Chapter 8

Exercise VIII.7.1
Prove that if the holomorphic function f has an isolated singularity at z0 ,
then the principle part of the Laurent series of f at z0 converges in C \ {z0 }.
P∞Let f be a holomorphic
n
function with an isolated singularity at z0 and
n=−∞ an (z − z0 ) be the Laurent series of f at z0 .
If z0 is removable, then the principle part of the Laurent series is just 0
and if z0 is a pole, then the principle part is just a rational function defined
for all z ∈ C{z0 }.
So, the only case we might worry about convergence for is if z0 is an
essential singularity.
There exists some R > 0 such that the Laurent series converges for z ∈
A0,R (z0 ). Let 0 < r < R and notice if z is outside A0,R (z0 ), then

r < |z − z0 | ⇒ |a−n | |z − z0 |−n < |a−n | r−n .

Since the principle part of the Laurent series converges at r, it must also
converge for z outside A0,R (z0 ). ■

1
Exercise VIII.7.2
Locate and classify the isolated singularities of the following functions.

(a)
z5
1 + z + z2 + z3 + z4
This function has singularities at ∞ and the roots of the polynomial
p(z) = 1 + z + z 2 + z 3 + z 4 .
First lets analyze the roots of p(z) as singularities. Notice for z ̸= 1,
z5 − 1
p(z) = . So, the roots of p(z) are zk = cis (2πk/5), k = 1, 2, 3, 4.
z−1
For an arbitrary root zk , the original function can then be written as
1 z5
Q
(z − zk ) j̸=k (z − zj )
| {z }
h(z)
Y
Since (z − zj ) is holomorphic and nonzero at zk , there exists an open disk
j̸=k
Dk centered at zk where h(z) is holomorphic. Let ∞ n
P
n=0 an (z − zk ) represent
h(z) for z ∈ Dk .
So, for z ∈ Dk \ {zk } we can write the original function as the Laurent
series ∞
a0 X
+ an (z − zk )n−1 .
z − zk n=1
Since the principle part is nonzero, finite and zk is an arbitrary root of p(z),
all the roots of p(z) are poles.
Finally lets analyze ∞ as a singularity. Plugging in z1 into the original
function we have for z ̸= 0
z −5 1 1
= = .
1 + z −1 + z −2 + z −3 + z −4 z(1 + z + z 2 + z 3 + z 4 ) zp(z)
Since p(z) is non-zero and holomorphic on an open disk D0 centered at 0,
1
p(z)
is also holomorphic on D0 and has a power series representation centered
X∞
at 0. Let bn z n be this power series.
n=0

2
 ∞
1 b0 X
For z in the punctured disk D0 \{0} we can write = + bn z n−1 .
zp(z) z n=1
Since the principle part of this power series is nonzero and finite, ∞ is a pole
of our original function.

(b)
1
sin2 (z)
This function has singularities at ∞ and the zeros of sin2 (z).
Notice sin2 (z) = 0 if and only if z = πk for k ∈ Z. Let Dr denote the
closed disk centered at the origin of radius r > 0. Since C \ Dr contains zeros
of sin2 (z) for any r > 0, ∞ is not an isolated singularity of our function.
So, the only isolated singularities are z = πk for k ∈ Z. Since cos(πk) ̸= 0,
all of these points are simple zeros of sin(z). We can therefore write sin(z) =
(z − πk)hk (z) for some holomorphic function hk (z) where hk (πk) ̸= 0.
1
Let ∞ n
P
n=0 an (z − πk) represent . The Laurent series of our orig-
(hk (z))2
inal function at πk is then
∞
a0 a1 X
+ + an (z − πk)n−2
(z − πk)2 z − πk n=2

Since the principle part is finite, the roots of sin(z) are poles of our function.

(c)
 
1
sin
z
The singularities of this function are 0 and ∞.
sin(z) is entire and so we can express it as a power series.
∞
X (−1)n 2n+1
sin(z) = z
n=0
(2n + 1)!

