Solutions: Homework 3

February 7, 2020

Problem 1. Suppose f is analytic on B(0; 1) and satisfies |f (z)| < 1 for |z| = 1. Find the
number of solutions (counting multiplicities) of the equation f (z) = z n where n is an integer
larger than or equal to 1.

Proof. Let g(z) = f (z) − z n for all z ∈ B(0; 1), and let h(z) = z n . Then since |f (z)| < 1 for
|z| = 1, g has no zeroes or poles on the unit circle. Obviously, h also has no poles or zeroes
on the unit circle. Note that, for |z| = 1,

|g(z) + h(z)| = |f (z)| < 1 = |z n | < |h(z)| + |g(z)|

Then, by Rouché’s theorem, Zg − Pg = Zh − Ph . Since g and h are analytic on B(0; 1),

Pg = Ph = 0. Hence Zg = Zh = n, since h has a zero of multiplicity n at 0 and has no other
zeroes. So, the equation f (z) = z n has n solutions, counting multiplicities.

Problem 2. Let f be analytic in a neighbourhood of D = B(0; 1). If |f (z)| < 1 for |z| = 1,
show that there is a unique z with |z| < 1 and f (z) = z. If |f (z)| ≤ 1 for |z| = 1, what can
you say?

Proof. By the problem above, the equation f (z) = z has exactly one solution in B(0; 1).
Now, suppose that |f (z)| ≤ 1 for |z| = 1. Suppose that there does not exist z with |z| = 1
and f (z) = z. Then, applying the Rouché’s theorem to f (z) − z and z as in the previous
problem, we see that f (z) = z still has a unique solution in B(0; 1). If f (0) = 0, and if
there exists z ∈ B(0; 1) \ {0} such that f (z) = z, then by Schwarz lemma, we know that
f (z) = z for all z ∈ B(0; 1). Otherwise, we cannot conclude anything for sure. For example,
if f (z) = z n , then it has a unique solution in B(0; 1), but if f (z) = (z 2 + 1)/2, then it has
no solution in B(0; 1).

Problem 3. Prove the following Lemma: If (S, d) is a metric space then

d(s, t)
µ(s, t) =
1 + d(s, t)

is also a metric on S. A set is open in (S, d) iff it is open in (S, µ); a sequence is a Cauchy
sequence in (S, d) iff it is a Cauchy sequence in (S, µ).

1
Proof. µ(s, t) = 0 iff d(s, t) = 0 which happens iff s = t. Clearly, µ(s, t) = µ(t, s) and
µ(s, t) ≥ 0 for all s, t ∈ S. Let s1 , s2 , s3 ∈ S. Let a = d(s1 , s2 ), b = d(s1 , s3 ) and c = d(s3 , s2 ).
Then b ≤ a+c ≤ a+c+ac, where the first inequality is due to the triangle inequality applied
to d. So 1 + b ≤ 1 + a + c + ac = (1 + a)(1 + c). Again by the triangle inequality, we have
a ≤ b+c, so a−c ≤ b. Combining there two inequalities, we have (a−c)(1+b) ≤ b(1+a)(1+c),
which upon rearranging gives,
a c b
− ≤
1+a 1+c 1+b
Putting back the values of a, b, c in terms of d(., .), we have

µ(s1 , s2 ) ≤ µ(s1 , s3 ) + µ(s3 , s2 )

for all s1 , s2 , s3 ∈ S. This proves the triangle inequality. Hence, µ is a metric on S.

Now, let U be open in (S, d). Let x ∈ U. Then there exists  > 0 such that Bd (x; ) ⊂ U. So,
d(y,x) 
for all y ∈ S such that d(y, x) < , y ∈ U. Now, note that if µ(y, x) = 1+d(y,x) < 1+ , then
d(y, x) < . So, for all y ∈ S such that µ(y, x) < /(1 + ), y ∈ U. Since x ∈ U was arbitrary,
U is open in (S, µ). Now, let V be open in (S, µ). Let x ∈ U. Then there exists 1 >  > 0
such that Bµ (x; ) ⊂ V. So, for all y ∈ S such that µ(y, x) < , y ∈ V. Now, note that if
µ(y,x) 
d(y, x) = 1−µ(y,x) < 1− , then µ(y, x) < . So, for all y ∈ S such that d(y, x) < /(1−), y ∈ V.
Since x ∈ V was arbitrary, V is open in (S, d).
Now, let us define the function i : (S, d) → (S, µ) where i(x) = x for all x ∈ S. Then
µ(i(x), i(y)) ≤ d(x, y) for all x, y ∈ S, and hence i is Lipschitz, and so uniformly continuous.
This implies that if {xn } is a Cauchy sequence in (S, d), it is still a Cauchy sequence in (S, µ).
Now, let {xn } be a Cauchy sequence in (S, µ). Let  > 0. Then there exists N ∈ N such that

µ(xn , xm ) < 1+ for all n, m ≥ N. But this implies that d(xn , xm ) <  for all n, m ≥ N. So
{xn } is Cauchy in (S, d).

