Math 302: Solutions to Homework

Steven Miller
December 14, 2010
Abstract
Below are detailed solutions to the homework problems from Math 302
Complex Analysis (Williams College, Fall 2010, Professor Steven J. Miller,
sjm1@williams.edu). The course homepage is
http://www.williams.edu/Mathematics/
sjmiller/public html/302
and the textbook is Complex Analysis by Stein and Shakarchi (ISBN13: 978-
0-691-11385-2). Note to students: its nice to include the statement of the
problems, but I leave that up to you. I am only skimming the solutions. I
will occasionally add some comments or mention alternate solutions. If
you nd an error in these notes, let me know for extra credit.
Contents
1 Homework #1: Yuzhong (Jeff) Meng and Liyang Zhang 2
2 Homework #2: Nick Arnosti and Thomas Crawford 6
3 Homework #3: Carlos Dominguez, Carson Eisenach, David Gold 11
4 Homework #4: Pham, Jensen, Kolo glu 16
4.1 Chapter 3, Exercise 1 . . . . . . . . . . . . . . . . . . . . . . . . 16
4.2 Chapter 3, Exercise 2 . . . . . . . . . . . . . . . . . . . . . . . . 18
4.3 Chapter 3, Exercise 5 . . . . . . . . . . . . . . . . . . . . . . . . 20
4.4 Chapter 3 Exercise 15d . . . . . . . . . . . . . . . . . . . . . . . 23
4.5 Additional Problem 1 . . . . . . . . . . . . . . . . . . . . . . . . 24
4.6 Chapter 3 Exercise 17a, Extra Creadit . . . . . . . . . . . . . . . 25
1
5 Homework #5: Pegado, Vu 26
6 Homework #6: Kung, Lin, Waters 41
7 Homework #7: Thompson, Schrock, Tosteson 50
8 Homework #8: Xiong, Webster, Wilcox 56
1 Homework #1: Yuzhong (Jeff) Meng and Liyang
Zhang
Due by 10am Friday, September 17: Chapter 1: Page 24: #1abcd, #3, #13.
Problem: Chapter 1: #1: Describe geometrically the sets of points in the
complex plane dened by the following relations: (a)
1
=
2
where

1
,
2
; (b) 1/ = ; (c) Re() = 3; (d) Re() > (resp., ) where .
Solution: (a) When
1
=
2
, this is the line that perpendicularly bisects the
line segment from
1
to
2
. When
1
=
2
, this is the entire complex plane.
(b)
1

2
. (1.1)
So
1

2
= = 1. (1.2)
This is the unit circle in .
(c) This is the vertical line = 3.
(d) This is the open half-plane to the right of the vertical line = (or the closed
half-plane if it is ).
Problem: Chapter 1: #3: With =

, where 0 and , solve the

equation

= in where is a natural number. How many solutions are there?

Solution: Notice that
=

=
(+2)
, . (1.3)
2
Its worth spending a moment or two thinking what is the best choice for our
generic integer. Clearly is a bad choice as it is already used in the problem; as
we often use for the imaginary part, that is out too. The most natural is to use
(though would be another ne choice); at all costs do not use !
Based on this relationship, we have

=
(+2)
. (1.4)
So,
=
(+2)

. (1.5)
Thus, we will have unique solutions since each choice of {0, 1, . . . , 1}
yields a different solution so long as = 0. Note that = yields the same
solution as = 0; in general, if two choices of differ by then they yield the
same solution, and thus it sufces to look at the specied values of . If = 0,
then we have only 1 solution.
Problem: Chapter 1: #13: Suppose that is holomorphic in an open set .
Prove that in any one of the following cases:
(a) Re() is constant;
(b) Im() is constant;
(c) || is constant;
one can conclude that is constant.
Solution: Let () = (, ) = (, ) + (, ), where = + .
(a) Since Re() = ,

= 0,

= 0. (1.6)
By the Cauchy-Riemann equations,

= 0. (1.7)
Thus, in ,

() =

= 0 + 0 = 0. (1.8)
3
Thus () is constant.
(b) Since Im() = ,

= 0,

= 0. (1.9)
By the Cauchy-Riemann equations,

= 0. (1.10)
Thus in ,

() =

= 0 + 0 = 0. (1.11)
Thus is constant.
(c) We rst give a mostly correct argument; the reader should pay attention to
nd the difculty. Since =

2
+
2
is constant,
{
0 =
(
2
+
2
)

= 2

+ 2

.
0 =
(
2
+
2
)

= 2

+ 2

.
(1.12)
Plug in the Cauchy-Riemann equations and we get

= 0. (1.13)

= 0. (1.14)
(1.14)

. (1.15)
Plug (1.15) into (1.13) and we get

2
+
2

= 0. (1.16)
So
2
+
2
= 0 or

= 0.
4
If
2
+
2
= 0, then, since , are real, = = 0, and thus = 0 which is
constant.
Thus we may assume
2
+
2
equals a non-zero constant, and we may divide
by it. We multiply both sides by and nd

= 0, then by (1.15),

= 0, and
by Cauchy-Riemann,

= 0.

= 0. (1.17)
Thus is constant.
Why is the above only mostly a proof? The problem is we have a division
by , and need to make sure everything is well-dened. Specically, we need to
know that is never zero. We do have

= 0 except at points where = 0, but

we would need to investigate that a bit more.
Lets return to
{
0 =
(
2
+
2
)

= 2

+ 2

.
0 =
(
2
+
2
)

= 2

+ 2

.
(1.18)
Plug in the Cauchy-Riemann equations and we get

= 0

= 0. (1.19)
We multiply the rst equation and the second by , and obtain

= 0

+
2

= 0. (1.20)
Adding the two yields

+
2

= 0, (1.21)
or equivalently
(
2
+
2
)

= 0. (1.22)
5
We now argue in a similar manner as before, except now we dont have the an-
noying in the denominator. If
2
+
2
= 0 then = = 0, else we can divide
by
2
+
2
and nd / = 0. Arguing along these lines nishes the proof.
One additional remark: we can trivially pass from results on partials with
respect to to those with respect to by noting that if = + has constant
magnitude, so too does = = +, which essentially switches the roles of
and . Though this isnt needed for this problem, arguments such as this can be
very useful.
Its worth mentioning that (a) and (b) follow immediately from (c). For exam-
ple, assume we know the real part of is constant. Consider
() = exp(()) = exp((, )) exp((, )).
As () = exp((, )), we see that the real part of being constant implies
the function has constant magnitude. By part (c) this implies that is constant,
which then implies that is constant.
2 Homework #2: Nick Arnosti and Thomas Craw-
ford
Due in my mailbox by 10am Friday, September 24: Chapter 1: Page 24:
#16abc, #24, #25ab. Chapter 2: (#1) We proved Goursats theorem for tri-
angles. Assume instead we know it holds for any rectangle; prove it holds
for any triangle. (#2) Let be the closed curve that is the unit circle cen-
tered at the origin, oriented counter-clockwise. Compute

() where
() is complex conjugation (so ( + ) = ). Repeat the problem for

()

for any integer (positive or negative), and compare this answer

to the results for

; is your answer surprising? (#3) Prove that the four

triangles in the subdivision in the proof of Goursats theorem are all similar to
the original triangle. (#4) In the proof of Goursats theorem we assumed that
f was complex differentiable (ie, holomorphic). Would the result still hold if
we only assumed f was continuous? If not, where does our proof break down?
6
Problem: If is a curve in , show that

() =

().
Parameterize by = () for in [, ], and dene () = ( + ).
Then () is a parameterization of on the interval [, ] (note that () =
(), () = ()). Additionally,

() =

( + ). It follows that

() =

(())

() =

(( + ))

( + ).
Making the substitution = + , we get that

=
(( + ))

( + ) =

=
(())

()
=

=
(())

(). (2.1)
But

=
(())

() =

(),
which proves the claim.
Problem: If is a circle centered at the origin, nd

.
We start by parameterizing by =

, 0 < 2, so =

. Then

2
0

) =
+1

2
0

(+1)
.
If = 1, this is
0

2
0
= 2. Otherwise, we get

+1

2
0

(+1)
=

+1
+ 1

(+1)

2
0
= 0.
7
Problem: If is a circle not containing the origin, nd

.
If = 1, the function () =

has a primitive (namely

+1
+1
), so by
Theorem 3.3 in Chapter 1 of our book,

() = 0.
If = 1, we parameterize by =
0
+

, 0 < 2, so =

.Then

2
0

0
+

2
0

1 +

.
Note that because our circle does not contain the origin,
0
> , so

<
1. Thus, we can write this expression as a geometric series:

2
0

1 +

2
0

=0
(

.
Interchanging the sum and the integral, we see that this is just

=0
(

0
)
+1

2
0

(+1)
=

=0
(

0
)
+1

(+1)
+ 1

2
0
= 0.
Why may we interchange? We can justify the interchange due to the fact that
the sum of the absolute values converges.
Problem: If is the unit circle centered at the origin, nd

.
We start by parameterizing by =

, 0 < 2, so =

and
=

. Then

2
0

) =

2
0

(1)
.
If = 1, this is

2
0
= 2. Otherwise, we get
8

2
0

(1)
=

(1)
1

2
0
= 0.
Note that instead of doing the algebra, we could have observed that on the unit
circle =
1
, so

. Applying our work from Problem 2, we

get the answer above.
Problem: Where in the proof of Goursats theorem do we use the fact that the
function is holomorphic? Is it sufcient to know that is continuous?
Start by recapping the main ideas behind the proof. We began by continually
splitting our triangle into smaller triangles. These triangles converge to a point
in the limit, and we called this point
0
. We then established the bound

()

()
()

.
Our goal was to show that this quantity tends to zero as
0
.
To do this, we Taylor expanded () around the point
0
: () = (
0
) +

(
0
)(
0
) + ()(
0
). Note that (
0
) divides (), so () 0 as

0
.

