The University of Sydney

School of Mathematics and Statistics

Solutions to Week 7 Tutorial and Homework

MATH3979/4079: Complex Analysis (Advanced) Semester 1, 2026
Lecturer: Dmitry Badziahin

1. Let f be an entire function such that Re (f (z)) ≤ M for some M ∈ R. Prove that
f is constant.
Hint: Consider the function ef (z) .
Solution: The function F (z) = ef (z) is entire function and |F (z)| ≤ eM . The
Liouville theorem then implies that F is constant, and f (z) will take values in the
discrete set of the form {α + 2kπ | k ∈ Z}. The continuity of f then implies that
it is constant.

2. Let f be an entire function which takes values in the upper half-plane: Im (f (z)) >
0 for each z ∈ C. Show that f is constant.
Hint: Find a Möbius transformation that maps the upper half-plane into a disk.
z−i
Solution: The Möbius transformation φ(z) = maps the real line to the
z+i
unit circle centered at the origin. Moreover, all points z in the upper-half plane
are closer to i than to −i, i.e. |z − i| < |z + i|, so φ(z) is inside the circle.
It follows that φ ◦ f is an entire function which maps the plane into the unit circle,
so it is constant by the Liouville theorem. Thus, f must be constant.

3. Prove that the range of an entire non-constant function is dense in C.

Solution: Suppose that f is an entire function, and a ∈ C such that no point
z with |z − a| < ε is in the range of f , i.e. |f (z) − a| ≥ ε for all z ∈ C. Function
1 1
g(z) = is thus entire and |g(z)| ≤ , so it is also bounded. Thus, g is
f (z) − a ε
constant, so f also must be constant.

4. Suppose f be an entire function and M > 0 such that |Re (f (z))| ≥ |Im (f (z))| for
all z with |z| ≥ M . Prove that f is constant.
Solution: Suppose f is not constant. Then, according to Question 3, f (C) is
dense in C, i.e. f (C) = C. Denote:
K = {z ∈ C | |z| ≤ M } and Ω = {z ∈ C | |Re (f (z))| ≥ |Im (f (z))|} .
We have f (C \ K) ⊂ Ω. Since Ω is closed and f continuous, f (C \ K) ⊂ Ω. Also,
since K is compact and f continuous, f (K) is closed and bounded.
Now we have:
C = f (C) = f (K) ∪ f (C \ K) = f (K) ∪ f (C \ K) ⊂ f (K) ∪ Ω.
Thus, C \ Ω ⊂ f (K), which is not possible since C \ Ω is unbounded.

Copyright © 2026 The University of Sydney 1

5. Let f be an entire function, K > 0 a real number, and n > 0 an integer such that
|f (z)| ≤ K|z|n for sufficiently large |z|. Prove that f is a polynomial.
Solution: Let R > 0 be a real number such that the inequality is satisfied
whenever |z| ≥ R.
Take r ≥ R and m ≥ n. We have:
Z
(m) m! f (z)
|f (0)| = dz by Cauchy’s integral formula
2πi |z|=r z m+1
|f (z)|
Z
m!
≤ |dz|
2π |z|=r |z|m+1
Z
m!
= |f (z)||dz|
2πrm+1 |z|=r
Z
m!
≤ K|z|n |dz|
2πrm+1 |z|=r
Z
m!K
= |dz|
2πrm−n+1 |z|=r
m!K
= m−n .
r
Since f is entire, this is valid for all r ≥ R, so we may let r → ∞, which gives
f (m) (0) = 0 for each m > n.
The Taylor series representation then implies that f is a polynomial of degree not
greater than n.

