Chapter 4

Exercise IV.1.1
Find the real and imaginary parts of the function exp (ez ).
Let z = x + iy.

exp(ez ) = exp(ex (cos(y) + i sin(y)))

= exp(ex cos(y)) exp(iex sin(y))
= exp(ex cos(y))(cos(ex sin(y) + i sin(ex sin(y))))
= exp(ex cos(y)) cos(ex sin(y)) + i exp(ex cos(y)) sin(ex sin(y))

Exercise IV.3.1
Suppose the function f is holomorphic in C and satisfies f ′ = f . Prove that
f is a constant times ez . (Suggestion: Consider the function e−z f (z)).
Let g(z) = e−z f (z) where f (z) = f ′ (z) for any z ∈ C.

g ′ (z) = −e−z f (z) + e−z f ′ (z)

= −g(z) + e−z f (z) = −g(z) + g(z) = 0.

So, there’s a constant C ∈ C where g(z) = e−z f (z) = C. Since for any z,
ez ̸= 0, we have f (z) = Cez .

Exercise IV.5.1
Describe the images under the map z 7−→ ez of the line y = x and of the
strip
π π
x− <y <x+
2 2
z
y = x under e is a spiral emanating out from around the origin.
In the case of the strip, for any value of x we have a semicircle centered at
0 of radius ex . As x increases this semicircle rotates and the radius increases,
tracing out a thick spiral emanating from the origin.

1
Exercise IV.5.2
Describe the curves |f | = constant and arg f = constant for the function

f (z) = exp(z 2 )

With z 2 = x2 − y 2 + 2ixy, exp(z 2 ) = exp(x2 − y 2 ) cis (2xy).

If |f | = r ∈ R, then exp(x2 − y 2 ) = r. So, we have x2 − y 2 = ln r which
is a hyperbola.
If arg f = θ ∈ [0, 2π), then 2xy = θ. So, we have another hyperbola.

2
Exercise IV.5.3
Repeat Exercise 4.5.2 for the function
 
z+1
f (z) = exp
z−1

z+1 z+1
ϕ(z) = ↔ ϕ−1 (z) =
z−1 z−1
Since ez sends the coordinate grid for rectangular coordinates to the coor-
dinate grid for polar coordinates, if |ez | is constant, then the preimage of this
curve under ez is a vertical line, and if arg ez is constant, then the preimage
is a horizontal line.
ϕ is an LFT that is it’s own inverse. So, the image of horizontal lines
and vertical lines under ϕ correspond to the curves of arg exp(ϕ(z)) = c and
|exp(ϕ(z))| = c respectively.

Exercise IV.6.2
Derive the identities

cosh(z + w) = cosh(z) cosh(w) + sinh(z) sinh(w)

sinh(z + w) = sinh(z) cosh(w) + cosh(z) sinh(w).

cosh(z) cosh(w) + sinh(z) sinh(w)

1
1 w
1 z
1 w
= ez + e−z e + e−w + e − e−z e − e−w


2 2 2 2
1  z+w  1
 + e−z−w +  + e−z−w

= e + ez−w ew−z
 +  ez+w − 
ez−w ew−z
 − 
4 4
1 z+w
+ e−z−w


= e
2
= cosh(z + w)

3
 sinh(z) cosh(w) + cosh(z) sinh(w)
1
1 w
1 z
1 w
= ez − e−z e + e−w + e + e−z e − e−w


2 2 2 2
1  z+w  − e−z−w +
 1
 − e−z−w

= e +ez−w
 − ew−z ez−w
ez+w −  ew−z
 + 
 
4 4
1 z+w
− e−z−w


= e
2
= sinh(z + w)

4
Exercise IV.8.1
Find all the roots of the equation cos(z) = 2.
cos(z) = (eiz − e−iz ) /2.

0 = eiz − eiz = e−iz e2iz − 1

= e2iz − 1

So, cos(z) = 0 when e2iz = 1 ⇔ z = πk for k ∈ Z.

Exercise IV.8.3
Describe the images of the lines Re z = constant and Im z = constant under
the map z 7→ cos z.
Using the identities established in 4.8.2, fixing the real part of z yields a
hyperbola, while fixing the imaginary part of z yields a circle.

Exercise IV.9.1
Find all values of cosh (log(2)).

log(2) = ln(2) + i2πk, k ∈ Z

1
cosh (log(2)) = (exp(ln(2) + i2πk) + exp(− ln(2) − i2πk))
2 
1 1 5
= 2+ =
2 2 8

5
Exercise IV.9.3
In what sense is it true that log(zw) = log z + log w for complex numbers z
and w?
In the same sense that arg(zw) = arg z + arg w.

log zw = ln |zw| + i arg(zw)

= ln |z| + i arg z + ln |w| + i arg w
= log z + log w

Particular branches of log do not maintain this equality everwhere, how-

ever the set of logarithms of zw is the same as the set

{ ln |z| + i Arg z + ln |w| + i Arg w + i2πn | n ∈ Z }.