()
()

()
(
0
) +

(
0
)(
0
)

()
()(
0
)
The rst integrand in this sum has a primitive, so the value of this integral is
zero. Let

= max
on
() (). Then ()

, and
0
diam(
()
).
Hence, the value of the second integral is at most perim(
()
) diam(
()
)

.
Since the perimeter and diameter of
()
both decay at a rate of 2

, we estab-
lish the bound that

()
()

for some constant . Hence,

is an upper-bound for

()

, and since () 0 as
0
,

0 as
desired.
Now let us see what happens if we dont know that is differentiable. Using
only continuity, we can approximate () by (
0
) + ()(
0
). Dening

as before, we can still bound our integral by

. We want to say that

tends to 0, but lim

0
() = lim

0
()(
0
)

0
, which may not exist if is not
differentiable (and certainly may not tend to zero). Thus, this approach fails.
9
We could also try the expression () = (
0
) + (), and then () 0 as

0
. Unfortunately, without the factor of (
0
), our bound on

()
()
will simply be perim(
()
)

= 2

. Thus, our bound for

()
()
is 4

= 2

. Even though

tends to 0, the factor of 2

may
overwhelm it, so this approach fails. From these attempts, it seems that knowing
that was differentiable was a fairly important step in the proof.
Problem: Prove Goursats theorem for triangles using only the fact that it holds
for rectangles.
Note that it sufces to prove that the integral along any right triangle is zero,
since any triangle can be divided into two right triangles by dropping an altitude.
Given a right triangle ABC, by drawing a series of rectangles inside the trian-
gle, we can reduce the desired integral to the integral along a series of congruent
triangles similar to ABC, each of which border the original hypotenuse (as shown
in the gure).
Since is continuous on the original triangle ABC (a compact set) we know
that is uniformly continuous on the region of interest.
Thus, given any > 0, there exists a > 0 such that for any two points ,
in ABC with < , () () < . If is the length of the hypotenuse
of ABC, choose large enough so that the diameter of each small triangle, /,
is less than . Then for any triangle

and any point

on that triangle write

() = (

) + (), so that
10

() =

) + () =

) +

()
Since (

) is a constant, it has a primitive, so the rst integral is zero. Mean-

while, since any point on triangle

is within / of

, and we chose to
be such that / < , we know that () = () (

) < . Thus,

() < perim(

) . But perim(

) < 3/, so the integral of ()

along triangle

is at most 3/. Summing over all triangles, we see that the

integral of () along the entire curve is at most 3. Since this technique works
for arbitrarily small , this implies that the integral of along any right triangle is
zero, proving the claim.
3 Homework #3: Carlos Dominguez, Carson Eise-
nach, David Gold
HW: Due in my mailbox by 10am Friday, October 1 (even if this is Mountain
Day): Chapter 2, Page 64: #1, #8. Also do: Chapter 2: (#1) In the proof of
Liouvilles theorem we assumed f was bounded. Is it possible to remove that
assumption? In other words, is it enough to assume that () < () for
some real-valued, non-decreasing function ? If yes, how fast can we let
grow? (#2) a) Find all where the function () = 1/(1 +
4
) is not holomor-
phic; b) Let , , , and be integers such that = 1. Find all where
the function () = ( + )/( + ) is not holomorphic. (#3) Compute the
power series expansion of () = 1/(1 ) about the point = 1/2 (it might
help to do the next problem rst, or to write 1 as 1/2 ( 1/2)). (#4)
Do Chapter 1, Page 29, #18.
1. Let
1
denote the straight line along the real line from 0 to ,
2
denote the
eighth of a circle from to

4
, and
3
denote the line from

4
to 0.
Then by Cauchys theorem,

1
+
2
+
3

2
= 0.
11
We can calculate

2
=

(
/4
)
2

/4

=
/4

=
/4

0
cos (
2
) + sin (
2
)
=
/4

0
cos (
2
) sin (
2
)
So we can calculate the Fresnel integrals by calculating

1
+
2

2
, tak-
ing , dividing by
/4
, and looking at the real and negative imagi-
nary parts. First we show the integral over
2
goes to zero:

/4
0

/4
0

2
cos 2

/41/log
0

2
cos 2
+

/4
/41/log

2
cos 2

4

1
log
)

2
cos(

2
log
)
+
1
log

2
sin(
2
log
)
+
1
log
The
1
log
term goes to zero as goes to innity. So we need to show that
the rst term goes to zero. Note that sin /2 for positive sufciently
close to 0, since sin 0 = 0 and

sin 1/2 for sufciently small . So

for sufciently large the rst term is less than or equal to

1
log
=

4

log

log
,
which goes to zero as goes to innity. So, as , the contribution
from
2
goes to zero. And we know that as ,

2
=

/2.
12
So, nally,

0
cos (
2
) sin (
2
) =

2

1

2/2 +

2/2
=

2
4

2
4

as desired.
8. Since , is holomorphic in an open circle of radius centered at , 0
< < 1. And by Cauchys inequality,

()
()
!

Case 1: 0. For some 0 < < 1,

+
thus,
() (1 + + )

(1 + +)

by both the given and the triangle inequality. And in Cauchys inequality R
is just . So by combining results from above

()
()
!

(1 + +)

(1 + + + )

(1 + )

(1 +)

. (3.1)
Now let

=
(!)

(1 + )

thus,

()
()

(1 +)

.
Case 2: < 0. For some 0 < < 1,

13
by the reverse triangle inequality. When we rearrange the inequality we see
that
= +
Since is negative, our goal is to minimize (1+z) in order to get an upper
bound. Now, by combining our result above with the Cauchy inequality we
get that:

()
()
!

(1 +)

(1 + )

(1 )

(1 +)

. (3.2)
Now let

=
(!)

(1 )

thus,

()
()

(1 +)

.
q.e.d.
1. In the proof of Liouvilles theorem, we had that

(
0
)

where was an upper bound for . It only matters that is an upper bound
for in a disc of radius about
0
, however. Let

be the smallest upper

bound for in a disc of radius about
0
. Liouvilles theorem still holds
if

as long as

/ 0 for every choice of

0
. Alternatively,
we just need to grow slower than linear; say () is less than
1
or
/ log or anything like this (for those who have seen little-oh notation,
() = () sufces).
2. (a) is holomorphic wherever its derivative exists:

() =
4
3
1 +
4
That is, whenever
4
= 1. This gives =
/4
,
3/4
,
5/4
, and

7/4
, or

2
2
+

2
2
,

2
2
+

2
2
,

2
2

2
2
, and

2
2

2
2
.
14
(b) The = 1 condition prevents from being a mostly-constant
function with an undened value at = /. (That is, if = 0,
then / = /, and so the function would simply collapse to the value
of /.) So

() =
( + ) ( + )
( + )
2
=
1
( + )
2
The function is then not holomorphic at = /.
3. Just use the geometric series formula:
1
1
=
1
1/2 ( 1/2)
=
2
1 2( 1/2)
=

=0
2
+1
( 1/2)

.
4. Let () =

=0

. Then
() =

=0

(
0
+ (
0
))

=0

=0
(

)
(
0
)

0
]
=

=0
(
0
)

0
)
.
The inner sum converges by the root test:
limsup

)
=
1

lim

)
=
1

where R is the radius of convergence of the original power series for and
second limit is evaluated by noting 1

)

/
and lim

/
=
15
1. Since the inner sum has the same radius of convergence as the original
sum,
0
still lies in the disc of convergence in the inner sum; hence all the
coefcients of
0
converge, and has a power series expansion about

0
.
4 Homework #4: Pham, Jensen, Kolo glu
HW: Due in my mailbox by 10am Friday, October 8 (even if this is Mountain
Day): Chapter 3, Page 103: #1, #2, #5 (this is related to the Fourier trans-
form of the Cauchy density), #15d, #17a (hard). Additional: Let () =

=5

and () =

=2

be the Laurent expansions for two

functions holomorphic everywhere except possibly at = 0. a) Find the
residues of () and () at = 0; b) Find the residue of () + () = 0;
c) Find the residue of ()() at = 0; d) Find the residue of ()/() at
= 0.
4.1 Chapter 3, Exercise 1
Exercise 4.1. Using Eulers formula sin =

2
, show that the complex
zeros of sin are exactly the integers, and that they are each of order 1. Calulate
the residue of
1
sin
at = .
Solution: To show that the complex zeros of sin are exactly the integers,
we will show that

0
2
= 0 if and only if
0
.
First prove the forward direction. We see that

0
2
= 0 gives

0
=

0
. (4.1)
Since
0
= + with , ,

. (4.2)
For complex numbers to be equivalent, their magnitudes must be the same. Thus,

. (4.3)
This implies that = , so = 0. The angles corresponding to Equation 4.2
must be congruent modulo 2 as well. Thus,
mod 2, (4.4)
16
which means 0 or . So we have
2 mod 2 0, (4.5)
which implies that is an integer. Thus . Since = 0, we have
0
= ,
implying
0
.
To prove the backward direction, consider
0
for
0
even,
sin
0
=

