we have to deﬁne the angle between two such curves. If γ (t) ∈ C exists at some point and is not zero. Note that diﬀerent branches of the logarithm assign values to log z that diﬀer by addition of 2πin.e. By the formula for the derivative of an inverse.. Note that eb = exp b. then f is called a branch of the logarithm on G. z 2 doubles angles at the origin.. Consider the angle between two line segments with a common vertex in the complex plane. Conformality. cis θ = exp(iθ). exp z = cos z + i sin z. (Not true if the derivative is zero. near z0 f simply behave like rotation by θ followed by dilation.
which. ∂z ∂x Thus. y) − f (x. then f is analytic on Ω. ∂f 1 := ∂z ¯ 2 ∂f ∂f +i ∂x ∂y . the Cauchy–Riemann equations are exactly equivalent to ∂f = 0. y) ∂f = lim = . ∂x ∂y ∂v ∂u =− . then the partial derivaties of f exist and are continuous on Ω. y) ∂f f (z + ih) − f (z) = lim −i = −i . If the partial derivatives of f exist and are continuous on Ω. y + h) − f (x. h∈R ih h ∂y h→0 Thus complexdiﬀerentiability of f at z implies not only that the partial derivatives of f exist there. Abuse notation slightly by writing f (x.) In terms of these. and the CauchyRiemann equations are satisﬁed there. if z and z ¯ ¯ ¯ were independent variables. (These are motivated by the equations x = (z + z )/2. ∂x ∂y Another convenient notation is to introduce ∂f 1 := ∂z 2 ∂f ∂f −i ∂x ∂y . ∂y/∂z = −i/2. h∈R h h ∂x h→0 h∈R h→0 and f (z) = lim h∈R h→0 f (x. etc. would give ∂x/∂z = 1/2. y) as an alternative for f (x + iy). The converse is true as well: Theorem. but also that they satisfy the Cauchy–Riemann equation ∂f ∂f = −i . y = (z − z )/(2i). ∂x ∂y If f = u + iv. then this equation is equivalent to the system ∂u ∂v = .7 Cauchy–Riemann Equations. If f (z) exists for some z = x + iy ∈ Ω. then f (z) = lim f (z + h) − f (z) f (x + h. if f is analytic on Ω. Let f : Ω → C with Ω ⊂ C open. ∂z ¯ which is also equivalent to ∂f ∂f = . . and the CauchyRiemann equations are satisﬁed there.
8 Proof. y + t∗ ) . y + t) − f (x. y + t) − (x. then γ is the boundary of a subset G of . while not analytic. which is independent of h. ∂z ∂ z ¯ ∂ z ∂z ¯ 4 This shows that any analytic function is harmonic (equivalently. y) → 0 ∂x ∂x as h → 0. y) + (x. then u is harmonic. is harmonic. does there exist a harmonic conjugate. f (z + h) − f (z) f (x + s. Note that ∂f ∂f (x + s∗ . for a proof. y + t) s + = s s + it t s + it ∂f s ∂f t = (x + s∗ . This is the Looman–Menchoﬀ Theorem. h ∂x s + it ∂y s + it where limh→0 R(h) = 0. by the mean value theorem. then the the ﬁrst two terms on the right hand side sum to ∂f /∂x(x. y) + R(h). ∂x s + it ∂y s + it where s∗  < s and t∗  < t. y). Thus f (z + h) − f (z) ∂f s ∂f t = (x. if there exists v it is determined up to addition of a real constant (or f is determined up to addition of an imaginary constant).e. Complex Analysis in One Variable. The book gives an elementary construction of the harmonic conjugate in a disk and in the whole plane. See Narasimhan. if u is harmonic on a region Ω. Remark. y + t) − f (x. so the limit exists and is equal to ∂f /∂x(x. Now if the Cauchy–Riemann equations hold. since otherwise we could take f to be the characteristic function of the coordinate axes. If u(x. y) t f (x + s. a function v on Ω such that f = u + iv is analytic? Clearly. Let z = x + iy ∈ Ω. y + t) + (x. y). Conversely. then the complex derivative of f exists on Ω (which is equivalent to f being analytic on Ω. Moreover s/(s + it) stays bounded (by 1). We do need at least continuity. i.. It also shows that the conjugate of an analytic function. We sketch the idea of a more general proof. and similarly for the second partial derivative. We require that the domain satisfy the condition that if γ is any piecewise simple smooth closed curve in Ω. Note that ∂ ∂ ∂ ∂ 1 = = ∆. its real and imaginary parts are harmonic). y) = Re f (x + iy). The theorem can be weakened to say that if f is continuous on Ω and the partial derivatives exist and satisfy the Cauchy–Riemann equations there (without assuming that the partial derivatives are continuous). We must show that f (z) exists. Let r be small enough that the disk of radius r around z belongs to Ω and choose h = s + it with 0 < s + it < r. Then. y) = h s + it f (x. y + t) − f (x.
by the divergence theorem. y0 ) towards (x. −f1 ) · n ds = · (f2 . But f · τ ds = γ γ (f2 .g. b. If S and T are MTs. To prove the “if” part. d be complex numbers with ad = bc. o Remarks. c. It is easy to check that For a nonsimply connected region. a b To each MT we associate the nonsingular matrix of its coeﬃcients. In other words. This is evident from the az + b . y) ∈ Ω. It is essential that this quantity doesn’t depend on the choice of path. but cannot be extended to an analytic function on C \ {0}. 3) The linear (but nonconstant) polynomials are MTs. 1) Note that if ad = bc the same expression would yield a constant. there may exist no harmonic conugate. and let τ be the unit tangent to the path pointing from (x0 . Deﬁne v(x. let γ(x. With the understanding that S(∞) = a/c and S(−d/c) = ∞ (or. namely the transform with the inverse coeﬃcient matrix. o Deﬁnition. E. 5) We can deﬁne S(−d/c) = ∞ and S(∞) = a/c. −f1 ) dx dy = 0. if u = log z on C \ {0}. so S can be viewed of as a map from C∞ into itself (which is. y0 ) in Ω and for any (x. then so is S ◦ T . 2) The coeﬃcents aren’t unique. which is c d determined up to a nonzero multiple. Any MTs admits an inverse. cz + d z ∈ C. which holds on simplyconnected domains: let f : Ω → R2 be a C 1 vectorﬁeld.y) be a piecewise smooth path in Ω from (x0 . as we will now see. 11 and onto). then u + i arg z is analytic on C \ {z ≤ 0}. since we can multiply them all by any nonzero complex constant. Then there exists v : Ω → R such that f = v if and only if ∂f1 /∂y = ∂f2 /∂x. if c = 0. that f · τ ds = 0 γ for all piecewise smooth simple closed paths γ. the MT maps C∞ 11 onto itself. namely the ones for which c = 0. 4) If z = −d/c then S(z) ∈ C and S (z) = (ad − bc)/(cz + d)2 = 0. ﬁx a point (x0 . Our proof follows directly from the following result.y) f · τ ds. Let a. equivalently.. M¨bius transformations. y).9 Ω. or. y0 ) to (x. y). y) = γ(x. its coeﬃcient matrix being the product of the coeﬃcient matrices of S and T . then S(∞) = ∞. The “only if” part is obvious. Ω has no holes (is simplyconnected). v = f. . z = −d/c. Then the mapping S(z) = is a M¨bius transformation.
Since the same property is evident for translation. z4 to 1. z3 . we can map one to the other by a MT. Substituting 1/z for z and doing some simple manipulations gives the equation z − 1/p = (kq/p)z − 1/q. Since the geometric action of translation. there is a unique MT taking the ﬁrst onto the second. Then it is easy to see that there is a unique MT taking these points to 1. so z1 − z3 z2 − z4 (z1 . with respect to the image orientation. If S is a nonlinear MT we can write it as b − ad az + b =a+ . Consider the equation z − p = kz − q where k > 0. Example: z → (1 − z)/(1 + z) takes the right half plane onto the unit disk. z4 ) for any MT T . which will be omitted. z3 . z4 ) = . p. We now show that inversion maps lines and circles to other lines and circles. 0. z2 . respectively. which is another such circle. namely z − z3 z2 − z4 . z4 be distinct points in C. which we call C1 .10 existence of its inverse S −1 (z) = (dz − b)/(−cz + a). the composition of M¨bius o transformations is again one. Moreover. z4 ) (cross ratio) for the image of z1 under the transformation taking z2 . so they form a group under composition. the region to the left of the circle will map to the region to the left of the image circle. z3 . . and rotation. The book makes heavy use of the the notation (z1 . (T z1 . 2) A MT preserves the property of two points being symmetric with respect to a circle. Sz = z − z4 z2 − z3 We can also handle. We infer from this and the invertibility of the MTs that given two ordered triples of distinct points in C∞ . respectively. i. z2 . dilation. This is (x − a)2 + (y − b)2 = k 2 [(x − c)2 + (y − d)2 ] which is clearly the equation of a circle if k = 1 and the equation of a line if k = 1. and rotation are quite evident. showing that inversion maps circles in the sphere onto other circles in the sphere. z2 . and such that the geometric mean of their distances from the center equals the radius. Thus we have 1/z ∈ C1 ⇐⇒ z ∈ C2 . the possibility that one of the zi is ∞. and can arrange that the image of any three points on the boundary of one go to three points on the boundary of the other. Let z2 . and ∞. rotations and dilations (multiplication by a complex number). z1 . 1) A MT is orientiation preserving in the sense that.e. or one of each. z1 − z4 z2 − z3 From the deﬁnition. 0. if we traverse a circle in the order of three distinct points on it. dilation. z3 . z2 . lying on the same ray from the center. Some further analysis. In either case we have a circle in the Riemann sphere. z3 . we conclude that this property holds for all MT. establishes two further properties. T z2 . and inversion (reciprocals). C2 . It takes the unit disk D to C∞ \ D. as a special case. let’s ¯ consider inversion. z3 . z+d z+d This shows that any MT can be written as a composition of translations. and ∞.. T z4 ) = (z1 . q ∈ C. T z3 . If we have any two disks or halfplanes.