Problem 4. Suppose {fn } is a sequence in C(G, Ω) which converges to f and {zn } is a

sequence in G which converges to a point z in G. Show lim fn (zn ) = f (z).

Proof. Since fn converges to f in C(G, Ω), fn converges to f uniformly on any compact

subset of G. Let K = {zn |n ∈ N} ∪ {z}. Then K is compact. Hence fn → f uniformly on
K. Let  > 0. Then there exists N1 ∈ N such that d(fn (y), f (y)) < /2 for all n ≥ N1 and
for all y ∈ K. Since f is continuous, there exists N2 ∈ N such that d(f (zn ), f (z)) < /2 for
all n ≥ N2 . Then, for all n ≥ max(N1 , N2 ),

d(fn (zn ), f (z)) ≤ d(fn (zn ), f (zn )) + d(f (zn ), f (z)) < 

So, lim fn (zn ) = f (z).

Problem 5. (Dini’s Theorem) Consider C(G, R) and suppose that {fn } is a sequence
in C(G, R) which is monotonically increasing (i.e., fn (z) ≤ fn+1 (z) for all z in G) and
lim fn (z) = f (z) for all z in G where f ∈ C(G, R). Show that fn → f.

2
Proof. Replacing fn with f −fn , we can assume WLOG that f ≡ 0 and {fn } is monotonically
decreasing. Let K ⊂ G be compact. Also, note that fn ≥ 0 for all n ≥ 1 and lim fn (z) = 0 for
all z ∈ G. Fix  > 0. Let z0 ∈ K. Choose N (z0 ) ∈ N such that fn (z0 ) < /2 for all n ≥ N (z0 ).
Choose δ(z0 ) > 0 such that |fN (z0 ) (z) − fN (z0 ) (z0 )| < /2 for all z ∈ B(z0 ; δ(z0 )). Then, we
have fN (z0 ) (z) <  for all z ∈ B(z0 ; δ(z0 )). Since the fn ’s are monotonically decreasing, this
implies that fn (z) <  for all z ∈ B(z0 ; δ(z0 )) and for all n ≥ N (z0 ). Since K is compact,
there exists finitely many zi ’s, say, for 1 ≤ i ≤ r such that K ⊂ ∪ri=1 B(zi ; δ(zi )). Let
N = max{N (zi ) : 1 ≤ i ≤ r}. Then, for all n ≥ N, fn (z) <  for all z ∈ K. So fn → 0
uniformly in K. Since K is an arbitrary compact subset in G, this implies that fn → 0 in
C(G, R).
P∞ n
Problem Pn6. (a) kLet f be analytic on B(0; R) and let f (z) = n=0 an z for |z| < R. If
fn (z) = k=0 ak z , show that fn → f in C(G; C).
(b)
P∞Let G =nann(0; 0, R) and Plet f be analytic on G with Laurent series development f (z) =
n k
n=−∞ an z . Put fn (z) = k=−∞ ak z and show that fn → f in C(G; C).

Proof. (a) Let K ⊂ B(0; R) be compact. Then there exists r < R such that K ⊂ B(0; r).
Since r is less than the radius of convergence of this power series, by Theorem 1.3(c) of
Chapter III, we see that fn → f uniformly on B(0; r), and hence obviously on K. Since K
is an arbitrary compact subset of G, this implies that fn → f in C(G; C).
(b) Let K ⊂ ann(0; 0, R) beP compact. Then there exists P∞that K−n⊂
P∞ 0 < nr1 < r2 < R such
n k
ann(0; r1 , r2 ). Let fn,+ (z) = k=0 ak z . Let f+ (z) = n=0 an z and f− (z) = n=1 a−n z .
Then, by Theorem 1.11 in Chapter V, fn,+ → f+ uniformly on ann(0; r1 , r2 ), and hence in
K. Note that fn = f− + fn,+ . So fn → f− + f+ = f uniformly on K. Since K is an arbitrary
compact subset of G, this implies that fn → f in C(G; C).