Solutions to Question Sheet 11.

v1 2021-22
Work on these Questions in Week 12 and the solutions will be released at
the end of that week.

1. Let f (x) = x3 on [0, 1] and let Pn be the arithmetic partition that splits
[0, 1] into n equal subintervals.
Evaluate U (Pn , f ) and L (Pn , f ).
Thus show that f is Riemann integrable on [0, 1] and find the value of
Z 1
x3 dx.
0
Pn
You may need to recall i=1 i = n2 (n+1)2 /4.
3

Solution The arithmetic partition of [0, 1] is

 
i
Pn = :0≤i≤n .
n

The function f (x) = x3 is increasing on R, so

   3
i−1 i i
Mi = sup f (x) : ≤x≤ = ,
n n n
   3
i−1 i i−1
mi = inf f (x) : ≤x≤ = ,
n n n
Thus
n  3 n
X i 1 1 X 3
U (Pn , f ) = = 4 i,
i=1
n n n i=1

n  3 n n−1
X i−1 1 1 X 1 X 3
L (Pn , f ) = = 4 (i−1)3 = 4 j ,
i=1
n n n i=1
n j=1

on writing j = i − 1. Sum these arithmetic series using the given

recollection to get
2
1 n2 (n+1)2

1 1
U (Pn , f ) = = 1+ ,
n4 4 4 n
2
1 (n−1)2 n2

1 1
L (Pn , f ) = 4
= 1− .
n 4 4 n

1
 From the theory of integration we have,
Z b Z b
L (Pn , f ) ≤ f≤ f ≤ U (Pn , f )
a a

or, in our case,

 2 Z 1 Z 1  2
1 1 1 1
1− ≤ f (x) dx ≤ f (x) dx ≤ 1+
4 n 0 0 4 n

for all n ≥ 1. Let n → ∞ to see that we must have equality in the

R1 R1
centre, that is 0 f (x) dx = 0 f (x) dx. Thus f (x) = x3 is Riemann
integrable over [0, 1]. The common value is 1/4 so
Z 1
1
x3 dx = .
0 4

2. i) Integrate f (x) = x2 over [1, 2] by using the arithmetic partition of

[1, 2] into n equal subintervals.

ii) Integrate f (x) = x2 over [1, 2] by using the geometric partition

Qn = 1, η, η 2 , η 3 , ..., η n = 2 ,


where η is the nth -root of 2.

Solution i. The arithmetic partition of [1, 2] is
 
i
Pn = 1+ : 0 ≤ i ≤ n .
n

Since f (x) = x2 is increasing on R we have

   2
i−1 i i
Mi = sup f (x) : 1+ ≤x≤1+ = 1+ ,
n n n
   2
i−1 i i−1
mi = inf f (x) : 1+ ≤x≤1+ = 1+ ,
n n n

2
Since the expression for Mi is slightly simpler then that for mi we
consider first the Upper Sum:
n  2 n 
i2

X i 1 1X 2i
U (Pn , f ) = 1+ = 1+ + 2
i=1
n n n i=1
n n
 
1 2 n(n+1) 1 n (n+1) (2n+1)
= n+ + 2
n n 2 n 6

6n3 + 6n2 (n+1) + n (n+1) (2n+1)

=
6n3

14n2 + 9n + 1
= .
6n2
For the Lower Sum we wish to reuse work and so attempt to relate the
Lower Sum to the Upper Sum.
n  2 n−1  2
X (i−1) 1 X j 1
L (Pn , f ) = 1+ = 1+
i=1
n n j=0
n n

n  2  2
X j 1 0 1  n 2 1
= 1+ + 1+ − 1+
j=1
n n n n n n

1 4
= U (Pn , f ) + −
n n

14n2 + 9n + 1 3
= −
6n2 n

14n2 − 9n + 1
= .
6n2
As in the last question the theory gives
Z 2 Z 2
14n2 − 9n + 1 14n2 + 9n + 1
≤ f (x) dx ≤ f (x) dx ≤ .
6n2 1 1 6n2

Let n → ∞ to deduce that the Riemann integral exists and

Z 2
7
x2 dx = .
1 3

3
ii. Let
Qn = η i : 0 ≤ i ≤ n

√
n
with η = 2, be the geometric partition of [1, 2]. Then

2
Mi = sup f (x) : η i−1 ≤ x ≤ η i = η i ,


2
mi = inf f (x) : η i−1 ≤ x ≤ η i = η i−1 ,


Again the expression for Mi is slightly simpler than that for mi , so

consider
n n
i 2
X
i i−1 −1

X
i
η3



U (Qn , f ) = η η −η = 1−η
i=1 i=1

η3
−1 3n


= 1−η η −1
η 3 −1

on summing the geometric series,

7η 3
= 1−η −1 since η n = 2,
η 3 −1

7 (1−η) η 2 7η 2
= = .
(1−η) (1+η+η 2 ) 1+η+η 2
Note in evaluating U (Qn , f ) do not argue as
n n n
X
2 i i−1
X
2 X
2
U (Qn , f ) = ηi η −η = ηi ηi − η i η i−1 .
i=1 i=1 i=1

Having two summations simply doubles the chance of making an error.