then γ F (z) dz = F (β) − F (α) where α = f (a). z − z0 Local results and consequences. ones for which γ is only of bounded variation. β = f (b). Using this deﬁnition it can be extended to rectiﬁable paths. In a neighborhood of γ the function Log(z − z0 ) is a primitive of 1/(z − z0 ). This is simply the line integral of f along γ. From the fundamental theorem of calculus we see that if F is any analytic function on a neighborhood of γ. then ρ f (z) dz = γ f (z) dz.11 III. the point on the circle with the same imaginary part as z0 and negative real part. By direct calculation we have γ z −1 dz = 2πi. let z0  < r and consider γ z−z0 dz. z ± → z1 . so is the Riemann– Stieltjes integral of f ◦ γ with respect to γ. d] onto [a. Let z ± be the point on the circle z − z0  = r with arg(z ± − z0 ) = ±(π − ). Complex Integration and Applications to Analytic Functions If γ : [a. If γ is a closed path. as → 0. 1 More generally. if γ = reiθ . and f is a continuous complexvalued function deﬁned on the image of γ. Clearly. If φ is a piecewise smooth increasing function from [c. Theorem (Cauchy’s integral formula for a disk). n = −1. Let f be analytic on a closed disk and let γ be a path around the boundary of the disk. As a consequence. z − z0 ↓ 0 gives γ Taking the limit as 1 dz = 2πi. Then f (z) = 1 2πi f (w) dw w−z γ . i. So γ 1 dz = Log(z + − z0 ) − Log(z − − z0 ) = log r + i(π − ) − log r + i(π − ). then z n dz = 0 γ for all n ∈ Z.. b] and we let ρ = γ◦φ (so ρ is the same path as γ but with a diﬀerent parametrization). deﬁne b f (z) dz = γ a f (γ(t))γ (t) dt.e. this is zero. Now let γ be the curve running from z − to z + on the circle. b] → C is piecewise smooth. It can be deﬁned analogously to the Riemann integral as the limit of sums of the form (f ◦ γ)(τk )[γ(tk ) − γ(tk−1 )].
12 for all z in the open disk. the same sort of convergence holds for f (w) f (w) (z − a)n = . Consider the function g(t) = γ f (z + t(w − z)) dw.) Theorem. w−z As an easy corollary. Then f (z) = some an ∈ C. (w − a)n+1 w−z Therefore we can interchange summation and integraton to get 1 f (w) f (z) = dw = an (z − a)n . (w − a)n+1 . continuously complex diﬀerentiable) function has a power series expansion around each point in its domain. Let γ be the path around w − a = r.. Since f is bounded on the circle. R). γ f (w) dw = g(1) = g(0) = f (z) w−z γ 1 dw = 2πif (z). w−z f (z + t(w − z)) dw. γ Proof. Therefore. so the Weierstrass Mtest implies that the series 1 1 (z − a)n = . Fix z and choose r < R with z − a < r.e. for Proof. Then f (w) dw = 0. (And we already know the converse is true. Fix z in the open disk and apply Cauchy’s formula to the function w → f (w)(w − z). Let f be analytic on a closed disk and let γ be a path around the boundary of the disk. 2πi γ w − z where an = γ f (w) dw. Then (z − a)/(w − a) < 1. g (t) ≡ 0. n+1 (w − a) w−z converges absolutely and uniformly in w for w − a = r. since F (w) = f (z + t(w − z))/t is a primitive for f (z + t(w − z)). Proof. We can now show that an analytic (i. an (z − a)n for z − a < R. we have: Cauchy’s Theorem for a disk. Let f be analytic in B(a. Then g (t) = γ and.
if the function achieves its maximum modulus at a point. The set of such a is closed.R) f  . An analytic function on a region is completely determined by its value on any set of points containing a limit point. Corollary. then f (n) (a) ≤ n! maxB(a. Then pass to limit. Proof. and using the Taylor expansion. Cauchy’s estimate for n = 1 implies f ≡ 0. Corollary (Maximum Modulus Principle). Proof. Every nonconstant polynomial has a root in C. If f is analytic on a region Ω and not identically zero. it achieves it everywhere on a disk containing the point. By the Cauchy estimate for n = 0. Theorem. Proof. Corollary. A bounded entire function is constant. Under the same hypotheses f (n) (a) = n! 2πi f (w) dw. then the roots of f are isolated. An analytic function on a region which achieves a maximum modulus in the region is constant.13 Corollary (of proof ). unless f is identically zero on the region. The number p is the multiplicity of a as a root of f . is open. and g(z) = n=p cn (z − a)n−p . . f (z) = cn (z − a)n with cn = f (n) (a)/n!. such that f (z) = (z − a)p g(z). R). Evident if R is replaced by r < R. It follows from the Cauchy–Riemann equations that the function itself is constant. It is also clearly closed. Rn Proof. If f is analytic on a region and not identically zero. If f is analytic on a region then there does not exist a point a such that f (n) (a) = 0 for all n. Corollary (Liouville’s Theorem). analytic in a neighborhood of a and nonzero at a. then for each a ∈ Ω there exists a unique integer p ∈ N and a function g. Corollary. Let p be the least integer with cp = 0 ∞ (which exists by the preceeding theorem). the set of points were the maximum modulus is achieved is open. (w − a)n+1 γ Corollary (Cauchy’s estimate). If f is analytic in B(a. Thus. Proof. Hence the function has constant modulus. Corollary (Fundamental Theorem of Algebra).
t)) . This is an equivalence relation. 1) = γ1 . s)) ∂Γ (t. so that Γ maps each of the squares [j/n. 1] → G be two closed paths in a region G ⊂ C. s)] dt = 0. s)] dt = ∂s ∂t 1 0 ∂ ∂Γ [f (Γ(t. 1. s) dt. We say that the paths are homotopic in G if there exists a continuous function Γ : [0. γ1 : [0. ∂t ∂s ∂t∂s The proof without assuming smoothness is based on Cauchy’s theorem for a disk and a polygonal approximation argument. Therefore. It is easy to conclude from Cauchy’s theorem for a disk that f (z) dz = ρk ρk+1 f (z) dz. k/n). . γ1 f (z) dz = ρn f (z) dz. Let ρk be the closed polygonal path determined by the points Γ(j/n. j = 0. 1] → G such that for each s Γ( · . Theorem. ∂t ∂s since both integrands equal f (Γ(s. If we assume that the homotopy Γ is C 2 . . Since Γ is continuous on the compact set [0. there exists > 0 so that G contains a closed disk of radius about each point in the range of Γ. 0) = γ0 and Γ( · . the result is easy: let 1 g(s) = 0 f (Γ(t. written γ0 ∼ γ1 in G. 1] × [0. s)) (t. (k + 1)/n] into the disk of radius about Γ(j/n. n. k/n). Moreover. and also that f (z) dz = γ0 ρ0 f (z) dz. Then f (z) dz = γ0 γ1 f (z) dz. t)) ∂Γ ∂Γ ∂2Γ + f (Γ(s.14 Homotopy of paths and Cauchy’s Theorem. Deﬁnition. . (j + 1)/n] × [k/n. . Suppose that γ0 and γ1 are two homotopic piecewise smooth closed curves in G and that f is analytic on G. it is uniformly continuous. ∂t Then 1 g (s) = 0 ∂ ∂Γ [f (Γ(t. Let γ0 . s)) (t. 1]. 1] × [0. we can ﬁnd an integer n. Sketch of proof. s) is a closed path with Γ( · . .
then f (z) dz = 0. Then the index of γ with respect to a.15 Deﬁnition. Intuitive meaning of winding number: if a pole is planted in the plane at a point a and a closed curve is drawn in the plane not meeting a. Cauchy’s Theorem. 1]. Corollary. If γ is a piecewise smooth path which is homotopic to 0 in G. Winding numbers and Cauchy’s Integral Formula. and f is analytic in G. It is constant on each connected component of C \ γ. Proof. Let γ be a piecewise smooth closed curve in C and a ∈ C \ γ. If it contracts around the pole planted at a. if the curve were made of welllubricated rubber and were to be contracted as small as possible. Deﬁnition. Assuming γ is parametrized by [0. A region G is simplyconnected if every closed path in G is homotopic to 0. As an immediate consequence of the theorem above. γ Deﬁnition. γ(s) − a . or the winding number of γ about a is deﬁned to be n(γ. a) = 1 2πi 1 dz. we have Cauchy’s theorem (previously proved only for G a disk). If G is is a simplyconnected region in the complex plane then f (z) dz = 0 γ for every function f analytic in G and every piecewise smooth closed curve γ in G. A closed path in G is homotopic to 0 in G if it is homotopic to a constant path. z−a γ Demonstrate how the winding number can be calculated by drawing a negatively directed ray from a and integrating over pieces of the curve between successive intersections of the curve with the ray. its winding number is the number of times it circles the pole (signed by orientation). if it contracts to a point other than a its winding number is 0. Proposition. The index n(γ. deﬁne t g(t) = 0 γ (s) ds. a) is an integer for all a ∈ C \ γ.
