Unofficial answers to M337/S 1995 page 1 of 1

Based on tutorials of dr. Paul Reed and the OUSA revision weekend 1998, and my own
Typed by I.R. van de Stadt BSc. (Hons)

Part I
Question 1
(a) (i)
(a) (ii)
(b)

(i + 1) = 1 (i + 1) Arg w = 3
4
(i 1)(i + 1) 2
14
w = 12 (i + 1) = 1 e i3 4 (2 1 8 e i3 4 ) = 2 1 8 e i3 16 = 2 1 8 (cos 316 + i sin 316 )

w=

( )

i
i = ei 2 i i = ei 2 = e 2 (Principal argument)

Question 2
(a)

(b)

A {1} is not a region. A B is not a region.

(c)

B A is compact. A is not compact.

Question 3
(a)
The domain of the function f is C . F is continuous by the Composition Rule, since the functions Imz
and z 2 are basic continuous functions.
(b)

= 0, = 1 + i (t ) = (1 + i )t , t [0,1] (t )(1 + i ) , Im z 2 Im(1 + i )2 t 2 = 2t 2

So,

1 2
21
f (z )dz = 0 2t (1 + i )dt = (1 + i )[23 t ]0 = 23 (1 + i )

Question 4
(a)

Taylor series is f (0 ) + zf (0 ) +

z2
z3
z4
f (0 ) +
f (0 ) +
f (0 ) + L
2!
3!
4!

Since f ( z ) = e z cos z , we have

f (0 ) = 1

f ( z ) = e cos z e sin z

f (0 ) = 1

f (z ) = e z cos z e z sin z e z sin z e z cos z = 2e z sin z

f (0 ) = 0

f (z ) = 2e sin z 2e cos z

f (0 ) = 2

f ( z ) = 2 e z sin z + e z cos z + e z cos z e z sin z = 4e z cos z

f (0 ) = 4

(i)

So up to z we have
2
4
f (z ) = 1 + z z3 z 4 + L
3!
4!
Since f (z ) is analytic for all discs 0 z r , the Taylor series represents f (z ) for all discs 0 z r
(i.e. all of C )
f (z )
1
2 3 4 4
4
1 1 2

dz =
+ z z 2 + Ldz

1 + z z z + Ldz =
z =1 z 2
z =1 z 2
z =1 z 2
3!
4!
z 3!
4!

(only this term is important since

from it we find the residue at z = 0 )
So,

z =1

f (z )
z2

dz = 2iRes( f ,0 ) = 2i 1 = 2i

(ii)

Unofficial answers to M337/S 1995 page 2 of 2

Based on tutorials of dr. Paul Reed and the OUSA revision weekend 1998, and my own
Typed by I.R. van de Stadt BSc. (Hons)
d
Note that g (z ) =
f (z ) . Since f (z ) is represented by its Taylor series in all discs we have
dz
d
2 3 4 4

2 4 3
g (z ) =
1 + z z z + L = 1 z z + L
dz
3!
4!
3!

(by the Differentiation Rule). The result is the first 3 terms of the Taylor series for g ( z ) .

Question 5
(a)

e3iz
e3iz
f (z ) = 2
=
Note this has simple poles at 2i
z + 4 (z 2i )(z + 2i )
To find Residue use Cover up rule (Theorem 1.1 unit C1)
Residue at 2i
e3iz
e 6 ie 6
=
=
(z 2i )(z + 2i ) 4i
4
z 2i
Residue at -2i
lim (z 2i )

e3iz
e6
ie 6
=
=
(z 2i )(z + 2i ) 4i 4
z 2i
[Note the cover up rule only works for simple poles. For multiple poles which have a term like
lim (z + 2i )

(z )n ,

(b)

n = 2,3,4,K in the denominator then other methods need to be used]

We have
e3ti

=
Re
dt
dt
t 2 + 4
t 2 + 4

Theorem 3.4 (C1) satisfied.