For the Lower Sum we first express mi in terms of Mi so we can write the
Lower Sum in terms of the Upper Sum and then reuse the calculation
above. (No need to do the same work twice.) Thus

2
2
mi = η i−1 = η −2 η i = η −2 Mi .
So
n
X n
X
−2
L (Qn , f ) = mi (xi − xi−1 ) = η Mi (xi − xi−1 )
i=1 i=1

7
= η −2 U (Qn , f ) = .
1 + η + η2

4
 Hence
2 2
7η 2
Z Z
7
≤ f (x) dx ≤ f (x) dx ≤ .
1 + η + η2 1 1 1 + η + η2

Let n → ∞ when η → 1 and we again deduce that the Riemann integral

exists and Z 2
7
x2 dx = .
1 3

3. Integrate f (x) = 1/x3 over [2, 3] by using the geometric partition

Qn = 2, 2η, 2η 2 , 2η 3 , ..., 2η n = 3 ,


where η is the nth -root of 3/2.

Solution Let
Qn = 2, 2η, 2η 2 , 2η 3 , ..., 2η n = 3 ,


where η is the nth -root of 3/2. Then for 1 ≤ i ≤ n we have

[xi−1 , xi ] = 2η i−1 , 2η i .
 

The function f (x) = x−3 is decreasing so

1
Mi = sup f (x) : 2η i−1 ≤ x ≤ 2η i =

,
(2η i−1 )3

1
mi = inf f (x) : 2η i−1 ≤ x ≤ 2η i =

.
(2η i )3
Since the expression for mi is slightly simpler we look first at the Lower
Sum.
n n
X
i −3 2 −1

X
i
i i−1
η −2



L (Qn , f ) = 2η 2η −2η = 3 1−η
i=1
2 i=1

1
η −2
1−η −1 1−η −2n


= −2
4 1−η

on summing the geometric series,

1
1 5
= 1−η −1 2 since η n = 3/2,
4 η −1 9

5 η−1 1 5
= = .
36 η (η+1) (η−1) 36η (1+η)

5
 For the Upper Sum we have

1 η3
Mi = = = η 3 mi .
(2η i−1 )3 (2η i )3

Thus
n
X n
X
U (Qn , f ) = Mi (xi − xi−1 ) = η 3 mi (xi − xi−1 )
i=1 i=1

5η 2
= η 3 L (Qn , f ) = .
36 (1+η)

Let n → ∞ when η → 1 and we again deduce that the Riemann integral

exists and Z 3
dx 5
3
= .
2 x 72

4. i) If the function h : [a, b] → R is bounded, Riemann integrable and

satisfies h(x) ≥ 0 for all x ∈ [a, b], show that
Z b
h(x) dx ≥ 0.
a

Hint What does h(x) ≥ 0 for all x ∈ [a, b] say about any Lower
Sum? What does it then say about the Lower Integral of h? Use
also the fact that h is Riemann integrable implies that the lower
and upper integrals both exist and are equal.
ii) Prove that if the functions f and g, are bounded on [a, b], and
satisfy f (x) ≤ g(x) for all x ∈ [a, b], then
Z b Z b Z b Z b
f≤ g and f≤ g.
a a a a

iii) Prove that if the Riemann integrable functions f and g satisfy

f (x) ≤ g(x) for all x ∈ [a, b], then
Z b Z b
f≤ g.
a a

6
Solution i. For any partition P of [a, b], the fact that h (x) ≥ 0 for all
x ∈ [a, b] means that L (P, h) ≥ 0. So
Z b Z b
h = h since h is integrable,
a a

Z b
= lub {L (P, h) : P partition} , by definition of ,
a

≥ 0.

ii. Let a partition P of [a, b] be given. On any interval [xi−1 , xi ] , the

inequality f (x) ≤ g(x) means that

Mif = lub f (x) ≤ lub g(x) = Mig ,

[xi−1 ,xi ] [xi−1 ,xi ]

mfi = glb f (x) ≤ glb g(x) = mgi .