Then n(γ. which implies that n(γ. dt so. exp[−2πin(γ.. and let f be an analytic function on G. For a ∈ G \ γ ﬁxed. so it is identically 0 on the unbounded component.e. that n(γ. or. a) = 1 2πi j=1 f (w) dz. g(z) = z−a f (a). i. The book proves Cauchy’s formula under the assumption that γ is homologous to 0. so the book’s result is a bit stronger. Cauchy’s Integral Formula. it tends to 0 as a tends to ∞. Proposition. If γ0 and γ1 are paths which are homotopic in C \ {a} for some point a.. Clearly also. z ∈ G \ {a}. exp[−g(t)][γ(t) − a] = γ(0) − a.16 An easy direct calculation shows that d exp[−g(t)][γ(t) − a] = 0. it is constant on connected components. indeed. The book also goes on to consider chains. a) is indeed integral. i. Let γ a piecewise smooth curve in a region G which is null homotopic there. e. a)] = 1. a) = n(γ1 . a) = 0 for all a ∈ C \ G. f (a) j=1 n(γj . then n(γ0 . This presentation is quite a bit diﬀerent from the book’s. a)][γ(1) − a] = γ(0) − a. z) = 0 for all z ∈ C \ G. Taking t = 1. Applying Cauchy’s theorem to g gives the desired result immediately. It is easy to see that homotopic to 0 implies homologous to 0. Remark. It is true that all piecewise smooth closed curves in region are homologous to zero iﬀ G is simplyconnected (see Theorem 2. However we can deal with this solution in other ways (by adding a line connecting the circles). z = a.. w−z k if the chain γ1 + · · · + γn is homologous to zero.2 of Chapter 8). z−a a ∈ G \ γ. j=1 n(γj . exp[−2πin(γ. a)f (a) = Proof. It is easy to show that g is continuous (even analytic!) on G (show the continuity as a homework exercise). This sort of result is useful. a). which are unions of curves and proves that k k γj 1 2πi γ f (z) dw. .g. but the reverse is not true. let f (z) − f (a) . The next result is an immediate corollary of the invariance of path integrals of analytic functions under homotopy. and being integer. The function is obviously continuous. in the case of an annulus.e.
17 Zero counting. . Then f (z) 1 1 1 g (z) = + + ··· + . This is an immediate corollary. . . f (z) z − a1 z − a2 z − am g(z) and the theorem follows. Proof. . Then 1 2πi f (z) dz = f (z) m n(γ. ak ). Proof. then there exists . and the fact that it is integer valued. δ > 0 such that for all 0 < ζ − α < δ. . That the number of roots is m comes from the continuity of the integral as long as ζ remains in the same connected component of C \ f ◦ γ as α. am repeated according to multiplicity. If f is analytic at a and f (a) = α with ﬁnite multiplicity m ≥ 1. A nonconstant analytic function is open. the equation f (z) = ζ has precisely m roots in z − a < and all are simple. k=1 γ Note that we may as well write the integral as 1 2πi 1 dw. Theorem. (z − am )g(z) where g(z) is analytic and nonzero on G. then 1 2πi f (z) dz = f (z) − α m n(γ. Let G be a region. ak ). a2 . γ a curve homotopic to 0 in G and f an analytic function on G with zeros a1 . If the ai are the solutions of the equation f (z) = α. w−α f ◦γ Theorem. We can insure that the roots are simple by taking small enough to avoid a root of f . Open Mapping Theorem. We can write f (z) = (z − a1 )(z − a2 ) . . k=1 γ Proof. . Open Mapping Theorem.
z0 ] 0 f (z0 + t(z − z0 ))dt. then f is analytic. we consider another result that uses an argument similar to Morera’s Theorem. . Let G be a region and f : G → C complex diﬀerentiable at each point of G. T Proof. (As an interesting example. Only if is clear. Theorem. Proof. If G is a simply connected domain. For the if direction. with boundaries Tn such that  T0 f  ≤ 4n  Tn f . 0 ∈ G. Proof. Let a be the center of the disk. Let ∆0 be a closed triangle with boundary T0 . ﬁx a ∈ G and deﬁne F (z) = γ(a. Reasoning as in Morera’s theorem F = f. so is analytic. If G is a region and f : G → C a continuous function such that f dz = 0 for each triangular path T in G. we can assume G is a disk.18 Morera’s Theorem and Goursat’s Theorem. If 0 ∈ G. then a branch of the logarithm may be deﬁned there. Let G be a region and f : G → C a continuous function. Thus Proposition. . In particular. Before proceeding to Goursat’s Theorem. Then f is analytic. then 1/z admits a primitive. Then f admits a primitive on G if and only if γ f = 0 for all piecewise smooth closed curves γ in G. By hypothesis F (z) − F (z0 ) 1 = z − z0 z − z0 1 f (w) dw = [z. z) is any piecewise smooth path from a to z. We will show that the integral of f vanishes over every triangular path. . and deﬁne F (z) = [a. take G to be the complement of the spiral r = θ. Goursat’s Theorem. Using repeated quadrisection we obtain triangles ∆0 ⊂ ∆1 ⊂ .z) f where γ(a. if G is a simplyconnected region then every analytic function admits a primitive. It is then easy to check by diﬀerentiation that exp(F (z))/z ≡ 1. Morera’s Theorem.z] f . so f has a primitive. Continuity then implies that F (z0 ) = f (z0 ). Adjusting the primitive by a constant / we can assure that it agrees with a logarithm of z0 at some point z0 .) We can at long last show that a function which is complex diﬀerentiable in a region (but without assuming it is continuously diﬀerentiable) is analytic. Since it is enough to show f is analytic on disks.
g(a) = 0. Using this we see that f has a pole at a if and only if f (z) = (z−a)−m g(z) for some integer m > 0 and g analytic in a neighborhood of a. some n. z↓0 lim e1/z = 0. Let z0 be in the intersection (it’s unique). From the diﬀerentiability of f at z0 we have: given >0 f (z) − f (z0 ) − f (z0 )(z − z0 ) ≤ z − z0  for all z in some disk around z0 . Theorem. an essential singularity. IV. it must vanish. The ﬁrst case is that of a removable singularity. Since z dz = 0. Then diam Tn = d/2n . or does not exist as a ﬁnite number or inﬁnity. Suppose f has an isolated singularity at a. Proof. since was arbitrary. the third. or even f (z) ≤ Cz −a−(1− ) for some > 0. If there exists g analytic on some disk centered at a which agrees with f except at a. For “if” let g(z) = (z − a)f (z). z↑0 . Thus ∞ f (z) = n=−m an (z − a)n in a punctured neighborhood of a. or even f (z) is bounded in a punctured neighborhood of a. or in the interior). Therefore  T0 f  ≤ dp. but not in B(a. Using Morera’s theorem (separating the cases a inside the triangle. Tn  Tn Tn f (z) dz = f =  Tn [f (z) − f (z0 ) − f (z0 )(z − z0 )]dz ≤ diam Tn perim Tn = dp/4n . It is removable if and only if limz→a (z − a)f (z) = 0. The function e1/z has an essential singularity since lim e1/z = +∞. The second is called a pole. We say f has an isolated singularity at a. By compactness ∆n = ∅. on the boundary. perim Tn = p/2n . R). “Only if” is clear. z = a. If f is analytic in B(a. then 1/f (z) has a removable singularity there and extends to a with value 0. therefore for all z ∈ ∆n . we say f has a removable singularity. then f has a removable singularity. equals +∞. Singularities of Analytic Functions Deﬁnition. and. If limz→a f (z) ∈ C. Therefore g(z) = (z − a)h(z) for h analytic. Corollary. R) \ {a}. Now either limz→a f (z) exists as a ﬁnite real number. p = perim T0 . for some a ∈ C. Note that if f has a pole at a. it is easy to see that g is analytic.19 Let d = diam T0 . and h is the desired extension of f . some C. R > 0.
