13. Residue Theorem & Applications.

Theorem 249 (Cauchys Residue Theorem) Let f be holomorphic inside and on a simple
closed curve except at points a
1
, . . . , a
n
inside . Then
_

f(z) dz = 2i
n

k=1
res(f(z); a
k
).
Proof. On a disc D(a
k
,
k
) about a
k
we can expand f(z) as a Laurent expansion
f(z) =

n=0
c
(k)
n
(z a
k
)
n
. .
g
k
(z)
+
1

n=
c
(k)
n
(z a
k
)
n
. .
h
k
(z)
= g
k
(z) + h
k
(z),
so that h
k
(z) is the principal part of f about a
i
. Note also that the sum dening h
k
(z) converges
except at a
k
. Hence
F(z) = f(z)
n

k=1
h
k
(z)
is holomorphic in except for removable singularities at a
1
, . . . , a
n
. As we have seen each
F(a
k
) may be assigned a value so as to extend F to a holomorphic function in . By Cauchys
Theorem and the Deformation Theorem we have
_

f(z) dz =
n

k=1
_

h
k
(z) dz =
n

k=1
_
(a
k
,r
k
)
h
k
(z) dz
where 0 < r
k
<
k
. As we saw in the proof of Laurents Theorem, the sum dening h
k
(z)
converges uniformly on (a
k
, r
k
) and hence
_
(a
k
,r
k
)
h
k
(z) dz =
_
(a
k
,r
k
)
1

n=
c
(k)
n
(z a
k
)
n
dz =
1

n=
c
(k)
n
_
(a
k
,r
k
)
(z a
k
)
n
dz = 2ic
(k)
1
and the result follows.
Remark 250 Standard Contours. As we shall see, the Residue Theorem is a powerful
means of calculating certain real integrals. There are three standard contours that we will use
and we shall also introduce certain embellishments here and there when slightly more nesse is
needed. If ever a contour is needed other than the standard ones below then you will typically
be advised what contour to use.
Integrals between 0 and 2. We have already seen integrals of this type: as moves
from 0 to 2 then z = e
i
moves positively around the unit circle (0, 1).
RESIDUE THEOREM & APPLICATIONS. 95
Integrals from 0 to or to . We will standardly use a semicircular contour
in the upper half-plane made up of
+
(0, R) and the line segment [R, R], as pictured
below,
R R
Re
Im
with the aim of showing that as R the contribution from
+
(0, R) tends to 0. The
contribution from [R, R] then tends to the integral on (, ) which we are seeking.
If the integrand is even then the integral on (0, ) will be half that on (, ) .
Innite sums. We will use a square contour
N
with vertices at (N + 1/2) (1 i)
where N is a positive integer and aim to show that the integral around the contour tends
to 0 as N .
Full details follow in examples below.
13.1 Integrals on R
Example 251 Evaluate
_

dx
1 + x
2
+ x
4
.
Solution. We will consider f(z) = (1 + z
2
+ z
4
)
1
. As
_
1 + z
2
+ z
4
_ _
z
2
1
_
= z
6
1
then f(z) has simple poles at e
i/3
, e
2i/3
, e
4i/3
, e
5i/3
. If we use the contour which consists
of the positively-oriented contour consisting of the interval [R, R] and the upper semicircle

+
(0, R) then only the simple poles = e
i/3
and
2
= e
2i/3
lie inside and we have
res(f(z); ) =
1
4z
3
+ 2z

z=
=
1
4
3
+ 2
=
1
4 + 2
; res(f(z);
2
) =
1
4
6
+ 2
2
=
1
4 + 2
2
.
By Cauchys Residue Theorem, and noting
2
+ = i

3,
2
= 1 we see
_

1
1 + z
2
+ z
4
dz = 2i
_
1
2 4
+
1
2
2
+ 4
_
=
2i (2
2
+ 2)
4(1 2
2
+ 2 4
=

3
. (13.1)
Note that if |z| = R then

1 + z
2
+ z
4

z
4

z
2
+ 1

z
4

z
2

|1| = R
4
R
2
1,
96 RESIDUE THEOREM & APPLICATIONS.
so that, by the Estimation Theorem, we have