2
=
1 1
2
= 0. (4.6)
Similarly for
0
odd,
sin
0
=

2
=
1 + 1
2
= 0. (4.7)
Thus sin
0
= 0 if and only if
0
. So the zeros of sin are exactly the
integers.
Next we must show that each zero has order 1. We refer to Theorem 1.1 in
Stein and Shakarchi.
Theorem 4.2. Suppose that f is holomorphic in a connected open set , has a
zero at a point
0
, and does not vanish identically in . Then there exists a
neighborhood of
0
, a non-vanishing holomorphic function g on U, and a
unique positive integer n such that () = (
0
)

() for all .
Since sin is analytic, take its Taylor series about
0
. We add zero to write
as
0
+
0
. Using properties of the sine function, we claim
sin = sin (+
0

0
) = sin (
0
) cos
0
+cos (
0
) sin
0
. (4.8)
Note this statement does require proof, but will follow from standard properties of
the exponential function (or from analytic continuation). The reason some work
needs to be done is that
0
need not be real, but the relation above does hold
when is real. What we are trying to do is understand the behavior of the function
near
0
from knowledge near 0 (as
0
is close to zero). This is a common trick,
but of course what makes this tractable is that we have the angle addition formula
for sine.
17
When
0
is an integer, we always have sin
0
= 0. If
0
is odd then cos
0
is -1 while if
0
is even it is 1. Thus for odd
0
,
sin =

1!
(
0
)
1
+

3
3!
(
0
)
3

5
5!
(
0
)
5
+ (4.9)
and for even
0
,
sin =

1!
(
0
)
1

3
3!
(
0
)
3
+

5
5!
(
0
)
5
. (4.10)
We thus see that all zeros are simple.
We now turn to nding the residue at = for 1/ sin. From our Taylor
expansion above, we have
1
sin
=
1
sin ( ) cos
=
1
cos
1
sin ( )
. (4.11)
The problem is now solved by using the Taylor expansion of sine and the geomet-
ric series. We have cos = (1)

, so
1
sin
= (1)

1
(
0
)
1
3!
(
0
)
3
+
=
(1)

0
1
1
(
1
3!
(
0
)
2
+
)
=
(1)

0
(
1 +
(
1
3!
(
0
)
2
+
)
+
(
1
3!
(
0
)
2
+
)
2
+
)
.
(4.12)
Note that each term in parentheses in the last line is divisible by (
0
)
2
, and
thus none of these will contribute to the residue, which is simply (1)

.
4.2 Chapter 3, Exercise 2
Exercise 4.3. Evaluate the integral

1 +
4
.
18
Solution: Consider the function () =
1
1+
4
. This function has poles at
1/() = 0
1 +
4
= 0
=
(

4
+

2
)
. (4.13)
Consider the contour of the semicircle in the upper half plane of radius ,
denoted . Denote the part of the contour along the real line
1
and the part along
the arc
2
. Note that two of the poles of () lie inside this contour. Thus by
Cauchys residue theorem,
1
2

= Res

(
/4
) + Res

(
3/4
). (4.14)
To nd the residues, write
() =
1
1 +
4
=
(
1

4
)(
1

3
4
)(
1

5
4
)(
1

7
4
)
.
Thus
Res

(
/4
) =
(
1

3
4
)(
1

5
4
)(
1

7
4
)
=

3
4
(
1
1
)(
1
2
)(
1
1 +
)
=
1 +
4

2
(4.15)
and similarly
Res

3
4
) =

9
4
(
1
1 +
)(
1
1
)(
1
2
)
=
1
4

2
(4.16)
Thus we have
1
2

=
1 +
4

2
+
1
4

2
=

2

2
. (4.17)
19
Now, note that

1
+
2
=

1
+

2
. (4.18)
Observe that

1
=

1
1 +
4

and that

2
=

1
1 +
4

1
1 +
4

max

1
1 +
4

=
1

4
1
. (4.19)
Thus
lim

lim

4
1
= 0. (4.20)
Hence, as ,

2
0. Therefore as we get our nal result;
lim

1
1 +
4
+ lim

2
=

1
1 +
4
=

2
. (4.21)
4.3 Chapter 3, Exercise 5
Exercise 4.4. Use contour integration to show that

+

2
(1+
2
)
2
=

2
(1 +
2)
2
for all real.
Solution: Let () =

2
(1+
2
)
2
=

2
(+)
2
()
2
. We see that () has poles of
order 2 at = . Thus
res

0
() = lim

(
0
)
2
(). (4.22)
20
Alternatively, we could write our function as
() =
()
(
0
)
2
, (4.23)
and then we need only compute the coefcient of the
0
term of .
Now consider the residue at
0
= :
res

0
=
() = lim

(
2
( + )
2
)
= lim

(2
2
( + )
2
2
2
( + )
3
)
=
1
2

1
4

2
. (4.24)
For
0
= , we have:
res

0
=
() = lim

(
2
( )
2
)
= lim

(2
2
( )
2
2
2
( )
3
)
=
1
2

2
+
1
4

2
. (4.25)
Now let us rst consider the case when < 0. We will use the contour of a
semicircle oriented counterclockwise in the upper half-plane with radius . Call
the portion of along the real line
1
and the arc portion
2
. Note that there is a
pole inside at
0
= . By the residue formula, we have that

() = 2
(
1
2

1
4

2
)
=
2

2
+
1
2

2
. (4.26)
We also know that

+

() = lim

1
(). (4.27)
Along
2
, =

and =

, where = cos + sin . Thus

2
() =

(1
2

2
)
2
. (4.28)
21
Then it follows that

2
()

2cos

2sin

(1
2

2
)
2

2sin
(1
2
)
2

(1
2
)
2
=

(
2
1)
2
. (4.29)
Taking the limit as goes to innity, we have
lim

2
()

lim

(
2
1)
2
= 0. (4.30)
Thus
lim

2
() = 0. (4.31)
So lim

() = lim

1
(). It thus follows from Equation 4.26 that

2
(1 +
2
)
2
=
2

2
+
1
2

2
=

2
(1 + 2)
2
(4.32)
Now consider 0. We will use the contour of a semicircle oriented
counterclockwise in the lower half-plane with radius . Call the portion of
along the real line
1
and the arc portion
2
. Note that there is a pole inside at

0
= . By the residue formula, we have that

() = 2
(
1
2

2
+
1
4

2
)
=
2

1
2

2
. (4.33)
Also note that,

() = lim

1
(). (4.34)
Along
2
, =

and =

, where = cos + sin . Thus,

2
() =

2cos

2sin

(1
2

2
)
2
. (4.35)
22
Accordingly,

2
()

2
()

2sin
(1
2

2
)
2

(1
2
)
2

=

(1
2
)
2
(4.36)
Taking the limit as goes to innity, we have
lim

2
()

lim

(
2
1)
2
= 0. (4.37)
And thus,
lim

2
() = 0. (4.38)
So lim

() = lim

1
(). It thus follows from Equation 4.33 that

2
(1 +
2
)
2
=
(

1
2

2
)
=

2
(1 + 2)
2
(4.39)
Thus for all real,

2
(1 +
2
)
2
=

2
(1 + 2)
2
(4.40)
4.4 Chapter 3 Exercise 15d
1
For any entire function , lets consider the function
()
. It is an entire function
and furthermore we have the real part of is bounded so:

=
+
=

1
Hint from Professor Miller
23
Hence

is bounded and therefore, by Louisvilles Theorem,

is constant. It
then follows that is constant .
Alternatively, we could argue as follows. We are told the real part of is
bounded. Lets assume that the real part is always at most 1 in absolute value.
Then if we consider () = 1/( ()) we have () 1. To see this, note
the real part of () is at least 1. We again have constructed a bounded, entire
function, and again by Liouvilles theorem we can conclude (and hence ) is
constant.
4.5 Additional Problem 1
2
Let:
() =

=5

() =

=2

1. We have:
res
0
=
1
res
0
=
1
2. We have
() + () =
3

=5

=2
(

So res
0
( + ) =
1
+
1
.
3. We have 1 = 5 +4 = 4 +3 = 3 +2 = 2 +1 = 1 +0 = 0 1 =
1 2 so:
res
0
( ) =
5

4
+
4

3
+
3

2
+
2

1
+
1

0
+
0

1
+
1

2
2
Hint from Professor Miller
24
4. We have (assuming
2
= 0):
()
()
=

=5

=2

=
1

=2

=2

=
1

=2

3

1 (
1

=1

2

)
. (4.41)
As 0 the nal quantity in parentheses tends to zero, and thus we can
expand using the geometric series formula. We only care about the constant
term of this fraction, as it is multiplied by 1/
2
and thus only the constant
term contributes to the pole. This is a very useful observation. It means that,
when we expand with the geometric series, we can drop many terms, as we
only need to keep terms that contribute to the constant term. Remember, we
are not trying to nd the Taylor expansion of this function, but rather just
one particular term. We can thus write:
()
()
=
1

=2

=0
(

=1

)
)

=
1

[
(
5

2
)
(
1

2
(
0

2
+ ) +
1

2
2
(
1

2
+ ) +
)
+(
4

1
)
(
1

2
(
1

1
+ ) +
)
+ (
3

0
) (1 + ) + .
]
(4.42)
So:
res
0
(

) =
1

2
[

5
(

0

2
+

1

2
2
) +
4
(

2
) +
3
]
.
4.6 Chapter 3 Exercise 17a, Extra Creadit
Exercise 4.5. Let be non-constant and holomorphic in an open set containing
the closed unit disc. Show that if () = 1 whenever = 1, then the image of
contains the unit disc.
25
Solution: Suppose () does not have a zero in the unit disc, . Then 1/()
is holomorphic in . Note that since () = 1 whenever = 1, 1/() =
1/() = 1 whenever = 1 as well. But () is holomorphic in , imply-
ing () 1 in by the maximum modulus principle since () = 1 on the
boundary of . We nd 1 () 1 in the unit disk, which implies that our
function is constant as its modulus is constant (from an earlier exercise), contra-
dicting the assumption that is not constant!
Let
0
. Consider the constant function () =
0
. On the unit circle,
() = 1 >
0
= () for all = 1. Thus by Rouches theorem, () and
()+() have the same number of zeroes inside the unit circle (ie, in ). But we
have shown that () has at least one zero, thus for some