We easily see that f has a removable singularity at ∞ ⇐⇒ ∃ limz→∞ f (z) ∈ C ⇐⇒ f is bounded in a neighborhood of inﬁnity. Consider the function g(z) = [f (z) − c]/z. . Deﬁne g(w) dw. If the conclusion is false. Fix z ∈ γ. Proof. n We say that a series of the form converges if both the series n=−∞ an (z − a) ∞ ∞ −n n a−n (z − a) and an (z − a) converge (and their sum is the value of the n=1 n=0 doubly inﬁnite summation). Thus f is diﬀerentiable at each z ∈ C \ γ. Let γ be a piecewise smooth curve and g a continuous complexvalued function on γ. there is a punctured disk around 0 in which f (z) stays a ﬁxed positive distance away from some number c ∈ C.20 Theorem (CasoratiWeierstrass). ∞ Lemma. Then we say that f has an isolated singularity at ∞. If f has an essential singularity at a. which is not possible if f has an essential singularity. z m f (z) → 0. z ∈ γ. then the image under f of any punctured disk around a is dense in C. Therefore. w−z w=r where R2 > r > z (it doesn’t matter which such r we take. Therefore z m [f (z) − c] → 0 as z → 0 for suﬃciently large m. Suppose f is analytic on {z : z > R} for some R. WLOG. Note that if f has a removable singularity at ∞ then we can expand f in a power series in 1/z. Deﬁne f2 (z) = 1 2πi f (w) dw. so it has a pole at 0. Then / g(w) ζ−w − ζ−z g(w) z−w =− g(w) g(w) →− (ζ − w)(z − w) (z − w)2 as ζ → z with the convergence uniform in w ∈ γ. assume a = 0. Deﬁnition. Proof. Then f2 is analytic in z < R2 and so has a convergent power series there: ∞ f2 (z) = n=0 an z n . It is removable if f (1/z) has a removable singularity at 0. Let f be deﬁned and analytic in an annulus R1 < z < R2 . Laurent series. / f (z) = γ z−w The f is analytic on C \ γ. It tends to ∞ as z → 0. convergent for z > R uniformly and absolutely for z ≥ ρ > R. by Cauchy’s theorem).
in which case this is the order of the 0. since we can easily recover the an : am = 1 2πi an n∈Z z=r zn z m+1 dz = 1 2πi z=r f (z) dz z m+1 where r is any number in (R1 . The convergence is uniform and absolute on compact subsets of the annulus and the coeﬃcients an are uniquely determined by the formula above. We deﬁne orda (f ) = inf{n ∈ Z : an = 0}. Thus ∞ f1 (z) = n=1 a−n z −n . . Note that such an expansion is unique. orda (f ) < 0 but ﬁnite ⇐⇒ f has a pole. R2 ) for some 0 ≤ R1 < R2 ≤ ∞. with the convergence absolute and uniform on compact subsets of the annulus. then ∞ f (z) = n=−∞ an z n for all z in the annulus. If f has an isolated singularity at a. 0 < z − a < R. Now. Summarizing: Theorem (Laurent Expansion). This shows that ∞ f (z) = n=−∞ an z n .21 Also. it has a Laurent expansion in a punctured disk at a: f (z) = n=−∞ ∞ an (z − a)n . in which case − orda (f ) is the order of the pole. Thus f has a removable singularity ⇐⇒ orda (f ) ≥ 0. Then this is analytic in z > R1 and tends to 0 at ∞. orda (f ) = +∞ ⇐⇒ f is identically 0 near a. R2 ) (result is independent of choice of r since two diﬀerent circles are homotopic in the annulus). orda (f ) = −∞ ⇐⇒ f has an essential singularity at a. If f is an analytic function on the annulus G = ann(0. it is easy to use Cauchy’s integral formula to see that f = f1 + f2 in the annulus. let f1 (z) = − 1 2πi w=r f (w) dw w−z where now r is taken with R1 < r < z. and it extend to a as 0 ⇐⇒ orda (f ) > 0. R1 .
a). The f − gi is analytic on K (after removing the removable singularities at the ai ). a) = g(a) (and thus Cauchy’s therem may be viewed as the special case of the residue theorem where f has a simple pole). a2 . Then {a ∈ En(γ. then m 1 f (z) dz = Res(f . Let F : [0. If f has a double pole at a. Residue integrals. f analytic on G except for isolated singularies. and γ a nullhomotopic piecewise smooth closed curve in G which doesn’t intersect E. so γ f − gi = 0. z−a=r Give pictoral proof that if G is simplyconnected domain. then g(a) g (a) + + ··· . however. if f has an isolated singularity at a. . The result follows easily. how to practically compute residues. a) = 0. First. then for r suﬃciently small. For essential singularities it can be very diﬃcult. the homotopy takes place in the the complement of a. so has a ﬁnite intersection with E. and let gi be the singular part of f at ai . . we consider the application of the e residue theorem to the computation of deﬁnite integrals. am be the points in the intersection. If f has a simple pole at a. for all functions f which are analytic in G \ E. then f (z) = g(z)/(z −a) with g analytic at a and Res(f . and γ is a simple closed curve that encloses singularities at a1 . . If a ∈ E does not belong to this intersection. . f (z) = g(z)/(z − a)2 = 2 (z − a) z−a . 2πi γ j=1 a simple closed curve in G . 1] × [0. a) = 1 2πi f (z) dz. . Note that. First. denoted Res(f . Let a1 . a) Res(f . the argument principle and Rouch´’s theorem. a).22 We call a−1 the residue of f at a. showing that the n(γ. aj ). Proof. E a discrete subset of G. We will proceed to two easy but useful consequences of the residue theorem. Let G be an open set. More precisely: Theorem (Residue Theorem). but for poles it is usually easy. Res(f . . am (necessarily ﬁnite). a) = 0} is ﬁnite and 1 2πi f (z) dz = γ a∈E n(γ. The image K of F is compact. 1] → G be a homotopy from γ to a constant. . .
) On the other hand. we 1 expand g as a Taylor series around a and can read oﬀ Res(f . then along a vertical segment to X + iX. The answer is π. Similarly. and then down a vertical segment back to −X. z X The imaginary part of the quantity inside the limit is 2 δ (sin x)/x dx. This approach 2i applies to rational functions in sin and cos. with γ the unit circle. so its contour integral is 0. even though its imaginary part. It is easy to see that the integral is bounded by 1/X on each of the vertical segments and by 2e−X on the horizontal segment. 1) and r(t) is a rational function of degree ≤ −2 with at worst a simple pole at 0. then a semicircle in the upperhalf plane to δ. to get the answer 3π/4. . but it is easier to remember the procedure than the formula. Note: it is essential that when taking the limit we excise a symmetric interval about zero. Same technique works for all rational functions of degree ≤ −2. 2π Example 2: evaluate 0 4 sin4 (x) dx. and integrate around this cut and over a large circle centered at the origin (avoid origin with a small ∞ circle). it is not necessary that we cut oﬀ the integrand at symmetric points −X and X (we could have used −X1 and X2 and let them go independently to zero). 1 Write as γ iz z+1/z dz. if f has a pole of order m with f (z) = g(z)/(z − a)m . a) = g (a). (The ordinary integral doesn’t exist: eix /x has a nonintegrable singularity near the origin. t1/2 + t3/2 Use a branch of the z −1/2 with the positive real axis as the branch cut. a). (It is (m−1)! g (m−1) (a). ∞ Example 4: evaluate 0 sin x dx x We will evaluate X→+∞ δ↓0 −δ X lim + −X δ eiz dz. The integrand ez /z has no poles inside this curve. then back along a horizontal segment to −X + iX. We use a somewhat fancy path: starting at −X we follow the xaxis to −δ. ∞ Example 3: evaluate 0 1 dt. then along the xaxis to X. (sin x)/x doesn’t.) ∞ Example 1: evaluate −∞ 1 (x2 + a2 )(x2 + b2 ) dx. 0 < a < b.23 so Res(f . On the other hand. Use a semicircle centered at the origin in upper half plane and let radius tend to inﬁnity to get the answer of π/[ab(a + b)]. This approach applies to integrals of the form 0 tα r(t)dt where α ∈ (0.
is bounded by 2 on the horizontal segments. f a meromorphic function on G. Then 1 2πi f = f n(γ. The residue theorem then immediately gives Theorem (Argument Principle). A complex function on an open set G is called meromorphic if it is analytic on G except for a set of poles. Use a rectangular path with opposite corners ±(n+1/2+ni). It follows easily that the integral goes to 0 with n. a) orda (f ) a ∞ γ where the sum is over Z ∪ P and contains only ﬁnitely many terms. (n − a)2 sin2 πa n=−∞ To state the argument principle we deﬁne: Deﬁnition. Then Res(f /f . whence ∞ 0 sin x π dx = .e.  cot x + iy = sin2 x + sinh2 y 2 so the  cot  equals 1 on the vertical segments and. which is behind the name (how much does the argument change as we traverse this curve). on a neighborhood of a). Indeed. Note that the pole set of a meromorphic function is discrete (but may be inﬁnite). n a positive integer and evaluate γ (z − a)−2 cot πz dz. a) = orda (f ). in which case the integral counts the zeros of f . Since sin x + iy = sin x cosh y + i cos x sinh y and cos x + iy = cos x cosh y − i sin x sinh y. f (z) = (z − a)m g(z) with m = orda (f ) and g(a) = 0. Our ﬁnal application of the residue theorem is Rouch´’s theorem. For meromorphic f we count the poles as negative zeros. Let G be an open set. so 1 π2 = . and γ a nullhomotopic piecewise smooth closed curve in G which doesn’t intersect either the zero set Z or the pole set P of f . Suppose f is meromorphic at a (i. We saw this theorem previously in the case where f was analytic. a ∈ Z. A meromorphic function is continuous when viewed as taking values in C∞ . we get cos2 x + sinh2 y . and f /f = m/(z − a) + g /g. x 2 1 . Note that the left hand side is nothing other than n(f ◦ γ. (n − a)2 n=−∞ ∞ Example 5: evaluate a ∈ R \ Z. 0).. e .24 eiz /z = 1/z + analytic and the integral of 1/z over the semicircle (integrated clockwise) is easily seen to be −πi. if n is large. The poles of the integrand are at a and the integers and / 2 the residue is −π/ sin (πa) and 1/[π(n − a)2 ]. We conclude that the limit above equals πi.