Exercise IV.10.1
Prove that there is no branch of arg z in the region 0 < |z| < 1.
Suppose such a branch α exists. Let γ(t) = reit where 0 < r < 1 and
t ∈ R. If 0 ≤ t1 < t1 ≤ 2π, then there are m, n ∈ Z where α(γ(t1 )) = t1 +2πm
and α(γ(t2 )) = t2 +2πn. Since α(γ) is continuous, we must be able to restrict
the distance between t1 and t2 so that

|α(γ(t1 )) − α(γ(t2 ))| < 2π

|t1 + 2πm − t2 − 2πn| < 2π
|t1 − t2 | < 2π − 2π |m − n| .

If m ̸= n, then |t1 − t2 | < 0 which isn’t possible. So, m = n implying

α(γ(t1 )) < α(γ(t2 )). Hence, α(γ) is increasing on [0, 2π]. However, this
means
α(γ(0)) < α(γ(0 + 2π)) ⇔ α(r) < α(r).
So, no α exists.

6
Exercise IV.12.1
Prove that the logarithmic derivative of the product of two holomorphic
functions equals the sum of their logarithmic derivatives.
Let f ̸= 0 and g ̸= 0 be complex differentiable functions, then the loga-
rithmic derivative of f g is:

(f g)′ f ′g + f g′ f ′ g′
= = + .
fg fg f g

Exercise IV.13.2
1 h′ f′
Prove that if h is a branch of f n , then h is holomorphic and h
= nf
.
g 1
Let g be a branch of log f . The function h = e n is a branch of f n .
g n ′ g
h′ e g′ f′
= ng = =
h en n nf

Exercise IV.13.3
Let G be the open set one obtains by removing from C the interval
r [−1, 1]
z+1
on the real axis. Prove that there is a branch of the function in G.
z−1
z+1
Let ϕ(z) = . ϕ maps the real axis to the real axis and the unit circle
z−1
to extended imaginary axis. Since ϕ(0) = −1, ϕ must put the interval [−1, 1]
onto the negative real axis.
The range of ϕ is then the open set C \ (−∞, 0], which is the domain of
the principle
√ branch of log. Let g be a branch of log ϕ, then exp(g/2) is a
branch of ϕ.

7
Exercise IV.13.4
Let
√ G be as in Exercise 4.13.3. Prove that there is a branch of the function
z 2 − 1 in G.
Let f (z) = z 2 − 1. The problem is to find a continuous function h so that
for any z ∈ G, (h(z))2 = f (z).
If reiθ = √
f (z), then for any k ∈ Z, reiθ+i2πk = f (z). So, the valid values
√ iθ/2
iθ/2+iπ
for h(z) are re = ± re .
f is continuous
√ iθ/2 and non-zero in G, so we can always write the square root
of f (z) as re which is also continuous, yielding the branch.

Exercise IV.14.2
Prove that a branch of the inverse of a holomorphic function is always uni-
valent.
Let f : G 7→ C be a holomorphic function and g be a branch of it’s
inverse.
If g is not onto, then there is at least one point, w, in the codomain of g
that g does not map to. However, g(w) ∈ G, so g(f (w)) = w.
If g is not one-to-one, then there are at least two points w1 ̸= w2 where
for some z ∈ G, z = g(w1 ) = g(w2 ). Since w1 and w2 are in the range of
f , there must be two points z1 ̸= z2 where f (z1 ) = w1 and f (z2 ) = w2 .
However, f (z) = f (g(f (z1 ))) = f (g(f (z2 ))), implying that w1 = w2 .

8
Exercise IV.15.1
Find a relation between the branches of arctan z and the logarithm function.
Let z = tan(w).
1
2i
(eiw − e−iw )
z= 1
2
(eiw + e−iw )
zeiw + ze−iw = −ie iw
+ ie −iw

ze2iw + z = −ie2iw + i
−z + i
e2iw =
z+i
So, the branches of arctan z are expressed by
 
1 −z + i
log .
2i z+i

Exercise IV.16.1
Find all the values of (1 + i)i .

(1 + i)i = exp (i log(1 + i))

= exp (i ln |1 + i| − Arg (1 + i) − 2πn) , n∈Z
 √ π 
= exp i ln 2 − − 2πn
4
√
 
8n + 1
= exp i ln 2 − π
4