[xi−1 ,xi ] [xi−1 ,xi ]

Thus
L (P, f ) ≤ L (P, g) and U (P, f ) ≤ U (P, g) (1)
for all P.
Rb
By definition g is an upper bound for all L (P, g) as P varies. From
a
Rb
(1) we then get that a g is an upper bound for {L (P, f ) : P}. Yet by
Rb
definition a f is the least of all upper bounds of this set, and so

Z b Z b
f≤ g. (2)
a a

Rb
Similarly, f is a lower bound for all U (P, f ) as P varies. Again from
a
Rb
(1) we then get that a f is a lower bound for {U (P, g) : P}. Yet by
Rb
definition a g is the greatest of all lower bounds of this set, and so

Z b Z b
g≥ f.
a a

7
 iii. The fact that f and g are Riemann integrable gives
Z b Z b
f = f since f is Riemann integrable,
a a

Z b
≤ g by (2) ,
a

Z b
= g since g is Riemann integrable.
a

5. Integrate f (x) = x2 −x over [2, 5] by using

i) the arithmetic partition of [2, 5] into n equal length subintervals

and
ii) the geometric partition of [2, 5] into n intervals.

Solution i. Let f (x) = x2 −x and

 
3i
Pn = 2 + :0≤i≤n ,
n

an arithmetic partition of [2, 5] . The function f is increasing for x > 1/2

and thus on this interval. Hence
 
3 (i−1) 3i
Mi = sup f (x) : 2 + ≤x≤2+
n n
 2  
3i 3i
= 2+ − 2+ ,
n n
 
3 (i−1) 3i
mi = inf f (x) : 2 + ≤x≤2+
n n
 2  
3 (i−1) 3 (i−1)
= 2+ − 2+ ,
n n

8
Consider first the Upper Sum:
n
( 2  )
X 3i 3i 3
U (Pn , f ) = 2+ − 2+
i=1
n n n

n 
9i 9i2

3X
= 2+ + 2
n i=1 n n
 
3 9 n(n+1) 9 n (n+1) (2n+1)
= 2n + + 2
n n 2 n 6

(12n3 + 27n2 (n+1) + 9n (n+1) (2n+1))

=
2n3

57n2 + 54n + 9
= .
2n2
For the Lower Sum
n
( 2  )
X 3 (i−1) 3 (i−1) 3
L (Pn , f ) = 2+ − 2+ .
i=1
n n n

Change variable from i to j = i − 1 so the sum now runs from 0 to

n−1: (
n−1 2  )
X 3j 3j 3
L (Pn , f ) = 2+ − 2+ .
j=0
n n n

Next, express this in terms of U (Pn , f ),

( 2  )
3×0 3×0 3
L (Pn , f ) = U (Pn , f ) + 2+ − 2+
n n n
( 2  )
3×n 3×n 3
− 2+ − 2+
n n n

6 60
= U (Pn , f ) + −
n n

57n2 + 54n + 9 54
= −
2n2 n

57n2 − 54n + 9
= .
2n2
9
From the theory we have
5
57n2 − 54n + 9
Z
≤ f (x) dx
2n2 2

5
57n2 + 54n + 9
Z
≤ f (x) dx ≤ .
2 2n2

Let n → ∞ to deduce that the Riemann integral exists and

Z 5
57
x2 −x dx = .


2 2

ii) Let
Qn = 2η i : 0 ≤ i ≤ n


p
with η = n 5/2, be the geometric partition of [2, 5]. Then, since f is
increasing on [2, 5] ,

2
Mi = sup f (x) : 2η i−1 ≤ x ≤ 2η i = 2η i − 2η i ,


2
mi = inf f (x) : 2η i−1 ≤ x ≤ 2η i = 2η i−1 − 2η i−1 .


10
Thus
n  
X
2
2η i −2η i 2η i −2η i−1


U (Qn , f ) =
i=1

n n
!
3 i 2 i
X X
−1




= 1−η 8 η −4 η
i=1 i=1

η3 η2
 
−1 3n 2n




= 1−η 8 3 η −1 − 4 2 η −1
η −1 η −1

on summing the geometric series,

η3
η2
= 117 1−η −1 − 21 1−η −1
η 3 −1 η 2 −1

since η n = 5/2,

117 (1−η) η 2 η
= 2
− 21 (1−η)
(1−η) (1 + η + η ) (1−η) (1+η)