, 0 = (

)+(

) =
(

)
0
. Thus for all
0
, there exists

such that (

) =
0
. Thus the
image of () contains the unit disc.
5 Homework #5: Pegado, Vu
HW: Due in my mailbox by 10am Friday, October 22 (as there is no class on
Tuesday): Chapter 5: Page 155: #6, #7, #9 (extra credit: what is the combi-
natorial signicance of this problem?). Chapter 3: Page 104: #10. Additional
Problems: (1) Find all poles of the function () = 1/(1
2
)
4
and nd the
residues at the poles. (2) Consider the map () = ( )/( + ). Show that
this is a 1-to-1 and onto map from the upper half plane (all = + with
> 0) to the unit disk. (3) Calculate the Weierstrass product for cos() (this
is also problem #10b in Chapter 5, and the answer is listed there), and for
tan().
6. Prove Walliss product formula

2
=
2 2
1 3

4 4
3 5
. . .
2 2
(21) (2 + 1)
. . . .
[Hint: Use the product formula for sin at = /2.]
6. We know (from p. 142) the product formula for the sine function is
sin()

=1
(
1

2

2
)
.
26
Let = 1/2. Then,
sin(/2)

=
1
2

=1
(
1
(1/2)
2

2
)
.
Using sin(/2) = 1, we simplify this equation:
1

=1
(
1

2

2
)
(5.1)
1

=
1
2

=1
(
1
1
(2)
2
)
(5.2)
2

=1
(
(2)
2
1
(2)
2
)
(5.3)
2

=1
(
(2 + 1)(2 1)
(2)
2
)
. (5.4)
(5.5)
But this implies that

2
=

=1
(
(2)
2
(2 + 1)(2 1)
)
,
proving the identity.
7. Establish the following properties of innite products.
(a) Show that if

2
converges, and

= 1, then the product

(1 +

)
converges to a non-zero limit if and only if

converges.
(b) Find an example of a sequence of complex numbers {

} such that

con-
verges but

(1 +

) diverges.
(c) Also nd an example such that

(1 +

) converges and

diverges.
7. a) Let

2
converge with
1
= 1.
() First assume

converges to a nonzero limit. Without loss of gener-

ality we may assume that each

satises

1/2; this is clearly true in the

27
limit (as the sum converges, the summands must tend to zero). We assume this
to facilitate expanding with logarithms. Consider the product

(1 +

). Taking
logs, we see log
(
(1 +

)
)
=

log(1 +

). Setting =

and using the

Taylor expansion
log(1 + ) = +

2
2


3
3
+

4
4
,
we see that
log
(

(1 +

)
)
=

2
+

3

3

)
.
In general, notice that

=2

2
2
+

3
3

=2

,
or

2
(1 + +
2
+ . . . )

2
2
+

3
3

2
(1 + +
2
+ . . . ).
If a sum

converges to a nonzero limit, we know that converges to zero;
thus we may assume (without changing convergence) that
1
2
. Thus using the
geometric expansion, we see that 1 + +
2
+ =
1
1
. Because
1
2
,
we have that
1
1
2. Hence we have that
2
2

2
2
+

3
3

2
2
.
Recall that we were looking at log
(
(1 +

)
)
=
(

2
+

3

3

)
.
Since

converges, we know eventually we must have

< 1/2, so we can

assume

< 1/2 without changing convergence, and thus use the simplication
involving the geometric series expansion developed in the previous paragraph.
Thus we write
log
(

(1 +

)
)
=

2
+

3

3

)

+ 2

2
)
=

+ 2

2
.
(5.6)
28
AQUICKWORDOFWARNING. THEABOVEEQUATION, ANDTHE
ONES BELOW, AREALITTLE ODD. REMEMBER THATOURSEQUENCE
NEED NOT BE JUST REAL NUMBERS. AS SUCH, WE MUST BE CARE-
FUL WITH THE DEFINITION OF ABSOLUTE VALUE. WE ABUSE NO-
TATION A BIT WHEN WE WRITE + , THIS MEANS THE DE-
SIRED RELATION IS TRUE UP TO A LINEAR RESCALING. REALLY
WHAT WE MEAN IS = UP TO AN ERROR AT MOST . WE RE-
ALLY SHOULD WRITE THINGS LIKE , BUT IN A HOPE-
FULLY OBVIOUS ABUSE OF NOTATION....
Since by assumption both

and

2
converge, we must have that

+2

2
is nite, call it . Thus log
(
(1 +

)
)
, so

(1 +

, which is again nite. Thus the product converges.

() Next assume

(1 +

) converges to a nonzero limit. Since

(1 +

)
is converging to a nonzero limit, the terms in the product must be converging to
1, so we must have

approaching zero and we can assume

< 1/2 without

affecting convergence.
We now write:
log
(

(1 +

)
)
=

2
+

3

3

)

2
2

3
3

)

3

)
.
(5.7)
As before, we substitute in using the geometric series expansion:
log
(

(1 +

)
)

3

)
=

2
(1 +

2
+ )
)

2
)
=

2
.
(5.8)
Thus we see that log
(
(1 +

)
)
+2

. Since

(1 +

) and

2
converge, we must have that log
(
(1 +

)
)
+2

2
are both nite.
29
Thus our sum

is bounded by nite terms, and so the sum must also be nite

itself. Hence the sum

must converge to a nite limit.

b) Let {

} = {

1
,
1

1
,

2
,

2
, . . . }. The sum

converges by the alter-

nating series test, since the absolute value of the terms approaches zero (one can
show this by showing that rst the odd terms tend to zero in absolute value and
then that the even terms do as well).
Consider now the product
(
1 +

)
. For an arbitrary integer , look at the
2-th partial product:
2

=1
(
1 +

)
=
(
1 +

1
)(
1

1
)

(
1 +

2
)(
1

2
)
=
(
1

2

1
2
)

(
1

2

(2)
2
)
=
(
1 +
1
1
)

(
1 +
1
2
)
=
(
2
1
)

(
2 + 1
2
)
= 2 + 1.
(5.9)
Thus when we evaluate at an even term 2, we see that
lim
2
2

=1
(1 +

) = lim
2
(2 + 1) = ,
so the product diverges. Hence the product diverges at even terms and thus cannot
converge in general.
c) For a trivial example, let {

} = {1, 1, 1, 1, . . . }. The sum

does
not converge because the limit of the th partial sum as tends to innity does
not converge; it alternates between 0 and 1. However, the product will clearly
converge:

= (1 + 1)(1 1)(1 + 1)(1 1) = (1)(0)(1)(0) = 0.

30
For an example in which the sum diverges but the product converges to a nonzero
limit, consider the sequence {

21
= 1/

,
2
= 1/(1+

)}

=1
. Group-
ing the pairs 2 and 2 1 together, we see that

=1

=1
(
1

1
1 +

)
=

=1
1
+

.
Well show that this series diverges. Notice that for every n,

=1
1
+

=1
1
2
and since the series on the RHS diverges, by comparison test, so does the series
on the LHS. So

diverges. However, grouping again the even and odd pair

terms, for even , we have

=1
(1 +

) =
/2

=1
(1 +
1

)(1
1

+ 1
)
=
/2

=1
(1 +
1

+ 1

1

+
)
=
/2

=1
(1

+ 1 1

+
)
=
/2

=1
1 = 1
and for odd N,

=1
(1 +

) = (1 +
1

)
which converges to 1 as . Thus,

=1
(1 +

) = 1.
31
Hence {

} is the desired sequence.

9. Prove that if < 1, then
(1 + )(1 +
2
)(1 +
4
)(1 +
8
) =

=0
(
1 +
2

)
=
1
1
.
9. Consider the product (1 +)(1 +
2
)(1 +
4
)(1 +
8
) . Suppose we tried
to multiply this product out: to get one term, we would need to choose either the 1
or the power of in each term to multiply by. For example, one term we could get
out is simply , where we would choose the in the rst term and the 1 in every
succeeding term; another way to say this is to write = 1 1 . To write
out the entire product, we would have to make sure we evaluated every possible
choice of ones and powers of .
But this isnt so bad if we think of choosing terms as counting in binary. In
binary counting, a number is written entirely in terms of 0s and 1s. For any given
number, each digit represented a choice between the digit 0 and the digit 1. If
we think of selecting the power of in a term as picking 1 for a given digit in
binary counting, and selecting the 1 in a term as picking 0 for a given digit in
binary, we can identify a bijective correspondence between integers written in
binary and products from our term (with the exception that 000000000 = 1 in
our product). For example, the binary number 101 = 000101 = 2
2
1 +2
1

0 + 2
0
1 = 5, and if choose the terms ()(1)(
4
)(1)(1) , we see that we get
the product
5
.
To evaluate our product we must sum over all such possible choices. Since
all possible binary numbers together yield precisely the nonnegative integers, this
bijective correspondence importantly tells us that the sum over all such products
will be the sum over all nonnegative powers of z, or 1+ +
2
+
3
+. . . . Thus we
have (1+)(1+
2
)(1+
4
)(1+
8
) = 1++
2
+
3
+. . . . Since < 1, we can
use the geometric expansion of to write (1+)(1+
2
)(1+
4
)(1+
8
) =
1
1
,
as desired.
Signicance for combinatorics: notice the way in which our solution invokes
combinatorics (such as seeing how many ways we can choose our terms to make
a product).
Alternatively, we can truncate the product and multiply by 1 . Note that
(1 )(1 + ) = (1
2
), then (1
2
)(1 +
2
) = (1
4
), and so
(1 )(1 + )(1 +
2
)(1 +
4
) (1 +
2