The Laurent series expansion of sin(1/z) at 0 is then

∞
X (−1)n 2n+1
z
n=0
(2n + 1)!

3
Since infinitely many terms of the principle part are nonzero, we conclude
that 0 is an essential singularity of sin(1/z).
To analyze ∞ we look at 0 after plugging in 1/z into our original function
we are just left with sin(z). The Laurent series expansion of sin(z) at zero
has a principle part equal to 0. Therefore, ∞ is a removable singularity of
our original function.

4
Exercise VIII.7.3
Prove that a rational function has no essential singularities.
Let r(z) = f (z)/g(z) be a rational function. The singularities of r(z) are
∞ and (if any) the zeros of g(z).
First assuming g is nonconstant, lets analyze the zeros of g(z). Let w be
a zero of g(z). There are two cases: 1. w is also a zero of f (z) 2. w is not a
zero of f (z).
Case 1. w is a zero of f (z). Let p be the order of a zero that w is of f and q
(z − w)p F (z)
be the order of a zero that w is of g. We can then write r(z) =
(z − w)q G(z)
where F and G are holomorphic functions in an open disk centered at w such
that f (z) = (z − w)p F (z), F (w) ̸= 0, g(z) = (z − w)q G(z), and G(w) ̸= 0.
F (z)
Since is holomorphic on an open disk centered at w, it is also analytic
G(z)
here meaning for z in this disk,
∞
X
p−q
r(z) = (z − w) cn (z − w)n .
n=0

So, if p ≥ q, r(z) is represented by a Laurent series with a principle part

equal to 0 and w is removable. However, if p < q, then w is a pole of order
p − q of r(z).
Case 2. w is not a zero of f (z). Let G be a holomorphic function in a
disk centered at w and q be a positive integer such that g(z) = (z − w)q G(z)
and G(w) ̸= 0. Since G(w) ̸= 0, there is an open disk centered at w where
1
is analytic. So we can write for z in this disk,
G(z)
∞
1 X
r(z) = dn (z − w)n
(z − w)q n=0

Similar to case 1 we have that w is then a pole of order q of r(z).

So, the zeros of g(z) are not essential singularities of r(z). Finally lets
analyze ∞ as a possible essential singularity. f and g are just polynomials
so we can write,

am z m + am−1 z m−1 + · · · + a0
r(z) = .
bn z n + bn−1 z n−1 + · · · + b0

5
where am ̸= 0 and bn ̸= 0. To determine the behavior of r(z) at ∞ we look
at r(1/z).
am z −m + am−1 z −m−1 + · · · + a0
r(1/z) =
bn z −n + bn−1 z −n−1 + · · · + b0
If n < m, we can scale r(1/z) by z m /z m and see

a0 z m + a1 z m−1 + · · · + am
r(1/z) =
b0 z m + b1 z m−1 + bn z m−n
This is a rational function with 0 as a zero of the polynomial in the denom-
inator. So, we can apply essentially the same reasoning when analyzing the
zeros of g(z) to conclude that 0 is either a removable singularity or a pole of
r(1/z).
Now if n ≥ m, because limz→0 r(1/z) is a constant, the continuity of
r(1/z) implies that r(1/z) is bounded and by the criterion for a removable
singularity, we conclude that 0 is removable.
So, the only singularities of r(z) are either removable or a pole meaning,
r(z) has no essential singularities. ■

6
Exercise VIII.7.5
Let the holomorphic function f be defined in C \ F where F is a finite set.
Assume f has no essential singularities. Prove that f is a rational function.
Since F is finite, each point in it must be isolated. We can extend f
holomorphically to the points in F that are removable singularities using
Cauchy’s formula for a circle leaving only poles for analysis.
Suppose f has n poles Let w1 be a pole of f , then we can write for a
nonnegative integer m1 , f (z) = (z − w1 )−m1 g1 (z) for a holomorphic g1 (z)
such that g1 (w1 ) ̸= 0.
Let w2 be a pole of f (z)(z − w1 )m1 , and repeat this for all poles of f
leaving,

f (z) = (z − w1 )−m1 (z − w2 )−m2 · · · (z − wn )−mn gn (z),

where gn is an entire function with no poles except possibly at ∞.