6. Let f be an entire function such that |f (z)| ≤ | cos(z)| for each z ∈ C. Prove that
f (z) = k cos(z) for some constant k.
f (z)
Solution: Consider g(z) = . g(z) is bounded wherever it is defined: |g(z)| ≤
cos z
1. cos(z) is an entire function, so its zeroes are isolated. Thus the singularities of
g(z) are also isolated.
At each singular point α, we have lim g(z)(z − α) = 0, since g is bounded. Thus,
z→α
by the theorem on removable singularities, there is an entire extension ĝ of g,
lim g(z) = ĝ(α). That extension is thus also bounded by 1, so it must be constant.
z→α

7. Define F : C3 → C3 by:

F (u, v, w) = (u + v + w, uv + vw + wu, uvw).

Is F onto and one-to-one?

Solution: F is not injective, since, for example, F (1, 0, 0) = F (0, 1, 0). In fact
F is invariant with respect to permutations of (u, v, w).
Let (A, B, C) ∈ C. Consider the polynomial z 3 −Az 2 +Bz−C. By the Fundamental
Theorem of Algebra, that polynomial has three zeroes α, β, γ in C. Vieta’s formulas
then imply F (α, β, γ) = (A, B, C), so F is onto.

2
 8. Let f and g be two entire functions and k ∈ R such that Re (f (z)) ≤ k · Im (g(z))
for all z. Show that there are constants a, b such that f (z) = ag(z) + b.
Solution: Let h(z) = f (z) + ikg(z). Then h is entire and Re (h(z)) ≤ 0. Now
consider r(z) = eh(z) . It is also entire but on top of that, |r(z)| ≤, i.e. it is bounded.
Hence r(z) = const and h(z) is a constant as well.

1
9. Let f be an entire function such that |f (z)| ≤ p , for each z which is not on
|Re z|
the imaginary axis. Prove that f is constant.
1 √
Solution: For |Re z| ≥ , we have |f (z)| ≤ 2.
2
1
Let z0 be a point such that |Re z0 | ≤ , and S the square with vertices i Im z0 ±1±i.
2
Point z0 is in the interior of that square, so the Cauchy integral formula gives:
Z
1 f (z)
f (z0 ) = dz.
2πi ∂S z − z0
Here, we integrated along the sides of the square in the counterclockwise direction.
1
Since |z − z0 | ≥ for z ∈ ∂S, thus
2
f (z) 2
≤p , z ∈ ∂S.
z − z0 |Re z|

If z is on one of the vertical edges, we have:

2
p ≤ 4,
|Re z|

while integrating along a horizontal edge gives:

Z 1
1
2 p dx = 8.
−1 |x|

Hence,
1 16
|f (z0 )| ≤
(8 + 8 + 8 + 8) = ,
2π π
proving that f is bounded. By the Liouville theorem, it will be constant.

10. Let γ be positively oriented circle {z | |z| = 1}, a and b two distinct points in C
which do not belong to γ, and m, n positive integers. Compute:
Z
dz
m n
.
γ (z − a) (z − b)

Solution: If |a| > 1 and |b| > 1, the integrand is analytic inside γ, so the integral
equals 0.

3
Suppose now that only one of the two points, say a is inside γ, and set
1
f (z) = .
(z − b)n

That function is analytic for |z| < |b|. Then the required integral equals, by the
Cauchy differential formula:
Z
f (z) 2πi
m
dz = f (m) (a)
γ (z − a) (m − 1)!
2πi (m + n − 1)! 1
= (−1)m
(m − 1)! (n − 1)! (a − b)n+m
(m + n − 1)! (−1)m
= 2πi · ·
(n − 1)!(m − 1)! (a − b)n+m

If both a and b are within γ, then let γa , γb be small positively oriented circles
centred at a, b, disjoint with each other, both within γ. By the general form of the
Cauchy theorem, we have that the required integral equals:
Z Z
dz dz
m n
+ m n
,
γa (z − a) (z − b) γb (z − a) (z − b)

which can then be calculated by application of the Cauchy differential formula, as

in the previous case. The result is:

(−1)m (−1)n
 
(m + n − 1)!
2πi · · + .
(n − 1)!(m − 1)! (a − b)n+m (b − a)n+m