+
(0,R)
dz
1 + z
2
+ z
4

R
R
4
R
2
1
= O
_
1
R
3
_
0 as R .
Finally, letting R in (13.1) we have
_

dx
1 + x
2
+ x
4
=

3
.
Example 252 Let a 0. Evaluate
_

0
cos ax
1 + x
2
dx.
What is the value of the integral when a < 0?
Remark 253 At rst glance this looks no more complicated than the previous example. How-
ever note that

e
iz

= e
Imz
and cos z =
e
iz
+ e
iz
2
.
So on
+
(0, R) we see that e
iz
and hence cos z are very large we would, in due course,
have found that the integral around
+
(0, R) did not converge. So rather than integrating
cos az/ (1 + z
2
) we will use e
iaz
/ (1 + z
2
), noting that e
iaz
is small on
+
(0, R). When done we
can note cos ax = Re e
iax
for real x.
Solution. We will consider the function
f(z) =
e
iaz
1 + z
2
,
which has simple poles at i and i, and use the contour which denotes the positively-oriented
contour consisting of the interval [R, R] and the upper semicircle
+
(0, R) . Only the simple
pole i lies inside and we have
res(f(z); i) =
e
iaz
z + i

z=i
=
e
a
2i
.
By Cauchys Residue Theorem
_

e
iaz
1 + z
2
dz = 2i
_
e
a
2i
_
= e
a
.
Note that
if z = Re
i
then

e
iaz

= e
a Re z
= e
aRsin
1;
if |z| = R then

z
2
+ 1

z
2
(1)

z
2

|1| = R
2
1,
INTEGRALS ON R 97
so that, by the Estimation Theorem, we have

+
(0,R)
e
iaz
1 + z
2
dz

R
R
2
1
= O
_
1
R
_
0 as R .
(Note that this part of the argument relies on a being positive so that e
aRsin
1.) Now,
letting R in the identity
_
R
R
e
iaz
1 + z
2
dz +
_

+
(0,R)
e
iaz
1 + z
2
dz = e
a
we arrive at
_

e
iax
1 + x
2
dx = e
a
.
Comparing real parts we have
_

cos ax
1 + x
2
dx = e
a
.
Using the evenness of the integrand we arrive at
_

0
cos ax
1 + x
2
dx =

2e
a
.
We clearly couldnt have argued quite this way if a had been negative. One alternative would
have been to use the lower semicircle

(0, R) and repeat a similar calculation, but we can be

smarter than that and note, because of the evenness of cos ax as a function of a that
_

0
cos(a)x
1 + x
2
dx =
_

0
cos ax
1 + x
2
dx =

2e
a
and so for general a R we have
_

0
cos ax
1 + x
2
dx =

2
e
|a|
.
13.2 Innite Series
Example 254 Determine

n=1
1
1 + n
2
and

n=1
(1)
n
1 + n
2
.
Before we start this example we will need to introduce some relevant theory. The contour
we shall use is the positively-oriented square contour
N
with vertices
_
N +
1
2
_
(1 i) and
we shall use the integrands
f(z) =

(1 + z
2
) tan z
, g(z) =

(1 + z
2
) sin z
.
Quite why becomes apparent with the next two lemmas.
98 RESIDUE THEOREM & APPLICATIONS.
Lemma 255 Suppose that the function (z) is holomorphic at z = n Z with (n) = 0. Then
(a) (z) cot z has a simple pole at n with residue (n);
(b) (z) csc z has a simple pole at n with residue (1)
n
(n).
Proof. Note that tan z and sin z have simple zeros at z = n and hence (z) cot z and
(z) csc z have simple poles there. So
res
_
(z)
tanz
; n
_
=
(n)
sec
2
n
= (n); res
_
(z)
sin z
; n
_
=
(n)
cos n
= (1)
n
(n).
Lemma 256 There exists a C > 0 such that for all N and for all z
N