V. analytic. f : G → C. γ f /f = g /g and the theorem follows from the argument principle. and γ a e nullhomotopic piecewise smooth closed curve in G which doesn’t intersect Zf ∪Pf ∪Zg ∪Pg . . (f + φ) + (−f ) = φ < f  ≤ f + φ +  − f . a) orda (f ) = a∈Zf ∪Pf a∈Zg ∪Pg n(γ. endowed with the topology of uniform convergence on compact subsets. and Montel. Moreover fn converges to f uniformly on compact subsets of G. Montel’s theorem identiﬁes the precompact subsets of the space of analytic functions. Consider the space of continuous functions on a region in the complex plane. Proof. . then the equation z exp(λ − z) = 1 has exactly one solution in the unit disk (take f = z exp(λ − z). . fn : G → C. Proof. Hence. (In fact this is a metric space topology: we can deﬁne a metric such that a sequence of continuous functions converge to another continuous function in this metric. . If fn → f uniformly on compact subsets of G.) Weierstrass’s theorems states that the space of analytic functions on the region is a closed subspace and diﬀerentiation acts continuously on it. Theorem (Weierstrass). Let G ⊂ C be open. z ∈ γ.25 Theorem (Rouch´). If f (z) + g(z) < f (z) + g(z). Let G be an open set. φ = −1). if and only if the sequence converges uniformly on compact subsets. Let l be a branch of the logarithm on C \ [0. Then l(f /g) is a primitive of (f /g) /(f /g) = f /f − g /g which is analytic in a neighborhood of γ. then n(γ. then f + φ has the same number of zeros as f in the region bounded by γ. . then f is analytic. Proof. The hypothesis means that on γ the quotient f (z)/g(z) is not a nonnegative real number. 2. f = lim T n→∞ fn = 0. T so f is analytic by Morera’s theorem. Suppose f and φ are analytic e on a simplyconnected region G and that γ is a simple closed piecewise smooth curve in G. Hurwitz. a) orda (g). Example: If λ > 1. ∞). If T is a triangular path. Further results on analytic functions The theorems of Weierstrass. f and g meromorphic functions on G. n = 1. γ Corollary (Classical statement of Rouch´’s theorem). If φ < f  on γ.
fn . If f is not identically constant. d(a)/2). d(a)).26 Let γ be the boundary of a disk around z contained in G. This contradicts the injectivity of the fn . . r) ⊂ G. then we can choose r > 0 so that B(a. etc. Then there exists a subsequence which converges uniformly on compact subsets. For the diagonal sequence gj = fmj (j) cn (gj ) converges for all n. and B(a. f analytic functions on G with fn → f uniformly on compact subsets. then f is either injective or identically constant. and a is the only zero of f on this disk. r). First we show that we can ﬁnd a subsequence that converges uniformly on B(a. but there exists distinct a. r) ∩ B(b. Corollary. fn (z) = f (a) for some z ∈ B(a. Proof. r). Suppose that the functions f1 . b with f (a) = f (b). for some z ∈ B(b. Let G ⊂ C be open. fn . the argument principle then tells us that γ fn /fn = 0. ¯ Proof. so we have a contradiction. but γ f /f = 2πi orda (f ) = 0. lim fn (z) = lim 1 2πi fn (w) 1 dz = 2 (w − z) 2πi f (w) dz = f (z). ∂G)/2. If each of the fn is nowherevanishing. Now ∞ fm (z) = n=0 cn (fm )(z − a)n . and. ¯ z ∈ B(a. r) = ∅. by the same argument. Now fn (z) − f (a) converges to f (z) − f (a) uniformly on compact subsets of B(a. where cn (f ) = f (n) (a)/n!. By the Cauchy estimates cn (fm ) ≤ M/d(a)n ¯ where M is the uniform bound for the fm on B(a. we can ﬁnd a subsequence fm1 (j) for which c1 (fm1 (j) ) converges as j → ∞. But. Proof. (w − z)2 γ γ Combining Weierstrass’s Theorem and the Argument Principle we get: Theorem (Hurwitz). Then we can ﬁnd a further subsequence so that c2 (fm2 (j) ) converges. so for all suﬃciently large n. but f (a) = 0. Theorem (Montel). Fix a ∈ G and let d(a) = dist(a. F a family of analytic functions on G. we can choose r > 0 so that B(a. as well. r) ∪ B(b. Since c1 (fm ) is uniformly bounded in C. If each of the fn is injective. then f is either nowherevanishing or identically zero. r) and the latter function has a zero but is not identically zero there. f analytic functions on G with fn → f uniformly on compact subsets. . Let G be a domain in C. in view of Weierstrass’s Theorem the hypothesis tells us that fn /fn converges to f /f uniformly on γ. . r) ⊂ G. . If f is not identically zero. are uniformly bounded on each compact subset of G. f2 . Letting γ be the boundary of the disk. Let G be a domain in C. d(a)).
z ∈ B(a. d(a)/2). etc. N ∞ (n) gj (z) − gk (z) ≤ n=0 N gj (a) − gk (a)z − an /n! + n=N +1 (n) (n) 2M z − an d(a)n ≤ n=1 gj (a) − gk (a)z − an /n! + (n) (n) 2M . z ∈ B(a. 2N Given > 0 we can choose N large enough that 2M/2N ≤ /2 and can then choose j and (n) (n) k large enough that gj (a) − gk (a) ≤ n! /{2(N + 1)[d(a)/2]n } for 0 ≤ n ≤ N . Thus we can ﬁnd a subsequence fm0 (j) for which fm0 (j) (a) converges as j → ∞. Now fm (z) = By the Cauchy estimates (n) fm (a) ≤ M n!/d(a)n (n) ¯ where M is the uniform bound for the fm on B(a. This proves the claim. . supm fm (a) < ∞ for each n. d(a)/2). Claim: gj converges uniformly on B(a. Now. d(a)). d(a)). To prove this. we must show that for any > 0 gj − gk  ≤ on B(a. d(a)/2). n! n=0 ∞ (n) ¯ z ∈ B(a. First we show that we can ﬁnd a subsequence that converges uniformly on B(a. N ∞ gj (z) − gk (z) ≤ n=1 N cn (gj ) − cn (gk )z − an + n=N +1 2M z − an d(a)n ≤ n=1 cn (gj ) − cn (gk )[d(a)/2]n + 2M . d(a)/2).27 Claim: gj converges uniformly on B(a. for any N ∈ N. For the diagonal sequence gj = fmj (j) gj (a) converges for all n. ∂G)/2. d(a)/2). d(a)/2). d(a)/2). we must show that for any > 0 gj − gk  ≤ on B(a. This proves the claim and shows that we can ﬁnd a subsequence which converges uniformly on B(a. Fix a ∈ G and let d(a) = dist(a. Now. fm (a) (z − a)n . d(a)/2). To prove this. In particular. for any N ∈ N. 2N Given > 0 we can choose N large enough that 2M and then choose j and k large enough 2N that cn (gj ) − cn (gk ) ≤ /[2(N + 1)[d(a)/2]n ] for 0 ≤ n ≤ N . Then we can ﬁnd a further subsequence so that fm2 (j) (a) converges as well.
28 Now we may choose a countable sequence of points ai ∈ G such that G = B(ai . Proof. We then get: Theorem (Schwarz’s Lemma). consequence of the maximum modulus principle. The most diﬃcult part is to remember that the German analyst Herman Schwarz (actually Karl Herman Amandus Schwarz. all points in G with rational real and imaginary parts). we translate in the domain and range space so that a = f (a) = 0 and dilate in the domain and range so that = δ = 1. d(ai )/2) (e. and extends analytically to the disk with g(0) = f (0). Then we can apply the special case above to f ◦ φ−a to get f (φ−a (z)) = cz or f (z) = cφa (z). To keep the statement simple. d(ai )/2). We can use this to estimate the derivative f (a) and the stretching f (b) − f (a)/b − a. Then Schwarz shows that f (0) ≤ 1 with equality if and only if f (z) = cz some c = 1. Thus equality does indeed hold. because any compact subset of G is contained in a ﬁnite union of the B(ai . The function g(z) = f (z)/z has a removable singularity at the origin. then f (z) = cz for some c of modulus 1. Suppose f maps the open unit disk into itself leaving the origin ﬁxed. Suppose that an analytic map f maps a disk of radius about a into the disk of radius δ about f (a). We can use Schwarz’s lemma to show that this is most general onetoone analytic map of the disk onto itself. It satisﬁes g(z) ≤ 1/r on the circle of radius r < 1. d(ai )/2).” “Schwarz lemma. but very useful. and if equality holds for any nonzero point. The diagonal subsequence hj converges uniformly on each B(ai . 18431921).. Returning to the general case. d(a1 )/2). Schwarz’s Lemma. Also f (0) ≤ 1 and if equality holds. we may ﬁnd a subsequence fn1 (j) which converges uniformly on B(a1 . ∞ i=1 We complete the proof by showing that hj converges uniformly on all compact subsets. . we know that the most general M¨bius o transformations of the open unit disk D into itself is given by z→c z−a 1 − az ¯ with a in the disk and c = 1. First consider the special case f (0) = 0. no. uses a “t”. then f (z) = cz for some c of modulus 1. while the French analyst. so by the maximum modulus theorem it satisﬁes this condition on the disk of radius r. Applying Schwarz to f −1 gives 1/f (0) = (f −1 ) (0) ≤ 1. Then f (z) ≤ z for all z in the disk. etc. From a homework exercise (§III . Schwarz’s lemma is an easy. Equality implies g is constant.” spells his name without a “t”. Then we may take a subsequence of that sequence which converges uniformly on B(a2 . 10). Laurent Schwartz. inventor of the theory of distributions. Letting r tend to one gives the result.” and “Schwarz alternating method.3. say f (a) = 0. Given the sequence of functions fm uniformly bounded on compact sets. inventor of the “Schwarz inequality. d(a2 )/2). Indeed this is immediate. 1915–.g.