η2 η
= 117 2
− 21 .
1+η+η 1+η

For the Lower Sum we have

n  
X
2
2η i−1 − 2η i−1 2η i − 2η i−1


L (Qn , f ) =
i=1

n n
!
−1

8 X 3
i 4 X 2
i
= 1−η η − η
η 2 i=1 η i=1

1 1
= 117 − 21 .
1 + η + η2 1+η

Hence
Z 2
117 1
− 21 ≤ f (x) dx
1 + η + η2 1+η 1

2
η2
Z
η
≤ f (x) dx ≤ 117 2
− 21 .
1 1+η+η 1+η

11
 Let n → ∞ when η → 1 and we again deduce that the Riemann integral
exists and Z 2
117 21 57
x2 − x dx =


− = .
1 3 2 2

6. Definition If f is continuous on (a, b)and F is continuous on [a, b]

and and differentiable on (a, b) with F 0 (x) = f (x) for all x ∈ (a, b)
then F is a primitive for f .
Find primitives for
1 x 1
(i) √ , (ii) √ , (iii) √ .
1 − x2 1 − x2 1 + x2
x 1 x
iv) √ , v) , vi) .
1 + x2 1 + x2 1 + x2

Solution A primitive of
√
i. 1/ 1 − x2 is arcsin x, by Question 6ii, Sheet 9,
√ √
ii. x/ 1 − x2 is − 1 − x2
√
iii. 1/ 1 + x2 is sinh−1 x, by Question 8i, Sheet 9,
√ √
iv. x/ 1 + x2 is 1 + x2 .

v. 1/ (1 + x2 ) is arctan x, by Question 6iii, Sheet 9,

p
vi. x/ (1 + x2 ) is ln (1 + x2 ).

7. The Fundamental Theorem of RCalculus says, in part, that if f is

x
continuous on (a, b) then F (x) = a f (t) dt is a primitive for f (x) on
(a, b) .
Prove that ln x, defined earlier as the inverse of ex , satisfies
Z x
dt
ln x =
1 t
for all x > 0.
Hint: Find two primitives for f : (0, ∞) → R, x 7−→ 1/x and note that
primitives are unique up to a constant.

12
Solution From the notes we know that (as an example of differentiating
inverse functions)
d 1
ln x =
dx x
for x > 0 and so ln x is a primitive for 1/x in this range. But, since
1/t is Riemann integrable and continuous on (0, ∞) we know, from the
Fundamental Theorem of Calculus, that
Z x  Z 1 
dt dt
F (x) := =− if x < 1
1 t x t

is also a primitive for 1/x. Primitives are unique up to a constant, so

Z x
dt
ln x = +C
1 t
for some constant C. Put x = 1 to find that C = 0.

13
 Solutions to Additional Questions

8. Integrate f (x) = x2 − 6x + 10 over [2, 5] using the arithmetic partition

of [2, 5] into 3n equal length subintervals.

Note how we look at P3n and not Pn , ask yourself why.

Solution Let f (x) = x2 − 6x + 10 on [2, 5]. This time f 0 (x) = 2x − 6

so f increases for x > 3 and decreases for x < 3.
Look at the partition
   
3i i
P3n = 2 + : 0 ≤ i ≤ 3n = 2 + : 0 ≤ i ≤ 3n .
3n n

We have chosen 3n instead of n so that one of the points in the partition

is x = 3, (when i = n) where the function has a turning point. Note
that the width of the intervals in the partition is 1/n.

Because of the minimum at x = 3, i.e. i = n, we have

(
f (xi−1 ) for 1 ≤ i ≤ n
Mi = sup {f (x) : xi−1 ≤ x ≤ xi } =
f (xi ) for n + 1 ≤ i ≤ 3n.

Similarly
(
f (xi ) for 1 ≤ i ≤ n
mi = sup {f (x) : xi−1 ≤ x ≤ xi } =
f (xi−1 ) for n + 1 ≤ i ≤ 3n.

Note that
i2
 
i i
f (xi ) = f 2 + = 2 − 2 + 2,
n n n
and so
(i−1)2 (i−1)
f (xi−1 ) = 2
−2 + 2.
n n

14
Hence
n
!
X (i−1)2 (i−1) 1
U (P3n , f ) = 2
−2 +2
i=1
n n n

3n  2 
X i i 1
+ 2
−2 +2
i=n+1
n n n

n−1 n−1
1 X 2 2 X
= i − 2 i+2
n3 i=1 n i=1

3n 3n
1 X 2 2 X
+ i − i + 4.
n3 i=n+1 n2 i=n+1

We can combine two pairs of summations, noting that the i = n term

is missing in both. So
3n
! 3n
!
1 X 2 2 2 X
U (P3n , f ) = i −n − 2 i−n +6
n3 i=1 n i=1