) = 1
2
+1
;
32
as < 1 the latter tends to 1, and thus
(1 + )(1 +
2
)(1 +
4
) (1 +
2

) =
1
1


2
+1
1

1
1
.
Chapter 3
10. Show that if > 0, then

0
log

2
+
2
=

2
log .
[Hint: Use the contour in Figure 10.]
10. We will rst nd the residue at and then integrate over the given contour.
Let () =
log

2
+
2
, where we take the branch cut of the logarithm along for
all [0, ). Furthermore, is a zero of order 1. Finding the residue at , we
have
res

= lim

( )
log

2
+
2
= lim

(
log
+
)
=
log
2
=
log
2
+

2
Label the contours from the portion on the positive real axis
1
, the larger arc

2
, the portion on the negative real axis
3
, and the smaller arc
4
. Choose <
min{, 1}, > max{, 1}. Parametrize
1
with () = from to ,
2
with
() =

from 0 to ,
3
with () = from to , and
4
with () =

33
from to 0. Integrating over the
2
and taking absolute values, we have

2
log

2
+
2

0
log

)
2
+
2

log

2
+
2

log

2
+
2

log +

2
+
2

0
log +

2
+
2

0
log +

2
+

2

0
log +
+

2

log +
+

2

since , log > 0. Since , log + , +

2

, by LHopital,
lim

log +
+

2

= lim

1/
1

2

2
= lim

1


2

= 0.
Thus, as , the contribution along
2
vanishes to 0. Similarly, for
4
, we
34
have

4
log

2
+
2

log

)
2
+
2

log

2
+
2

log

2
+
2

log +

2
+
2

log +

2
+
2

log +

2
+

2

log +
+

2

log +
+

2

since , log > 0. Since 0, log + , +

2

, by LHopital,
lim
0
log +
+

2

= lim
0
1/
1

2

2
= lim
0
1


2

= 0.
Thus, as 0, the contribution along
4
also vanishes to 0. For the integral over

1
,
3
, we have

1
+
3
log

2
+
2
=

log

2
+
2
+

log

2
+
2
.
35
Letting = , we have

1
+
3
log

2
+
2
=

log

2
+
2
+

log
()
2
+
2
(1)
=

log

2
+
2
+

log

2
+
2

log

2
+
2
+

log +

2
+
2

= 2

log

2
+
2
+

2
+
2

= 2

log

2
+
2
+

arctan

Thus we have, as , 0 and as res

=
log
2
+

2
, we have
lim
,0
(
2

log

2
+
2
+

arctan

)
= 2
(
log
2
+

2
)
lim
,0
(
2

log

2
+
2

)
+ lim
,0
(

arctan

)
=
log

+

2

lim
,0
(
2

log

2
+
2

)
+

2

=
log

+

2

lim
,0
(
2

log

2
+
2

)
=
log

0
log

2
+
2
=
log
2
as desired.
Additional Problems
1. Find all poles of the function () = 1/(1
2
)
4
and nd the residues at
the poles.
36
Let () = 1/() = (1
2
)
4
= ((1 + )(1 ))
4
. We see that the zeros of
are 1, each with order 4. Hence, the residues are
res
1
() = lim
1
1
(4 1)!
(

)
41
( 1)
4
1
(1
2
)
4
= lim
1
1
6
(

)
3
1
(1 + )
4
= lim
1
1
6
(4)(5)(6)
1
(1 + )
7
= lim
1
20
(1 + )
7
=
20
2
7
=
5
32
and
res
1
() = lim
1
1
(4 1)!
(

)
41
( + 1)
4
1
(1
2
)
4
= lim
1
1
6
(

)
3
1
( 1)
4
= lim
1
1
6
(4)(5)(6)
1
( 1)
7
= lim
1
20
( 1)
7
=
20
2
7
=
5
32
Thus we have found the desired residues.
37
We sketch an alternative proof. We have
() =
1
( 1)
4
1
( + 1)
4
=
1
( 1)
4
1
( 1 + 2)
4
=
1
( 1)
4
1
2
4
1
(1 +
1
2
)
4
=
1
( 1)
4
1
16
(
1
1
2
+
( 1)
2
4

( 1)
3
8
+
)
4
. (5.10)
The difculty is we have to expand the factor to the fourth power well enough to
identify the coefcient of ( 1)
3
. A little algebra shows it is
5
2
( 1)
3
, and
thus (remembering the factor 1/16) the residue is just 5/32.
2. Consider the map () = ( )/( + ). Show that this is a one-to-one
and onto map from the upper half plane (all = + with > 0) to the unit
disk.
2. First well show that the range of is the unit disk. Writing = +
where , , > 0, then we have
( + ) =

+ ( 1)
+ ( + 1)

2
+ ( 1)
2

2
+ ( + 1)
2
and since > 0,

2
+ ( 1)
2
<

2
+ ( + 1)
2
, (+) < 1, so the range
of is the unit disk.
Now well show that f is injective. Suppose for
1
,
2
with imaginary part
positive, (
1
) = (
2
). Then

1
+
=

2

2
+
(
1
)(
2
+ ) = (
2
)(
1
+ )

2
+
1

2
+ 1 =
1

1
+
2
+ 1
2(
1

2
) = 0

1
=
2
. (5.11)
38
Heres another, faster way to do the algebra. We add zero:

1
+
=

2

2
+

1
+ 2

1
+
=

2
+ 2

2
+
1
2

1
+
= 1
2

2
+ 1
; (5.12)
it is clear that the only solution is when
1
=
2
.
Now well show that is surjective. Given any , setting = ( +
1)/(1 ), we see that
() =
(+1)
(1)

(+1)
(1)
+
=
( + 1) (1 )
( + 1) + (1 )
= .
Now well show that has positive imaginary part. Writing = + with
, ,
2
+
2
< 1, we have
=
( + 1) +
(1 )
=
2 + (1
2

2
)
(1 )
2
+
2
.
So the imaginary part is 1 (
2
+
2
) > 0, so has positive imaginary part.
3. Calculate the Weierstrass product for cos() (this is also problem 10b in
Chapter 5, and the answer is listed there) and for tan().
39
3. By the Euler formulas for sine and cosine, we see that
cos() =

2
=

2
(

)
2
=
(
(+
1
2
)
+
(
1
2
)
)
2
=

(
1
2
)

(
1
2
)
2
= sin((
1
2
))
and since the zeros of sin occur only at the integers, the zeros of cos occur
at +
1
2
for all . Thus, dene the sequence {
21
= +
1
2
,
2
=
( +
1
2
)}

=1
, which are precisely the zeros of cos . Furthermore, since the
zeros of sine are of order 1, the zeros of cosine are also of order one. Thus we have,
for

() =

=1

, grouping together the pairs 2 and 21, the Weierstrauss

product of cos is, up to a factor of
()
for some entire function ,

=0
(1

)
()
=

=0
(1

+
1
2
)(1

( +
1
2
)
)

=1

()
=

=0
(1

2
( +
1
2
)
2
)

=1
()
=

=0
(1
4
2
(2 + 1)
2
)

=1
()
.
Considering

=0
(1
4
2
(2+1)
2
), well show this product converges. Note that

=1
1

2
=

=1
1
(2)
2
+

=0
1
(2 + 1)
2
3
4

=1
1

2
=

=0
1
(2 + 1)
2
so since the sum on the RHS is bounded

=1
1

2
, a convergent series, the series
on the RHS converges as well, and as the convergence is absolute, the product
40
converges. Thus (up to the exponential of an entire function) the Weierstrauss
product of cos is

=0
(1
4
2
(2+1)
2
).
Next, notice that tan(), has poles at odd integer multiples of

2
, and so by
denition does not have a Weierstrass product.
6 Homework #6: Kung, Lin, Waters
HW: Due Friday, November 5: (1) Evaluate

cos(4)/(
4
+ 1). (2) Let
U be conformally equivalent to V and V conformally equivalent to W with
functions f: U > V and g: V > U. Prove g f (g composed with f) is a
bijection. (3) The Riemann mapping theorem asserts that if U and V are
simply connected proper open subsets of the complex plane then they are
conformally equivalent. Show that simply connected is essential. In other
words, nd a bounded open set U that is not simply connected and prove that
it cannot be conformally equivalent to the unit disk. (4) Chapter 8, Page 248:
#4. (5) Chapter 8: Page 248: #5. (6) Chapter 8: Page 251: #14.
1. Evaluate

cos(4)/(
4
+ 1). Evaluate

cos(4)/(
4
+ 1)
First observe that cos(4) =
1
2
(
4
+
4
), since
4
= cos(4) + sin(4)
and
4
= cos(4) sin(4). We can rewrite this integral, then, as

cos(4)

4
+ 1
=
1
2

4
+
4

4
+ 1
=
1
2
(

4
+ 1
+

4
+ 1

)
and we can evaluate both halves separately.
For both halves, observe that the poles are located at =
1
4

,
3
4

,
5
4

,
7
4

,
since those are the solutions to
4
+ 1 = 0. We can now choose a contour over
which to integrate and apply the residue theorem. Our choice of contour is moti-
vated by the decay of the functions. We need to work in the upper half plane for
exp(4) to decay, and in the lower half plane for exp(4) to decay.
41
For