Indeed. The structure of the proof is as follows: ﬁrst we use the hypotheses to exhibit an injective analytic map of h0 of G onto a domain G0 which satisﬁes the hypotheses of Koebe’s lemma. r(z) > z. Deﬁne q(z) = φ−a ([φ−b (z)]2 ) (where φα (z) := (z − α)/(1 − αz)). Two regions in the complex plane are called conformally equivalent if they are analytically isomorphic. note that φa is an analytic map of G into D which never vanishes. that is. We now turn to a lemma of Koebe. Next. let g be primitive of f /f . and q(r(z)) ≡ z. which we may determine uniquely by insisting that g(0) = b. z ∈ G \ {0}. so Schwarz’s lemma implies q(z) < z. we have r(0) = 0. it is not 11 on D). Deﬁning r = φb ◦ g. r(z) = 0 and z = q(r(z)) < r(z).29 The Riemann Mapping Theorem. if there exists a onetoone analytic mapping of the ﬁrst domain onto the second (whose inverse is then automatically analytic). 1) and let G D be a region containing the origin and having the square root property. so admits an analytic square root g(z). Recall that every analytic function on a simplyconnected region admits a primitive. then there exists an analytic function g with [g(z)]2 = f (z). and so adjusting g by an additive constant. Let a ∈ D \ G and let b be one of the square roots of −a. 0 = z ∈ D. C be simplyconnected. then there exists an analytic function F on that domain such that f = exp(F ). It follows that if f is a nowherevanishing function on a simplyconnected domain. The only consequence of simpleconnectivity that we shall use in the proof is that G has the squareroot property. Proof. Then G is conformally Proof. we can just take g = exp(F/2). Let D = B(0. and q is not just multiplication by a ¯ constant (in fact. Then there exists an injective analytic map r : G → D such that r(0) = 0. Clearly q(0) = 0. It follows immediately that f exp(−F ) is constant. f = exp(F ). Thus if 0 = z ∈ G. For future use we remark that if there exists an analytic isomorphism of a region G1 onto a region G2 . Lemma. We begin by deriving some easy consequences of simpleconnectivity. Then we use an extremal problem to deﬁne an injective analytic map f of G0 onto a domain . then G1 has the squareroot property if and only if G2 does. Let G equivalent to the open unit disk. Indeed. Riemann Mapping Theorem. Another consequence is the squareroot property: if f is a nowherevanishing analytic function on the domain.
so we need only show that f maps onto D. 1/r) is analytic and injective. Let a ∈ C \ G and let g be an analytic square root of z − a. and by the maximum modulus principle. but I will quickly state some key results. Therefore B(−w0 . Then set h0 (z) = r[h(z) − h(z0 )]/3 where z0 ∈ G. G1 := f (G0 ) D then (since G1 inherits the squareroot property from G0 which inherits it from G) Koebe’s lemma gives an injective analytic function r : G1 → D with r(0) = 0. if G satisﬁes the square root property. r) ⊂ g(G) (Open Mapping Theorem). Then r ◦ f ∈ F. Complements on Conformal Mapping. so f is not constant. If. then g = f2 ◦ f1 is an analytic isomorphism of the disk onto itself with g(0) = 0. If G C is a simplyconnected region and z0 ∈ G is arbitrary we may take any analytic isomorphism of G onto the disk and follow it with a suitably chosen M¨bius transformation o (z → c(a − z)/(1 − az). and since the fn are injective. Indeed. and so is simplyconnected. ¯ Set h(z) = [g(z) + w0 ]−1 so that h : G → B(0. c = 1. Obviously f (0) = 0 and f (w0 ) = α > 0. then either G = C. fn (w0 ) → α. g (0) > 0. On the other hand. . then it can be proved that a conformal map of the region onto the open disk extends to a topological homeomorphism of the closure of the region on to the closed disk. Note α ∈ (0. but r(f (w0 )) > f (w0 ). so is f .30 G1 contained in the disk. and it easily follows that g is the identity and f1 = f2 . −1 If f1 and f2 were two such analytic isomorphisms. G1 satisﬁes the hypotheses of Koebe’s lemma and we use the lemma to contradict the extremality of f . which is simplyconnected. We already know that f is analytic and injective. and let α = supf ∈F f (w0 ). a contradiction. f  < 1. or G is isomorphic to the disk (by the proof of the Riemann mapping theorem above). to obtain an analytic isomorphism f of G onto D ¯ satisfying f (z0 ) = 0. let fn ∈ F. Since the fn are uniformly bounded (by 1). If the boundary of a simplyconnected region is a simple closed curve. It is easy to see that g is injective and that if w ∈ g(G) then −w ∈ g(G). f (z0 ) > 0. If G1 ﬁlls the entire disk we are done. Indeed we have seen that simpleconnectivity implies the squareroot property. a < 1. Let F be the set of all injective analytic functions f : G0 → D with f (0) = 0. Thus f ∈ F and achieves the supremum. Fix 0 = w0 ∈ G0 . This map is an analytic isomorphism of G onto a connected open set G0 D containing 0. Pick w0 ∈ g(G) and then r > 0 / such that B(w0 . We claim that the supremum is achieved. Extension to boundary. Montel’s theorem assures us that there is a subsequence fnk which converges uniformly on compact subsets to some analytic function f on G0 . There are many interesting questions in conformal mapping which we will not have time to investigate. otherwise. We now note that the squareroot property is equivalent to simpleconnectivity. r) ∩ g(G) = ∅. Clearly f  ≤ 1 on G0 . to the contrary. 1]. We complete the theorem by showing that f is an isomorphism of G0 onto D (so f ◦ h0 is the desired analytic isomorphism of G onto D). r(z) > z for z = 0.
r). applied to a circle passing through −1 and containing 1 in its interior.e. Conformal is used to obtain explicit solutions to problems involving analytic and harmonic functions. but we just state. Cauchy’s integral formula then gives f (a) = Taking real parts we get: ¯ Mean Value Theorem. the result for doubly connected regions (regions whose complement in the Riemann sphere consists of two connected components. but they are not diﬃcult to compute numerically (with a computer). Harmonic Functions A realvalued C 2 function on an open subset of C is called harmonic if its Laplacian vanishes. In particular. VI. a ∈ G. and since D is simplyconnected u = Re f on D for some analytic function f on D. In particular. i. The resulting region. If r > 0 is such that B(a. roughly regions in the plane with one hole). one cannot usually compute the parameters analytically.e. For example. called a Joukowski airfoil. one can calculate things like the drag and lift of the airfoil. Joukowski used the conformal mapping f ◦ g where 1+z . 0 .. For lack of time we will not study conformal mapping of multiply connected regions. It can be shown (with basically the tools at our disposal) that any such region is conformally equivalent to the annulus 1 < z < R for some R > 1. r) ⊂ G. r) ⊂ D ⊂ G. The formula involves a few complex parameters which are determined by the vertices of the polygon. and by adjusting the circle. 0 u(a + reiθ )dθ. We have seen previously that the real part of an analytic function is harmonic and that if the domain is simplyconnected every realvalued harmonic function u admits a harmonic conjugate. Given the vertices. i. then we can choose ¯ a disk D with B(a. then (MVP) 1 u(a) = 2π 2π 1 2πi z−a=r f (z) 1 dz = z−a 2π 2π f (a + reiθ )dθ. and so this was a useful approach to airfoil design. Using the explicit conformal map to the disk and explicit formulas for the solution of Laplace’s equation on (the exterior of) a disk (similar to those that will be derived in the next section. looks like a crosssection of a wing. we see that every harmonic function is C ∞ . for which u+iv is analytic. ¯ Suppose u is harmonic on G. two such annuli with diﬀerent values of R1 are not conformally equivalent. The Schwarz–Christoﬀel formula is an explicit formula for the conformal map of a given polygon onto the disk. An important class of regions are the polygons. Multiply connected regions. one can adjust the exact shape of the airfoil.31 Constructive conformal mapping.. a realvalued function v. The value of R is uniquely determined. without proof. f (z) = 2 1−z g(z) = z−1 z+1 2 . If u : G → R is harmonic and B(a.