Since there are no simple poles on the real axis, we have

cos 3t

e3ti

2i

t 2 + 4 dt = 2i{sum of residues in upper half plane}

So we have (using the result of part (a)
ie 6
cos 3t

= 1 e 6
=
dt
Re
2
i

4 2
t 2 + 4

Question 6
(a)

We have e z = e x +iy = e x . Since z = 1 we have e z e 1

[Here you are expected to know that e < 3 ]

(b)

So we have e z e 1 > 13
Rouchs Theorem (page 19 Unit C1)
We have f e z = 13 z 4 = 13
So we have f e z = 13 e z , for z z = 1
So f has the same number of zeros inside z = 1 as e z . However e z has no zeros and hence f has no
zeros inside z = 1 . This is Rouchs Theorem.

(c)

f (z )dz = 0

This is because f (z ) is analytic an non-zero inside . So 1 f (z ) is analytic and so the integral is zero
by Cauchys Theorem.

Unofficial answers to M337/S 1995 page 3 of 3

Based on tutorials of dr. Paul Reed and the OUSA revision weekend 1998, and my own
Typed by I.R. van de Stadt BSc. (Hons)
Question 7
(a)
(b)

(c)

q (z ) =

2
is analytic on C {0} so q represents a model fluid flow (See HB. 1.14 p.38)
z
2
A primitive of q ( z ) = is (z ) = 2 Log z , z C {x R : x 0}
z
is a complex potential function for the flow.
i
1 + i
Streamlines satify the equation
Im (z ) = Im 2 Log z = Im(2 log e z + i 2 Arg z ) = 2 Arg z = k
o
0
2
2(1 + i )
2
q (i ) = = 2i ; q (1 + i ) = ____ =
= (1 + i )
i
(1 i )(1 + i )
1+ i

F = Im C q (z )dz = Im(2i 2)= 4

by Cauchys Integral Formula, since the function f (z ) = 2 is

analytic on C . C is a simply-connected region and the unit circle isis a simple-closed contour in C
and 0 is a point inside the unit circle.
Question 8
(a)

See HB 2.1 p. 41. zn +1 = 2 zn 2 zn2 , with a = 2, b = 2, c = 0 d = 2 0 + 12 2 14 22 = 0

So the sequence is conjugate ot the iteration sequence wn +1 = w2 .

Then the conjugating function is h(z ) = 2 z + 1 w0 = 2 z0 + 1 = 2 1 + 1 = 3

As required.

(b) (i)

P1 i (0 ) = 1 i; P1 (0 ) = ( 1 i )2 1 i = 1 + i; P 1i (0 ) = ( 1 + i )2 1 i = 1 3i
3
1i

(b) (ii)

(0)

10 > , so c M (See HB 4.5 p. 42).

i is in the large cardioid. Hence we check HB 4.9 (a) p. 43:

2
i

8 4 32 + 8 (i ) = 12 2 = 1 < 3

So by HB 4.3 p 43 c M .

Unofficial answers to M337/S 1995 page 4 of 4

Based on tutorials of dr. Paul Reed and the OUSA revision weekend 1998, and my own
Typed by I.R. van de Stadt BSc. (Hons)

Part II
Question 9
(a) (i) sin ( + iy ) = sin x cos

+ cos x sin iy = sin cosh y + i cos sinh z

) (

sin (x + iy ) = sin 2 x cosh 2 y + cos 2 x sinh 2 z = sin 2 x 1 + sinh 2 y + 1 sin 2 x sinh 2 y

2

= sin x + sin x sinh y + sinh y sin x sinh y = sin x + sinh y

as required.
cos( x + iy ) = cos x cos iy + sin x sin iy = cos x cosh y + i sin x sinh z
2

)(

cos(x + iy ) = cos2 x cosh 2 y + sin 2 x sinh 2 z = cos 2 x cosh 2 y + 1 cos 2 x cosh 2 x 1

2

= cos 2 x cosh 2 y + cosh 2 x 1 cos 2 x cosh 2 y + cos 2 x = cosh 2 x 1 + cos 2 x