4
+1
, consider the contour
1
that traverses the semicircle of radius
in the upper half-plane and the real axis, with standard orientiation. This con-
tour will enclose only the poles at =
1
4

,
3
4

, so it sufces to nd the residues

at those two points in order to apply the residue theorem.
The simplest way to compute the residues is to note that we have simple poles
and we may write () = ()/() with () having simple zeros and ,
holomorphic. Then the residue of at a pole
0
is just (
0
)/

(
0
). For us,
(
0
) = exp(4
0
), while

(
0
) = 4
3
0
.
At
1
4

2
2
+

2
2
, the residue will be
exp(4 exp(/4))
4 exp(/4)
3
=
exp(2(

2 +

2))
4 exp(3/4)
=
exp(2

2 + 2

2))
2

2 + 2

2
.
We can compute this another way as well:
lim

1
4

(
1
4

)

4

4
+ 1
=

2

2+2

2
(
1
4

3
4

)(
1
4

5
4

)(
1
4

7
4

)
=

2

2+2

2
2

2(1 + )
.
At
3
4

, the residue will be

lim

3
4

(
3
4

)

4

4
+ 1
=

2

22

2
(
3
4

1
4

)(
3
4

5
4

)(
3
4

7
4

)
=

2

22

2
2

2(1 + )
.
Thus, the countour integral over
1
is equal to
2
(

2

2+2

2
2

2(1 + )
+

2

22

2
2

2(1 + )
)
.
Now let the radius R tend to innity, and observe that the portion of gamma
that is not on the real axis (i.e. the semicircle of radius R) will make a zero
contribution to the integral. In the upper half-plane, the integral is at most the
maximum value of the integrand on our contour times the length of the contour.
Since the length of the contour is , then, we have

lim

1
,semicircle

4
+ 1

<

lim

4
1

= 0
42
(note we need
4
1 and note
4
+1 in the denominator, as the upper bound occurs
when the denominator is as small as possible in absolute value; this happens when

4
is negative, which occurs for = exp(/4)).
Only the portion of the contour integral that lies on the real axis makes any
non-zero contribution to the integral, then, so

4
+ 1
= 2
(

2

2+2

2
2

2(1 + )
+

2

22

2
2

2(1 + )
)
.
As our denominator is non-zero and decays rapidly, and exp(4) = cos(4) +
sin(4), we see we may drop the integral from the sine term. The reason is that
this is an odd, rapidly decaying function integrated over a symmetric region, and
thus it gives zero. We therefore nd

cos 4

4
+ 1
= 2
(

2

2+2

2
2

2(1 + )
+

2

22

2
2

2(1 + )
)
.
WE MAY STOP HERE! There is no need to evaluate the other contour, as it
will simply give us another calculation of our desired integral. For completeness,
we include how the calculation would go in the lower half plane, but again, there
is no need to do this!
For

4
+1
, we can repeat the same process, but we must use a different
contour. For this function,

4

4
+1
wont vanish as for in the upper half-
plane, since 4 will have a large positive real component, but it will vanish in
the lower half-plane. Use the contour
2
consisting of the semicircle of radius R
in the lower half-plane and the real axis; it is very important to note that we are
traversing the real axis in the opposite orientation, running fromto . Now,
with restricted to the lower half-plane, our integrand will again vanish, so we
have
lim

2
,semicircle

4
+ 1
= 0,
and we see that

4
+ 1
=

4
+ 1
=

4
+ 1
.
43
Our contour
2
encloses the poles at exp(
5
4
) and exp(
7
4
), so we need to nd
the residues at those two points.
At
5
4

, the residue will be

lim

5
4

(
5
4

)

4

4
+ 1
=

2

2+2

2
(
5
4

1
4

)(
5
4

3
4

)(
1
4

7
4

)
=

2

2+2

2
2

2(1 )
.
At
7
4

, the residue will be

lim

7
4

(
7
4

)

4

4
+ 1
=

2

22

2
(
7
4

1
4

)(
7
4

3
4

)(
7
4

5
4

)
=

2

22

2
2

2(1 )
.
The integral over
2
, then, is equal to
2
(

2+2

2
2

2(1 )
+

2

22

2
2

2(1 )
)
.
As , this equals the integral over the real line; however, remember that
we are proceeding with the opposite orientation, running from to as we
are using a semi-circle in the lower half plane, and thus we traverse the real line
in the opposite orientation as usual. To x this and restore the correct orientation
requires a minus sign, and we nd

4
+ 1
= 2
(

2+2

2
2

2(1 )
+

2

22

2
2

2(1 )
)
.
We then argue as before, namely that exp(4) = cos(4) sin(4), and
the sine integral does not contribute as it leads to an odd integral over a symmetric
region. Arguing along these lines, we nd the same answer as before.
2. Let U be conformally equivalent to V and V conformally equivalent to
W with functions f: U > V and g: V > U. Prove g f (g composed with f) is
a bijection.
44
To prove that g f is a bijection, we need to show that g f is one-to-one and
onto.
One-to-one: Consider an arbitrary
1
,
2
in and assume that (
1
) =
(
2
). We need to show that
1
=
2
. First observe that, since g is one-to-one,
(
1
) = (
2
) implies (
1
) = (
2
). Since f is also one-to-one, we have
that
1
=
2
, and we are done.
Onto: Consider an arbitrary . Since is onto, there is some such
that () = . Since is also onto, there is some such that () = .
Therefore, () = , so is onto.
Then is one-to-one and onto, so it is a bijection.
3. The Riemann mapping theorem asserts that if U and V are simply con-
nected proper open subsets of the complex plane then they are conformally
equivalent. Show that simply connected is essential. In other words, nd a
bounded open set U that is not simply connected and prove that it cannot be
conformally equivalent to the unit disk.
Solution: Consider the punctured unit disc, {0}, a bounded open set that is
not simply connected. Consider function f(z)=1/z on a circle of radius 1/2. Then
f(z) is holomorphic on the set, since the origin is not included.
If a conformal map exists from to the punctured disc, then the function
() will map to a holomorphic function on , and the circle will be mapped to a
closed curve in . (Technically we proved Cauchys theorem, which well use in
a moment, only for simple, non-intersecting curves. One can show that the image
of our closed curve is also a simple, non-intersecting closed curve. If it intersected
itself, that would violate the 1-1 property of our conformal map between the two
regions.)
We rst compute
1
2

=1/2
(). As () = 1/, a brute-force computa-
tion (or use the Residue Theorem) tells us that this is just 1.
What if we look at the inverse image of the circle of radius 1/2 in the unit
disk? Lets call the inverse image , so () = { : = 1/2}. Then, using the
change of variables formulas, if = () (recall is our assumed conformal map
45
from to the punctured disk), then =

() and
1
2

=1/2
() =
1
2

(())

().
As and are holomorphic, so too is (())

(). As we are integrating a

holomorphic function over a closed curve, it is just zero.
Weve thus computed the integral two different ways, getting 1 as well as 0.
As 1 = 0, we have a contradiction and thus the unit disk and the punctured unit
disk are not conformally equivalent.
4. Chapter 8, Page 248: #4. Does there exist a holomorphic surjection from
the unit disc to the complex plane ?
Solution: From 8.1.1 in the book, we know that there exists a conformal map
from the disc to the upper half-plane:
() =
1
1 +
(6.1)
Now we just map this image to the complex plane. We can do so by moving
it down i units and then squaring it. The upper half-plane represents complex
numbers with positive imaginary part (z=x+iy, y>0); however, a better way to
view this is to note that the upper half plane are all numbers of the form exp()
with > 0 and 0 < < . If we were just to square this as is, we would get every
angle we need but = 0 and every radius we need but = 0. The problem is that
the upper half plane is an open set and does not include its boundary, the real axis.
We may rectify this by mapping the image of the unit disk under , namely the
upper half plane, down units. We now include the entire real line as well. While
our resulting map wont be 1-1, it will be onto. Now our region includes all 0
and all [0, ]. Squaring this gives all 0 and all [0, 2], as desired.
Thus our next maps are
() = (6.2)
and
() =
2
(6.3)
The functions , , and are all holomorphic surjections on the complex
plane, so ((())) is a holomorphic surjection that will map .
46
((())) = ((
1
1 +
))
= (
1
1 +
)
= ((
1
1 +
1))
= (
2
1 +
)
=
4
2
(1 + )
2
. (6.4)
5. Chapter 8: Page 248: #5. Prove () =
1
2
( +
1
) us a conformal
map from the half-disk { = + : < 1, > 0} to the upper half plane.
First, we check that () is holomorphic. We have that
1
() =
1
2
(1
1

2
)
and so it is as = 0. We next check that this mapping will give us a value in the
upper half plane. We take = + . Because is in the upper half disk, > 0.
Thus,
() =
1
2
( + +
1
+
)
=
1
2
(
+ +

2
+
2
)
.
Because < 1, we have that
2
+
2
< 1, and thus the imaginary part inside
the parentheses above is negative, and thus becomes positive upon multiplication
by 1/2. Thus () is in .
We now show that () is onto. That is, given in the upper half plane, we
must nd a in the upper half disk such that () =
1
2
( +
1

) = . Thus, we
have to solve
+
1

= 2
2 =
2
+ 1

2
+ 2 + 1 = ( + )
2
(
2
1) = 0
( + )
2
=
2
1. (6.5)
47
Therefore,
=

2
1
and so proving () is onto is equivalent to showing that

2
1 is in the
upper half disk whenever is in the upper half plane. NEED MORE INFO
HERE.
We now show that () is one-to-one. To do this, we take () = (), with
, in the upper half disk. Thus, we have