For 0 ≤ r < 1. and it is obviously closed). Elementary (use the power series expansion for the integral and the formula Pr (θ) = for the limits). r↑1 θ ∈ R \ 2πZ. Pr (θ) is a smooth. The next theorem shows how to solve the Dirichlet problem for D using the Poisson kernel: 1 − r2 (1 − r)2 + 2r(1 − cos θ) . continuous on G. [Proof: Let M = supz∈G u(z) and take ¯ zn ∈ G with u(zn ) → M . then u1 − u2 is harmonic and vanishes on the boundary. and coincident with f on ∂G. For 0 ≤ r < 1. Moreover if θ > 0. even. Given a bounded domain G and a function f : ∂G → R. and if δ > 0. i. 1 − 2r cos θ + r2 Pr (θ) = Re(1 + 2 n=1 r e n inθ )= n=−∞ rn einθ . Pass to a subsequence with zn → z0 ∈ G. positive. The reverse inequality similarly shows that a harmonic function satisﬁes the Minimum Principle. so K ≥ lim supz→z0 u(z) ≥ limn u(zn ) = M . the Dirichlet problem consists ¯ ¯ of ﬁnding a function u : G → R such that u is harmonic on G. then lim Pr (θ) = 0. θ ∈ R let Pr (θ) = Re deﬁne the Poisson kernel. 2πperiodic function of θ with mean value 1. continuous on G. When applying the maximum principle to functions which are not known to extend continuously to the boundary this easytoprove consequence is useful: If u satisﬁes the maximum principle on a bounded open set and there is a number K such that lim supz→z0 u(z) ≤ K. An important consequence of the maximum and minimum principles is uniqueness for the Dirichlet problem. Otherwise z0 ∈ ∂G.e. but it takes its maximum and minimum on the boundary. u1 = u2 . so is constant and obviously M ≤ K. If z0 ∈ G. then u ≤ K on G. Proof. then u takes its maximum there.32 A harmonic function satisﬁes the Maximum Principle: it does not assume its maximum on a region unless it is constant there. so it is identically zero. Note that ∞ ∞ 1 + reiθ 1 − reiθ = 1 − r2 . Lemma.] The Poisson kernel. or even just satisfying the inequality u(a) ≤ 2π (2π)−1 0 u(a + reiθ )dθ (since this implies that the set of points were the maximum is achieved is open. and agree on ∂G. Now let D be the open unit disk and ∂D its boundary. If u1 and u2 are both harmonic on ¯ G. then the convergence is uniform over θ ∈ R such that θ − 2πn ≥ δ for all n ∈ Z. Indeed the conclusion is true for any continuous function with the meanvalue property (MVP)..
deﬁne u(z) = f (z). the entire integrand does so as well. namely θ − t0  ≤ δ/2. The integral deﬁnes an analytic function of z (since the integrand is continuous in t and z and analytic in z). 0 Now suppose that θ is taken suﬃciently near t0 . θ − t ≤ t − t0  + θ − t0  ≤ π + δ/2 Thus Pr (θ − t) → 0 uniformly for t ∈ I2 as r ↑ 1. Thus if θ − t0  ≤ δ/2 we may take r suﬃciently close to 1 that the integral over I2 is bounded by . t0 −π so it suﬃces to show that the last integral can be made arbitrarily small by taking θ suﬃciently close to t0 and r < 1 suﬃciently close to 1. is harmonic. Proof. so u. . Since f is bounded. being the real part of an analytic function. θ − t ≥ t − t0  − θ − t0  ≥ δ/2. Now z∈D z→eit0 t0 +π u(re ) − f (e iθ it0 1 )= 2π Pr (θ − t)[f (eit ) − f (eit0 )]dt. Split the interval of integration as {t : t − t0  ≤ π} = I1 ∪ I2 := {t : t − t0  ≤ δ} ∪ {t : δ ≤ t − t0  ≤ π}. Given f : ∂D → R continuous. so the integral is bounded by 2π 2π Pr (θ − t)dt ≤ I1 2π Pr (θ − t)dt ≤ . θ ∈ R. On I1 .33 Theorem. 0 ≤ r < 1. Now given > 0 we can choose δ > 0 so that f (eit ) − f (eit0 ) ≤ if t − t0  ≤ δ. It is now easy to deduce that the mean value property characterizes harmonic functions. u(reiθ ) = 1 2π 2π 0 z ∈ ∂D. ¯ The u is continuous on D and harmonic on D. It remains to show that lim u(z) = f (eit0 ) for t0 ∈ R. u(re ) = Re = Re or u(z) = Re 1 2π 0 iθ 1 2π 1 2π 2π f (eit ) 0 2π 1 + rei(θ−t) dt 1 − rei(θ−t) eit + reiθ dt eit − reiθ f (eit ) 0 2π f (eit ) eit + z eit − z . Pr (θ − t)f (eit ) dt. the integrand is bounded by Pr (θ − t) . z ∈ D. Then for t ∈ I2 .
U is ¯ harmonic on a neighborhood of B(0. the mean value property. R). 1). R) and nonnegative. we get 1 + r/(R − ) u(reiθ ) ≤ u(0). Choose > 0 small enough that r < R − . The diﬀerence between this function and u satisﬁes the mean value property and is zero on the boundary of the disk. From the Poisson kernel construction we can construct a harmonic function with the same boundary values as u on this disk. 1 − r/(R − ) Letting ↓ 0 this gives u(reiθ ) ≤ R+r u(0). If a sequence of harmonic functions converges uniformly on compact subsets. Thus we have proven: Theorem (Harnack’s Inequalities). 1+r 1−r Now suppose u is harmonic on B(0. Corollary. and reiθ is an arbitrary point of B(0. if u is harmonic and nonnegative on the disk of radius R about a then u(z) ∈ [u(a)/3. and set U (z) = u((R − )z). then it is harmonic. R). For example. then R−r R+r u(a) ≤ u(a + reiθ ) ≤ u(a) R+r R−r for 0 ≤ r < R. . so 1 u(re ) = U ([r/(R − )]e ) = 2π iθ iθ 2π Pr/(R− ) (θ − t)U (eit ) dt. so it vanishes on the disk. and the upper bound for Pr . From the expressions Pr (θ) = we have 1 − r2 1 − r2 = (1 + r)2 − 2r(1 + cos θ) (1 − r)2 + 2r(1 − cos θ) 1−r 1+r ≤ Pr (θ) ≤ . It is enough to show that u is harmonic on each disk with closure in G. R−r We obtain a lower bound in a similar fashion. 3u(a)] on the disk of radius R/2 about a. Of course we can translate the disk so it is centered at an arbitrary point a. Proof. 0 Using the nonnegativity of u.34 Theorem. then the limit function is harmonic. If u : G → R is a continuous function satisfying the mean value property. If u is nonnegative and harmonic on B(a.
Deﬁnition. In the second case. the numbers un (a) remain bounded as n → ∞ if and only if un (z) remains bounded for all z ∈ D. be a sequence of harmonic functions on a region G. n ≥ m. Again. Other simple properties are: • harmonic functions are subharmonic • the sum of two subharmonic functions is subharmonic • the pointwise maximum of two subharmonic functions is subharmonic The next lemma states that subharmonicity is a local property: Lemma. and the above estimate shows that the convergence is uniform on a covering set of discs. Then either limn un (z) = +∞ uniformly on compact subsets of G or there is an harmonic function u with limn un (z) = u(z) uniformly on compact subsets of G. un (z) → +∞ for all z ∈ G. As mentioned earlier. Hence its maximum on D is achieved on the boundary.35 Theorem (Harnack). so it is either the empty set or all of G. the sequence is uniformly Cauchy on D so converges uniformly on D. un (a) ≤ Cun (z). and since un (z) − um (z) ≤ C[un (a) − um (a)]. If D = B(a. z ∈ D. z ∈ D. This shows that the set of points where un stays bounded is both open and closed. r) ⊂ G. . limn un (a) ∈ C for all a. uniform convergence on compact sets follows. From the Poisson kernel representation. If 1 u(a) ≤ 2π 2π u(a + reiθ ) dθ 0 ¯ whenever B(a. Let u be a continuous realvalued function on a region G. In the ﬁrst case. 2π 0 Now u − u is subharmonic in a neighborhood of each point of D. . and it follows by the ˜ usual connectedness argument that the points where it achieves its maximum on D is ¯ either empty or all of D. Then u is subharmonic on G. we say that u is subharmonic. For any point a ∈ G let D be a disk centered at a contained in G. hence on all compact subsets. Let 0 ≤ u1 ≤ u2 ≤ . a subharmonic function satisﬁes the maximum principle. so is 0. Subharmonic functions and the solution of the Dirichlet problem. ¯ ¯ ¯ Proof. we have 2π 1 u(a + reiθ ) dθ. Consequently. Proof. This shows that u(a) ≤ u(a) as desired. ˜ u(a) = ˜ One more property we shall use: . Suppose that u is subharmonic in a neighborhood of each point of a region G. Then there exists a constant C > 1 with un (z) ≤ Cun (a). let u be the continuous function on D which agrees with u on ˜ ∂D and is harmonic on D. r) ⊂ G.