tan z

C and

sin z

C.
Proof. For the squares top and bottom we note

tan z

cos (x i (N + 1/2))
sin (x i (N + 1/2))

exp(ix (N + 1/2)) + exp (ix (N + 1/2))

exp (ix (N + 1/2)) exp (ix (N + 1/2))

exp ( (N + 1/2)) + exp ( (N + 1/2))
|exp( (N + 1/2)) exp ( (N + 1/2))|
= coth((N + 1/2) )
coth (3/2) .
On the sides we have (using the periodicity of tan)

tan z

tan(/2 + iy)

cot iy

= |tanh y| .
Similarly

sin z

=

|sin (x i (N + 1/2))|
=
2
|exp (ix (N + 1/2)) exp (ix (N + 1/2))|

2
|exp ( (N + 1/2)) exp ( (N + 1/2))|
=
2
sinh((N + 1/2) )

2
sinh(3/2)
and

sin z

sin(/2 + iy)

cos iy

= |sechy| .
INFINITE SERIES 99
Set C to be the maximum of these three numbers.
Solution. (To Example 254) Consider the positively-oriented square contour
N
with vertices
_
N +
1
2
_
(1 i) and the functions
f(z) =

(1 + z
2
) tan z
, g(z) =

(1 + z
2
) sin z
.
We know from Lemma 255 that f and g have simple poles at n Z and that
res (f(z); n) =
1
1 + n
2
, res (g(z); n) =
(1)
n
1 + n
2
.
Further there are simple poles at i with
res (f(z); i) = res (f(z); i) =

2i tan i
=

2 tanh
;
res (g(z); i) = res (g(z); i) =

2i sin i
=

2 sinh
.
Hence, by Cauchys Residue Theorem, we have
1
2i
_

N
f(z) dz =
N

n=N
1
1 + n
2


tanh
= 2
N

n=1
1
1 + n
2
+ 1

tanh
;
1
2i
_

N
g(z) dz =
N

n=N
(1)
n
1 + n
2


sinh
= 2
N

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

sinh
.
Now by the Estimation Theorem and Lemma 256 there exists C > 0 such that

N
f(z) dz

4 (2N + 1) C sup
z
N

1
1 + z
2

4 (2N + 1) C
_
N +
1
2
_
2
1
= O
_
1
N
_
0 as N .

N
g(z) dz

4 (2N + 1) C sup
z
N

1
1 + z
2

4 (2N + 1) C
_
N +
1
2
_
2
1
= O
_
1
N
_
0 as N .
Letting N we then have
2

n=1
1
1 + n
2
+ 1

tanh
= 0 =

n=1
1
1 + n
2
=
coth 1
2
.
2

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

sinh
= 0 =

n=1
(1)
n
1 + n
2
=
csch 1
2
.
Remark 257 Should we wish to sum

n=1
1
n
4
,

n=2
(1)
n
n
2
1
,
we would need to note that our chosen would have poles that mingle with the poles of cot z
and csc z. This is not in itself problematic save that it means such poles need to be treated
separately.
100 RESIDUE THEOREM & APPLICATIONS.
13.3 Renements
Remark 258 For integrals such as
_

xsin x
1 + x
2
+ x
4
dx
_

cos x
1 + x
6
dx
we would have no great trouble arguing along the lines of Example 252. These integrals converge
relatively quickly and we would see that the contribution from
+
(0, R) was O(R
2
) and O(R
5
)
respectively. For integrals such as
_

sin x
x
dx,
_

0
xsinx
1 + x
2
dx
without subtler reasoning we would estimate the contribution from
+
(0, R) as O(1). The
problem is that these integrals do not converge fast enough. (To be technical these integrals
in fact only exist as improper integrals.) We need the following inequality due to Jordan to
strengthen our estimate.
Lemma 259 (Jordans Lemma)
2

<
sin

< 1 for 0 < < /2.

