M.A./ M.Sc.

Semester II-Complex Analysis

This is the remaining part of the Syllabus.

1 Hurwitz’s Theorem
Proposition 1 ((Hurwitz’s Theorem)) Let U = B(0, R) and supose < fn >
be a sequence in H(U ) converges to f . Let 0 < r < R be such that f has no
zero on the circle C of radius r at 0. Then there is n0 such that for n ≥ n0 ,
each fn has the same number of zeros inside C as f does.

Proof. Since C is compact and |f (z)| > 0 on C, there is δ > 0 such that

|f (z)| ≥ δ > 0 for z on C. Let n0 be such that |fn (z)| ≥ δ/2 for all z ∈ C and
n ≥ n0 . Then for z ∈ C and for n ≥ n0 ,

1 1 |fn (z) − f (z)| 2

− = ≤ 2 |fn (z) − f (z)|
fn (z) f (z) |fn (z)| |f (z)| δ
which proves that < 1/fn > converges uniformly to 1/f on C. Further, since
< fn0 > convergesR uniformly to f 0 on C, we have < fn0 /fRn > converges to f 0 /f
1
on C. Hence, 2πi C
(fn0 /fn ) converges uniformly to 2πi
1
C
(f 0 /f ) or
Z Z
1 1
lim (fn0 /fn ) = (f 0 /f )
n→∞ 2πi C 2πi C
which by the Argument Principle proves the result.

2 Residue at the point at infinity

If a is an isolated singularity of f , then there is a circle of radius r > 0 such
that f is holomorphic on {z : 0 < |z − a| < r} and
Z
1
f (ζ)dζ = Res(f, a).
2πi C
Keeping this in mind, if ∞ is an isolated singularity of f , we define
Z
1
Res(f, ∞) = − f (ζ)dζ, (1)
2πi C

where f is holomorphic outside B(0, R) except ∞.

Proposition 2 If f has only finitly many poles p1 , p2 , ..., pn , then

n
X
Res(f, ∞) + Res(f, pj ) = 0.
j=1

1
Proof. Let R > 0 be such that all the poles are inside the circle |z| = R. Then
by Residue theorem
Z n
1 X
f (ζ)dζ = Res(f, pj )
2πi C j=1

which by (1) proves the result.

Proposition 3 If f has an isolated singularity at ∞, then Res(f, ∞) = −Res(g, 0),

where
g(z) = 1/z 2 f (1/z).

1
R
Proof. Let R > 0 be such that Res(f, ∞) = − 2πi C
f (ζ)dζ, where C is the
circle |z| = R and f has no singularity outside C except ∞. If we take ζ = 1/w,
then dζ = −1/w2 dw and the circle C is tranformed by this inversion to the
circle C1 : |w| = 1/R oriented negatively. Hence,
Z
1
Res(f, ∞) = − f (ζ)dζ
2πi C
Z
1
1/w2 f (1/w)dw


= −
2πi C1
Z
1
= − g(w)dw,
2πi C1
1
R
where g has no singularity inside C1 except 0.Thus 2πi C1
g(w)dw =Res(g, 0)
which proves the result.

With the use of the results proved in the Propositions 2 and 3, we may find
integrals of the functions having large number of poles.
dz
R
Example 1 Evaluate I = C (z−5)(z 17 −1) , where C is the circle of radius 2 at

the origin.
17
1
P
Let f (z) = (z−5)(z 17 −1) . Then by Residue theorem I = 2πi Res(f, pj ),
j=1
where p0j s are 17 , 17th roots of unity. Obviously this sum is not easy to
compute but in view of the abve Propositions, we have
17
X
Res(f, ∞) + Res(f, pj ) + Res(f, 5) = 0
j=1

and
Res(f, ∞) = −Res(g, 0),
where
z 16
g(z) =
(1 − 5z) (1 − z 17 )

2
and
Res(g, 0) = 0.
Thus
17
X 1
Res(f, pj ) = −Res(f, 5) = −
j=1
517 − 1

and
2πi
I=− .
517 −1

3 Analytic Continuation
Let f1 and f2 be two functions analytic, respectively, analytic in the domains
D1 and D2 and let in the region D1 ∩ D2 , f1 (z) = f2 (z), then f1 and f2 are
called the analytic continuation of each other from one domain to another.
P∞ 1
For example: Let f1 (z) = n=0 z n (|z| < 1) and f2 (z) = 1−z (z 6= 1) . Then
the function f2 is an analytic continuation of f1 (z).
Analytic continuation is a property of analytic functions. Using this prop-
erty, we have following results:

Theorem 4 If f (z) is analytic in a domain D and let f (z) = 0 at some point

or at some part in D, then f (z) = 0 throughout D.

Proof. Let z0 be a point in D such that f (z0 ) = 0. Then in a Taylor’s series of

(n)
f (z) in some nbh. N0 (⊂ D) of z0 coefficients an = f n!(z0 ) = 0 ∀n and hence,
f (z) = 0 at each poit of N0 . In this way, we may see that f (z) = 0 throughout
D.

Theorem 5 There can not be more that one continuations in the same domain.

Proof. Let f (z) be analytic in D and f1 and f2 be two analytic continuations

of f in the same domain D1 . That is we have f = f1 in D∩ D1 and also f = f2
in D∩ D1 which implies that f1 = f2 in D∩ D1 , where D∩ D1 is a part of D1 .
Thus by (i), f1 = f2 throughout D1 . This proves the uniqueness property on
analytic continuation.
The best method of analytic continuation is known as the power series
method of analytic continuation which is described as follows:
Let
∞
X n
f1 (z) = an (z − z1 ) (2)
n=0

be the Taylor’s series of the function f1 (z) at z1 . Then it is convergent within

1/n
the circle of radius r1 = lim |an | . Let D1 := (z : |z − z1 | < r1 ) . Then f1
n→∞
is analytic in D1 . Let L be a curve joining z1 to another point zn outside D1 .

3
We perform analytic continuation of f1 from D1 to Dn := (z : |z − zn | < rr ) as
follows: Let z2 be any point on L lying within D1 . Then from (2), we may find
∞ ∞
(k)
X n! n−k
X (m + k)! m
f1 (z) = an (z − z1 ) = am+k (z − z1 )
(n − k)! m=0
m!
n=k

and hence,
(n) ∞
f1 (z2 ) X (m + n)! m
= am+n (z2 − z1 ) =: bn .
n! m=0
m!n!

Thus a Taylor’s series of the function f1 (z) at z2 is given by

∞
X n
bn (z − z2 )
n=0

1/n
which converges to f2 (z) (say) within the circle of radius r2 = lim |bn | . Let
n→∞
D2 := (z : |z − z2 | < r2 ) . Then f2 is an analytic continuation of f1 from D1 to
D2 . Clearly, f1 = f2 in the common region D1 ∩ D2 . Continuing in this way,
we can get a Taylor’s series of the function f1 (z) at zn which converges to the
function fn within the disc Dn . The function fn is an analytic continuation of
f1 from D1 to Dn along the curve L.

Some times the continuation of a power series is not possible beyond its
region of convergence through any small arc of its circle of convergence, in that
case the circle of convergence is called a natural boundary.

Example 2 The circle of convergence of the power series

∞
X n
f (z) = 1 + z + z 2 + z 4 + z 8 + ... = 1 + z2
n=0

is a natural boundary.

For any assistance, please contact at phone : 9453601277, 7991200647.

E-mail: [email protected]