   
1 37 2 1 2 9 2 1
= 9n + n + n − 2 n + n +6
n3 2 2 n 2 2

12n2 + 5n + 1
=
2n2
Similarly
n  2 
X i i 1
L (P3n , f ) = 2
−2 +2
i=1
n n n

3n
!
2
X (i−1) (i−1) 1
+ 2
−2 +2
i=n+1
n n n

12n2 − 5n + 1
=
2n2
It matters not that we have 3n in place of n in
Z 5 Z 5
L (P3n , f ) ≤ f≤ f ≤ U (P3n , f ) .
2 2

15
 Thus
5 5
12n2 − 5n + 1 12n2 + 5n + 1
Z Z
≤ f (x) dx ≤ f (x) dx ≤ .
2n2 2 2 2n2

Let n → ∞ to deduce that the Riemann integral exists and

Z 5
x2 − 6x + 10 dx = 6.


2

R3 R5
Note In this proof we have essentially calculated 2 f , 3 f and added
the results together. That you can do this is a result we have not had
time to cover in the course.

9. Let f : [0, 1] → R be given by f (0) = 0 and, for x ∈ (0, 1],

1 1
f (x) = where n is the largest integer satisfying n ≤ .
n x

Draw the graph of f . Show that f is monotonic on [0, 1] .

Deduce that f is Riemann integrable on [0, 1].

Find Z 1
f.
0

Hint. First calculate the integral over [1/N, 1] for any N ≥ 1. Then
use this in evaluating the upper and lower integrals of f over [0, 1].

Solution Let 0 ≤ x < y ≤ 1 be given. Write nx for the largest integer

nx ≤ 1/x so f (x) = 1/nx . Similarly ny is the largest integer ≤ 1/y.
Then
1 1 1 1
x < y =⇒ < =⇒ ny ≤ nx =⇒ f (x) = ≤ = f (y) .
y x nx ny

Hence f is a monotonic (in fact, increasing) function.

16
Graph of y = f (x):
y
1/1

1/2

1/3
1/4
1/5

x
1 1 1 1 1
5 4 3 2 1

It can be shown that any monotonic function is Riemann integrable.

Here, though, we will not assume this but first note that f is Riemann
integrable over the interval [1/N, 1] for any N ≥ 1. In fact
Z 1 N −1 Z 1 N −1  
X j 1 X1 1 1
f (x) dx = = −
1/N j=1
1
j+1
j j=1
j j j +1

N −1 N −1
X 1 X 1
= − .
j=1
j2 j=1
j (j +1)

Here we have a ‘telescoping’ series,

N −1 N −1  
X 1 X 1 1
= −
j=1
j (j + 1) j=1
j j +1
       
1 1 1 1 1 1 1 1
= − + − + − + ... + −
1 2 2 3 3 4 N −1 N

1
= 1− .
N

So
Z 1 N −1
X 1 1
f (x) dx = 2
−1+ .
1/N j=1
j N

We cannot justify letting N → ∞, instead we examine the upper and

lower integrals of f .

17
First f ≥ 0 implies
Z 1 Z 1 Z 1
f (x) dx ≥ f (x) dx = f (x) dx,
0 1/N 1/N

the last step following from f being Riemann integrable over the inter-
val of integration.
For an upper bound we note that if 0 < x < 1/N then N < 1/x. So
if Nx is the largest integer ≤ 1/x we have Nx ≥ N . Yet by definition
f (x) = 1/Nx and so f (x) ≤ 1/N . That is,
1 1
0<x< =⇒ f (x) ≤ .
N N
Hence
Z 1 Z 1/N Z 1
f (x) dx = f (x) dx + f (x) dx
0 0 1/N

Z 1/N Z 1 Z 1
1 1
≤ + f (x) dx = 2 + f (x) dx
0 N 1/N N 1/N

Combining we have
Z 1 Z 1 Z 1 Z 1
1
f (x) dx ≤ f (x) dx ≤ f (x) dx ≤ 2 + f (x) dx.
1/N 0 0 N 1/N

That is,
N −1 Z 1
X 1 1
− 1 + ≤ f (x) dx
j=1
j2 N 0

Z 1 N −1
X 1 1 1
≤ f (x) dx ≤ 2
−1+ + 2.
0 j=1
j N N

Now let N → ∞, concluding that the lower and upper integrals agree
and so f is Riemann integrable over [0, 1]. Further, the value of the
integral is the common value,
Z 1 ∞
X 1 π2
f (x) dx = 2
− 1 = − 1.
0 j=1
j 6

18