Proof. Let f() = sin / on (0, /2) . Note that
f

() =
cos sin

2
.
Now, for 0 < < /2 let
g() = cos sin ,
so that
g

() = sin + cos cos = sin < 0,

so that g is decreasing on (0, /2). As g(0) = 0 then g() < 0 for 0 < < /2. So
f

() =
g()

2
< 0 for 0 < < /2
so that f is decreasing on (0, /2) . As f() 1 as 0 then
2

< f() < 1 for 0 < < /2

Example 260 Determine
_

0
xsin x
1 + x
2
dx.
REFINEMENTS 101
Solution. We will consider the function
f(z) =
ze
iz
1 + z
2
,
which has simple poles at i and i, and use the contour which denotes the positively-oriented
contour consisting of the interval [R, R] and the upper semicircle
+
(0, R) . Only the simple
pole i lies inside and we have
res(f(z); i) =
ze
iz
z + i

z=i
=
ie
1
2i
=
1
2e
.
By Cauchys Residue Theorem
_

ze
iz
1 + z
2
dz =
2i
2e
=
i
e
. (13.2)
Note that
if z = Re
i
then

e
iz

= e
Re z
= e
Rsin
;
if |z| = R then

z
2
+ 1

z
2
(1)

z
2

|1| = R
2
1,
so that, by the Estimation Theorem, we have

+
(0,R)
ze
iz
1 + z
2
dz

_

0
Re
i
e
iRe
i
1 + R
2
e
2i
iRe
i
d

_

0
Re
Rsin
R
2
1
Rd
=
2R
2
R
2
1
_
/2
0
e
Rsin
d

2R
2
R
2
1
_
/2
0
e
2R/
d [by Jordans Lemma]
=
2R
2
R
2
1

_

2R
_

_
e
2R/

/2
0
=
R
2
_
1 e
R
_
R(R
2
1)
= O
_
1
R
_
0 as R .
Letting R in (13.2) we obtain
_

xe
ix
1 + x
2
dx =
i
e
.
Taking imaginary parts we nd
_

x sin x
1 + x
2
dx =

e
.
Finally using the even-ness of the integrand we obtain
_

0
xsin x
1 + x
2
dx =

2e
.
102 RESIDUE THEOREM & APPLICATIONS.
Remark 261 With some integrals the natural integrand to take may leave us with a singularity
on the contour. This isnt a problem when that singularity is a simple pole as we may indent
at such singularities (see lemma below). It is impossible to sensibly indent at other
singularities and have any hope of convergence; it is only possible to indent at simple poles
because the contour has length O() and the integrand grows as O(
1
).
Lemma 262 Let f be a holomorphic function on D

(a, ) with a simple pole at a. Let 0 < r < ,

and let
r
be the positively oriented arc

r
() = a + re
i
< < .
Then
lim
r0
_

r
f(z) dz = ( ) i res(f(z); a).
Proof. Let R = res(f(z); a). There is a holomorphic function h on D

(a, ) such that

f(z) =
R
z a
+ h(z).
Now
_

r
dz
z a
=
_

ire
i
d
re
i
= i
_

d = i ( ) .
Further, as

D(a, /2) is compact and h is continuous, then h is bounded on

D(a, /2), and so
on every
r
with r < /2, say by M. So, by the Estimation Theorem

r
h(z) dz

L(
r
) M = M ( ) r 0 as r 0.
Finally then
lim
r0
_

r
f(z) dz = lim
r0
_

r
Rdz
z a
+ lim
r0
_

r
h(z) dz = iR( ) + 0 = iR( )
as required.
Example 263 Determine
_

0
sin
2
x
x
2
dx.
Solution. Note that
sin
2
x
x
2
=
1 cos 2x
2x
2
and so for an integrand we will use (1 e
2iz
) / (2z
2
). This has a simple pole at 0 as
1 e
2iz
2z
2
=
1
_
1 + 2iz +
(2iz)
2
2!
+
_
2z
2
=
i
z
+ 2 +
REFINEMENTS 103
So we shall indent our usual contour so that we now have a contour made up of line segments
[R, ] and [, R], a positively oriented semicircle
+
(0, R) and a negatively oriented indent