If gn is 0 everywhere, then f would also be 0 everywhere and have no
poles. Otherwise gn has a finite number of zeros. Let z1 , z2 , · · · , zk denote
the zeros of gn . We can then write

gn (z) = (z − z1 )n1 (z − z2 )n2 · · · (z − zn )nk h(z),

where h(z) is entire and nonzero. Now, we look at two cases for gn .
1. If ∞ is a pole of gn , then lim 1/gn (z) = 0 meaning, 1/gn is bounded
z→∞
in some neighborhood of ∞. However, since 1/h(z) is entire, then as shown
in exercise 7.11.1, 1/h(z) has to be a polynomial. So, we are left with

(z − z1 )n1 (z − z2 )n2 · · · (z − zk )nk 1

f (z) = m m m
·
(z − w1 ) 1 (z − w2 ) 2 · · · (z − wn ) n h(z)

2. If ∞ is just a removable singularity of gn , then gn is bounded in some

neighborhood of ∞. However, in this case gn would be a bounded entire
function, meaning, it’s constant by Louiville’s theorem. So, if gn (z) = c
where c is some constant, leaving
c
f (z) = .
(z − w1 )m1 (z − w2 )m2 · · · (z − wn )mn

7
Exercise VIII.8.1
Show that Liouville’s theorem is a corollary of the criterion for a removable
singularity.
Suppose f is entire and bounded. We can then write for any z ∈ C,
∞
X
f (z) = an z n
n=0

Notice ∞ is an isolated singularity of f (z). Since f (z) is bounded, f ( z1 )

is also bounded in C \ {0}.
The criterion for a removable singularity implies that the principle
P∞  part
1 1
 an z −n

of the Laurent series expansion of f ( z ) is 0. So, f ( z ) = a0 + 
n=1

meaning, f (z) is constant for all z ∈ C. ■

8
Exercise VIII.8.2
Let the holomorphic function f have an isolated singularity at the point z0
of C. Assume that there are postive numbers M and ϵ and a positive integer
m such that |f (z)| ≤ M |z − z0 |−m for 0 < |z − z0 | < ϵ. Prove that z0 is
either a removable singularity of f or a pole of order at most m.
Let g(z) = (z − z0 )m f (z). g has an isolated singularity at z0 and is
bounded on A0,ϵ (z0 ). So, the criterion for a removable singularity implies
X∞
that z0 is a removable singularity of g and therefore g(z) = an (z − z0 )n .
n=0
We can then write the Laurent series expansion of f as
∞
a0 a1 am−1 X
+ + · · · + + an (z − z0 )n−m .
(z − z0 )m (z − z0 )m−1 z − z0 n=m

Worst case a0 ̸= 0 and z0 is an order m pole and if a0 = a1 = · · · = am−1

z0 is a removable singularity. ■

9
Exercise VIII.12.1
Let the function g and h be holomorphic in an open set containing the point
z0 , and assume that h has a simple zero at z0 . Prove that

g g(z0 )
resz0 = ′ .
h h (z0 )

Since h has simple zero at z0 there exists a holomorphic function f such

g
that f (z0 ) ̸= 0 and h(z) = (z −z0 )f (z). The function is holomorphic on an
f
open diskP∞ D centered z0 and hence has a power series expansion centered at
n
z0 . Let n=0 an (z − z0 ) be this power series. We then have for z ∈ D \ {z0 }
∞
g 1 X
= an (z − z0 )n .
h z − z0 n=0

g(z0 )
So, the residue is just a0 which is . Notice h′ (z0 ) = f (z0 ) meaning,
f (z0 )
g g(z0 )
resz0 = ′ . ■
h h (z0 )