1
2
(
+
1

)
=
1
2
(
+
1

)
+
1

= +
1

2
+ =
2
+

2

2
+ = ( )( 1) = 0.
We then see that, if ( 1) = 0, then = 1, so = 1/. Because is in the
upper half disk, < 1. This would cause 1/ = > 1. Because we know
that is in the upper half disk as well, this cannot be the case, and so 1 = 0.
This means then that = 0, and so = . Therefore, () is one-to-one, and
so () is a conformal map from the upper half disk to the upper half plane.
6. Chapter 8: Page 251: #14. Prove all conformal maps of the upper half
plane to the unit disk are of the form

( )/( ) for real and in the

upper half plane.
We rst see that, given and two conformal maps from to , we then
have that
1
: and
1
is a conformal map from to . That is,

1
is an automorphism of . From the book, we know that
1
is of
the form

1
for some in the unit disk. In order to then solve for a general
form for , we can use the inverse of any function : . In other words,
() = (
1
)(). We choose () =

+
.
48
We nd
() = (
1
)()
=

()
1 ()
=

+
1

+
=

++
+
++
+
=

+ +
=

(1 ) (1 + )
(1 ) + (1 + )
. (6.6)
We have to be a bit careful in simplifying the above. Note the goal is to get a
rotation times over . We thus need to have just a plus or minus a
constant in the numerator and denominator. We therefore pull out a 1 from
the numerator and a 1 from the denominator. Note these two quantities have
the same norm, and thus their ratio is of size 1. We can thus write their ratio as
exp(

) for some

, and hence exp() exp(

) = exp(

). We nd
() = exp(

)
(1 + )(1 )
1
+ (1 + )(1 )
1
.
If we set
= (1 + )(1 )
1
then clearly we do have
= (1 + )(1 )
1
.
We thus have
() = exp(

)


;
all that remains is to show that is in the upper half plane. NEED TO DO THIS.
49
7 Homework #7: Thompson, Schrock, Tosteson
HW: Due Friday, November 12: DO ANY FIVE OUT OF THE FOLLOW-
ING SIX: IF YOU DO MORE, THATS GOOD BUT ONLY THE FIRST
FIVE WILL BE GRADED. (1) Consider the functions

() = /(1 +
2
)
where n is a positive integer. Prove that each

is uniformly continuous on
the real line. Is the family {

: n a positive integer} equicontinuous on com-

pact sets? (2) Consider a 2x2 matrix M with integer entries and top row (a,b)
and bottom row (c,d) such that ad-bc = 1; we denote the set of all such ma-
trices by SL(2,Z). Consider the map

() = ( + )/( + ) with z in the

upper half plane. Is the family {

: M in SL(2,Z)} uniformly bounded on

compact sets of the upper half plane? Hint: I think each map is bounded
on compact subsets of the upper half plane, but you can nd a sequence of
matrices such that no bound works simultaneously. (3) Let

() = 1
for 0 <= x <= 1/n and 0 otherwise, and let F = {

: n a positive integer}. Prove

that lim

exists and determine it. (4) Consider the family from (3). Prove
it is not normal (the problem is that the convergence is not uniform). Specif-
ically, to be normal not only must it converge, but given any epsilon there is
an N such that, for all n > N,

() () < epsilon (or this must hold for

a subsequence). (5) Evaluate the integral from -oo to oo of
2
/(
4
+
2
+ 1).
(6) Integrate from 0 to 2pi the function 1 / (a + b sin theta) where a and b are
real numbers. What restrictions must we place on a and b in order for this to
make sense?
(1) Consider the functions

() = /(1 +
2
) where n is a positive inte-
ger. Prove that each

is uniformly continuous on the real line. Is the family

{

: n a positive integer} equicontinuous on compact sets?

We must showthat, given any > 0 that there exists a such that, for any ,
and any

in our family that whenever < then

()

() < .
Suppose < . Then, by the Mean Value Theorem,

()

() =

() <

()
So, all we need to show is that

is bounded. Why? If

() for all , then

the above gives

()

() < < .
If we take < /( + 1) then we see that, whenever < then

()

() < , as desired.
50
We now show

is bounded. We easily nd that

() =
2
2

(1 +
2
)
2

() =
2
2
(1 +
2
)
2

8
2

2
(1 +
2
)
3
Now, setting the second derivative to zero to get

2
=
1
3
so there are at most two local extrema. Notice that
lim

() = lim

2
2

(1 +
2
)
2
= lim

2
2

4
( +
1

2
)
2
= 0,
which implies that the maximum of

cannot occur as . Thus the

maximum value of

occurs at both of = 1/

3, and this is the desired

bound.
Alternatively, we could argue as follows. We have

() = 2
2

(1 +
2
)
2
.
Once 1/ the denominator exceeds the numerator; as /(1 +
2
)
2
is
continuous on [1/, 1/], it is bounded on this interval. Thus

is bounded.
(2) Consider a 2x2 matrix M with integer entries and top row (a,b) and
bottom row (c,d) such that ad-bc = 1; we denote the set of all such matrices
by SL(2,Z). Consider the map

() = ( + )/( + ) with z in the upper

half plane. Is the family {

: M in SL(2,Z)} uniformly bounded on compact

sets of the upper half plane? Hint: I think each map is bounded on compact
subsets of the upper half plane, but you can nd a sequence of matrices such
that no bound works simultaneously.
If we let be an arbitrary compact subset of the upper half plane, we know
each has its imaginary part bounded above and below, and similarly for
51
the real parts. To show that our family is not uniformly bounded, we must nd
a sequence of matrices and points such that the maps applied to these bounds
become arbitrarily large in absolute value.
Were studying maps of the form

() =
+
+
For problems like this, it is often useful to try and analyze special cases, where the
algebra is simpler. Wouldnt it be nice if the denominator were just one? Well, to
get that and satisfy the conditions, we would have to study matrices of the form
(
1
0 1
)
,
which are in our family. These lead to

() = +. Clearly, as increases, this

is not bounded (as the real and imaginary parts of are bounded, so by sending
we see it is unbounded.
(3) Let

() = 1 for 0 <= x <= 1/n and 0 otherwise, and let F = {

:
n a positive integer}. Prove that lim

exists and determine it.

Let
0
= 0 be a point on the positive real line. Then for all > , where
> 1/
0
, we have

(
0
) = 0. This is because

() =
{
1 0
1

0 otherwise
and
>
1

0
>
1

(
0
) = 0.
So as ,

where

() =
{
1 x=0
0 otherwise
Of course, we havent said anything about lim

(0); however, as each

(0) = 1,
it is clear that the limit is 1 as well. Finally, what happens for negative? Well,
as

() = 0 for < 0 by denition, then lim

() = 0 for negative.
52
(4) Consider the family from (3). Prove it is not normal (the problem
is that the convergence is not uniform). Specically, to be normal not only
must it converge, but given any epsilon there is an N such that, for all n > N,

() () < epsilon (or this must hold for a subsequence).

Take = , = 0. Then obviously . But for such that
1

< :

()

() = 1.
So, not normal.
(5) Evaluate the integral from -oo to oo of
2
/(
4
+
2
+ 1).
Using the quadratic formula we nd that the equation
2
+ +1 = 0 has roots
at
2/3
and
4/3
. Therefore the function () =
4
+
2
+ 1 has roots at
/3
,

2/3
,
4/3
and
5/3
. Thus we can rewrite our integral as

2
(
/3
)(
2/3
)(
4/3
)(
5/3
)
.
For our contour we will take a semicircle in the upper halfplane of radius R cen-
tered at the origin. In this region we have poles at =
/3
and =
2/3
. To
nd what these residues are at the poles, we recall that if we can write a function
() as a ratio of two entire functions () and (), with () having a simple
zero at the point
0
, then the residue of at
0
is simply (
0
)/

(
0
). Using this
we see the residue of () at
/3
is:

2/3
(
/3

2/3
)(
/3

4/3
)(
/3

5/3
)
=
1
12
(3

3).
Similarly, the residue of () at
2/3
is:

4/3

2/3

/3
)(
2/3

4/3
)(
2/3

5/3
)
=
1
12
(3

3).
The sum of the residuals is therefore

3/6 = /(2

3). We now show that

the integral over the circular portion of the contour, call it
2
, contributes nothing
in the limit as . Since the length of
2
is , we have:

4
+
2
+ 1

2
1
0.
53
Therefore in the limit we have:
1
2

4
+
2
+ 1
= /(2

3),
which gives

4
+
2
+ 1
=

3
.
(6) Integrate from 0 to 2pi the function 1 / (a + b sin theta) where a and b are
real numbers. What restrictions must we place on a and b in order for this to
make sense?