= sinh 2 x + cos2 x
(a) (ii)

tan z

sinh 2 x + sin 2 x
=
1 sinh 2 x + sin 2 x sinh 2 x + cos 2 x
2
2
2
sinh
x
+
cos
x
cos z
sin z

sin 2 x cos2 x 14 x 14

(b)

f is defined on C . u (x, y ) = x 2 + by 2 ; v(x, y ) = 2axy . Now,

u
v
v
u
= 2by
= 2ax;
= 2ax;
= 2 x;
y
x
y
x
all exist on C , are all continuous on C , since they are all multiples (Multiple Rule) of the basic
continuous function x. The Cauchy-Riemann equations are statisfied on all z C when
2 x = 2ax a = 1; 2ax = 2by b = 1
Then f is differentiable on all of C when a = 1 and b = 1 , and, because its domain C
analytic on C . (See HB 2.3 p.19 and HB 1.3 p.18)

is a region f is

Unofficial answers to M337/S 1995 page 5 of 5

Based on tutorials of dr. Paul Reed and the OUSA revision weekend 1998, and my own
Typed by I.R. van de Stadt BSc. (Hons)
Question 10
(a)

(b) (i)

f (x ) =

sin z
has a removable singularity at z = 0 and simple poles at z = 3i and z = 3i . So
z (z 3i )( z + 3i )
by the Cover-Up Rule
sin 0
sin 3i
1 i sinh 3
Res( f ,0 ) =
= 0; Res( f ,3i ) =
= 18
( 3i )(3i )
3i 6i
sin ( 3i ) sin 3i 1
Res( f ,3i ) =
=
= i sinh 3
3i ( 6i ) 3i 6i 18
f is analytic on the simply-connected region C except for 3 singularities. is a simple-closed contour
in C , not passing through any of the singularities. So, since only 0 is inside and the other
singularities are outside

f (z )dz = 2i Res( f ,0) = 0

(see part (a))

by Cauchys Residue Theorem

(b) (ii) f is analytic on the simply-connected region C except for 3 singularities. is a simple-closed contour
in C , not passing through any of the singularities. So, all three singularies are inside

f (z )dz = 2i(Res( f ,0) + Res( f ,3i ) + Res( f ,3i )) = 2i(0 181 i sinh 3 + 181 i sinh 3) = 0

(c)

by Cauchys Residue Theorem

f is analytic on the simply-connected region C

(see part (a))

except for 3 singularities. Let = z 3i = 1 .Then is

a simple-closed contour in C , not passing through any of the singularities and only 3i is inside and
the other singularities are outside . Hence, by Cauchys Residue Theorem
1 i sinh 3 = sinh 3 (see part (a))
f (z )dz = 2i Res( f ,3i ) = 2i 18

9
as required.

(d)

zf (z )

z2 + 9
in analytic on {z : z < }, except for a simple pole at 0
sin z

02 + 9
1
Res
,0 =
= 9,
zf ( z ) cos 0
d sin z
by the g/h Rule, since z 2 + 9 and
= cos z are analytic at 0 and cos 0 = 1 0
dz

Unofficial answers to M337/S 1995 page 6 of 6

Based on tutorials of dr. Paul Reed and the OUSA revision weekend 1998, and my own
Typed by I.R. van de Stadt BSc. (Hons)

Question 11
(a) (i)

( )

( ( ))

exp e z = exp Re e z = exp e x (cos y ) as required.

( )

(a) (ii) The interval R = {z : 1 < x < 1, < y < } is a bounded region. The function f ( z ) = exp e z is
analytic on R and continuous and non-zero on R , then by the Maximum Principle (HB. p. 31) there is a
maximum is on R

[ (
)] (
)
On {z : x = 1, y }, max [exp(e (cos y ))] = exp(e (cos 0 )) = e
On {z : 1 x 1, y = }, max [exp(e (cos( )))] = exp( e ) = e
On {z : 1 x 1, y = }, max [exp(e (cos( )))] = exp( e ) = e
So max {exp(e ) : 1 Re z 1, Im z } is e which is attained at z = 1
On {z : x = 1, y }, max exp e1 (cos y ) = exp e1 (cos 0 ) = ee
1

(b)

e 1

e 1

Let h( z ) =

r
, z C {r}
rz
Now f and g are Basic Taylor Series (HB 3.5 p. 25)
f (z ) =

( )

1
z
= 1 rz ,
r
n = 0

g (z ) =

n =1

n =0

z < r f (z ) =
n

( )

r
= h(z ),
(r z )