+
(0, ). As the integrand is holomorphic inside the contour then, by Cauchys Theorem, we
have
_

1 e
2iz
2z
2
dz = 0.
On the large semicircle

1 e
2iz
2z
2

=
1 e
2Rsin
2R
2

1
2R
2
so that, by the estimation theorem,

+
(0,R)
1 e
2iz
2z
2
dz

R
2R
2
= O
_
1
R
_
0 as R .
We also have from the previous lemma that

+
(0,)
1 e
2iz
2z
2
dz = i res
_
1 e
2iz
2z
2
; 0
_
= (i) (i) = .
We have the identity
_

+
(0,R)
1 e
2iz
2z
2
dz +
_

R
1 e
2ix
2x
2
dx
_

+
(0,)
1 e
2iz
2z
2
dz +
_
R

1 e
2ix
2x
2
dx = 0
in R and . If we let R , 0 and take real parts then we obtain
+ 2 lim
R
0
_
R

1 cos 2x
2x
2
dz = 0.
Hence
_

0
sin
2
x
x
2
dx =

2
.
13.4 Further Contours
Example 264 Let 1 < a < 2. Use a keyhole contour to determine
_

0
x
a
1 + x
3
dx.
Solution. We will consider the function
f(z) =
z
a
1 + z
3
.
104 RESIDUE THEOREM & APPLICATIONS.
We need to dene a branch of z
a
on the cut-plane C\[0, ). Any z in the cut-plane can be
written z = re
i
where 0 < < 2; we then set
z
a
= r
a
e
ia
.
Our keyhole contour comprises two line segments just above and just below [0, R]. Alterna-
tively, we can consider these two line segments both as [0, R] but with one on the lower branch
with = 0 and one of the upper branch with = 2. (Cauchys Residue Theory is valid on
Riemann surfaces so this viewpoint is perfectly valid.) We also have as part of our keyhole
contour a small cut circle around the origin but as the integrand is O(
a
) and the contour has
length O() then the contribution from this is zero in the limit.
Re
Im
Note that f is holomorphic inside except at 1, e
i/3
and e
i/3
. Note also that we should
write e
i/3
= e
5i/3
for the purpose of the branchs denition. So
res(f; 1) =
z
a
3z
2

z=1
=
1
3
e
ia
;
res(f; e
i/3
) =
z
a
3z
2

z=e
i/3
=
1
3
e
i(a2)/3
;
res(f; e
i/3
) =
z
a
3z
2

z=e
5i/3
=
1
3
e
5i(a2)/3
.
By the Residue Theorem we have
_

z
a
dz
1 + z
3
=
2i
3
_
e
ia
+ e
i(a2)/3
+ e
5i(a2)/3
_
.
The contribution from the outside circle (0, R) satises

_
(0,R)
z
a
dz
1 + z
3

2R
1+a
R
3
= O(R
a2
) 0 as R .
FURTHER CONTOURS 105
Hence we have, letting R , that
_

0
x
a
dx
1 + x
3
e
2ia
_

0
x
a
dx
1 + x
3
=
2i
3
_
e
ia
+ e
i(a2)/3
+ e
5i(a2)/3
_
.
Rearranging we have
_