10
Exercise VIII.12.2
Determine the residues of the following functions at each of their isolated
singularities in C.

(a)
zp
f (z) = (p, q positive integers)
1 − zq
f (z) has singularities at ∞ and the zeros of 1 − z q .
z p+q
First we look at ∞, multiplying f (1/z) by p+q we have for z ̸= 0,
z
q q
z −z 1
p q
= p
z (z − 1) z 1 − zq
So, f (1/z) is represented for 0 < |z| < 1 by the Laurent series,
∞ ∞
−z q X q n X (n+1)q−p
(z ) = −z
z p n=0 n=0

If there exists a nonnegative integer n such that (n + 1)q − p = −1, then the
residue of f (z) at ∞ is −1. Otherwise, the residue is 0.
Now we look at the zeros of 1 − z q . Since (1 − z q )′ = −z q−1 , all zeros of
1 − z q are simple. Exercise 8.12.1 guarentees that the residue of f (z) at a
zero w of 1 − z q is −wp /wq−1 .

(b)
z5
(z 2 − 1)2
This function has singularities at ∞ and the zeros of (z 2 − 1)2 .
First we look at ∞. In exercise 8.7.3 it was shown for a rational function
that if the degree of the polynomial in the numerator is greater than the de-
gree of the polynomial in the denominator, then ∞ is a removable singularity
meaning, the residue of this function at ∞ is 0.
Now we look at the zeros of (z 2 − 1)2 . The zeros are −1 and 1. To find
the residue at −1, notice the Laruent series at −1 can be written as,
∞
z5 1 z5 1 X
= · = an (z + 1)n .
(z 2 − 1)2 (z + 1)2 (z − 1)2 (z + 1)2 n=0

11
 ′
z5

The residue is a1 which is at z = −1. Computing this we have
(z − 1)2

5z 4 (z − 1)2 − 2z 5 (z − 1) 5 · 22 − 2 · 2 16
= = = 1.
(z − 1)4 z=−1 24 16

Essentially the same reasoning lets write the residue at 1 as

5z 4 (z + 1)2 − 2z 5 (z + 1) 5 · 22 − 2 · 2
= = 1.
(z + 1)4 z=1 16

(c)
cos(z)
f (z) =
1 + z + z2
This function has singularities at ∞ and the zeros of 1 + z + z 2 .
Notice limz→∞ f (z) = 0 meaning, by the criterion for a removable singu-
larity ∞ is a removable singularity of f (z). So, the residue of f (z) at ∞ is
0. √
The zeros of 1 + z + z 2 are 21 ± i 21 3 and since (1 + z + z 2 )′ = 1 + 2z, these
zeros are simple zeros allowing the use of exercise 8.12.1. So, the residue at
a zero w of 1 + z + z 2 , is cos(w)/(2w + 1).

(d)
1
sin(z)
This function has singularities at ∞ and the zeros of sin(z).
Similar to exercise 8.7.2, ∞ is not isolated leaving just the zeros of sin(z)
for analysis. sin(z) = 0 when z = πk for k ∈ Z. Since (sin(z))prime = cos(z),
each zero of sin(z) is a simple zero allowing the use of exercise 8.12.1. So, the
residue at a zero πk is 1/cos(πk) which is 1 if k is even and −1 otherwise.

12
Exercise VIII.12.3
Let the function f be holomorphic in an open set containing the point z0 and
have a zero at z0 of order m. Prove that
f′
resz0 =m
f
Since f has a zero of order m, there exists a holomorphic function g
defined on the same open set where f is such that g(z0 ) ̸= 0 and f (z) =
(z − z0 )m g(z).
Notice f ′ (z) = m(z − z0 )m−1 g(z) + (z − z0 )m g ′ (z) meaning, for a circle C
centered at z0 we have
Z ′
g ′ (z)
Z
f (z) m
dz = + dz = 2πim.
C f (z) C z − z0 g(z)

So, the residue is m. ■

13