z ∈ G. Note that P(f. then the maximum principle implies that v ≤ u in G. and therefore on G. and then deﬁne Vn to be Vn oﬀ D and to be harmonic on D and ¯ ˜ continuous on D. r) ⊂ G. If u is a harmonic function on G continuous up to the boundary. ˜ ˜ ¯ Proof. Then u is subharmonic on G.36 ¯ ¯ Lemma. If v ∈ P(f. Vn converges to an harmonic function V on D. z ∈ ∂G. and v is any subharmonic function that is less than u on the boundary. G) contains all constants not exceeding min f . G) then v ≤ sup f on G. Now deﬁne the Perron function sup u(z) = v∈P(f. So it suﬃces to show that for each point a of ∂D. Let ˜ Vn = max(v1 . ˜ so V . so u ≤ u on D. vn ). ˜n (a) ↑ u(a). Clearly it is continuous and subharmonic in D and G \ D. Choose functions vn ∈ P(f. G) with vn (a) → u(a). Hence it is reasonable to try to solve the Dirichlet problem by seeking the largest subharmonic function that doesn’t exceed the given boundary values. From the maximum principle we obtain: Lemma. Perron. G) so are bounded by u.G) v(z). Thus ˜ u(a) = u(a) ≤ ˜ 1 2π 2π u(a + reiθ ) dθ ≤ 0 1 2π 2π u(a + reiθ ) dθ. All belong to P(f. Now u − u ˜ ˜ is subharmonic on D and 0 on ∂D. We know that Vn and Vn are subharmonic. u is bounded by its mean on any circle around a. This is the approach of O. . Deﬁne u ˜ ¯ as the continous function on D which agrees with u on ∂D and is harmonic on D. it extends continuously to G with value f on ∂G. G) as the set of subharmonic functions v on G for which ( ) lim sup v(z) ≤ f (a) z→a for all a ∈ ∂G. Let f be a bounded (at this point not necessarily continuous) realvalued function on ∂G. We will show that (1) the Perron function is harmonic on G. . Deﬁne the Perron family P(f. but vn ≤ Vn ≤ Vn . ¯ on D ¯ and set u = u on G \ D. . Thus we will have exhibited a solution to the Dirichlet problem. f (z). ˜ 0 Note that Let G be a bounded region. Suppose that u is subharmonic in a region G and D = B(a. and clearly the Vn and thus ˜ ˜ the Vn are nondecreasing. By Harnack’s Theorem. so it is not empty. Let D be any disk with closure in G and let a be any point in the disk. and (2) under mild restrictions ¯ on G. . Proof that the Perron function is harmonic.
. even on a disk minus a segment. This function is analytic on C\[0. Let M denote the maximum of f  on ∂G and m denote to the minimum of the barrier function φ on G \ D. so V − W is a nonpositive harmonic function on D which achieves its maximum at a. This will show that u is harmonic in D as desired. then there is a barrier function for G at z0 . and on many other domains. there is a harmonic function w on G.37 We know that V (a) = u(a). G). Let G be a bounded region and z0 ∈ ∂G. Theorem. suppose that there is a line through z0 such that G \ {z0 } lies inside one of the open halfplanes determined by the line. A standard example of a domain without a barrier at some boundary point is a punctured disk. and set wn = max(φn . G) with φn (b) → u(b). 1] and extends continuously to 1 with value 0. Wn ≥ Vn . ¯ For example. First we show that for any > 0. w(z) > M ≥ f (z). let D be a disk about z0 such that f (z) − f (z0 ) < for z ∈ ∂G ∩ D. Therefore V ≡ W on D. and is the only point of intersection of the interval [0. Thus barrier functions exist at all boundary points of smooth domains. then limz→z0 u(z) = f (z0 ). Then wn ≥ vn . 0]. Proof. Set w(z) = φ(z) φ(z) [M − f (z0 )] + f (z0 ) + = − 1 [M − f (z0 )] + M + . vn ) ∈ P(f. 1] with G. so m > 0. ¯ Deﬁnition. and u the corresponding Perron function. and its real part is everywhere positive on C \ [0. Let G be a bounded domain. ˜ ˜ Construct Wn and Wn from wn in analogy with Vn and Vn . We now turn to the continuity of the Perron solution at the boundary. f a bounded function on ∂G. continuous ¯ on G. Similarly. Choose functions φn ∈ P(f. If z0 is a point in ∂G possessing a barrier function and f is continuous at z0 . Next consider the function 1 − 1/z where we use the principal branch of the square root. w(z) ≥ f (z0 ) + > f (z). m m Then for z ∈ ∂G ∩ D. A continuous function φ on G ¯ \ {z0 } which is harmonic on G is called a barrier function for G at z0 if it is positive on G and zero at z0 . We may write the halfplane as Im[(z − z0 )/b] > 0 for some b ∈ C. so indeed V (b) = u(b). if z0 is any boundary ¯ point for which there exists a line segment which intersects G only at z0 . Indeed. 1]. polygonal domains. Also W (a) = u(a) = V (a). we can use Re 1 − 1/z as a barrier function for G at z0 . We will now show that if b is any other point of D then V (b) = u(b). and Wn converges to a harmonic function W ≥ V with W (b) = u(b). Thus if z0 = 1 ¯ belongs to ∂G. while for z ∈ ∂G \ D. ˜ ˜ Wn ≥ Vn . and so Im[(z − z0 )/b] is a barrier function for G at z0 . which satisﬁes w > f on ∂G and w(z0 ) = f (z0 ) + . deﬁned on C\(−∞.
but it is also easy to verify this directly using Morera’s theorem. they coincide. then f admits a unique extension to a harmonic function on all of G. The extension is given by u(z) = −u(¯) for z ∈ G− . lower halfplane. we have u ≤ w on G. so it suﬃces to show that it is harmonic in a neighborhood of z0 ∈ G0 . Reﬂection Principle for Analytic Functions. We could base a proof on the previous result applied to Im f and harmonic conjugates. If u is a continuous realvalued function on G+ ∪ G0 . D. z Proof. and the entire boundary of the lower halfdisk. Since both U and u are harmonic in the halfdisks. then v ≤ w on G. To state the reﬂection principles. From the Poisson kernel representation we have U = 0 on the real interval (−1. 1). and real axis. Thus U = u on the entire boundary of the upper halfdisk. ¯ Reﬂection Principle for Harmonic Functions. then u admits a unique extension to a harmonic function on all of G. z Proof. Let G be a bounded region in C which possesses a barrier at each point of its boundary. so it suﬃces to show that the stated extension deﬁnes a harmonic function in G. respectively. like U is harmonic in the whole disk. Let G be a region which is symmetric with respect to the real axis and deﬁne G+ . m m Now we get v < f on ∂G and v(z0 ) = f (z0 ) − . and obviously w(z0 ) = f (z0 ) + . G− . Since was arbitrary. Deﬁne U on D as the solution of the Dirichlet problem with boundary data u∂D (where u has been extended to G− by the formula above). we introduce the following terminology: a region G is symmetric with respect to the real axis if z ∈ G implies z ∈ G. Letting tend to zero we have lim inf z→z0 u(z) ≥ f (z0 ). and let f be a continuous function on ∂G. If such an extension exists it is certainly unique. Since v ∈ P(f. It follows that if v ∈ P(f. G). Let G. u ≥ v on G. G− . and G0 as the intersection of G with the upper halfplane. and so lim supz→z0 u(z) ≤ w(z0 ) = f (z0 ) + . By translating and extending we can assume that z0 = 0 and that G contains ¯ ¯ the closed unit disk. which is analytic on G+ and real on G0 . If f is a continuous complexvalued function on G+ ∪ G0 . The extension is given by f (z) = f (¯) for z ∈ G− . G+ . Then the Perron function for f on G solves the Dirichlet problem. and so lim inf z→z0 u(z) ≥ f (z0 ) − . Next let v(z) = f (z0 ) − − φ(z) φ(z) [M + f (z0 )] = 1 − [M + f (z0 )] − M − . Since this is true of all such v. and thus u. . lim supz→z0 u(z) ≤ f (z0 ). The Schwarz Reﬂection Principle. G). It certainly deﬁnes a continuous extension which is harmonic in G+ ∪ G− .38 So w > f on the boundary as claimed. Corollary. and G0 be as above. which is harmonic on G+ and zero on G0 .
and real on the line. it can be extended to be analytic on the whole domain. so it is not necessary that f be real on the symmetry line.) The extended function is analytic everywhere except at the points symmetric to the zeros of f . continuous on its closure. As an application of the reﬂection principle consider an analytic function on the unit disk which extends continuously to the closure taking the unit circle into itself. (We obtain this by applying the standard reﬂection principle to g = φ−1 ◦ f ◦ φ where φ(z) = (1 + iz)/(1 − iz). and real on the boundary. deﬁned and analytic on one side of the line. where it has a pole. We may z extend this function to C∞ by f (z) = 1/f (1/¯) for z > 1. . Inﬁnity is either a removable singularity or a pole according to whether f is 0 at 0 or not. Let f be analytic on the unit disk. f (w) = f (1/w) for w in the exterior of the disk. not just the real axis. Thus if G is symmetric with respect to any line. It follows that f is rational. Then f admits a unique extension to the entire plane. The extension is given by f (z) = f (1/¯) for z > 1. Let g(z) = f ((1 + iz)/(1 − iz)). so g maps the upper half plane into C and is real on the real axis. We can also translate and rotate the image.39 Of course the line of symmetry could be any line. An important application is to analytic continuation. For example. The proof can be found by translating and rotating the domain to the standard case. The reﬂection principles can be applied for symmetries with respect to circles as well as for lines. Thus the extended function is meromorphic on C∞ with a ﬁnite number of poles. or. Theorem. z Proof. Applying the reﬂection principle to g gives an extension of f with f 1 + iz 1 − iz =f 1 + i¯ z 1 − i¯ z for z in the lower half plane. equivalently. we can apply the reﬂection principle repeatedly to extend f to an entire function. if f maps a rectangle into itself analytically and takes the edges into the edges. it is suﬃcient that it map it into some other line.