0
x
a
dx
1 + x
3
=
2i
3

_
e
ia
+ e
i(a2)/3
+ e
5i(a2)/3
1 e
2ia
_
=
2i
3

_
1 + e
i(2a2)/3
+ e
i(2a+2)/3
e
ia
e
ia
_
=
2i
3

_
e
i(2(a+1))/3
+ 1 + e
i(2(a+1))/3
e
i(a+1)
e
i(a+1)
_
=
2i
3

_
e
i(2(a+1))/3
+ 1 + e
i(2(a+1))/3
e
i(a+1)
e
i(a+1)
_
=
2i
3

_
1
e
i(a+1)/3
e
i(a+1)/3
_
=

3
csc
_
(a + 1)
3
_
.
The cancellation to the penultimate line uses the identity (x
2
+ xy + y
2
) / (x
3
y
3
) = 1/ (x y) .
Remark 265 This integral could also have been done by using the contour in the next example
which goes around the sector.
Example 266 Evaluate
_

0
dx
1 + x
3
+ x
6
.
Solution. The denominator has roots at
e
2i/9
, e
4i/9
, e
8i/9
, e
10i/9
, e
14i/9
, e
16i/9
as 1 + x
3
+ x
6
= (x
9
1) / (x
3
1) . Let R > 1 and let = e
2i/9
. We will consider the
positively oriented contour made up of the line segment from 0 to R, the circular arc centred
at 0 connecting R to R
3
and the line segment from R
3
to 0. Let f (z) = (1 + z
3
+ z
6
)
1
.
Then f(z) has six simple poles at the above roots with only and
2
being inside . By
Cauchys Residue Theorem
_

dz
1 + z
3
+ z
6
= 2i
_
res
_
1
1 + z
3
+ z
6
;
_
+ res
_
1
1 + z
3
+ z
6
;
2
__
=
2i
3
_
1

2
+ 2
5
+
1

4
+ 2
_
=
2i
3
_
(2
2
+ 2
2
+ +
1
)
(1 + 2
3
) (
3
+ 2)
_
.
106 RESIDUE THEOREM & APPLICATIONS.
If we write = 2/9 then this simplies as
2i
3
3
_
4 cos(2) + 2 cos
(
3
+ 2) (
3
+ 2)
_
=
2i
3
3
_
4 cos(2) + 2 cos
|2 +
3
|
2
_
=
4i
9
3
[2 cos(2) + cos ] .
The contribution to this integral from the circular arc satises

_
arc
dz
1 + z
3
+ z
6

2R
3
sup
arc

1
1 + z
3
+ z
6

2R
3 (R
6
R
3
1)
= O
_
R
5
_
0 as R .
Hence, letting R , we have
_

0
dx
1 + x
3
+ x
6
+
_
0

3
dx
1 + x
3
+ x
6
=
4i
9
3
[2 cos(2) + cos ] ,
giving
_

0
dx
1 + x
3
+ x
6
=
4i
9
[2 cos(2) + cos ]

6
=
4
9

3
[2 cos(2) + cos ]
Example 267 Show that
_

0
3

xlog x
1 + x
2
dx =

2
6
and
_

0
3

x
1 + x
2
dx =

3
.
Remark 268 Note that its not necessary to use a keyhole contour here as along the negative
real axis our integrand at x = t < 0 becomes
3

t log(t)
1 + (t)
2
=
e
i/3
3

t (log t + i)
1 + t
2
which contributes, in its real and imaginary parts, towards both integrals.
Solution. Let R > 1 and let U be the cut complex plane with the negative imaginary axis
removed. Consider the positively oriented contour comprising [R, R] and
+
(0, R) and the
function
f(z) =
3

z log z
1 + z
2
(z U)
where
log z = log |z| + i arg z,
3

z = exp
_
1
3
log z
_
and

2
< arg z <
3
2
. By these denitions
3

i = e
i/6
and log i = i/2. So
_

z log z
1 + z
2
dz = 2ires (f; i) = 2i
_
3

i log i
2i
_
=

2
i
2
e
i/6
.
FURTHER CONTOURS 107
Now

+
(0,R)
3

z log z
1 + z
2
dz

R sup
z
+
(0,R)