2
0

+ sin
=

= 1/, = , = /

2
0

+ sin
=

( + ( 1/)/2)
=

2
2 + (
2
1)
where is (the circle bounding the unit disk).
The following lines are the original write-up of the solution; these are
based on the previous line having a factor of 2 instead of 2
2
.
This has poles at

0
=

2
)
where the only one inside the unit circle is the plus root. This gives residue:

2
So

2
0

+ sin
=
2

2
as long as
2
>
2
.
Unfortunately, the above cannot be correct, as a simple test shows. If we
double and , then the original integral decreases by a factor of 2, while our
answer here does not change. Thus there must be an algebra error. Below is
the corrected argument.
54
Consider the integral

2
0

+ sin
.
Making the change of variables =

= 1/, = , = /,
we nd

2
0

+ sin
=

( + ( 1/)/2)
=

2
(2(/) +
2
1)
where is (ie, is the unit circle centered at the origin). From the quadratic
formula, we see that the integrand has poles at

0
=
(

)
2
1
)
where the only one inside the unit circle is the plus root. To compute the residue,
we use the following fact: if () = ()/() and () is a holomorphic
function with a simple zero at
0
and () is holomorphic, then the residue at
0
is just (
0
)/

(
0
). This gives a residue of
2

)
2
1
So

2
0

+ sin
=
2

2
as long as
2
>
2
(remember that the residue formula requires the integral to be
multiplied by 1/2, thus in our case we must multiply the residue by 2 as our
integral was unadorned).
Alternatively, if we factor out a from the denominator we have
1

2
0

(/) + sin
.
This is solved exactly like the problem on the midterm, except instead of having
+ sin we now have (/) +sin , with an extra factor of 1/ outside. Thus the
answer is just
1

(/)
2
1
=
2

2
.
55
Notice this solution has all the desired properties. It doesnt make sense for <
. For xed and it converges to 2/, et cetera. It is always good to do
these quick consistency checks.
8 Homework #8: Xiong, Webster, Wilcox
HW: Due Tuesday, November 23rd: DO ANY FIVE OUT OF THE FOL-
LOWINGSIX: IFYOUDOMORE, THATS GOODBUTONLYTHEFIRST
FIVE WILL BE GRADED. (1) Let Omega be the subset of the complex plane
of all z = x+iy with |x| < |y|. Does there exist a logarithm on Omega? If yes,
what does the image of Omega under the logarithm look like? (2) Let Omega
be the region from (4); is Omega conformally equivalent to the unit disk?
Prove your assertion. Hint: remember the full denition of what it means to
be simply connected. (3) Let Omega be the subset of the complex plane of all
z = x + iy with |x| < |y| + 1. Conformally map Omega onto an open subset of
the disk you must give an explicit form for the map. (Note: The Riemann
Mapping Theorem asserts that you can get a map that is onto the disk; here
you are just being asked to get a map that is holomorphic and 1-1). (4) Evalu-
ate

2
0
cos()

, where m is a positive integer. (5) Evaluate

1
0
(1
2
)

,
where n is a positive integer. Hint: Let x = Sin[theta], dx = Cos[theta] d theta.
(6)

0
log()/(1 +
2
).
Problem: Let = { = + : < }. Does there exist a logarithm
on . If yes, what does the image of under the logarithm look like?
Solution: First note that contains the imaginary axis save the origin. Now
imagine the lines = in the plane. looks like the open hourglass strictly
between the lines and containing the imaginary axis. Noting that (, 0],
we can take the principal branch of the logarithm and use the restriction of this
logarithm to (see Chapter 3, Theorem 6.1 in the book). To see the image of
under this logarithm, write =

with > 0 and < , and then we have

log = log + where log is the standard logarithm on the positive reals. Note
that for example would be written 1
/2
. Evaluating at some test points, we
see
log() = log(
1+/2
) = 1 + /2
log(
1+/4
) = 1 + /4
log() = 1 /2,
56
etc. Evaluating points along the boundary lines shows us that the image of
under this logarithm is two horizontal strips, dened by
log[] =
{
= +

(
3
4
,

4
)

4
,
3
4
)}
.

Problem: Let be the region from Problem 1. Is conformally equivalent to the

unit disk ? Prove your assertion.
Solution: is not conformally equivalent to since is not simply connected
due to the hole at the origin. Citing Hw 6 #3, being simply connected is necessary
for to be uniformly equivalent to .
A simple illustration: we may choose one point (the upper-half
complex plane) and another point in the lower half. Suppose that is con-
formally equivalent to , then a conformal map : . If we draw a
continuous curve connecting () and () in , we may use the reverse con-
formal map
1
: to map into and still obtain a continuous curve.
However, since excludes the origin, it is impossible to draw a continuous curve
in connecting and .
Problem: Let be the set from Problem 1. Conformally map onto an open
subset of the (unit) disk you must give an explicit form for the map. (Note: The
Riemann Mapping Theorem asserts that you can get a map that is onto the disk;
here you are just being asked to get a map that is holomorphic and 1-1).
Solution: Dene a map
1
:
1
() by

1
() = + 2010.
Clearly,
1
is conformal, and
1
() lies outside the unit disk. Dene a map
2
:
be dened

2
() =
1

,
which conformally maps regions outside the disk onto regions inside of the disk.
Thus,
2
(
1
()) lies inside the disk. Let
() =
2
(
1
()) =
1
+ 2010
.
57
As desired, maps conformally onto
2
(
1
()) , which is open because
is open and f is continuous.
Problem: Evaluate

2
0
(cos())

,
where
+
.
Solution: Begin by setting = and applying Eulers formula for cosine.

2
0
(cos())

2
0
(

2
)

Let =

. Notice that as goes from 0 to 2, travels around the unit circle.

We then have
=

=
1
,
and substituting yields

2
0
(

2
)

=1
(
+
1
2
)

=

2

=1
( +
1
)

=

2

=1

=0
(

21

= 2

=

2
1

,
where

is the coefcient of
1
in

=0
(

21
.
58
If m is odd, then

= 0 and our integral is 0. If m is even, then

=
(

/2
)
,
and our integral is

2
1
(

/2
)
.

Problem: Evaluate

1
0
(1
2
)

,
where
+
.
Solution: Note that we could expand using the binomial formula and power
through some algebra. Instead, we translate to the unit circle and use complex
analysis and recurrence relations to simplify the computation. So let = sin
where [0,

2
], and then = cos and note 1
2
= cos
2
. Substituting
yields

1
0
(1
2
)

=

2
0
(cos )
2+1
.
Now write

=

2
0
(cos )
2+1

for , and note that

0
=

2
0
cos
= sin

2
0
= 1.
59
Then for 1, we have
(cos )
2+3
= (cos )
2+1
(1 (sin )
2
),
thus

+1
=

2
0
(cos )
2+3

=

2
0
(cos )
2+1


2
0
(cos )
2+1
(sin )
2
,
and observe

2
0
(cos )
2+1
=

.
Now let = sin and = (cos )
2+1
sin , then = cos and
=
(cos )
2+2
2 + 2
.
Then using integration by parts, we have

2
0
(cos )
2+1
(sin )
2
=

=
[
sin
(cos )
2+2
2 + 2
]
2
=0
+
1
2 + 2

2
0
(cos )
2+3
.
Evaluating the left term we nd

[
sin
(cos )
2+2
2 + 2
]
2
=0
= 0,
and we are left with
1
2 + 2

2
0
(cos )
2+3
=

+1
2 + 2
.
60
In the end we have,

+1
=

+1
2 + 2
,
so
+1
=
2 + 2
2 + 3

and nally,

=
2
2 + 1

2 2
2 1

2
3

0
=
(2)!!
(2 + 1)!!
.
While we could try to solve this problem by replacing the cosines with and
1/, that unfortunately leads to a bit of algebra as the resulting expression wont
just be integrals of

for an integer, but rather for 2 (or 4, depending on how

clever we are with our choice of variables) an integer.

Problem: Evaluate

0
log
1 +
2
.
Solution: Rather than resort to the branch cut method, whose details have al-
ready been exposed
3
. We apply the Cauchy residue formula more directly, using
the indented semicircle contour. Note that this requires that we take advantage
of the symmetry of our integrand. In particular, it is easily integrable along the
negative real axis. So, let
() =
log
1 +
2
,
(where we take the branch cut for the log to be along the negative imaginary axis)
and let =
1
+
2
+
3
+
4
, where
1
is the line segment along the positive real
axis from to ,
2
is the upper half semicircle of radius with counterclockwise
orientation,
3
the segment from to and
4
the upper half semicircle of
3
A very similar integral is done via the branch cut method on pages 10 and 11 of the wikipedia
printout handed out in class if youre interested.
61
radius and with clockwise orientation. We will show that the integrals along
2
and
4
go to zero in the limit 0 and . Then we may evaluate the
integral by collecting terms and computing residues. We jump to the punchline
rst.
Note that has a simple pole at = , and (applying Theorem 1.4 from
Chapter 3) the residue there is
lim

( )() = lim

log()
+
=
log()
2
=
log(1) +

2
2
=

4
.
And when applying the residue theorem, we will need
2

poles
0
res

(
0
) = 2

4
=

2
2
.
Now consider the integral over
3
, and note that for , with < 0, we have
log = log + where the log on the right hand side is taken to be the standard
logarithm for the positive reals. Then

3
() =

log
1 +
2

=

0
log +
1 +
2

=

0
log
1 +
2
+

0
1
1 +
2

=

0
log
1 +
2
+

2
2
.
Then if we can show that the integrals over
2
+
4
vanish in the limit, we will
have by Cauchys residue theorem that

2
2
=

1
+
3
() = 2

0
log
1 +
2
+

2
2
,
62
at which point we may conclude

0
log
1 +
2
= 0.
Now on to show that the required integrals vanish. We begin with
2
. We have

2
() =

2
log
1 +
2
,
and then letting =

, =

and [0, ]

2
() =

0
log(

)
1 +
2

.
Then we note that for our branch of the logarithm, log(

) = log + where
the log on the right is the standard, and next take absolute values to see

2
()

max
0

log +
1 +
2

log +
1
2
.
Now citing the fact that lim
0
log = 0, we see that this integral vanishes in
the limit. Moving more quickly through
4
, we see

4
() =

0
log(

)
1 +
2

,
and so in absolute value we have

4
()

log +

2
1
.
Thus the integral grows like (log )/ in the limit as , and by lHopitals
rule, this goes to 0. So the integrals over
2
and
4
vanish as claimed, and we now
have that the original integral is also 0.

63