1
+ 1 = 1 rz + 1,

z <r

z >r

z
z
z+rz
r
=
= h( z ), z > r
+1 =
+1 =
(r z )
(r z ) (r z )
(z r )
f , g and h are analytic on their domains.
The region z < r , the domain of f overlaps with the region z C {r} , the domain of h. Also the
g (z ) =

region z > r , the domain of g overlaps with z C {r} . Hence f and h are direct continuations of each
other (or f is a direct continuation of h by Taylor series (HB 2.1 p.33) and so are h and g likewise. The
regions z < r and z > r do not overlap, and so f and g are indirect continuations of each other. (HB
1.1 p. 33 and HB 2.3 p.33)

Unofficial answers to M337/S 1995 page 7 of 7

Based on tutorials of dr. Paul Reed and the OUSA revision weekend 1998, and my own
Typed by I.R. van de Stadt BSc. (Hons)
Question 12
(a) (i) The inverse mappings are to the standard triple of points 0,1, (HB 2.11 p.36)
= i 0
=11
=i
So f 1 corresponds to

(z + i )(1 i ) = (z + i )(1 i )2 = i(z + i ) = iz + 1

f 1 (z ) =
(z i )(1 + i ) (z i )(1 + i )(1 i ) (z i )
z i
and so the extended Mbius transformation f that maps 0 to i, 1 to 1 and to i is
iz 1 iz + 1
f (z ) =
=
z i
z+i
(a) (ii) The extended imaginary axis is the set {z : Re z = 0} {}
We use the three point trick. The imaginary axis passes through the points i, 0 and i.
Now
i2 + 1 2
i 0 +1 1
i2 + 1 0
= = , f (0 ) =
=
=0
f ( i ) =
= = i and f (i ) =
i+i
0
0+i
i
i+i
2i
So the image of the imaginary axis is the generalized circle that passes throught the extended points
, -i, and 0. Thus the image is the imaginary axis, as required.
Maybe the following is a safer proof.
The Apollonian form of the imaginary axis is {z : z + 1 = z 1 } {}
Now from part (a) (i) we know already that
iw + 1
z = f 1(w) =
wi

Since w f (Im axis ) f 1(w) Im axis . So, w = , w = i or

iw + 1 w + i
iw + 1 + w i
iw + 1
iw + 1

=
1
+1 =
wi
wi
wi
wi

(1 i )w + 1 i

(b) (i)

= (1 + i )w + 1 + i = 1 i w + 1 = 1 + i w + 1 w + 1 = w 1

Therefore the image is {w : w + 1 = w 1 } {} , which is the imaginary axis as required.

HB 1.10 p. 12 and HB 5.3 p. 14 .

) ( z )ei Arg z 2

Let h(z ) =

z = z1 2 = exp 12 Log z = exp 12 (log z + i Arg z ) =

So, if

i Arg z1 2

( z )e
1

( z )e
2

i Arg z 2 2

z1 ei Arg z1 z2 ei Arg z 2 z1 = z2

Hence h(z ) is a one-one function. h(z ) is analytic on {z : Re z > 0} , with h(z ) = 1 0 for all z in

{z : Re z > 0} .

2 z

By HB 2.6 p. 36 the Mbius transformation f is one-one on {z : Re z > 0} since

i i 1 1 = 2 0 and by HB 2.2 f is conformal. Now g (z ) = h( f (z )) =

f (z ) , so by HB 4.5 p. 37 g is

a one-one conformal mapping.

De domain is at the left of the points i, 0 and i, which f maps to 0, -i, and with the image at the
left. This shows that f maps {z : Re z > 0} to {z : Re z > 0} . Using the three point trick to show
where h(z ) maps the imaginary axis onto. The imaginary axis passes through the points i, 0 and i with
the domain at the left, which h(z ) maps to 1 (1 i ), 0, 1 (1 i ) , with the image at the left, which

shows that h(z ) maps {z : Re z > 0} to z : 14 < Arg z < 14

(b) (ii) f and h are one-one and conformal on these regions. h 1( z ) = z 2 Hence since (h o f )1 = f 1 o h 1 ,
g 1 (z ) =

iz 2 + 1
z2 i

z2 + i
iz 2 + 1