z log z
1 + z
2

R
4/3
(log R + )
R
2
1
0 as R .
For z = x where x > 0 we have
3

x =
3

xe
i/3
and log (x) = log x + i. Hence, letting
R we have
_

0
3

xe
i/3
(log x + i)
1 + x
2
dx +
_

0
3

xlog x
1 + x
2
dx =

2
i
2
e
i/6
which simplies to
ie
i/3
_

0
3

x
1 + x
2
dx +
_
1 + e
i/3
_
_

0
3

xlog x
1 + x
2
dx =

2
i
2
e
i/6
.
As e
i/6
/
_
1 + e
i/3
_
= 1/
_
e
i/6
+ e
i/6
_
= 1/ (2 cos /6) = 1/

3 then
ie
i/6

3
_

0
3

x
1 + x
2
dx +
_

0
3

xlog x
1 + x
2
dx =

2
i
2

3
. (13.3)
Taking imaginary parts we nd

2
_

0
3

x
1 + x
2
dx =

2
2

3
=
_

0
3

x
1 + x
2
dx =

3
and then taking real parts in (13.3) we nd
_

0
3

xlog x
1 + x
2
dx = Re
_

ie
i/6

3
__

0
3

x
1 + x
2
dx =
_

2

3
__

3
_
=

2
6
.
Example 269 By considering a rectangular contour with corners at R, R + i, R + i,
R, and with an appropriate indentation, determine
_

0
x
sinh x
dx.
Solution. We will consider the function f(z) = z/ sinh z on the suggested contour. Then f(z)
has a removable singularity at 0 (which we can remove, and so treat the function as holomorphic
there) and also a simple pole at z = i which we need to indent around.
R R
Ri Ri
Re
Im
108 RESIDUE THEOREM & APPLICATIONS.
As f(z) is holomorphic in the rectangle then
_

f(z) dz = 0
by Cauchys Theorem. The contribution from the rectangles sides satisfy

_

0
R + iy
sinh(R + iy)
i dy

_

0
2 (R + iy)
e
R+iy
e
Riy
dy

_

0
2 (R + iy)
|e
R
e
R
|
dy

R +
2
/2
sinhR
0
as R . The residue at i is
res(f(z); i) =
i
coshi
= i.
Also note along the top of the rectangle that the function is
f(x + i) =
x + i
sinh(x + i)
=
x + i
sinh x
.
(This identity, coming from the periodicity of sinh x, is the very reason for choosing the rec-
tangular contour.) Hence letting R and shrinking the indentation to zero, we see
_

xdx
sinh x
+ lim
0
__

(x + i) dx
sinh x
+
_

(x + i) dx
sinh x
_
+ () (i) = 0.
Comparing real parts we nd
_

xdx
sinhx
=

2
2
.
Remark 270 We have not discussed in any detail precisely what integration theory we are
working within when calculating these integrals. In Mods Analysis III, all integrals discussed
were on closed bounded intervals. For those who go on to take Part A Integration they will see
that many of the integrals such as
_

xdx
sinhx
,
_

0
dx
1 + x
3
+ x
6
,
_

0
cos ax
1 + x
2
dx
exist as proper Lebesgue integrals. However integrals (usually those that require Jordans
Lemma to evaluate them) such as
_

cos x
2x 1
dx,
_

0
xsin x
1 + x
2
dx
only exist as improper Lebesgue integrals that is the limits
lim
R
_
R
R
cos x
2x 1
dx, lim
R
_
R
0
xsin x
1 + x
2
dx
FURTHER CONTOURS 109
but the integrals are not proper Lebesgue integrals. This is closely related to the dierence
between absolutely summable and conditionally summable series. The limit
lim
R
_
R
0
x|sinx|
1 + x
2
dx
does not exist and if the signed area in these improper integrals were calculated in a dierent
order to the usual one then dierent answers could be determined. The values we have assigned
to such an improper Lebesgue integral is called the principal value integral.
110 RESIDUE THEOREM & APPLICATIONS.